A Landscape Archaeology Approach to Understanding Household Water. Management Practices of the Ancient Lowland Maya

Transcription

1

2 A Landscape Archaeology Approach to Understanding Household Water Management Practices of the Ancient Lowland Maya A dissertation submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Geography of the College of Arts and Sciences by Jeffrey L Brewer M.A. University of Cincinnati B.A. University of Cincinnati March 2017 Committee Chair: Nicholas P. Dunning, Ph.D.

3 Abstract For the ancient Maya, the collection and storage of rainfall were necessary requirements for sustainable occupation in the interior portions of the lowlands in Mexico, Guatemala, and Belize. The importance of managing water resources at the household level, in the form of small natural or culturally modified tanks, has recently been recognized as a spatially and temporally widespread complement to a reliance on the larger, centralized reservoirs that occupied most urban centers. Emerging evidence indicates that these residential tanks functioned to satisfy a variety of domestic water needs beginning in the Middle Preclassic ( BC) period. The research presented in this dissertation aims to clarify the role of small topographical depressions in ancient Maya domestic water management utilizing a combination of satellite remote sensing and archaeological excavation to identify, survey, and evaluate small household tanks. The three research articles included here focus on the lidar identification and subsequent archaeological investigation of these features at the central lowland sites of Yaxnohcah in southern Campeche, Mexico and Medicinal Trail in northwestern Belize. In addition to clarifying the origin and functions of these reservoirs, their role within the broader mosaic of ancient Maya water management infrastructure and practices, particularly within the Elevated Interior Region (EIR) of the Yucatán Peninsula is also explored. This original research supplements existing archaeological, environmental, and remote sensing studies of ancient Maya civilization and contributes to the advancement of Maya studies by: (1) providing a scheme for identifying closed depressions as probable water features within lidar imagery; (2) placing household tanks within the larger framework of ancient Maya water systems and practices; and (3) identifying spatial and temporal linkages between household tanks and differing levels of urban development in the Maya lowlands. ii

4 iii

5 Acknowledgements I would first like to thank the Department of Geography at the University of Cincinnati for accepting me into the Doctoral Program, funding my graduate studies, and providing me the platform for conducting this dissertation research. Dr. Nick Dunning, who served as my Advisor and Ph.D. Committee Chair, shepherded me through the program, offering guidance, advice, support, and criticism when necessary. Thanks, Dr. D.! Thank you also to my Ph.D. Committee: Dr. Hongxing Liu, Dr. Robert South, Dr. Kevin Raleigh, and Dr. Vern Scarborough. Your time and efforts of my behalf are greatly appreciated. The research reported here was conducted through two archaeological projects: the Programme for Belize Archaeological Project (PfBAP) in Belize and the Yaxnohcah Archaeological Project (YAP) in Campeche, Mexico. The PfBAP is directed by Fred Valdez, Jr. from the University of Texas at Austin and Kathryn Reese-Taylor and Armando Anaya Hernández from the University of Calgary and the Universidad Autónoma de Campeche, respectively, co-direct the YAP. Thank you all for hosting and supporting my fieldwork. Finally, this work would not have been possible without the support of my friends and family, particularly my wife Megan, throughout my graduate school career. Thanks for the love, support, and patience! iv

6 Table of Contents 1. Introduction Statement of Problem Organization of the Dissertation Literature Review Study Area Water Availability Ancient Maya Water Management Practices: Place, Power, and Control Preclassic Period Water Management Classic Period Water Management Recent Interdisciplinary Approaches to Ancient Maya Water Management Research Article Research Article Research Article Summary and Conclusions Bibliography v

7 List of Figures Research Article 1 Figure 1: Map of the Maya region showing Yaxnohcah Figure 2: Lidar-derived hillshade image of study area Figure 3: Elevation profile of closed depression CD-101 at the Wo Group Figure 4: Alba Group and adjacent depressions Figure 5: Fidelia Group and Operation D Figure 6: Wo' Group with Operations D-1, D-2, and D-5 and associated watersheds Figure 7: Closed depression D-3 (Op. 16F) north profile drawing Figure 8: Closed depression D-4 north profile image Figure 9: Base of closed depression D-2 showing cut limestone blocks Research Article 2 Figure 1: Map of the Maya region showing Medicinal Trail and Yaxnohcah Figure 2: Medicinal Trail site map showing structure groups discussed in the text Figure 3: Map of Group A at Medicinal Trail Figure 4: South profile drawing of Operations 10Q, 10R, 10Y, and 10Z Figure 5: Image of cf. Onagraceae seeds Figure 6: Plan map of Group C at Medicinal Trail Figure 7: Map of Yaxnohcah showing the location of investigated depressions Figure 8: Operations 16B, 16C, and 16D south profile drawing Figure 9: Operation 16F west profile image showing remnants of plaster floor Figure 10: Image of Operations 16G, 16H, and 16I showing large placed stones vi

8 Figure 11: Operation 16J north profile drawing Figure 12: Pre-excavation image of the Yax-3 hinterland aguada Figure 13: North profile image of the Yax-3 group reservoir Research Article 3 Figure 1: Map showing the situation of the EIR and sites discussed in the text Figure 2: Diagram showing depression filling and stream damming reservoirs Figure 3a: Profile of south wall of Op. SB3C-1 in Aguada San Bartolo; Figure 3b: Lidar-derived hillshade image of Brisa Reservoir and associated ceremonial center, Yaxnohcah Figure 4: Cross-sectional drawing of Aguada Tintal, northeast of San Bartolo Figure 5: Cross-sectional drawing of La Milpa Aguada with buk'te Figure 6a: Idealized drawing of Puuc residential group with chultun; Figure 6b: Photo of chultun from the site of Yaxhom Figure 7a: Idealized drawing of residential tank; Figure 7b: Photo of a residential tank at Yaxnohcah prior to excavation; Figure 7c: Photo of excavation in Yax-3 Residential Group tank (Yaxnohcah) Figure 8: Space Shuttle photo of Uxmal and surrounding reservoirs, with inset photo of Aguada ChenChan vii

9 List of Tables Research Article 1 Table 1: Data summary and evaluation for investigated depressions Research Article 2 Table 1: Data summary and evaluation of all reservoirs investigated by the author Research Article 3 Table 1: Estimated capacity of selected EIR reservoirs viii

10 1. Introduction 1.1 Statement of Problem The study of water management practices of the ancient Maya has contributed significantly to scholarly understanding of the myriad environmental, social, political, and economic activities that characterized this resilient civilization for centuries prior to European colonization of the Americas. Our identification and understanding of the water management features of the Maya lowlands, in particular, continue to provide insight into the complex humanenvironment relationships that characterized much of the Late Preclassic (400 BC AD 150) and Classic (AD ) period Maya. At the community scale, a network of reservoirs (aguadas), seasonally inundated wetlands (bajos), canals, dams, wells, natural sinkholes (cenotes), and small subterranean cisterns (chultunes) characterized the hydrologic infrastructure of Maya settlements. At the household level, small topographical depressions that functioned as seasonal reservoirs served as significant sources of water for the region s inhabitants. Many natural depressions associated with ancient Maya settlements were lined with clay, lime plaster, and/or stone-facing to enhance their water retention capabilities. Archaeological evidence also shows that some of these residential tanks evolved from human-modified depressions such as ancient limestone, clay, or chert quarries that were subsequently transformed to store water. The combined effects of the karst hydrology of the Maya lowlands, the seasonality of available precipitation, and a lack of permanent natural water sources necessitated the collection and storage of rainwater, a system in which small reservoirs likely played a necessary and substantial role in satisfying domestic water needs. The objective of this dissertation is to gain a clearer understanding of the origins of residential water tanks and the role they played in the daily lives of the ancient lowland Maya. In 1

11 addition to broadening our understanding of water management activities, the analysis of domestic scale water management features provides an opportunity to supplement our knowledge of human-environment interactions in the Maya area. To fully understand the origin and function of these household tanks, an interdisciplinary approach incorporating aspects of archaeology, remote sensing, geology, hydrology, and climatology that considers the physical, spatial, and temporal characteristics of these features is necessary. 1.2 Organization of the Dissertation This dissertation begins with a review of the relevant literature on the physical environment of the Maya lowlands; the availability and nature of water resources in the region; an overview of ancient Maya water management practices from the Middle Preclassic through Classic periods, including a discussion of the hydrologic infrastructure of specific sites that serve as case studies for these eras; and a summary of interdisciplinary approaches to studying aspects of ancient Maya water management over the past few decades. The following chapters are composed of two journal articles detailing household water management investigations conducted at the central Maya lowland sites of Medicinal Trail, Belize, and Yaxnohcah, Mexico, between 2006 and 2015 and a portion of a book chapter summarizing household water collection and storage within the Elevated Interior Region (EIR) of the Maya lowlands. The first research article, written in collaboration with a team of researchers from the Yaxnohcah Archaeological Project (YAP), discusses the application of airborne lidar (Light Detection and Ranging) for detecting closed topographical depressions and the subsequent ground verification and archaeological testing of five of these features at the site of Yaxnohcah in southern Campeche, Mexico. Our goal in locating and analyzing these depressions was to 2

12 develop a scheme for identifying closed depressions in the lidar imagery and to determine through archaeological excavation if they functioned as water storage features. If they did serve in a water management capacity, we also sought to clarify the nature of these activities based on recovered material and environmental data. The second article presents the results of decentralized, domestic scale water management investigations from two central lowland sites: the hinterland agricultural community of Medicinal Trail in northwestern Belize and the urban center of Yaxnohcah in southern Campeche, Mexico. Research included a single reservoir and additional water features associated with two residential courtyard groups evaluated at Medicinal Trail and a series of four small tanks associated with residential structures throughout Yaxnohcah. A multidisciplinary approach consisting of landscape survey, excavation, and material analysis was undertaken to understand the water management practices in which the inhabitants of these two sites were engaged at the household level. The succeeding chapter contains a segment of a book chapter detailing ancient Maya water collection and storage in the Elevated Interior Region (EIR) of the Maya lowlands. Environmental characteristics of the EIR posed significant challenges for year-round occupation and urbanization, with the rarity of accessible surface and groundwater sources and a 5-month annual dry season necessitating the collection and storage of rainwater in order to concentrate human population. This paper presents a detailed overview of water management challenges, strategies, and infrastructure in the region. In addition to large urban and hinterland reservoirs operating from the Middle Preclassic through Classic periods, residential scale tanks and cisterns represented important adaptations for urban centers in the EIR throughout their occupation. Although considerable interregional variation existed in water management strategies, these 3

13 residential features have been less investigated than their larger counterparts. Particular focus is placed on a discussion of the construction and maintenance of these household systems, how their role in the daily social, economic, and political lives of their community s members is reflected in their location and scale, and the necessity of considering household water features as important components within the complex mosaic of water management adaptations in the EIR. The dissertation concludes with a brief synopsis and discussion of the research presented, including summaries of the preceding three chapters, and an outline of areas targeted for future research. 4

14 2. Literature Review 2.1 Study Area The Maya lowlands are an extremely diverse series of microenvironments subject to thin soils, variable rainfall, and an extended period of water stress largely without immediate access to permanent water sources. The main factors responsible for the diverse patterning of microenvironments in Mesoamerica are variation in rainfall, soil, geomorphic processes, and slope gradients and drainage caused by structural hydrology (Dunning and Beach 2000:181). In an area largely dominated by tropical and semitropical forests, the Maya managed a complex political economy based on a structured landscaping investment (Scarborough and Valdez 2003). The Maya systematically altered their surroundings over hundreds of years, producing an environmental setting profoundly different than what was originally settled or what has now been reclaimed. An overview of the key elements composing the environmental context of the Maya lowlands highlights the importance of water management and helps to clarify the role played by household reservoirs Geographic Location The lowlands are located entirely within the tropics, extending over 68,000 km2 between 14 and 22 N and 87 and 93 W. This area covers approximately 900 km from north to south and 550 km east to west along the continental divide, near the Pacific coast, and roughly 400 km across the Yucatán Peninsula (Gill 2000). Nearly half of the Maya area is found within modern Mexico, including much of Chiapas and Tabasco and the entirety of Campeche, Yucatán, and Quintana Roo. The remainder is comprised of Belize, the Department of the Petén in northern Guatemala, and the northern portions of Honduras and El Salvador. The area is generally 5

15 subdivided into the northern, central, and southern lowlands (Hammond and Ashmore 1981), defining both a cultural and physical region defined by internal variations in elevation, rainfall, drainage, soils, and other factors that create considerable environmental diversity. Within the lowlands, evidence of Maya culture and its impacts on the natural landscape date from the Early Preclassic period ( BC) to the present, punctuated by significant fluctuations over time (Brenner et al. 2003; Pohl et al. 1996) Geology and Physiology The majority of the lowlands are founded on the Cretaceous and Tertiary Yucatán platform, and most of the region is underlain by limestone (Beach et al. 2006; Dunning and Beach 2010). From the northern and coastal plains, this platform ranges from little surface gradient and low elevation to an area with about 250 m of regional relief shaped and cut by karst dissolution and fluvial erosion structured by a series of normal faults. Abrupt elevation changes occur along scarps near the fault lines and across much of the peninsula drainage is primarily subterranean with the karst interior essentially acting as a massive bedrock sponge for much of the region s rainfall (Dunning and Beach 2010). Belize, Tabasco, the Pacific lowlands, the Caribbean coast of northern Yucatán, and Quintana Roo are bands of geologically recent alluvial deposits. The northern third of the Yucatán Peninsula is Tertiary limestone and calcareous marl (sascab), while the southern portion and areas of Chiapas, Petén, and Belize are composed of older limestones, exhibiting greater elevation and relief and more developed drainage. The hills of the southern Petén, the Alta Verapaz, and the foothills of the Maya Mountains in southern Belize are formed from older and more elevated Cretaceous and Jurassic limestones, Triassic shales, and older sedimentary rocks lying still further south. The Maya Mountains themselves are 6

16 a northern isolated block of much older rocks, separated from the Guatemalan Highlands, and are made up of sedimentary and metamorphic deposits and Paleozoic granites (Gill 2000). Three areas of limestone, becoming successively younger in age and lower in elevation from south to north, form the land mass known as the Yucatán Platform. During most of its history it has remained a rigid mass, emerging from submersion by the surrounding seas as a stable tectonic unit. The result is a widespread area of low topography that makes up the Yucatán Peninsula and the Petén. The southern end of the platform, forced against the Guatemalan Highlands, has buckled slowly into high ridges that have eroded into steep, forested hills. The northern end is the site of the Cretaceous extraterrestrial impact, the Chicxulub crater, believed to have caused the massive ecological catastrophe responsible for the extinction of the dinosaurs, among other life forms, approximately 65 million years ago (Gill 2000). Four major types of land resources are found in the Maya lowlands, differing in potential productivity under conditions of low and high labor output (Fedick and Ford 1990). The first is the well-drained uplands of the central lowlands, consisting of rolling limestone hills and shallow, but fertile, soils. Next are the slow-drained lowlands, commonly found in low lying areas within uplands and often representing transitional zones between well-drained uplands and various forms of swamps. These slow-draining lowlands can be found in extensive tracts, representing the prevalent land type of large areas. The riverine-associated swamps of San Antonio (Bloom et al. 1985) and Pulltrouser Swamp (Turner and Harrison 1983), both located in northern Belize and the subject of interdisciplinary investigations of wetland agricultural features and the environment, represent a third type of land resource. Pulltrouser Swamp comprises three linked depressions between the New and Hondo rivers in northern Belize, with its southern end lying within 200 m of the New River. The swamps at San Antonio are directly associated with 7

17 the Hondo River. Both are strongly linked to local ground water levels and affected by the fluctuation rates of their associated rivers. Much of the current knowledge of these swamps is based on studies within the Bajo de Santa Fe located near the site of Tikal (Fedick and Ford 1990) and the Far West Bajo near La Milpa (Kunen 2004), and research from the rural site of Guijarral (Kunen and Hughbanks 2003). These seasonal swamps are closed karstic depressions filled with deep deposits of impermeable clays. They are not associated with river systems, but support an elevated water level high above the basal water table. These swamps hold significant amounts of standing water during the rainy season, but with the exception of occasional sinkholes, completely desiccate during the dry season (Dunning et al. 2002). The overall terrain of the region is strongly influenced by the underlying geology. Horst and graben topography exists due to tectonic movement, with the faults producing large escarpments and valleys exhibiting vast elevation changes (Dunning and Beach 2000). Wetlands are karstic features with roughly 40% of the southern lowlands currently occupied by bajos (Scarborough 2003). In the wetter, more elevated south, seasonal surface runoff and perennial spring discharge feed rivers including the Usumacinta and Candelaria to the west and the Hondo and Belize rivers to the east. Within the central interior of the peninsula, surface water is generally found only in surface ponds (aguadas) and a few deep caves during the dry season. Structural faulting, combined with limestone and gypsum dissolution, have created countless basins (bajos) of varying size in these elevated interior areas which internally drain into the bedrock and are sometimes linked to one another by seasonal surface channels (Dunning and Beach 2010). On the more arid northern plains, permanent freshwater is extremely limited and largely available intermittently in sinkholes (cenotes). 8

18 2.1.3 Weather and Climate The impacts of seasonality, rainfall, and climate on Maya civilization including their potential role in the Classic period collapse have not suffered for lack of scholarly discussion (e.g. Adams 1991; Brenner et al. 2003; Bullard 1960; Dunning et al. 1999, 2006a, 2014; Ford 1996; Gill 2000; Hansen et al. 2002; Harrison 1993; Lucero et al. 2011; McAnany 1990; Scarborough 1991a, 1993b, 2003; Scarborough and Gallopin 1991; Scarborough and Valdez 2003). The Maya area exhibits a mixture of tropical and semitropical climate, owing to its location on the northern edge of the Intertropical Convergence Zone (ITCZ). Weather in the lowlands is part of the North Atlantic weather system and Nortes (or northers), which occasionally blow in during the winter, are the only weather phenomena possibly originating in the Pacific. These brief cold air outbreaks originate within the frigid interior of the wintertime North American continent (Chmilar 2005; Gill 2000). The lowlands lie in the path of easterly trade winds, characterized by roughly six-month rainy and dry seasons. Climatically, based on this wet and dry seasonal pattern, the lowlands fall into is either Köppen Am (Tropical Monsoon) or Aw (Tropical Wet and Dry) classifications depending on location (Akpinar 2011). The primary difference between the two regimes is the length of the dry season, which becomes more pronounced moving south to north through the lowlands and into the northern Yucatán Peninsula (Foster and Turner 2004). Rainfall in the lowlands varies spatially as a result of latitude, elevation, and the effects of rain shadow (Beach et al. 2008), and interannually due to regional and global climatic dynamics (Brenner et al. 2003; Hodell et al. 2001; Mueller et al. 2009). Precipitation rates across most of the lowland region range from an average of 1,350 to 2,000 mm/year, with the majority falling during the rainy season between late May and middecember and the corresponding dry season receiving little rain (Dunning et al. 1999; Ford 9

19 1996; Scarborough 1991a, 1993b, 2003; Scarborough and Valdez 2003). In the Petén, for example, precipitation ranges from 900 to 2500 mm/year, with a regional annual mean of 1600 mm (Rosenmeier et al. 2002). In addition, annual precipitation is largely affected by shifting atmospheric belts and prevailing winds. The northward migration of the ITCZ and the BermudaAzores High, a large subtropical semi-permanent center of high atmospheric pressure typically found south of the Azores in the Atlantic Ocean, stimulates particularly heavy rains in the lowlands between the months of June and October. During November and December, the lowsun season, dry weather conditions and strong trade winds dominate due to the ITCZ and Bermuda-Azores shifting toward the Equator. A distinct dry season is in effect between the months of January and May throughout the region (Kueny and Day 2002; Piperno and Pearsall 1998; Rosenmeier et al. 2002) Soils As geology impacts terrain, the terrain at least in part determines the distribution of soil types in the lowlands. Lowland soils can be broadly characterized as having relatively high fertility and being well drained, clayey, and calcareous, though they tend to be shallow. Moving from north to south, soils generally become deeper and more poorly drained (Dunning and Beach 2010; Dunning et al. 1998). Soil formation is more apparent in sunken areas with greater deposition of sediment, directly contributing to the variety of soil types blanketing the Maya landscape. Soils associated with aguadas range from thin, highly erodible, and arable to heavy, impervious clays (Akpinar-Ferrand et al. 2012; Chmilar 2005; Dunning and Beach 1994; Dunning et al. 2007; Higbee 1948; Lundell 1933). Within bajos, weathering of the bedrock results in the deposition of montmorillonitic clays that produce an impermeable layer seen 10

20 ubiquitously in the formation of Vertisols. A mixture of poorly drained Vertisols and Mollisols grading toward Vertisols comprises the soils of the slow-draining lowlands. Vertisols expand and contract with wet and dry seasons due to their content of swelling-type montmorillonitic clays and, when wet, they are plastic and sticky. Conversely, these soils become extremely hard when dry, developing their characteristic cracks (Buol et al. 2003). While Vertisols (and Mollisols grading toward Vertisols) are relatively fertile, their agricultural development under hand cultivation is inhibited by poor drainage and workability. In addition, the saturated (or flooded) slow-drained lowlands are too wet for swidden cultivation during the rainy season (Fedick and Ford 1990). In addition to Vertisols, deeper and often fertile Mollisols and Histosols (organic mucks) are also present. Episodic shrink-swell activity (argilliturbation) and considerable drainage limitations characterize these soils. Mollisols, among the world s most naturally productive soils, are the dominant soil type of the well-drained uplands. These dark-colored tropical soils are rich in organic matter and, although shallow, are highly productive under certain hand cultivation systems. Despite lacking available phosphorous, modern Maya farmers consider them to be the most productive for traditional wet season swidden cultivation (Fedick and Ford 1990) Fertile, but shallow, clayey Mollisols and rendzinas are found in the limestone uplands where they are prone to erosion owing to their general location on sloping surfaces (Dunning et al. 1998; Siemens 1978). Lowland soils are thin, but fertile, on hills and well-drained areas, with thick clays dominating in low-lying areas (Buol et al. 2003; Dunning and Beach 2000; Fedick and Ford 1990; Scarborough 1991a, 1993b). Wetter conditions with hydromorphic clay and the formation of organic soils dominate in deeper depressions (Dunning et al. 1998). Within basins, deep layers 11

21 of anthropogenic sedimentary clays have been collectively referred to by scholars as Maya Clay. Deforestation and agricultural practices leading to accelerated soil erosion are believed to have generated this deposition (Rosenmeier et al. 2002). In addition, many of the perennially wet bajo systems are thought to have evolved into seasonal swamps through the combined effects of erosion and deposition associated with human disturbance (Dunning and Beach 1994; Dunning et al. 1998, 2006a) Vegetation The vegetation of the Maya lowlands is influenced by both rainfall patterns and the soil drainage properties of the region (Dunning et al. 1998). Lundell s (1933) early survey classified the dominant local vegetation as quasi-rainforest, and the area is characterized by a varied vegetative landscape owing to diverse surface topography and water availability. The semitropical jungles of the Maya area are marked by high species diversity, but with a low concentration of any given species in a localized area (Dunning et al. 2003; Scarborough 1996; Scarborough and Valdez 2003). Four distinct zones, each with a unique vegetative regime, define the Maya area (Chmilar 2005). Broad-leaved, closed-canopy forest, achieving a maximum height in excess of 50 m, characterizes the well-drained uplands of the central Maya lowlands. The dominant vegetation of slow-drained lowlands or low-lying bajos is high marsh forest with a broken canopy and a variety of palms and hardwoods. Vegetation communities of riverine-associated swamps such as Pulltrouser and San Antonio are dominated by a rush-sedge community, with variations relating to unique water regimes ranging through palm forest and grass communities. The vegetation of the fourth zone, closed depression seasonal swamps such as the Bajo de Santa Fe, is noticeably 12

22 different from that of the riverine-associated swamps of northern Belize. Although given to holding water for nearly six months out of the year, the mixed xerophytic forest they support is adapted to survive the dry season, during which soil moisture is held in an unavailable (hygroscopic) state by heavy clays (Fedick and Ford 1990). This distinctive vegetation community is also adapted to poor aeration and severe soil acidity. 2.2 Water Availability In addition to the lengthy annual period of desiccation, surface water is variably distributed across the region. Hydrologically, the area is defined as an open basin (Villaruso and Ramos 2000). Surface and spring-fed rivers drain the margins of the southern half of the Yucatán Peninsula, while perennial surface water is virtually absent and groundwater largely inaccessible due to increasing subsurface water depth in the elevated karst interior of the southern Maya lowlands (Dunning and Beach 2000; Hammond and Ashmore 1981; Scarborough 1993; Wilson 1980). Groundwater discharge along the margins of the elevated interior feeds coastal river systems. The primary river systems of the southern Maya lowlands are the Rio Hondo and Belize River to the east and the Usumacinta and San Pedro Rivers to the west (Scarborough 2003). The eastern and western perimeters of the lowlands are drained by perennial rivers, and a string of permanent lakes stretches across the fault zone of the central Petén. Surface water is absent from the interior Petén around Tikal, however, with no natural rivers or permanent streams. Drainage catchments are found in closed-depression swamps, such as the Bajo de Santa Fe, that hold standing water in the wet season due to their deep, impermeable clays but desiccate completely in the dry season to a useless cement-like hardpan (Ford 1996). 13

23 Aside from the seasonal availability of reservoirs (aguadas) and bajos, water is generally accessible year-round in civales (Hansen et al. 2002; Scarborough 2003). These features are treeless, wet areas of herbaceous vegetation much less extensive than bajos, yet almost always located within or adjacent to these seasonal swamps. The marshy civales usually remain wet throughout the dry season in most years, and many of the region s natural water holes are found within them (Hansen et al. 2002). Aguadas, bajos, and civales are karstic landscape features occurring throughout the Maya lowlands. Bajos alone cover between 40% and 60% of the interior southern Maya lowlands and occur in depressions relative to the general terrain of the area. The clay soils of bajos swell and flood during the rainy season, with the ground drying out and cracking during the dry months (Dunning et al. 2002, 2006a). Aguadas, widely distributed across the central and southern lowlands in particular, originate through multiple natural or human actions and can offer a dependable, multi-year water source (Akpinar 2011; Chmilar 2005). The availability of water to the ancient Maya, or in fact the lack thereof, has been cited as one possible explanation for the Classic period collapse. Many of the same scholars have alluded to climate change and draught as key factors in the reduction of water available to the lowland Maya (Dunning and Beach 2000, 2004; Dunning et al. 2003, 2014; Gill 2000; Hansen et al. 2002; Hodell et al. 1995; Lucero 2002; Scarborough 2003; Siemens 1978). Falling sea levels coupled with a decline in annual precipitation led to increasing scarcity of water, and paleoecological studies have indicated that climate change was at least partially responsible for environmental alterations in particular, the distribution of bajos and civales (Chmilar 2005; Dunning et al. 1997, 1998, 2006b). As a result, the ancient lowland Maya were undoubtedly 14

24 occupying an environment prior to the collapse that was markedly different, at least in terms of water distribution, from that which is present in the region today. 2.3 Ancient Maya Water Management Practices: Place, Power, and Control Humans require an absolute minimum of 2 to 3 liters of water a day in a settled environment under normal living conditions (White et al. 1972). This fact alone underlines the critical role water plays in human survival, regardless of levels of physical activity, differences in body type, or environmental conditions. Nearly as important, at least in terms of a society s reliance on water, is the idea that water availability limits the location of permanent populations (Scarborough and Gallopin 1991:658). Scarborough (1991a:103) writes that the natural environment partially conditions water use, and that the primary geomorphological variables conditioning water management are topography and soil permeability (the latter measured by rates of seepage). Furthermore, a pronounced dichotomy exists between the natural environments in which early complex cultures emerged. Ford (1996) suggests four major variables that contribute to the development of complexity: overall resource productivity, local resource control, resource diversity, and critical resource control. She contends that, although taken together these factors are key to the evolution of complex societies, the source of power largely depends on the effective control and management of critical resources. She defines critical resources as those directly related to subsistence and suggests the distribution of water in the Maya lowlands as a prime example. Historically, complicated experiments in water management occurred in arid regions with both rugged and low relief, as well as in humid settings with gentle topographic contours. Along with the Hohokam (AD ) of the U.S Southwest and cities such as Chan Chan (c. AD 1000) near the north coast of Peru, the Maya lowlands are one 15

25 of the principal regions in the New World where complex water systems evolved (Scarborough 1991a). In the southern lowlands, many of the largest centers and their densely-settled hinterlands are in fertile agricultural areas without permanent surface water such as lakes or rivers, but with seasonal water sources including aguadas and bajos (Fedick and Ford 1990; Lucero 1999). The major urban centers of Tikal, Caracol, Calakmul, Naranjo, and El Mirador exemplify these settings (Dunning et al. 2015; Folan et al. 1995; Matheny 1987; Scarborough et al. 1994, 1995, 2012). Although the Maya first settled along coastal areas and rivers, populations moved inland during the Middle Preclassic period (ca BC) to fertile agricultural areas located in the karstic uplands like the Petén region. These karstic hills also contain aguadas, some of which hold significant reserves of water throughout the year (Ford 1986; Lucero 1999). The sites of El Mirador and Nakbe are two well-known large Preclassic centers in the Petén, both of which were abandoned at or near the end of Preclassic period (ca. AD 150) possibly due to a failure of their water management systems in part caused by the buildup of silt (Matheny 1987; Scarborough 1993). Even today, a lack of permanent water sources combined with the desiccation of many reservoirs, streams, and other surface sources during the dry season is a major hindrance to modern settlement of many parts of the Petén (Bullard 1960; Lucero et al. 2011). Pioneer groups also pushed inland due to increasing population growth, following the minor streams and associated swamp margins and adapting to the lengthy dry season by living near natural depressions that might hold water for an extended period (Scarborough 1998). These natural depressions, or reservoirs, were open catchment basins originating from collapse or dissolution sinkholes commonly found along bajo margins and bedrock fractures in karst uplands (Akpinar-Ferrand et al. 2012; Siemens 1978). Many of these features in the vicinity of ancient 16

26 Maya settlements were modified to collect wet season rainfall and hold sporadic surface runoff or water from more permanent canalized sources and were found to have been lined with clay, lime plaster, and/or stone-facing to enhance their water retention abilities (Adams and Jones 1981; Akpinar-Ferrand et al. 2012; Brewer 2007; Wahl et al. 2007; Weiss-Krejci and Sabbas 2002). Reservoirs in the Maya area were of several varieties. Chultunes, large, bell-shaped cisterns of the dry northern Yucatán Peninsula, were designed to collect rainfall from a prepared platform surface (McAnany 1990). Depression reservoirs were also created through construction and extraction activities, as ancient limestone, clay, or chert quarries were later transformed to store water (Brewer forthcoming; Brewer et al. forthcoming; Bullard 1960; Dunning et al. 2007; Hester and Shafer 1984; Weiss-Krejci and Sabbas 2002). The exploitation of natural drainage gradients was a common practice in Maya hydrological engineering, and large seasonal reservoirs and aguadas were frequently modified from the naturally low-lying terrain throughout the lowlands (Matheny 1978). Water management in the Maya lowlands emphasized collection over diversion, source over allocation (Scarborough and Gallopin 1991:658). Examination of the ancient reservoir technology employed in the tropical wet-dry forests of the southern lowlands supports the notion that rainwater storage tanks were the primary source of water for many communities during the lengthy annual dry season. For example, the planning and placement required for the substantial reservoir construction at the major site of Tikal one of the best documented major sites in the Petén region reveals the significance of collection and storage and strongly suggests a program of centralized water management (Dunning et al. 2015; Scarborough and Gallopin 1991; Scarborough and Grazioso 2015). According to Scarborough and Gallopin (1991:659), this centralization, evidenced by the size, location, and density of reservoirs within the spatial core of 17

27 the city, also implies political and economic control by an elite, a previously unexamined urban perspective. This relationship between the control and management of water and the consolidation of power in the Maya area has been examined by multiple scholars (cf. Ford 1996; Lucero 1999, 2002, 2006; Scarborough 1991a, 1993ab, 1996, 1998; Scarborough and Gallopin 1991; Scarborough and Valdez 2003, 2009). As a critical and scarce resource during the lengthy dry season, water was politically manipulated by a Maya elite to centralize and express power during the Classic period. The essential need for water made the management of this resource in a delicate, water-stressed environment a powerful organizing force (Lucero 1999, 2006; Scarborough 1998). The creation and maintenance of water systems in the settlements and urban centers of the ancient Maya concentrated water in a quantity and quality that was unavailable naturally. By placing water and its management systems within the center of their elevated Classic period communities, the Maya allowed a controlling elite to monitor and direct the supply. Another key factor in the relationship between water control and expressions of power is the idea that the combined factors of seasonal drought and agricultural demands resulted in the seasonal nucleation around water sources and the seasonal dispersal of many farmers during the dry agricultural season a pattern particularly noticeable beginning in the Early Classic period (AD ) (Ford 1996; Lucero 1999, 2006). This necessary seasonal mobility made it even easier for a controlling elite to direct and manage the dispersal of water among the populace. In discussing the development of complexity and resource control among the Classic Maya, Ford (1996) proposes that the disproportionate investment in the public realm, represented by the monumental size of Tikal, demonstrates a significance difference in labor control by the local hierarchy when compared with surrounding areas of the lowland Maya region. Recent analyses of the site have identified a 18

28 centralized, multi-component water system consisting of large summit-ridge reservoirs, bajomargin reservoirs, and small residential tanks that would have operated in tandem with one another to provide Tikal with the volume of water necessary to sustain the highly-urbanized center throughout the year (Scarborough et al. 2012). As this system evolved over time, with advanced developments in the Late Classic (AD ) including a Temple cofferdam and sophisticated Palace Dam, planning and labor costs associated with the construction and maintenance of this system undoubtedly also became increasingly sophisticated and expensive (Scarborough and Grazioso 2015). A major clue in the question of how Tikal and other interior Petén centers consolidated their power so effectively may be found in the distribution of water resources throughout the region. In addition to the lengthy annual period of desiccation, surface water is variably distributed across the Petén. Perennial rivers drain the eastern and western margins of the lowlands and a string of permanent lakes stretch across the fault zone of the central Petén. The interior Petén surrounding Tikal, however, is devoid of surface water, containing no natural rivers or permanent streams. Drainage catchments are found in the closeddepression swamps that hold standing water in the wet season due to their deep, impermeable clays, but desiccate completely in the dry months to a useless cement-like hardpan (Ford 1996). Weather patterns affecting the Maya lowlands that make for a seasonal deficit of water at the height of the dry season also result in a serious drinking water shortage during this period. This annual water shortage clearly presented a major obstacle to ancient communities, especially in the interior Petén. Likely as a result of these conditions, this area of the lowlands was the last to be settled during the pioneering developmental period and the first to be abandoned during the Classic period collapse of Maya civilization (Ford 1996; Lucero 2002). Similarly, the increasingly pronounced effects of climate change, which would have significantly affected 19

29 weather patterns, agricultural productivity, and floral and faunal communities in addition to water resources were likely also determining factors in population aggregation and relocation during this period (Dunning et al. 2007, 2013a; Lucero 2006; Lucero et al. 2011). The dispersed nature of the elite hierarchy was an inherent weakness, and could have had a destabilizing effect on population integration. As a result, other mechanisms to effectively enforce controls on competing members of the elite had to exist. Consolidation of control in the interior Petén relates directly to the nature of critical resources, those that Ford (1996) describes as vital to subsistence yet discrete enough to be controlled by the elite. While the production and dissemination of agricultural products was also a focus of elite control for the Maya, as it was for complex societies in other parts of the world, the distribution of agrarian land in the Maya area is insufficiently concentrated to manage directly. The critical (and reliable) absence of drinking water during the dry season, however, provided an important mechanism for control (Ford 1996; Lucero 1999; Scarborough 1996). As a daily survival requirement for humans, water is vital to dry season subsistence and could be readily used as a control mechanism. Reservoirs, a visible component of the landscape of centers in the interior Petén, were also discrete and controllable particularly when located near, and incorporated into, the architecture of major and minor centers of the ancient Maya. As part of the social and spatial organization of centers, then, access to reservoirs would have been monitored and restricted without significant effort. These centralized water resources could have solidified ties among members of the elite, while also serving to integrate their constituent populations (Ford 1996; Lucero 1999, 2002, 2006). Lucero (2002) contends that the scale of water control correlates with the degree of political power, reflected in three levels of Maya civic-ceremonial centers regional, secondary, and minor. Regional centers such as Tikal, Calakmul, and Caracol were located in upland areas 20

30 with large pockets of dispersed fertile land difficult to monopolize and lacking permanent water sources. The need for an adequate water supply around these major centers is directly related to annual rainfall. Secondary centers such as Lamanai, Dos Pilas, and Xunantunich are typically found in upland areas with scattered pockets of agricultural land that supported local polities. Settlement density was typically high near centers and lower in hinterland areas. These centers arose as secondary polities in part because of their rulers participation in a royal interaction sphere established by regional leaders. The minor centers of Barton Ramie and Saturday Creek were located along the Belize River, in a landscape with extensive alluvium. These communities were established in lower elevation areas with higher annual rainfall than most regional centers (Ford 1996; Lucero 1999). Minor centers were also composed of relatively low-density dispersed farmsteads, a condition hardly conducive to aspiring leaders interested in monopolizing resources and acquiring surplus. Power derives from a complex interrelationship of center location, seasonal water supply, amount of agricultural land, and settlement density. Maya elites, in an effort to solidify and express power and control, monopolized artificial reservoirs and other water sources during annual droughts, providing the means to extract tribute from their subjects (Lucero 2002, 2006). 2.4 Preclassic Period Water Management Sizable communities anywhere in the world are necessarily located near sources of fresh water. In the lowlands, however, these sources were not widely available naturally; the Maya were forced to create them. The ancient Maya controlled water in at least three ways: by draining water from inundated lands, by conserving soil moisture, and by collecting and storing rainwater. Reservoir storage and canal diversion systems drew from small perennial drainages and seasonal 21

31 runoff catchments. To a large extent, the form and extent of water systems in the Maya lowlands define the communities in which they exist (Matheny 1978; Scarborough 1991a). Large-scale water systems first emerged in the Maya area during the Middle Preclassic period. The Maya were already landscaping their environment early in the first millennium, and by the beginning of the Middle Preclassic they had initiated runoff modifications (Gill 2000; Scarborough 1993). The earliest evidence for water manipulation in the Maya lowlands is reported from Albion Island, Belize (Bloom et al. 1985; Pohl 1990) and more recently from Cobweb Swamp, Belize (Jacob 1991). The two central elements of water management, channelizing or ditching and basin construction, appeared prior to monumental architecture, as might be expected if population growth is highly dependent upon the management of water resources (Matheny 1982; Scarborough 1993). By the Late Preclassic (400 BC AD 250), hydraulic technology focused on utilizing low-lying depressions or seasonal swamps (bajos) as water reservoirs (Dunning et al. 2002, 2006a, 2015). Scarborough has termed these features concave microwatersheds (Scarborough 1994). The earliest landscape engineering took place at the margins of these bajos, direct developments of the earliest villages associated with ponded water. As early as the Late Preclassic, the Maya had begun to alter the lower areas of gently sloping landscapes to catch and divert water into artificial channels and reservoirs. These modifications created concave microwatersheds that channeled and retained the natural runoff from surrounding higher elevations and held the collected water for extended periods through the dry seasons (Gill 2000; Scarborough 1994, 2003). Reservoirs in the Maya area were of several varieties. The large, bell-shaped cisterns (chultunes) of the dry northern Yucatán Peninsula caught rainfall from prepared platform surfaces (Dunning et al. 2013b; McAnany 1990; Scarborough 2003). These tanks were also 22

32 constructed with extremely narrow openings to limit evaporation. In contrast, open-surface reservoirs formed from natural karst sinkholes or as the result of quarrying for limestone, clay, or chert (Bullard 1960; Dunning et al. 2007; Hester and Shafer 1984; Weiss-Krejci and Sabbas 2002). In many cases, natural drainage gradients were also exploited to produce reservoirs and large depressions (aguadas) were frequently modified from the naturally low-lying terrain (Matheny 1978; Scarborough 2003). In addition, researchers have theorized that even the internally-draining bajos covering extensive tracts of the Maya area were altered to establish predictable water levels for supporting raised field agriculture (Beach et al. 2002; Dunning and Beach 2003; Dunning et al. 2006a, 2015; Harrison 1977). The sites of Cuello and Nakbe represent Preclassic examples of water storage in both chultunes and quarried depressions (Scarborough 1991a). By the conclusion of the Late Preclassic, Cerros, Edzná, and El Mirador present evidence of significant alterations to their immediate environment, with the excavation of canal and reservoir systems and the construction of causeways used to dam or divert water through depressed zones (Scarborough 1994) Preclassic Case Study: Cerros At the Late Preclassic community of Cerros, Belize, the volume of construction fill used in built structures compares favorably to that removed from the extensively quarried canal and reservoir depressions. Through the manipulation of sills horizontal blocks of stone extending into a tank or canal segment allowing access to the lower depths of the water-holding basin and dams, water was conserved in basin canals during the dry season, and the community s center was drained during the wet season (Scarborough 1983). Research suggests that approximately 200,000 m3 of quarried fill was removed from the site, with the systematic horizontal stripping of 23

33 the underlying limestone producing a manmade watershed (Scarborough 1983, 1991b). The core area of the site enclosed by the main canal extended over 37 hectares and contained 99% of the structural fill volume identified at the site, but registered only a 2-meter relief differential across its surface, excluding elevated structure height. Quarry fill was obtained predominantly from within the core area and removed systematically to produce the catchment surface and depression volume that defined the water management system (Scarborough 1991b). The main canal at Cerros was 1.2 km long, approximately 6 m wide, and 2 m deep, and graded less than 1 m from west to east (Scarborough 1993). Rainy season runoff was directed into tanks and basins throughout the community due to the system of reservoirs, feeder canals, sills, and dikes. The topography of the site indicates that high ground connected all residential and civic space, while also providing its inhabitants ready access to water. Another interesting feature of the site is a low relief area that features several raised fields separated from two reservoirs and numerous structure mounds by a paved causeway (sacbe). The sacbe was constructed primarily by quarrying the margins of the feature to create adjoining low-lying areas. In addition to connecting the residential area to the northwest with a ballcourt group to the southeast, the causeway also functioned as a dike. Acting as a dam, it separated the private and (presumably) potable water source of the reservoirs from the more public and agricultural waters source of the raised fields and main canal (Scarborough 1993, 1994, 2003). Cerros is a prime example of a Late Preclassic concave microwatershed, with its hydraulic system separated by an elevated road or dike into settlement-wide agricultural sources on one side and potable private household sources on the other. Canalization was particularly well developed in this slow-moving, internally draining hydrological system (Scarborough 1991ab). 24

34 2.4.2 Preclassic Case Study: Edzná At the much larger Late Preclassic site of Edzná in Campeche, the Maya utilized a system of linear canal reservoirs for potable water as well as for transportation (Matheny 1976). Matheny and colleagues (Matheny et al. 1983) have reported that over 1.75 million m3 of fill were removed in the construction of these monumental canal basins, a significant volume much greater than the volume of the Pyramid of the Sun at Teotihuacan. This water storage system was part of an extensive, planned urban area that covered nearly 17 km2. Much of the water system was constructed during the Late Preclassic, and there are several structure mounds built on top of the canal banks that also date to this period (Matheny 1982). These canals are massive constructions, with one of them involving the removal of nearly 900,000 m3 of water. The main canal at Edzná measures 12 km in length, with a depth of 1.5 m and a width of 50 m for at least half its length (Gill 2000; Scarborough 1993). The canal connects a large reservoir system at its northern end, which joins at least a dozen other canals converging on the urban core in a spokelike alignment associated with several residential tanks. The entire storage capacity of the canal basins was approximately 2,000,000 m3 (Scarborough 1993). Multiple researchers have cited Edzná as providing evidence of deliberate water control during the Preclassic (Gill 2000; Matheny 1976, 1978, 1982; Matheny et al. 1983; Scarborough 1993). Reservoirs in the northwest region of the site are located at higher elevations than the canals that connect them. As the water drained from these reservoirs to the south end of the canals near the site center, smaller feeder canals directed it to other reservoirs closer to public buildings and residences (Gill 2000). Interestingly, this is the reverse of the distribution process that would later be reported at Tikal (Scarborough and Grazioso 2015). In addition, there is evidence of silting tanks that were likely used to filter water collected in the northern canals. 25

35 Water was then directed into additional reservoirs from these retention and filtering tanks, presumably for consumption. (Matheny 1976, 1978; Scarborough 1993). The numerous canals and reservoirs at the site provided water in excess of agricultural requirements, and may also have been used for pot irrigation. Along with the hydraulic system, there were significant landscape alterations including banking, platform building, small dam construction, causeway building, and channel excavation (Matheny 1976, 1978, 1982). Edzná serves as an excellent example of a Preclassic community where the urban form and system of water management were closely integrated. The ancient Maya manipulated the upper Edzná Valley in the Late Preclassic period to meet their food production requirements and to provide water for a population that built hundreds of structures across an extensively urbanized area Preclassic Case Study: El Mirador Like at Cerros, sacbeob traversed low-lying bajo-like settings in proximity to the central precinct of Late Preclassic El Mirador (Dahlin 1984; Scarborough 1983, 1991ab, 1993a, 1994). Similar to other monumental Petén sites such as Calakmul and Tikal, El Mirador exhibited a dependency upon an extensive reservoir system. The lack of available water is the primary environmental factor limiting settlement in this region, and no perennial freshwater sources exist near El Mirador or the major site of Nakbe, located only 13 km to the southeast (Hansen et al. 2002). Water storage chultunes are fairly common to this area, however, and at least six have been identified at El Mirador (Matheny 1982). In addition, three large canal reservoirs within the urban core would have served to store and transport water from adjacent building complexes and, in one case, from a partitioned bajo. The site itself is located on the immediate margins of a sizable bajo that is traversed by four lengthy causeways radiating from the site center 26

36 (Scarborough 1993). Removal of quarried fill used in the construction of these causeways would have significantly modified the existing depression (Dahlin 1983). Given the quantity of clay and limestone used in the creation of structures at the site and the absence of large quarry scars (at least within the massive West Group), bajos are the likely source for much of the fill used in constructing El Mirador (Scarborough 1993). The sacbeob might also serve as accessways into the bajos, possibly representing the original, slightly elevated ground of the bajo surface. A desired gradient during the rainy season may also have been achieved by scraping the bed of the bajo floor. This process would not only provide construction fill for temples and courtyard groups, but portions of the bajo were likely also used for agricultural ends (Beach et al. 2002; Dunning and Beach 2003; Dunning et al. 1998, 2002, 2006a). Scarborough (1983, 1991b) has noted that the lateral scraping of caprock for construction fill has also been well documented at Cerros during this period. According to Matheny (1982), the reservoir system at El Mirador was likely carefully planned, with the barrows for structures and platforms later converted into water storage tanks. The success of these reservoirs in collecting runoff would likely have been dependent upon the strategic placement of building complexes that served as catchment areas. Investments in labor and time necessary to construct some sacbeob may also reflect attempts to impound runoff or divert water to location better suited for human use. These causeways then functioned as dikes or dams in later periods (Scarborough 1993). 2.5 Classic Period Water Management The underground (phreatic) water table was inaccessible to most of the lowland Maya, who were forced to adopt other strategies for water collection. The Classic period Maya relied 27

37 primarily on a combination of aguadas, chultunes, cenotes, and limited lakes and freshwater lagoons for their potable water sources (Gill 2000). Where the Late Preclassic Maya positioned their communities at the margins of bajos or similar large depressions, subsequent Classic period (AD ) cities adopted a land-use strategy requiring construction of their most expansive and densely concentrated civic centers at the summits of elevated hillocks or ridges (Scarborough 2003). The Classic built-urban environment was more deliberate than the concave microwatersheds of its predecessor and was characterized by sizable, but elevated, reservoirs. Excavated for the fill produced in constructing the monumental architecture of the major centers, these reservoirs were designed to receive large quantities of runoff flowing seasonally over the great plazas, pyramids, and palaces that populated the urban core. During the dry season, these convex microwatersheds permitted the directing of water downslope to residential consumers and bajo-margin agricultural plots (Dunning et al. 1999; Scarborough 2003; Scarborough et al. 1995). In contrast to Late Preclassic water management infrastructure, hydraulic technology of the Late Classic period allowed the Maya to construct settlements on higher ground, essentially using the city itself as the watershed for the reservoirs created from quarries and barrow pits. The major Classic period sites of Calakmul and Caracol each had artificial reservoirs constructed adjacent to monumental palaces and temples (Chase et al. 2011; Lucero 2002; Scarborough 1998). Calakmul is also surrounded by bajos and contains extensive canal systems in addition to at least 13 large reservoirs and aguadas (Folan et al. 1995). Caracol contains at least two major reservoirs next to pyramids and is surrounded by hillsides terraced for both agriculture and water control (Lucero 2002). At Late Classic lowland sites such as Tikal, La Milpa, and Kinal, runoff across the shaped surface of a convex microwatershed allowed for greatly improved control over the 28

38 resource than did the earlier lowland systems (Scarborough 1993, 2003; Scarborough and Gallopin 1991). The small hills on which these sites rest, removed from permanent streams or springs, were heavily paved with plastered surfaces aligned toward elevated reservoirs. The most elevated central precinct catchment at Tikal, for instance, covered an area of 62 hectares. Due to the impermeability of plaza pavements and plastered monumental architecture reducing seepage loss, this area alone may have collected more than 900,000 m3 of water (based on an average of 1,600 mm of annual rainfall) (Scarborough 2003). Combined runoff from this and bordering catchment areas would have easily filled the associated reservoirs and natural depressions, ultimately leading into the adjacent bajos Classic Case Study: Tikal By the beginning of the Late Classic, Tikal displayed one of the most sophisticated water control systems in the New World (Scarborough and Gallopin 1991). According to Scarborough (1993b), a heavy investment in upland water control during the drought-like conditions hypothesized for the Early Classic period permitted the environmental and cultural alterations of the Late Classic. Large, well-defined reservoirs were identified very early at the site, where the Maya constructed a complex watershed designed to capture rainfall from the thick and impermeable plaster surfaces covering the imposing architecture and spacious plazas. Building upon more than a millennium of landscape modifications, these Maya shaped the limestone hills to allow gravity to provide the core community with an adequate supply of water (Scarborough 1996; Scarborough and Gallopin 1991). The planning and placement required for the substantial reservoir constructions at Tikal demonstrates the significance of water collection and storage in a major urban area. The water system was designed to accommodate the seasonal availability of 29

39 rainfall, the absence of permanent streams or springs, and the gentle topographic relief of the region (Scarborough 2003; Scarborough and Grazioso 2015). Six major reservoir catchment areas or drainage divides, ranging in area from 9 62 hectares, drain the summit of this artificially modified watershed (Gill 2000; Lucero 1999; Scarborough 1993; Scarborough and Gallopin 1991). The directed runoff from the summit of the site core would have easily filled the associated reservoirs, and all sizable catchment areas ultimately terminated in bajo-margin reservoirs or natural aguadas, leading into the flanking bajos. In addition, three distinct reservoir types are documented within the catchment areas: (i) central precinct reservoirs, (ii) residential reservoirs, and (iii) bajo-margin reservoirs (Scarborough and Gallopin 1991). The six central precinct reservoirs are located near the summit of the city within the central core of the site. These tanks retained runoff from the largest completely paved catchment area, where more than 900,000 m3 of water could be collected annually (based on an average of 1,600 mm of annual rainfall) (Scarborough and Gallopin 1991). These central precinct reservoirs appear to have stored significant water reserves for the seasonal replenishment of the bajomargin reservoirs, and the controlled release of water from elevated tanks to downslope flanks and adjacent bajo margins would have provided potable water as well as moisture during the dry season (Scarborough 1993). All central precinct reservoirs were formed behind well-defined sacbeob, which connected various portions of the urban core while also damming water within the major catchment area. Residential reservoirs are located downslope from the central precinct within the most densely populated zone immediately outside the core s public architecture. These depressions are generally smaller than the other reservoir types identified, with many having a diameter of less 30

40 than 10 m and a depth less than 1 m, and tend to be located close to house mound groups (Scarborough 1996). These features, in addition to numerous pozas (small household tanks) also identified throughout the site, were likely utilized for domestic purposes. Neither the reservoirs nor smaller household tanks would have been replenished from the larger central precinct reservoirs during periods of drought. Reservoirs of this type are common components of nearly all settlements in the Maya area (Scarborough and Gallopin 1991). Near the foot of the ridge on which central Tikal sits, positioned approximately equidistant from each other and located roughly in the cardinal directions from the epicenter of the site, are the four bajo-margin reservoirs. These tanks are positioned adjacent to the expansive bajos bordering the site and include some of the largest reservoirs at Tikal. Placed away from the densest concentrations of house mounds in the central part of the site, at least two of these features are directly connected to the elevated central precinct reservoirs by drainage channels. These basins are positioned to receive most of the runoff issuing from the edges of the outcrop defining central Tikal and, given their size and placement, likely acted as holding tanks for the allocation of water to neighboring agricultural fields (Dunning et al. 2015; Scarborough 1996; Scarborough and Gallopin 1991). The Tikal system represents one of the most complex examples of water management in the Maya lowlands. Although one of the largest Classic period cities, Tikal is not the only southern lowland site with neither rivers nor springs located nearby. Due to the size, density, and location of reservoirs within the urban core of the city combined with this lack of permanent natural water sources the water management system is indicative of resource specialization (Ford 1996; Scarborough 1996). By utilizing a gravity-flow system, the Maya of Tikal were able 31

41 to provide significant residential populations within and adjacent to the central precinct with regularly replenished water stores during the dry season Classic Case Study: La Milpa The site of La Milpa in northwest Belize is a prime example of a convex microwatershed and illustrates the sophistication of a reservoir-based water system dating primarily to the Late Classic period (Scarborough 1993, 1998; Scarborough et al. 1995). Like Tikal, it has revealed a complicated water system at a site with no permanent water source for an extended portion of the year. The founders of La Milpa positioned the site on a natural hillock to take advantage of the quarried surface for construction fill and the resulting reservoir and rainfall catchment surfaces. The convex topographic relief allowed for greater control of runoff across this artificial microwatershed than was apparent at earlier Late Preclassic communities such as Cerros and Edzná, which were dependent on natural slope-margin runoff carried into a low-lying site center (Scarborough et al. 1995). The upper portions of local watersheds draining the site core at La Milpa were dammed to create reservoirs. The release of water from these reservoirs may have been controlled through the combined use of sluice gates and check dams (Scarborough et al. 1995). One function of this intricate water diversion strategy would likely have been to regulate soil moisture levels in pockets of flatter upland soils that appear to have been used as plots for intensive agriculture (Dunning et al. 1999). The central precinct of the site dominates the summit of a small hill, with three reservoirs positioned at the tops of three gently sloping arroyos providing natural drainage for the site. 32

42 Most runoff from the main plaza associated with monumental architecture at the site was directed into the northwest arroyo. Test excavations at the site revealed a 17.5 m long U-shaped dam at the plaza edge that would have aided in ponding and retaining water (Scarborough 1993). The second reservoir dam was a severely eroded earthen feature positioned to the south of the main plaza, where it would have received runoff from the southern third of the summit catchment. A series of sluices appear to have been intentionally cut through the embankment in an attempt to better control the passage of water. The third reservoir, located to the southeast, was relatively large and likely provided potable water to a complex of elite courtyard groups at its margins and further downslope. Each of these reservoirs drained into channels carrying water to the south and southwest areas of the site. Ridges occupied by courtyard groups and house mounds bordered the associated channels. Intermittent check dams placed on the ridge slopes controlled the velocity of the water as it flowed from the reservoir, limiting rainy-season erosion. After exiting the reservoirs and the inclined drainage channel, water would have fanned out over a series of flats presumed to be fields (Scarborough et al. 1995). A system of shallow earthen canals, at least one chultun, believed to have functioned as a cistern based on its location within the drainage margins of the field zone, and a modified aguada have also been uncovered at the site (Scarborough 1993). The aguada was likely altered to improve water retention and would have recharged during the rainy season through runoff from the south and southeast channelized arroyos. The system of water management at La Milpa, like that of Tikal, was extremely well developed. As demonstrated at other Classic period sites, the ancient Maya created a microwatershed at La Milpa to accommodate the lengthy seasonal drought. In addition to the water conservation measures associated with reservoirs, planned channelization, diversion weirs, 33

43 and the fields themselves, the importance of rainy-season erosion control is evident at this midsized Maya center (Scarborough et al. 1995). In addition, a variety of additional hydrological features located around the periphery of La Milpa have been identified and examined in recent years, demonstrating the necessity of water collection and storage to inhabitants outside of the urban core (Akpinar 2011; Brewer 2007; Chmilar 2005; Trachman 2007; Weiss-Krejci 2013; Weiss-Krejci and Sabbas 2002). Numerous small depressions modified to function as household reservoirs, a sizeable aguada with associated channel and berm features, and at least one chultun that may have been lined with a plaster sealant to aid in water retention compose this hinterland water management mosaic around La Milpa. Similar features have been documented in proximity to medium and large centers elsewhere in the Maya area (Akpinar 2011) Classic Case Study: Kinal The site of Kinal, located in the northeast Petén region, sits on a ridge dividing two massive bajos to the northeast and southwest. In contrast to the central precinct reservoir adaptation, in which water was collected from the summit catchment and held in large tanks for release downslope, the Kinal system was dependent solely on the summit catchment itself for the diversion of runoff (Scarborough 1993, 1996). The tanks at Kinal were positioned downslope from the central precinct, within the residential core and near the likely location of agricultural fields. The channel gradient that fed the site s reservoirs was steeper than those identified at Tikal and La Milpa, and required methods of slowing the movement of water into the reservoir. These included check dams, a one-piece diversion stone, and a pooling area, each designed to decelerate erosion (Scarborough 1993). A dam or weir approximately 8 m in length, serving to direct channel water from the central precinct into a very small silting tank before it entered the 34

44 main body of the reservoir, has been exposed at the Kinal West Reservoir. The silting tank would have acted to prevent large particulate matter from entering the potable water supply, where this reservoir was designed to systematically release water during the dry season. A V-shaped outlet was discovered within the reservoir, located at a depth in its embankment indicating that the entire volume of the tank could be drained (Scarborough 1993, 1996). Even though no sizable central precinct reservoirs have been identified at Kinal, water was likely captured from the numerous paved surfaces associated with the urban core of the site and directed into residential or bajo-margin reservoirs in a manner similar to that documented at Tikal and other Classic period centers. Although the Kinal reservoir system might be interpreted as representing a less centralized form of water management than the Tikal and La Milpa systems, the planning and labor required to control erosion and sedimentation was pronounced and well-defined at the site, reflecting the reservoir dependency of the Classic lowland Maya. The advancement of ancient Maya civilization in the lowlands was a consequence of multiple factors. One significant influence was the availability and enhancement of water collection and storage features. In addition to monumental construction, the Maya had another, more immediate, agenda in their landscape modifications; they engineered the natural terrain by systematically quarrying some locations and elevating others to create a carefully contoured cultural relief (Scarborough 1993). The lack of midden debris reported in the epicenter of Tikal is one indication of the importance of creating such water catchment surfaces (Harrison 1993). The dispersed household and public building pattern in Maya centers may reflect, in part, an attempt to maximize the quality of the water source, and the unintentional addition of household waste to the main water supply was clearly discouraged (Scarborough 1996). Maya watershed management processes were accretional, with the Late Preclassic concave landscapes evolving 35

45 into Classic period convex watersheds of urban proportions. Although these developments may have been partially triggered by climate change, increases in population combined with the fluctuating availability of resources, unstable political relationships, alliance formation, and exchange access were clearly key players impacting settlement growth and development (Lucero 1999, 2002; Lucero et al. 2011). Nevertheless, the seasonal water deficit represented a constant pressure upon an expanding state. The dispersed settlement pattern of the ancient Maya may have evolved as a consequence of the limited amount of permanent water available across this karstic landscape (Scarborough 1994). Operating in tandem with large-scale water management infrastructure within major centers, hinterland aguadas and seasonal reservoirs were focal points for household and small community development, but would have been inadequate to support sizable urban populations. 2.6 Recent Interdisciplinary Approaches to Ancient Maya Water Management Much of our knowledge of the infrastructure and practices that comprise ancient Maya water management derives from numerous survey and excavation projects undertaken throughout the lowlands that have analyzed a variety of hydrologic features including reservoirs, seasonally inundated wetlands (bajos), canals, dams, wells, natural sinkholes resulting from the collapse of limestone bedrock (cenotes), and subterranean bell-shaped cisterns (chultunes) in an archaeological context (e.g. Akpinar-Ferrand et al. 2012; Beach and Dunning 1997; Brewer 2007, forthcoming; Brewer et al. forthcoming; Bullard 1960; Chmilar 2005; Dunning et al. 1999; Harrison 1993; Isendahl 2011; Johnston 2004; Lohse and Findlay 2000; Matheny 1976, 1978, 1982; Scarborough 1991a, 1994, 1996, 2003; Scarborough and Gallopin 1991; Scarborough and Grazioso 2015; Scarborough et al. 1995; Weiss-Krejci 2013; Weiss-Krejci and Sabbas 2002). 36

46 Taken together, these studies provide detailed insights into the creation and maintenance of these features, as well as their use by the ancient Maya. As scholars continue to study the multifaceted components of Maya water management, it is becoming increasingly evident that these ancient systems were varied, complex, and unevenly distributed across the Maya world. In an ongoing effort to gain a clearer understanding of this mosaic, these complexities continue to be explored through a number of broader applications, including within the contexts of population estimates and archaeological variability (Becquelin and Michelet 1994; McAnany 1990), landscape modifications (Dunning and Beach 1994; Dunning et al. 1999), critical resource control (Ford 1996), paleoecology (Dunning and Beach 2010; Johnston et al. 2001; Wahl et al. 2007), political control and economy (Lucero 1999, 2002, 2006), social and environmental changes (Beach et al. 2015; Dunning et al. 2013a; Kunen 2004; Lucero et al. 2011), and satellite imagery and remote sensing analyses (Brewer et al. 2017; Chase 2016; Chase et al. 2011; Garrison 2010; Saturno et al. 2007; Weller 2006). At the northern lowland site of Sayil, McAnany (1990) quantitatively assesses ancient water storage facilities in particular, chultunes in an attempt to determine a population estimate for the site, as well as to understand the function of specific structures and rooms. Four types of structures (foundation braces, stone buildings, platforms, and basal platforms) are analyzed, with each determined to relate independently to chultun presence and frequency. Based on her formula combining water catchment, chultun capacity, human water requirements, and structural unit type and size termed a prime domestic ratio she concludes that these cisterns are associated with the larger, more permanently occupied structures at the site. In addition, the data suggest that while these large stone structures play a complex multifunctional role, the high frequency of small, single platforms lacking chultunes suggests seasonal habitations or 37

47 functionally specific structures. The positive relationship between basal platform area and chultun frequency is indicative of the important role that cisterns play in the finished size of the site s residential units. Similarly, the work of Becquelin and Michelet (1994) at the site of Xculoc reinforces the strong relationship between water and settlement in the Puuc region. Despite the presence of fertile soils, a lack of available water posed a significant disadvantage to dense settlement in the area. Although a pronounced dry season is present across most the Maya lowlands, localized conditions of rainfall, hydrology, and geomorphology collectively result in an acute seasonal water deficit in the Puuc Hills. At sites such as Xculoc, the real challenge was to make it through the dry season without exhausting all water reserves. The demographic implication of this deterministic relationship between available dry season water resources and the population of these sites is that the number of water features should directly relate to the size of the dry season population (Becquelin and Michelet 1994). With an understanding of the relationship between the capacity of available water features and human water requirements, related settlement and archaeological questions can be addressed, such as the variability in architectural units vis-à-vis chultun or reservoir frequency. On the eastern periphery of the central Maya lowlands, recent work has focused on the course of human-environment interactions, including both intentional and unintentional environmental changes (Dunning and Beach 1994; Dunning et al. 1999). Evolving landscapes at the major centers of La Milpa and Dos Hombres and surrounding lands reflect the cumulative processes of human action and environmental change and their impact on the hydrology, soils, and architecture of the region. By the Classic period, environmental degradation from deforestation and agricultural activities throughout the Middle Late Preclassic and Early Classic periods and the increasingly impactful effects of climate change placed even greater 38

48 stresses on water features, soils, and settlement (Dunning et al. 1999). The area s farmers faced other problems, including devastating rainy-season runoff and flooding and dry season desiccation and clay contraction. These problems were undoubtedly exacerbated by the combined effects of deforestation and the creation of a sunbaked, parched bajo landscape (Rice 1993). At Dos Hombres, a growing scarcity of water is suggested by evidence of an unusually dense concentration of late settlement along the margins of a large aguada, and the long-term abandonment of this and other nearby sites attest to the probable severity of environmental degradation, lack of available water, and depopulation that occurred in the Terminal Classic (ca. AD ) (Dunning et al. 1999). Dunning and Beach (1994) emphasize that increasing population density and agricultural terracing in the southern Maya lowlands appear to be part of a pattern of watershed modification and localized population concentration that was an element of Classic period urbanization and state development. While the spread and adaptation of terracing was likely linked to increasing populations during this time, like the management of sustainable water resources, it was spatially conditioned by environmental variability. As the authors suggest, the additional study of small watersheds where sedimentation and other paleoecological investigations can be strongly linked with the careful study of relict slope management and associated settlement features is necessary (Dunning and Beach 1994). In her examination of the resource base of the Classic period Maya, Ford (1996) suggests four major variables contributing to the development of complexity: overall resource productivity, local resource control, resource diversity, and critical resource control. Defining critical resources as those (such as water) relating directly to subsistence, she contends that reservoirs are associated with elite control and investment in public works construction and maintenance through the organization and control of labor. In this scenario, the governing 39

49 hierarchy monopolized by the elite has a vested interest in resources and labor, from which its power is derived. As such, consolidation of control visible in Petén region sites such as Tikal is visible directly related to these critical resources. The economic landscape of the central lowland Maya is a mosaic of dispersed subsistence resources, with the single most important resource in the region primary well-drained uplands occurring in small and large patches comprising 15-50% of the overall area of the region (Fedick and Ford 1990). Illustrating the relationship between occupation of this ecological niche, labor control, and expressions of authority, Tikal and other interior Petén centers were able to consolidate control so effectively due to the distribution of water in the region. Although the dispersed nature of the elite hierarchy was inherently weak and could have had a destabilizing effect on population and integration, the critical absence of drinking water during the dry season provided an important mechanism for control. Reservoirs are discrete and controllable, particularly when located and incorporated into the architecture of the major and minor centers of the ancient Maya, and access could have been monitored and restricted. As a risk management strategy, investment in public works is a key to controlling a populace. Ford (1996) argues that such public works can include construction of facilities to provide water for irrigation, drinking, or other important activities critical to subsistence, and that reservoirs are a dependable and widespread example of one such critical capital investment. Examination of the paleoecological record and the management of water resources is another recent avenue of investigation in the Maya area, with scholars noting fundamental relationships between climatic fluctuations, agricultural activity, settlement, and water availability (Dunning and Beach 2010; Johnston et al. 2001; Wahl et al. 2007). In their multidisciplinary study of Laguna Las Pozas in Guatemala, Johnston and colleagues combine 40

50 magnetic, palynological, and paleoecological data from the Río de la Pasión basin to conclude that some refugee populations migrated to previously unoccupied, geographically marginal nondegraded landscapes within the southern lowlands following the Classic period collapse (Johnston et al. 2001). One of the likely contributors to this episode of depopulation across the Maya area was a significant decrease in water resources, and the large Laguna Las Pozas would have represented a significant, and relatively stable, water resource for Early Postclassic colonization and settlement in the southwestern lowlands. Magnetic, organic carbon, geochemical, and lithological analyses from the site indicate that the basin was colonized and deforested, occupied by agriculturalists, and abandoned and reforested all during the Early Postclassic period (Johnston et al. 2001). As a result, the post-collapse population of the southern lowlands may have been larger and more territorially widespread than archaeologists have previously suspected. Palynological studies from Aguada Zacatal, a Late Classic reservoir located within a bajo near the site of Nakbe, are similarly used to isolate periods of agricultural activity, abandonment, and forest succession in the northern Petén (Wahl et al. 2007). The aguada s pollen record documents a period of agricultural activity during the Late Classic (AD ), where corn pollen is found in conjunction with disturbance indicators, before an era of abandonment is revealed around AD 840. The record from this period is dominated by aquatic pollen types and the absence of corn pollen, a profile that the researchers believe represents the end of the Classic period. The northern Petén of Guatemala is one of the more remote areas in the Maya lowlands and its intense dry season, lack of perennial water sources, and extensive seasonal wetlands form a considerable barrier to settlement. However, evidence of large populations from densely packed urban centers in the Middle and Late Preclassic periods, and modestly occupied Late 41

51 Classic communities, has led researchers to examine the possible role of environmental change in shaping regional cultural events. Whether Aguada Zacatal was used for irrigating dry season crops or if the collected water was simply used for domestic purposes, the reservoir was and continue to be today an important source of dry season water and focal point for cultural activity in the area. Early scholarly thinking on Maya civilization viewed the environment as static and culturally limiting, though recent scholarship has begun to recognize the complexity of the lowlands environment and the varied ways in which the Maya transformed it. In their review of ancient Maya landscapes that existed in the Preclassic and Classic Maya lowlands, Dunning and Beach (2010) highlight the variability produced by spatial differences in the underlying environment, several environmental changes, and Maya adaptations over time. They contend that spatial variation in adaptive strategies used by the Maya under different environmental conditions is expressed in whether Maya peoples chose to focus agriculture in wetlands or dry uplands, how they dealt with the annual problem of dry season water availability, and how diverse and intrinsically fertile soils were across the broader landscape. As evidenced from a study of Preclassic through Postclassic sites including La Milpa, San Bartolo, Caracol, Tikal, and Uxmal instability of regional landscape dynamics was introduced by both natural and human factors. The authors note that population growth over time often necessitated greater amounts of sustainable water sources and agricultural land, thus depleting forest resources and disturbing the soil cover. In addition, social or cultural factors associated with dynastic expressions of power, monumental constructions, and labor control could also have introduced instability and exacerbated ongoing land and water availability challenges. 42

52 Lucero (1999, 2002, 2006) takes a similar approach to link problems of water management to wider social and cultural factors, examining the role of water control in Maya politics, ritual, and the Classic period collapse. After first reviewing the interrelationship between water systems and Maya rulership at Tikal and other comparable centers (larger centers in areas without permanent water sources such as lakes or rivers) (Lucero 1999), she shifts her focus to the role of water control in the emergence and demise of Classic Maya political power. She suggests that the scale of water control correlates with the degree of political power, reflected in three levels of Maya civic-ceremonial centers regional, secondary, and minor (Lucero 2002). Recognizing that such power derives from a complex relationship among center location, seasonal water supply, amount of agricultural land, and settlement density, she argues that Maya kings monopolized artificial reservoirs and other water sources during annual drought, providing the means to extract tribute from their subjects. As water control and political power were inherently linked within this cycle, climate change in the form of increased, and more severe, periods of drought undermined political control at regional centers such as Tikal, Caracol, Copan, and Palenque. The collapse of rulers power at these regional centers in the Terminal Classic had differing impacts on smaller communities, and many secondary and minor centers not heavily dependent on water control survived the drought and collapse of regional centers. In the southern Maya lowlands, annual water shortages dramatically affected the livelihood of the Maya like no other natural resource shortage. Even in areas where water was plentiful, seasonal water issues impacted settlement decisions and agricultural practices. Lucero (2006) notes that the Maya met this challenge of an annual 4- to 5-month drought by building water catchment systems to provide water that would last until the rains began again. Although 43

53 elites organized the construction and maintenance of water retention features, many farmers were not necessarily reliant upon a particular ruler because they could choose to build their own smallscale water systems or contribute to the coffers of other nearby rulers. Issues of maintenance, drought, flooding, and agricultural impacts all served to link water with notions of power, control, and ritual. As a result, a major means that rulers used to bring together Maya farmers was traditional rituals writ large, including dedication, ancestor veneration and termination rites, prayers for abundant water, and other household and community rites (Lucero 2006). Among these commoner, elite, and royal ritual interactions across the southern lowlands from the Late Preclassic through the Terminal Classic periods, the more people kings integrated, the more powerful they became particularly in areas with plentiful agricultural land and noticeable seasonal vagaries. The links between water and changing social and environmental conditions continue to be explored by numerous scholars. In her look at social and environmental change in the wetlands of Belize, Kunen (2004) explores ancient Maya life in the Far West Bajo through a combination of mapping, survey, and excavation data from multiple agricultural zones and residential groups in the vicinity of La Milpa. In her quest to understand how this landscape was organized and utilized by residents of the bajo communities, Kunen applies a pair of models to explain the spatial and social structure of the ancient agricultural landscape. A variation on the garden infield-outfield model of agriculture and a model of preferential access to critical natural resources by community founders are used to illuminate the spatial organization of the Far West Bajo area during the Late Classic, after environmental degradation prompted residents to conserve depleted soil and water resources through an intensive program of terracing and berm construction (Kunen 2004). 44

54 Lucero and colleagues examine the linkages between people and water managers responding to long-term climate change, asserting that the basis for royal power rested in what rulers provided their subjects materially such as water during annual drought via massive artificial reservoirs, and spiritually through public ceremonies, games, festivals, feasts, and other integrative activities. In the face of kings losing their power due to drought, people left. Without their labor, support, and services, the foundation of royal power crumbled; it was too inflexible and little suited to adapting to change (Lucero et al. 2011). This evolution was particularly visible in the southern Maya lowlands as a semitropical region with noticeable wet and dry seasons where the critical importance of water is undeniable. Their work details not only how Classic Maya society dealt with the annual seasonal extremes, but also how kings and farmers responded differently in the face of a series of droughts in the Terminal Classic. Similarly, these social-environmental dynamics are further examined in the Maya lowlands through the contribution of environmental variability to cultural interruption and two distinct periods of widespread depopulation in the Preclassic and Classic periods (Dunning et al. 2013a). The concept of the Mayacene, a segment of the Early Anthropocene that occurred from years ago, is explored through Late Quaternary paleoenvironmental records that synthesize the evidence for Maya impacts on climate, vegetation, hydrology, and the lithosphere (Beach et al. 2015). This interdisciplinary team combines data from studies of soils, lakes, floodplains, wetlands, and other ecosystems to highlight the alterations Maya civilization made to local and regional ecosystems and hydrology by the Preclassic period. Their findings indicate that the Maya altered ecosystems with vast urban and rural infrastructure that included thousands of reservoirs, wetland fields and canals, terraces, field ridges, and temples. Notably, existing lowland forests are still influenced by ancient Maya forest gardening, particularly by the large 45

55 expanses of ancient stone structures, terraces, and wetland fields that form their substrates. A much larger body of research documents the Maya impacts on hydrology, in the form of dams, reservoirs, canals, eroded soils, and urban design for runoff, reflecting the Mayacene identity as a patchwork of cities, villages, roads, urban heat islands, intensive and extensive farmsteads, forests, and orchards. As the authors point out, although forests and wetlands cover much of the Maya area today, like so many places these are now under the onslaught of the deforestation, draining, and plowing of the current Anthropocene (Beach et al. 2015). The application of satellite imagery and remote sensing analyses has proven to be a revolutionary approach to the study of ancient Maya water resources. At the major center of Caracol in Belize, advances in remote sensing and space-based imaging have led to an increased understanding of past settlements and landscape use. The application of airborne lidar remote sensing, in particular, provides a detailed raster image that mimics a 3-D view of a 200 km 2 area of the site, penetrating the dense tropical canopy to provide visual data on the topography of the landscape as well as structures, causeways, and agricultural terraces (Chase et al. 2011). This work has demonstrated the utility of swath mapping lidar as a powerful time and cost effective tool to analyze past settlement and landscape modifications in the Maya area. As a method for identifying and mapping potential water features both large- and small-scale lidar data from both Caracol and the site of Yaxnohcah in the southern Yucatán has been used to conceptualize storage capacities, flow patterns, and spatial relationships with residential groups, as well as to target for archaeological excavation (Brewer et al. 2017; Chase 2016). In his study of ancient Maya rural populations, Garrison (2010) tests a methodology for estimating settlement size using QuickBird satellite imagery. Seeking a cost-effective solution for expanding the results of ground survey in jungle environments, the author tests a 25 km 2 46

56 research area near the site of San Bartolo in the lowland jungle of Guatemala. This localized area is composed of a mosaic of vegetation classes and geomorphological catenas (slopes soil profiles), and his findings indicate that the ancient Maya exploited the ecological niches present in the landscape but chose to construct their residences predominantly on well-drained uplands. Analysis of the imagery allowed for the accurate isolation of upland terrain and vegetation from the rest of the landscape, presented similar overall results to archaeological surveys conducted elsewhere in the Maya area (at a much greater cost), and allowed for comparable population estimates (Garrison 2010). Weller (2006) had previously investigated similar issues of survey and settlement in the region, focusing specifically on Late Classic Maya utilization of bajos at the sites of Tikal and Yaxha in Guatemala. These seasonal swamps comprise 40% of the ancient Maya landscape of lowland Guatemala and provide the template for her study of landscape modifications in and around bajos. Applying a combination of Landsat, IKONOS, AIRSAR, and SRTM imagery data, spatial analysis, and ground survey, Weller seeks to answer questions relating to settlement, subsistence, and population. Expanding her investigation to include the detection and modifications of canals, agricultural fields, and aguadas, the author finds that these satellite technologies allow for a more cohesive view of survey and settlement of the landscape by including the often-overlooked bajo environments. In their remote sensing investigation of the Petén region of northern Guatemala, Saturno and colleagues use remote sensing technologies combined with satellite imagery, regional survey, and archaeological excavation to locate and map ancient Maya sites that are currently threatened by accelerating deforestation and looting (Saturno et al. 2007). Similar to Weller s (2006) nearby work, this landscape archaeology approach results in the identification of a spectral vegetation signature that the researchers can 47

57 correlate with locations, boundaries, and dimensions of ancient Maya sites. This powerful tool is not only useful for a number of survey applications, including classifying specific vegetation types with various hydrological, topographical, and soil types, but also for defending against current deforestation and site destruction activities. At the Classic period center of Caracol in southwest Belize, Chase (2016) created a 200 km 2 Digital Elevation Model (DEM) from lidar data to explore the nature and extent of ancient rainwater capture particularly through the employment of residential reservoirs at the site. His analysis of the lidar data revealed nearly 1600 small depressions that fit the physical profile of household water storage features. Based on this ability of households to harness available water in their immediate environment, he questions earlier scholarly suggestions (cf. Lucero 1999, 2002, 2006; Scarborough 1998, 2003) that the control and management of water contained in large centralized reservoirs was a key manifestation of ancient Maya political and ritual authority. In his view, the ubiquity and decentralized nature of residential reservoirs at Caracol is suggestive of elite authority based on something other than the control of water resources an argument proposed earlier by water management researchers in northwestern Belize (Weiss- Krejci and Sabbas 2002). Although the likelihood of all 1590 depressions identified in the lidar functioning as reservoirs is undoubtedly an optimistic overestimation, Chase s research successfully demonstrates the increasing importance of lidar data for archaeological landscape analysis and provides additional insight into the nature of elite power among the Classic Maya. The study of water management in the Maya region constitutes an integral part of our understanding of ancient, and in some cases contemporary, Maya civilization. As a result of their unique environmental compositions, both the northern and southern Maya lowlands necessitated a variety of landscape modifications to support sustainable settlement. Chief among these was 48

58 the alteration of their physical environment to collect, store, and release water for a variety of needs. Although the management of water resources forms but one of the areas of ancient human-environment interaction among the Maya, understanding the necessity, nature, and scale of these activities has provided scholars with a broad range of insights into Maya civilization. Taken together, these approaches ranging from population and settlement and political control and economy to landscape modifications, environmental changes, and satellite imagery and remote sensing analyses continue to inform ancient Maya water management studies across a variety of spatial and temporal scales. 49

59 3. Research Article 1 EMPLOYING AIRBORNE LIDAR AND ARCHAEOLOGICAL TESTING TO DETERMINE THE ROLE OF SMALL DEPRESSIONS IN WATER MANAGEMENT AT THE ANCIENT MAYA SITE OF YAXNOHCAH, CAMPECHE, MEXICO Jeffrey L Brewer a, Christopher Carr a, Nicholas P. Dunninga, Debra S. Walkerb, Armando Anaya Hernándezc, Meaghan Peuramaki-Brownd, and Kathryn Reese-Taylore Submitted to Journal of Archaeological Science: Reports (October 29, 2016) Accepted: March 20, 2017 DO NOT CITE IN ANY CONTEXT WITHOUT PERMISSION OF THE AUTHOR(S) a Department of Geography, University of Cincinnati, Cincinnati, OH, USA Florida Museum of Natural History, Gainesville, FL, USA c Centro de Investigaciones Históricas y Sociales, Universidad Autonóma de Campeche, Campeche, Mexico d Centre for Social Sciences, Athabasca University, Athabasca, Alberta, Canada e Department of Anthropology and Archaeology, University of Calgary, Calgary, Alberta, Canada b 50

60 Abstract High-resolution airborne lidar has been employed in the Maya lowlands to examine landscape modifications, detect architectural features, and expedite and expand upon traditional settlement surveys. Another potentially beneficial and to-date underutilized application of lidar is in the analysis of water management features such as small reservoirs and household storage tanks. The urban center of Yaxnohcah, located within the Central Karstic Uplands of the Yucatan Peninsula, provides an ideal test case for studying how the residents of this important Maya community managed their seasonally scarce water resources at the household scale. We employ an integrative approach combining lidar-based GIS analysis of 24 km2 of the site area, ground verification, and excavation data from five small depressions to determine their function and the role they may have played in water management activities. Our research shows that some, but not all, small depressions proximate to residential structures functioned as either natural or human-made storage tanks and were likely an adaptive component of expanding Middle Preclassic to Classic period urbanization at the site. Thus, while lidar has revolutionized the identification of topographical features and hydrologic patterns in the landscape, a combination of ground verification and archaeological testing remains necessary to confirm and evaluate these features as potential water reservoirs. Keywords Ancient Maya; Lidar; Archaeology; Water Management; Hydrologic Analysis 51

61 1. Introduction In this paper we discuss the application of airborne lidar (Light Detection and Ranging) for the detection and hydrologic analysis of closed depressions supported by the subsequent ground verification and archaeological testing of five of these features at the ancient Maya site of Yaxnohcah in southern Campeche, Mexico. Our goal in studying these depressions was to develop a scheme for identifying closed depressions in the lidar imagery and to determine whether or not they functioned as water storage features. If they did serve in a water management capacity, we also sought to understand the nature of these functions based on data recovered through archaeological excavation. Lidar is a remote sensing technique that uses pulsed returns from an aircraft-mounted laser to measure distance to ground surface and can detect features either indistinguishable or, in some cases, inaccessible through traditional field survey methods. Since its initial application in the Maya lowlands as part of the Caracol Archaeological Project (Chase et al. 2011), lidar has proven to be an extremely useful component of interdisciplinary studies seeking to understand the complex interrelationship between settlement, urbanism, and ecological adaptations to local environments. Despite being employed in multiple contexts, including land-use and land cover mapping (McCoy et al. 2011; Parent et al. 2015), hydrological (Pe eri and Philpot 2007; Brzank et al. 2008; Turner et al. 2014), and archaeological applications (Devereux et al. 2008; Gallagher and Josephs 2008; Corns and Shaw 2009; Evans et al. 2013; Johnson and Ouimet 2014), not until its inclusion in the Caracol project did the benefits of lidar to studies of the ancient Maya landscape become realized. In particular, the ability of this technology to penetrate the dense tropical canopy and record the contour of the ground has proven extremely useful to Maya 52

62 archaeologists seeking to locate, identify, record, and investigate a variety of natural and cultural features. In addition to the ongoing work at Caracol (Chase et al. 2014), recent projects in the Maya area ranging from mapping the site of Mayapán (Hare et al. 2014), to measuring the effects of ancient Maya land use on contemporary forest canopy (Hightower et al. 2014), to assisting landscape archaeology studies (Hutson 2015; Prufer et al. 2015) have utilized lidar to examine various facets of ancient Maya settlement and ecology. Despite these advances, an underutilized aspect of lidar application in the Maya area is in the study of water management, specifically the analysis of how a society captures, controls, stores, and distributes water resources. Landscape archaeology studies, including those analyzing various aspects of water management, have traditionally relied upon topographic maps derived from field survey and mapping to study patterns and processes of water management, a labor intensive process that can take multiple years to complete (c.f. Carr and Hazard 1961; Carr et al. 2015; Scarborough and Gallopin 1991). A major advantage of employing lidar in such studies is its ability to supply the topographic detail necessary for greatly expedited multi-scale hydraulic analysis (c.f. Chase et al. 2010, 2011, 2014). A more holistic understanding of the physical, temporal, and spatial characteristics of these hydraulic features is necessary in order to understand their precise roles in the daily lives and activities of community inhabitants. The volume of spatial data provided by lidar is also extremely useful in compiling a catalogue of both large and small depressions that can then be examined on the ground (ground-truthed) and analyzed in terms of their water management capabilities. An important consideration in this study is that, as opposed to simply considering all small depressions as household-scale water reservoirs, we attempt to determine their most likely 53

63 function based on the physical and cultural data provided by archaeological excavation. Small depressions are ubiquitous throughout the Maya lowlands and have been shown to serve a variety of cultural functions in addition to water storage, including limestone quarries and clay or sascab (calcareous marl) mines (Folan 1982; Tourtellot and Rose 1993; Weiss-Krejci and Sabbas 2002); areas of specialized agriculture, horticulture, and apiculture (Folan 1983; Gomez-Pompa et al. 1990; Kepecs and Boucher 1996); and refuse dumps (Weiss-Krejci and Sabbas 2002). Depressions can also originate naturally as sinkholes (dolines), which are a common phenomenon in karst systems and often occur close together in high densities (Akpinar-Ferrand et al. 2012; Jennings 1985; Hughbanks 1995; Lene 1997; Siemens 1978). Following lidar analysis and ground-truthing of the small depressions, our primary focus was on determining the water management capabilities of these residential scale features. 2. Background: Yaxnohcah Yaxnohcah lies within the Elevated Interior Region of the Maya lowlands (Figure 1), a karst physiographic province where natural perennial surface and accessible groundwater sources are negligible, making capture and collection of rainfall essential for year-round ancient human settlement (Dunning et al. 2012). Hence, the Maya took advantage of natural depressions, or created them as needed. The permeability of such depressions varied considerably, with natural clay sediments occurring in some, but with many floored with highly permeable limestone making sealing necessary for water storage. Yaxnohcah is located at the southern end of the Calakmul Biosphere Reserve, situated beneath high tropical forests that form one of the highest and least disturbed tracts of continuous rainforest canopy in Mesoamerica. Karl Ruppert and John Dennison, Jr. initially visited the site 54

64 in 1933 as part of the Carnegie Institute s Second Campeche Expedition, documenting portions of the site that they visited briefly. Their report described three major platforms, multiple courtyards and mounds, three chultuns, and at least one aguada, identified as Aguada Monterey (Ruppert and Dennison 1943). Monterey served as the initial name for the site itself until 2004, when Iván Šprajc and colleagues revisited and investigated the site in detail, compiling fifteen maps in addition to assigning it the name Yaxnohcah meaning first big city (Šprajc 2008). Six large civic-ceremonial architectural complexes (designated Groups A F) and two expansive elite residential groups were identified in this survey. The primary civic-ceremonial site core, containing the Alba, Brisa, and Carmela architectural complexes, is situated on a ridge overlooking the Bajo El Tomatal, while the Dolores, Eva, and Esma groups are located between 1 km east and southeast, partially surrounded by the bajo, a large seasonal swamp. The Fidelia group is located roughly 2 km northeast of the site core and is surrounded on three sides by the sprawling Bajo El Laberinto that borders the north and northeast sides of the site and continues northwest to the site of Calakmul (Reese-Taylor and Anaya Hernández 2013). The primary structure within every major civic-ceremonial complex measures at least 50 m x 50 m in area and a minimum of 20 m in height. Stucture Alba-1 in the Alba group represents the tallest construction at Yaxnohcah; the superstructures sit on a base that is 85 m x 75 m in area and the tallest reaches 38 m in height (Reese-Taylor and Anaya Hernández 2013, 2014). Review of the lidar imagery and associated ground verification by the Yaxnohcah Archaeological Project (YAP) (Reese-Taylor et al. 2016a, 2016b) have also shown public buildings to be widespread across the entire 24 km2 lidar area. In addition, the project has identified and mapped three additional monumental architectural groups: Grazia, Helena, and Irma (Anaya-Hernández et al. 2016; Flores Esquivel 2014; Reese-Taylor 2014). 55

65 Based on ceramic data initially obtained by Šprajc (2008), the height of occupation at Yaxnohcah was estimated to have occurred between 600 B.C A.D. 400 and with a reflorescence from A.D Dates obtained from ongoing ceramic analyses by YAP support these findings, further indicating that the site was continuously occupied from the early Middle Preclassic through to the end of the Classic period, followed by at least some Postclassic activity, possibly associated with regional pilgrimmage (Walker 2016). However, the number of triadic-style main structures typical traits of the Middle and Late Preclassic (950 B.C. A.D. 150) suggest that most of the monumental buildings date to this period. Based on the quantity of civic-ceremonial architecture and the extent of the area over which it is spread, Yaxnohcah appears to be one of the larger Preclassic centers in the Maya lowlands. 3. Data and methods 3.1 Lidar data acquisition and post-processing Closed depressions (potential reservoirs), were identified by analyzing the ground elevations derived from the lidar overflight of the site. The list of closed depressions was the starting point for our field planning. Closed depressions are low spots in the landscape where water can pond. Features that had the appearance of a blown out dam (and which could be reclosed with the placement of a few stones) were also included on the potential list. The term closed refers to the elevation contour line forming a closed loop. The lidar data collection was managed and conducted by the National Center for Airborne Laser Mapping (NCALM). Full details of the data collection and processing for this work and the broader Yaxnohcah Archaeological Project are discussed elsewhere (Reese-Taylor et al. 2016a, 2016b). The lidar data was collected with an Optech Gemini terrain mapping system set to a pulse repetition frequency of 125 khz. The nominal shot density was 15 shots/m2. The 56

66 classify ground function of Terrasolid s TerraScan software was used to identify the ground points nominally 1.5 returns/m2. The ground returns were converted to a 0.5 m pixel DEM by Kriging. As is typically done in the Maya area, the ground points category includes both the ground surface and the ruins of ancient Maya structures. Our search for closed depressions began with this DEM. The DEM was used to make a hillshade image and a elevation contour map using the GIS software ArcMap version 10 (ESRI) (Figure 2). The hillshade and contour maps were visually inspected for closed depressions aided by computer-highlighting of contour lines between 35 and 200 m in length. A 3-D profile graph for each closed depression was also created to confirm that the contour lines represented a depression as opposed to a peak. The potential water holding capacity of each depression was calculated from the DEM derived area and depth. For closed depressions where a small dam would greatly increase the water holding capability, the capacity was calculated as if the eroded dam were replaced. The volume of the closed depressions ranged from 1 m3 to almost 4,000 m3 some clearly too small to viably serve as water storage features. It should be noted that these volumes are based on the current ground elevations and that the depressions have certainly filled in from more than a thousand years of soil erosion, sedimentation, and soil development. For example, the top of D-1 (Figure 3) is at m elevation, with a current low point of m, resulting in a depth of 0.8 m. Although the top contour line does not close the end of the depression, we hypothesize that the ancient Maya may have placed a small dam across the opening on the northeast side. The locations of the candidate aguadas (the closed depressions) and the lidar-derived hillshade map were transferred to a Garmin Map64 GPS for navigation in the field. 3.2 Ground verification plan 57

67 Our intent was to examine and evaluate closed depressions as potential reservoirs in areas where the project had, or was in the process of evaluating, data from residential structures (Brewer 2016; Peuramaki-Brown et al. 2016). We specifically wanted to examine the link between residential structures and adjacent water storage tanks. To that end, prior to entering the field we inspected the lidar area for closed depressions in four of the 500 m x 500 m project grid squares (squares 3E, 3F, 4E, and 4F; Figure 2). Within that area, we identified 39 closed depressions with a potential water holding capacity of greater than 20 m3. This cutoff volume was somewhat arbitrary and takes into consideration post-use sedimentation. The number of closed depressions greater than 20 m3 in volume in each of the four grid squares was 11 (3E), 14 (3F), 5 (4E), and 9 (4F). Many closed depressions have smaller capacities than this cutoff, but these were of secondary interest due to our assumption that they would have had limited use as household tanks. In a few cases, closed depressions representing potential water storage features were discovered where no depression was identified in the DEM. These were always small depressions (shallower than the minimal detection limit of this method; less than approximately 30 cm deep) and/or appear to be located in sections of the lidar where low, dense vegetation obscured the ground surface (Fernandez-Diaz et al. 2014; Reese-Taylor et al. 2016a) 3.3 Archaeological excavation plan Although the pre-survey lidar analysis of the area between the Alba and Fidelia groups returned a total of 39 closed depressions with volume greater than 20 m3, the practical and logistical limitations of a five week field season necessitated the selection of a representative sample to test archaeologically. We sought to examine a sample of these features in terms of 58

68 their physical dimensions as well as their spatial and (hypothesized) temporal characteristics. We used a GPS to guide us to the location of several (N=11) closed depressions for a visual inspection. From this reconnaissance we selected four areas for excavation depressions D-1 to D-3 and D-5. Depression D-4 was located outside of the four pre-surveyed lidar blocks but was selected for investigation based on ground survey in an area on the north side of the Fidelia complex where previous test excavations of a potential agricultural field had taken place (Figure 2 and Table 1). Although no residential data has been recovered from this area, the presence of an L-shaped mound adjacent to the northern edge of the depression supported our hypothesis that the feature may have served as a household reservoir. While tested depressions were of varying dimensions, each was considered a residential scale feature and located within 80 m of a house mound or residential group in the peri-urban zone between Complexes Alba and Fidelia (a 0.5 km x 1.8 km transect). This study area was selected due to our interest in potential spatial and temporal linkages between small domesticscale reservoirs and residential groups. Since our intent was to tie together data from the residential excavations to that from the closed depressions, we selected three areas where the project had, or was gathering, data on residential occupation. Based on recovered ceramics, preliminary excavations had established primarily Preclassic and Classic period occupations at the Fidelia and Alba groups, respectively (Walker 2016), and we sought to determine if potential water features spatially associated with these groups would exhibit similar date ranges for their construction and use. Finally, we intentionally selected depressions of varying dimensions including depth and surface area to investigate any apparent correlations between these physical variables and the functional nature of the features. For example, did deeper depressions originate as limestone 59

69 quarries prior to being utilized for water storage? Or did broader, shallower depressions appear to function in an agricultural capacity perhaps to distribute water to, or collect runoff from, an adjacent agricultural field? The nature of the recovered geoarchaeological and cultural data would provide information on the probable functions of these closed depressions. Depressions were excavated to bedrock or culturally-sterile sascab and all recovered ceramic material was assessed, assigned a preliminary date (if possible) based on the chronologic ceramic typology being developed for Yaxnohcah, and catalogued by the project ceramicist (Walker 2016). These data would prove useful in establishing a date range for the active lifespan of these potential water features, as well as drawing parallels between their use and the occupation of adjacent residential complexes. We also collected charcoal fragments from multiple depressions to submit for radiocarbon dating, although only one sample was considered substantial enough to test. Combined with ceramic material from the same unit, these data would be used to substantiate a chronology for the feature s active operation. 4.1 Assessment of depressions A combination of observations including the presence of catchments several times larger than the surface area, sufficient calculated capacity, and evidence of a clay or other sealant overlying the bedrock or sascab base of the depression were the key determinants in identifying depressions as domestic-scale reservoirs. This strategy is based on a similar scheme employed by Weiss-Krejci and Sabbas (2002) in their evaluation of small depressions as potential water features in the central lowlands and has since been applied in other studies of aguadas and small reservoirs in the Maya area (Akpinar-Ferrand et al. 2012; Brewer 2007, 2016). 60

70 After GPS points were recorded from the center of each unit, excavations were initiated either in the center (Depressions D-1, D-3, and D-4) or along the rim (Depressions D-2 and D-5) of each depression tested. Excavations placed in the center focused on establishing a stratigraphic profile for the depression, as well as seeking evidence of a clay plaster sealant overlying the bedrock or sascab base of the depression or visible within the profile. A relatively impervious clay layer, whether naturally or artificially deposited, has been demonstrated to effectively reduce the porosity of the underlying limestone and aid in water retention at several sites throughout the lowlands (e.g. Beach and Dunning 1997; Gill 2000; Siemens 1978; Scarborough et al. 1995). Remaining excavations were located along depression rims in an attempt to investigate the possible presence of a water retention or otherwise culturally-modified reservoir wall or of a channel that would have permitted the flow of water into or out of the reservoir. Such a channel was detected in an aguada near the site of La Milpa, where excavation revealed a natural depression modified to retain water that incorporated channel and berm features (Chmilar 2005). During excavations of the Aguada Los Loros near the site of San Bartolo, Akpinar-Ferrand and colleagues (2012) identified a well-defined berm within a depression, the position of which suggested its use as a siltation tank for filtering water into the reservoir. Additional examples of these and similar practices exist elsewhere in the Maya lowlands. Based on surface observations and recovered geoarchaeological data (Table 1), three of the five small depressions excavated at Yaxnohcah appear to have functioned as residential scale reservoirs at some point in their active lifespans. Depressions D-2, D-3, and D-4 were bordered by catchments that would have directed water into these tanks (Figure 4, Figure 5, Figure 6). Each had sufficient depth to store water, with measurements from present ground surface to 61

71 exposed bedrock ranging from 0.78 m to 1.96 m and approximate capacities between 95 and 453 m3. Reservoir capacity was estimated from two components: the current capacity of the closed depression based on the lidar DEM and the volume of sediment filling the postulated original ancient Maya depression. The potential capacity of the reservoir in ancient Maya time is the sum of the two capacities. The current capacity is estimated as one half the lidar-derived surface area times the lidar-derived maximum depth. The volume accounting for sedimentation is estimated by taking the formula for an elliptical cone, H(½) (A/2)(B/2), where H is the height and A and B are the length and width of the ellipse, respectively, and modifying it to accommodate the fact that small depressions tend to be more spherical in shape (Brewer 2007). An important consideration with this calculation is that the maximum excavated depth or the depth of a sealed plaster surface or impermeable clay layer identified within each depression defines H. This figure is based on the assumption that the ancient Maya would have excavated the depressions to their deepest extent in order to maximize their water storage capacity. In reality, each of the depressions investigated in this study contained a layer of in-washing sedimentation overlying the bedrock. The sealed plaster floors in Depressions D-3 and D-4 were placed on top of this sediment layer and would have represented the maximum utilized depth of these two features. This differs from the depth figure provided by the lidar (depth of present closed depression) and results in a significant difference between our lidar pre-screening volume estimates and our actual excavated volume calculation. This difference also accounts for more than one thousand years of sedimentation and soil formation taking place within the depressions. This is a notable distinction because it underlines the importance of excavating these depressions 62

72 in order to fully understand their physical characteristics, as opposed to merely relying on the topographical profile provided by the lidar. Importantly, evidence of former floors and superadjacent reservoir sediments was detected in the profiles revealed in Depressions D-3 (Figure 7) and D-4 (Figure 8). In D-4, the 3C horizon consists of the weathered remains of a floor constructed of hard-tamped clay mixed with sandy sascab (a technique still used today to line irrigation ditches and other hydraulic features in Yucatan). Above this floor, a thick layer of clayey reservoir sediments (2C horizon) accumulated over time along with broken ceramics and other artifacts. These sediments apparently experienced episodic seasonal desiccation as evident in the development of slickensides, a product of shrinking and swelling clay, which also produced distortion in the underlying floor. Either in the last years of its use, or post-abandonment, the depression began to fill with colluvial sediment (C-horizon) derived from the disintegration of the adjacent residential complex, atop which the modern soil has developed over a millennium. Depression D-3 likely originated as a quarry associated with the construction of the nearby enormous Alba Group monumental architecture, which subsequently partly filled with in-washing sediment (C3 and C4 horizons), before being sealed with a now highly weathered plaster floor (C2 horizon), then accumulated reservoir sediments (C1 and AC horizons). Dates associated with ceramics recovered in both depressions coincided with dates derived from neighboring residential groups, supporting the idea that these reservoirs were in use during the occupation of these areas of the site. Depression D-2 lacked evidence of a watertight surface. Based upon the exposed rock face on the depression rim and bedrock at the base of all three suboperations exhibiting cut marks on several limestone blocks, this feature appears to have initially functioned as a quarry 63

73 for building material (Figure 9). If limestone blocks were being mined for construction activity in the immediate vicinity, their most likely destination would be the Wo Group residential complex currently under investigation by YAP (Peuramaki-Brown et al. 2016). Indeed, initial comparative analyses of ceramic material from this depression and excavations at the Wo group indicate contemporaneity in Early and Late Middle Preclassic forms (Walker 2016). This apparent temporal (and type) correspondence between ceramics from the residential and depression excavations, combined with the accessibility of the depression s location, support the notion that the area could well have functioned as a quarry for Middle Preclassic period building material at the Wo Group before then serving as a water storage tank for the same community. Despite not being completely impermeable, the depression would nonetheless have been centrally located and readily accessible to serve local residents immediate possibly nonpotable water needs. The neighboring catchment area, in particular, would have been well placed to either shed runoff into or receive outflow from the reservoir, probably for agricultural purposes (Figure 6). Similar depressions initially opened as quarries prior to their use as reservoirs have been identified elsewhere throughout the Maya lowlands (Folan 1982; Weiss- Krejci and Sabbas 2002). In contrast, Depressions D-1 and D-5 appear to have originated naturally and presented little evidence of functioning in a significant water storage capacity. Both of these 1 m x 1 m units were excavated to solid bedrock. Neither was located adjacent to a large catchment area, nor did they contain the remains of a clay sealant layer. However, Depression D-1 did contain a circular indentation of loose, dark soil in the eastern edge of the north wall at a depth of approximately 45 cm below surface. With a diameter of 55 cm, this hole also contained considerable buried organic material. Such pockets are known to be favored locations for Maya 64

74 horticulture or arboriculture activities that would undoubtedly have taken advantage of localized water shedding into the depression downslope from the adjacent Wo Group.a The role of small depressions as household gardens or areas of agriculture, horticulture, and apiculture have also been recognized elsewhere in Maya communities (Folan 1983; Gomez-Pompa et al. 1990; Kepecs and Boucher 1996; Lohse and Findlay 2000). The chronological mix of recovered dateable ceramic material within Depression D-1 appears to indicate early and sustained (Preclassic through Classic period) activity in this area of the site coinciding with occupation of the Wo complex which supports the idea this feature may have served as a long-standing agricultural plot during this period. Depression D-5, located on the northern edge of the Wo Group, contained a fairly high density of ceramic sherds, although all were extremely eroded and unclassifiable. This depression appears to be an unmodified natural feature without any identifiable cultural function. 4.2 Significance of results We examined five closed depressions that we hypothesized functioned in a water storage capacity within our study area at Yaxnohcah, but their origin and varying functions could only be tested through excavation. Selected through a combination of lidar analysis, ground verification, and spatial association with ongoing residential excavations in the vicinity of the Alba and Fidelia Complexes, these depressions presented a wealth of data regarding water management and other cultural activities during the Preclassic and Classic periods in this area of the site. a This is only one possible interpretation. This buried depression may also result from a taproot or other natural subsurface feature. 65

75 Not every depression is a cultural feature; indeed, only three of the five investigated in this study presented a definite cultural component. A combination of variables, including surface observations, storage capacity, and geoarchaeological evidence, was used to determine the likelihood of a depression functioning as a reservoir. Three of the five features studied Depressions D-2, D-3, and D-4 present evidence of water storage activities. Conversely, Depressions D-1 and D-5 appear to be naturally occurring features that lack the clay, plaster, or stone facing needed to improve their ability to hold water. As a result, these two depressions may have served less defined and less reliable water management or other functions or perhaps no cultural role at all. Depression D-1 presented some evidence of having served as an area for localized agriculture. In the 1 km x 1 km area we examined most closely (Figure 2), we identified 39 closed depressions with a lidar-derived capacity greater than 20 m3. Extrapolating from this count, we would expect more than 900 closed depressions with volumes greater than 20 m3 throughout the entire 24 km2 lidar area. This very large number of closed depressions, potential reservoirs or household tanks should in hindsight not be unexpected. While many of the larger closed depressions appear to originate as natural karst sinkholes, we found that a greater number of the smaller depressions are holes left from quarrying limestone for building material. We assessed many of the closed depressions including the very small ones as quarries, based on exposed cut marks on the rim or walls and loose cut stones left in the hole. It is not surprising that the area is well covered with quarry holes, given the density of stone structures found across the upland surfaces in the lidar coverage area. Although there continues to be an increasing recognition of the importance of water management in the growth and sustainability of Maya civilization, small reservoirs or tanks 66

76 remain largely neglected as individual elements of systematic study within this larger system. Frequently occurring and widely dispersed throughout the Maya lowlands, these features have been shown to serve vital water collection, storage, and distribution functions at multiple scales within Maya communities. Confirming a water management function for these depressions, as well as determining additional cultural roles they may have served, provides essential knowledge in understanding the socioeconomic structure and day-to-day functionality of lowland Maya civilization. The comprehensive landscape-scale view provided by lidar is particularly beneficial to our understanding of these connections because it allows us to see the broader picture of water management beyond the spatial limitations of archaeological transects, for example. In addition to identifying individual reservoirs adjacent to residential structures in the midst of dense tropical forest, the lidar permits us to visualize how these residential tanks may have operated apart from, or in tandem with, larger reservoirs or canals as part of a complex hydraulic system at Yaxnohcah. 5. Conclusions Multiple studies conducted over the past few decades have emphasized the necessity of rainwater collection and storage as a critical aspect in the rise of Maya civilization (Adams 1991; Dunning et al. 1999; Scarborough 1993). More recently, lidar acquisition and analysis in the Maya area has begun to revolutionize landscape archaeology approaches to studying the urban and ecological adaptations of this adaptive, enduring culture (Chase et al. 2011; Chase et al. 2014; Johnson and Ouimet 2014; Prufer et al. 2015; von Schwerin et al. 2016). Although the use of lidar in archaeology has enabled the expedient acquisition of spatial data over large areas and the detection of architectural and landscape features including potential reservoirs this study 67

77 demonstrates that lidar analysis alone is unable to truly assess a community s residential-scale water management activities within the karst topography of the Maya lowlands. Despite executing a successful sampling strategy and acquiring critical spatial, functional, chronological, and cultural data, broader goals of understanding the complex patterns of water management activities at Yaxnohcah such as centralization versus decentralization and the degree to which water storage activities evolved during the Preclassic heyday of the community have yet to be fully achieved. Ongoing water feature investigations, including Dunning s excavations at the larger Brisa and Fidelia reservoirs (Dunning et al. 2016), additional sampling of household-scale reservoirs, and possible testing of chultuns and their role in water storage, will be necessary to gain a more complete picture of the unified and adaptive system of water management that was undoubtedly operational at Yaxnohcah throughout the Preclassic and Classic periods of ancient Maya civilization. Acknowledgements We would like to thank Jerry Murdock and Venture Capital Partners for generously funding the lidar acquisition for the Yaxnohcah Archaeological Project. The lidar data was collected by the National Center for Airborne Laser Mapping (NCALM; NSF Award No. BCS ). The work reported here was performed under permits extended to Kathryn Reese-Taylor and Armando Anaya Hernández for the YAP. Research was funded by National Science Foundation (NSF Award No ) and URGC (University of Calgary) grants, as well as ongoing support through the University of Calgary, Universidad Autonóma de Campeche, Athabasca University, and the University of Cincinnati. Special thanks go to the local field workers who assisted with the archaeological excavations detailed here, particularly Javier Cobos, Augustín Diaz, and Reyes Pérez Martínez. 68

78 References Adams, R. E. W Nucleation of Population and Water Storage Among the Ancient Maya. Science 251:632. Akpinar-Ferrand, Ezgi, Nicholas P. Dunning, David L. Lentz, and John G. Jones 2012 Use of Aguadas as Water Management Sources in Two Southern Maya Lowland Sites. Ancient Mesoamerica 23: Anaya Hernández, Armando, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor (editors) 2016 Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones. Mesoweb. Beach, Timothy and Nicholas Dunning 1997 An Ancient Maya Reservoir and Dam at Tamarindito, el Peten, Guatemala. Latin American Antiquity 8(1): Brewer, Jeffrey L 2007 Understanding the Role of a Small Depression in Ancient Maya Water Management at the Medicinal Trail Site, Northwest Belize. M.A. thesis, Department of Anthropology, Division of Graduate Studies and Research of the University of Cincinnati Investigaciones en Aguadas de Escala Residenciales. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan PeuramakiBrown, and Kathryn Reese-Taylor, pp Mesoweb. Brzank, Alexander, Christian Heipke, Jens Goepfert, and Uwe Soergel 2008 Aspects of Generating Precise Digital Terrain Models in the Wadden Sea from Lidar-Water Classification and Structure Line Extraction. Journal of Photogrammetry and Remote Sensing 63: Carr, Christopher, Eric Weaver, Nicholas P. Dunning, and Vernon L. Scarborough 2015 Bringing the University of Pennsylvania Maps of Tikal into the Era of Electronic GIS. In Tikal: Paleoecology of an Ancient Maya City, edited by David L. Lentz, Nicholas P. Dunning, and Vernon L. Scarborough, pp Cambridge University Press, Cambridge. Carr, Robert F. and James E. Hazard 1961 Map of the Ruins of Tikal, El Peten, Guatemala. Tikal Report No. 11. University Museum Monograph 21, University of Pennsylvania, Philadelphia. Chase, Arlen F., Diane Z. Chase, and John F. Weishampel 2010 Lasers in the Jungle: Airborne Sensors Reveal a Vast Maya Landscape. 69

79 Archaeology 64(4): Chase, Arlen F., Diane Z. Chase, John F. Weishampel, Jason B. Drake, Ramesh L. Shrestha, K. Clint Slatton, Jaime J. Awe, and William E. Carter Airborne LiDAR, Archaeology, and the Ancient Maya Landscape at Caracol, Belize. Journal of Archaeological Science 38: Chase, Arlen F., Diane Z. Chase, Jaime J. Awe, John F. Weishampel, Gyles Iannone, Holley Moyes, Jason Yaeger, Kathryn Brown, Ramesh L. Shrestha, William E. Carter, and Juan Fernandez Diaz 2014 Ancient Maya Regional Settlement and Inter-Site Analysis: The 2013 WestCentral Belize LiDAR Survey. Remote Sensing 6: Chmilar, Jennifer A Ancient Maya Water Management: Archaeological Investigations at Turtle Pond, Northwestern Belize. Unpublished M.A. thesis, Department of Anthropology, University of Cincinnati, Cincinnati. Corns, Anthony and Robert Shaw 2009 High Resolution 3-Dimensional Documentation of Archaeological Monuments & Landscapes Using Airborne LiDAR. Journal of Cultural Heritage 10S: Devereux, B. J., G. S. Amable, and P. Crow 2008 Visualisation of LiDAR Terrain Models for Archaeological Feature Detection. Antiquity 82: Dunning, Nicholas P., Vernon L. Scarborough, Fred Valdez, Jr., Sheryl Luzzadder-Beach, Timothy Beach, and John G. Jones 1999 Temple Mountains, Sacred Lakes, and Fertile Fields: Ancient Maya Landscapes in Northwestern Belize. Antiquity 73: Dunning, Nicholas P., Timothy Beach, and Sheryl Luzzadder-Beach 2012 Kax and Kol: Collapse and Resilience in Classic Maya Civilization. Proceedings of the National Academy of Sciences 109: Dunning, Nicholas P., Armando Anaya Hernández, and Helga Geovannini 2016 Operaciones 13 y 19: Investigaciones en los Reservorios. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan PeuramakiBrown, and Kathryn Reese-Taylor, pp Mesoweb. Evans, Damian H., Roland J. Fletcher, Christopher Pottier, Jean-Baptiste Chevance, Dominique Soutif, Boun Suy Tan, Sokrithy Im, Darith Ea, Tina Tin, Samnang Kim, Christopher Cromarty, Stéphane De Greef, Kasper Hanus, Pierre Bâty, Robert Kuszinger, Ichita Shimoda, and Glenn Boornazian 70

80 2013 Uncovering archaeological landscapes at Angkor using lidar. Proceedings of the National Academy of Sciences 110(31) Fernandez-Diaz, Juan Carlos, William E. Carter, Ramesh L. Shrestha, and Craig L. Glennie 2014 Now You See It Now You Don t: Understanding Airborne Mapping LiDAR Collection and Data Product Generation for Archaeological Research in Mesoamerica. Remote Sensing 6(10) Flores Esquivel, Atasta 2014 Trabajos de reconocimiento y mapeo en Yaxnohcah, Campeche. Temporada de In Proyecto Arqueológico Yaxnohcah, 2013 Informe de la Segunda Temporada de Investigaciones, edited by Kathryn Reese-Taylor and Armando Anaya Hernández, pp Mesoweb. November 1, pp. Folan, William J Mining and Quarrying Techniques of the Lowland Maya. Anthropology 6: Physical Geography of the Yucatan Peninsula. In Coba: A Classic Maya Metropolis, edited by William J. Folan, Ellen R. Kintz, and Laraine A. Fletcher, pp Academic Press, New York. Gallagher, Julie M. and Richard L. Josephs 2008 Using LiDAR to Detect Cultural Resources in a Forested Environment: an Example from the Isle Royale National Park, Michigan, USA. Archaeological Prospection 15: Gill, Richardson Benedict 2000 The Great Maya Droughts: Water, Life and Death. University of New Mexico Press, Albuquerque. Gómez-Pompa, Arturo, José Salvador Flores, and Mario Aliphat Fernández 1990 The Sacred Cacao Groves of the Maya. Latin American Antiquity 1: Hare, Timothy, Marilyn Masson, and Bradley Russell 2014 High-Density LiDAR Mapping of the Ancient City of Mayapán. Remote Sensing 6: Hightower, Jessica N., A. Christine Butterfield, and John F. Weishampel 2014 Quantifying Ancient Maya Land Use Legacy Effects on Contemporary Rainforest Canopy Structure. Remote Sensing 6: Hughbanks, Paul J Research at Guijarral (RB-18), In The Programme for Belize Archaeological Project: 1994 Interim Report, edited by Richard E. W. Adams and 71

81 Fred Valdez Jr., pp The University of Texas at San Antonio, San Antonio. Hutson, Scott R Adapting LiDAR data for regional variation in the tropics: A case study from the Northern Maya Lowlands. Journal of Archaeological Science: Reports 4: Jennings, Joseph N Karst Geomorphology. Basil Blackwell, Oxford. Johnson, Katharine M. and William B. Ouimet 2014 Rediscovering the lost archaeological landscape of southern New England using airborne light detection and ranging (LiDAR). Journal of Archaeological Science 43:9-20. Kepecs, Susan and Sylviane Boucher 1996 The Pre-Hispanic Cultivation of Rejolladas and Stone-Lands: New Evidence from Northeast Yucatán. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use, edited by Scott L. Fedick, pp University of Utah Press, Salt Lake City. Lene, Gene W Geology of the Blue Creek Area. In The Blue Creek Project: Working Papers from the 1996 Field Season, edited by W. David Driver, Heather L. Clagett, and Helen R. Haines, pp St. Mary s University, San Antonio, Texas. Lohse, Jon C. and Patrick N. Findlay 2000 A Classic Maya House-Lot Drainage System in Northwestern Belize. Latin American Antiquity 11(2): McCoy, Mark D., Gregory P. Asner, and Michael W. Graves 2011 Airborne lidar survey of irrigated agricultural landscapes: an application of the slope contrast method. Journal of Archaeological Science 38: Parent, Jason R., John C. Volin, and Daniel L. Civco 2015 A fully-automated approach to land cover mapping with airborne LiDAR and high resolution multispectral imagery in a forested landscape. ISPRS Journal of Photogrammetry and Remote Sensing 104: Pe eri, Shachak and William Philpot 2007 Increasing the Existence of Very Shallow-Water LIDAR Measurements Using the Red-Channel Waveforms. IEEE Transactions on Geoscience and Remote Sensing 45(5): Peuramaki-Brown, Meaghan, Shawn G. Morton, and Ana Castro 2016 Operacíon 17 en el Grupo Wo, Excavaciones de Prueba en un Lote 72

82 Doméstico Peri-Urbano. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor, pp Mesoweb. Prufer, Keith M., Amy E. Thompson, and Douglas J. Kennett 2015 Evaluating airborne LiDAR for detecting settlements and modified landscapes in disturbed tropical environments at Uxbenká, Belize. Journal of Archaeological Science 57:1-13. Reese-Taylor, Kathryn 2014 Resumen del Temporada In Proyecto Arqueológico Yaxnohcah, 2013 Informe de la Segunda Temporada de Investigaciones, edited by Kathryn ReeseTaylor and Armando Anaya Hernández, pp Mesoweb. November 1, Reese-Taylor, Kathryn and Armando Anaya Hernández (editors) 2013 Proyecto Arqueológico Yaxnohcah, 2011 Informe de la Primera Temporada de Investigaciones. Mesoweb. July 1, pp Proyecto Arqueológico Yaxnohcah, 2013 Informe de la Segunda Temporada de Investigaciones. Mesoweb. November 1, pp. Reese-Taylor, Kathryn, Armando Anaya Hernández, F. C. Atasta Flores Esquivel, Kelly Monteleone, Alejandro J. Uriarte Torres, Helga Geovannini Acuña, and Chris Carr. 2016a Verificacion en Campo del Reconocimiento de Lidar. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan PeuramakiBrown, and Kathryn Reese-Taylor, pp Mesoweb. Reese-Taylor, Kathryn, Armando Anaya Hernández, F. C. Atasta Flores Esquivel, Kelly Monteleone, Juan Carlos Fernández-Diaz, Alejandro Uriarte, Christopher Carr, Helga Geovannini Acuña, Meaghan Peuramaki-Brown, and Nicholas Dunning 2016b Boots on the Ground at Yaxnohcah: Ground Truthing Lidar in a Complex Tropical Landscape. Advances in Archaeological Practice 4(3): Ruppert, Karl and John H. Dennison 1943 Archaeological Reconnaissance in Campeche, Quintana Roo, and Petén. Carnegie Institution of Washington, Publication 543. Washington, D.C. Scarborough, Vernon L Introduction. In Economic Aspects of Water Management in the Prehispanic New World, edited by Vernon L. Scarborough and Barry L. Isaac, pp Research in Economic Anthropology, supplement No. 7, JAI Press, Greenwich, Connecticut. 73

83 Scarborough, Vernon L. and Gary G. Gallopin 1991 A Water Storage Adaptation in the Maya Lowlands. Science 251: Scarborough, Vernon L., Matthew E. Becher, Jeffrey L. Baker, Garry Harris, and Fred Valdez Jr Water and Land at the Ancient Maya Community of La Milpa. Latin American Antiquity 6(2): Siemens, Alfred H Karst and the Pre-Hispanic Maya in the Southern Lowlands. In Pre-Hispanic Maya Agriculture, edited by Peter D. Harrison and B.L. Turner II, pp University of New Mexico Press, Albuquerque. Šprajc, Ivan 2008 Reconocimiento Arqueológico en el Sureste del Estado de Campeche: Paris Monographs in American Archaeology 19. BAR International Series Archaeopress, Oxford. Šprajc, Ivan and Atasta Flores Esquivel 2008 Descripcíon de los sitios. In Reconocimiento arqueológico en el sureste del estado de Campeche, México, edited by Ivan Šprajc, pp BAR International Series Archaeopress, Oxford. Tourtellot, Gair III and John J. Rose 1993 More Light on La Milpa: Interim Report on the 1993 Season. Manuscript on file, Department of Archaeology, Boston University, Boston. Turner, D. D., R. A. Ferrare, V. Wulfmeyer, and A. J. Scarino 2014 Aircraft Evaluation of Ground-Based Raman Lidar Water Vapor Turbulence Profiles in Convective Mixed Layers. Journal of Atmospheric and Oceanic Technology 31: von Schwerin, Jennifer, Heather Richards-Rissetto, Fabio Remondino, Maria Grazia Spera, Michael Auer, Nicholas Billen, Lukas Loos, Laura Stelson, and Markus Reindel 2016 Airborne LiDAR acquisition, post-processing and accuracy-checking for a 3D WebGIS of Copan, Honduras. Journal of Archaeological Science: Reports 5: Walker, Debra S Apuntes Sobre La Secuencia Cerámica de Yaxnohcah. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan PeuramakiBrown, and Kathryn Reese-Taylor, pp University of Calgary, Calgary. Weiss-Krejci, Estella and Thomas Sabbas 2002 The Potential Role of Small Depressions as Water Storage Features in the Central Maya Lowlands. Latin American Antiquity 13(3):

84 Figure 1: Map of the Maya region showing Yaxnohcah. The site is located between the ancient Maya cities of El Mirador and Calakmul within the Elevated Interior Region of the Yucatán Peninsula, Mexico. 75

85 Figure 2: Lidar-derived hillshade image for our study area. Our excavations at D-1 through D-5 are adjacent to the Alba, Wo, and Fidelia structure groups. The thirty-nine closed depressions greater than 20m3 in volume within our primary study area, grid squares 3E, 3F, 4E, and 4F, are outlined with solid fill (blue in the online version of this article). 76

86 Figure 3: Wo' Group, closed depression CD-101, and our excavation at D-1. The graph shows the lidar-derived 3-D elevation profile across the plaza group and closed depression. The basemap is a lidar-derived hillshade with lidarderived contour lines (0.5 m interval). Architecture drawings here and in subsequent figures courtesy of the Proyecto de Reconocimiento Arqueológico en el Sureste de Campeche, directed by Ivan Šprajc; Atasta Flores Esquivel and Thomaž Podobnikar surveyors. 77

87 Figure 4: The Alba Group and adjacent depressions. The operation group D-3 (Op-16 E, F, G, H, I, K) was positioned on the north berm of closed depression CD

88 Figure 5: The Fidelia Group on top of an extensively modified "island" in Bajo Laberinto. Our operation at D-4 is located in the center of closed depression CD

89 Figure 6: The Wo' Group area with operations D-1, D-2, and D-5 and associated watersheds. The closed depression CD101 is associated with D-2 (Op-16 B, C, D). Closed depressions are not associated with D-1 (Op-16 A) and D-5 (Op-16 L). The six closed depressions in the image with volume greater than 20 m3 are labeled. 80

90 Figure 7: Closed depression D-3 (Op. 16F) north profile showing remnants of clay layer overlying colluvium. 81

91 Figure 8: Closed depression D-4 north profile showing remains of decomposing floor, colluvium and evidence of slickensides 82

92 Figure 9: Base of closed depression D-2 showing cut limestone blocks 83

93 Dep. No. Residential Group Distance to Nearest Structure (m) D-1 Wo 80 D-2 Wo 60 D-3 Alba 60 D-4 Fidelia 40 D-5 Wo 20 Chronology Preclassic Late Classic (1000 BC AD 800) Early/Middle Preclassic Late Classic (900 BC AD 800) Middle Preclassic Late Classic (700 BC AD 800) Middle Classic Postclassic (600 BC AD 1000) Unknown LidarDefined Catchment Area (m2) LidarDefined Surface Area (m2) LidarDefined Current Depth (m) LidarDefined Present Capacity (m3) Excavated (silted) Capacity (m3) Total Capacity (lidar plus excavation) (m3) Observations and Materials Postulated Origin 0 Excavated Depth Below Current Surface (m) ,828 Postulated Function Buried organic soil feature Natural Agri- or horticulture? Cut marks on exposed limestone Artificial Limestone quarry reservoir 19,245 1, (0.75 to plaster layer) 1, (375 to plaster layer) 1,834 (1,756 to plaster layer) Clay plaster layer (colluvium) Artificial Reservoir (1.65 to plaster layer) 9 95 (83 to plaster layer) 104 (92 to plaster layer) Clay plaster layer (colluvium) Artificial Reservoir (agricultural?) Exposed surface stones Natural Unknown Table 1: Data summary and evaluation for investigated depressions. The lidar-defined depth was determined by using the 3-D profile tool to measure the distance from the highest point along the depression rim to the lowest present ground surface within the depression based on the lidar-derived DEM. The lidar-derived depth represents the current silted-in elevation. Excavated depth represents total depth from present ground surface to the base of each excavated unit within the closed depression. In the cases of D-1 and D-5, the lidar did not define them as closed depressions, so surface area, depth, and capacity are listed as 0. 84

94 4. Research Article 2 HOUSEHOLDERS AS WATER MANAGERS: A COMPARISON OF DOMESTICSCALE WATER MANAGEMENT PRACTICES FROM TWO CENTRAL MAYA LOWLAND SITES Jeffrey L Brewera Submitted to Ancient Mesoamerica (November 22, 2016) DO NOT CITE IN ANY CONTEXT WITHOUT PERMISSION OF THE AUTHOR Department of Geography, University of Cincinnati, 409 Braunstein Hall, Cincinnati, OH brewerjy@mail.uc.edu Phone: Fax: a 85

95 ABSTRACT Multiple studies conducted over the past few decades have recognized the necessity of rainwater collection and storage as a critical aspect in the evolution of Maya civilization. Few of these efforts, however, have emphasized the importance of managing water resources at the household level. Data is presented from two central lowland sites the dispersed hinterland community of Medicinal Trail and the urban center of Yaxnohcah that elucidate the function of small reservoirs and associated landscape modifications in residential water management. Despite differing physical geographies and trajectories of urban development, residents of both communities were clearly engaged in water management activities based, at least in part, on the creation and use of small reservoirs. Decentralized household water management practices appear to have been widespread components, both temporally and spatially, of Maya civilization beginning in the Preclassic period. 86

96 The land and water management strategies that fostered the development of complex societies in southern Mesoamerica are incompletely understood. However, for the ancient Maya, environmental and resource studies provide a unique lens through which to view the integrative social, political, and economic activities that characterized this resilient civilization for centuries prior to the Conquest period. The study of water management features of the Maya lowlands, in particular, has recently proved invaluable in revealing some of the complex human-environment relationships of the Late Preclassic (400 BC AD 250) and Classic (AD ) period Maya. The interior portions of the lowlands in Mexico, Guatemala, and Belize were particularly challenging environments for ancient occupation, owing to a porous limestone (karst) landscape, a general lack of permanent surface water, and an acute 5-month annual dry season. These conditions combined to necessitate the seasonal collection and storage of rainwater not only for survival, but also for domestic, agricultural, and ritual activities across multiple societal levels. Our knowledge of ancient Maya water management is derived from numerous survey and excavation projects undertaken throughout the lowlands that have analyzed a variety of hydrologic features such as reservoirs, seasonally inundated wetlands (bajos), canals, dams, wells, natural sinkholes resulting from the collapse of limestone bedrock (cenotes), and small underground water cisterns (chultuns) in an archaeological context (e.g. Akpinar et al. 2012; Beach and Dunning 1997; Brewer 2007; Bullard 1960; Chmilar 2005; Dunning et al. 1999; Harrison 1993; Isendahl 2011; Johnston 2004; Lohse and Findlay 2000; Matheny 1976, 1978, 1982; Scarborough 1991, 1994, 1996, 2003; Scarborough and Gallopin 1991; Scarborough et al. 1995; Weiss-Krejci 2013; Weiss-Krejci and Sabbas 2002). Collectively, they provide many insights into the creation and maintenance of these features, as well as their use by the ancient Maya. As scientists continue to study the multifaceted components of Maya water management, 87

97 it is becoming increasingly evident that these ancient systems were complex, varied, and far from uniformly distributed across the Maya world. The complexities of Maya water management continue to be explored through a number of broader applications, including within the contexts of landscape modifications (Dunning et al. 1999), critical resource control (Ford 1996), paleoecology (Dunning and Beach 2010; Johnston et al. 2001; Wahl et al. 2007), social and environmental changes (Beach et al. 2015a; Dunning et al. 2013a; Kunen 2004; Lucero et al. 2011); political control and political economy (Lucero 1999, 2002, 2006); population estimates and architectural variability (Becquelin and Michelet 1994; McAnany 1990); and satellite imagery and remote sensing analyses (Chase 2016; Garrison 2010; Saturno et al. 2007; Weller 2006). At both the community and household levels, aguadas either natural or human-made ponds that still retain water for at least a portion of the year and topographical depressions that functioned as seasonal reservoirs served as significant sources of water for the region s inhabitants. Naturally occurring depressions originate from collapse or dissolution sinkholes that are usually found at the margins of bajos and along bedrock fractures in karst uplands (AkpinarFerrand et al. 2012; Siemens 1978). Many natural depressions in the vicinity of ancient Maya settlements were found to have been lined with clay, lime plaster, and/or stone-facing to enhance their ability to retain water (Adams and Jones 1981; Akpinar-Ferrand et al. 2012; Brewer 2007; Wahl et al. 2007; Weiss-Krejci and Sabbas 2002). However, both residential and communal tanks could also evolve from human-modified depressions such as ancient limestone, clay, or chert quarries that were later transformed to store water (Bullard 1960; Dunning et al. 2007; Hester and Shafer 1984; Weiss-Krejci and Sabbas 2002). As will be detailed below, the combined effects of the karst hydrology of the Maya lowlands, a lack of permanent natural water 88

98 sources, and the seasonality of available precipitation necessitated collection and storage of rainwater, a system in which aguadas and small reservoirs likely played a necessary and substantial role in the daily lives of the ancient Maya. In addition to broadening our understanding of water management activities, the careful study of small-scale water management features provides an opportunity to gather important information about human-environment interactions in the Maya area. To fully understand their origin and functions, an interdisciplinary, multi-scale approach incorporating archaeology, climatology, hydrology, geology, and remote sensing is necessary. For example, the record of pollen and other materials preserved within aguada sediments permits archaeologists and paleoecologists seeking to decipher the history of land use and environmental change in the Maya area important, highly localized glimpses of these past landscapes (Dunning and Beach 2010). Satellite imagery and remote sensing technology in concert with landscape archaeology approaches have recently been employed in the Maya area to investigate landscape modifications and answer questions relating to settlement, subsistence, and population by correlating spectral vegetation signatures with locations, boundaries, and dimensions of Maya sites (Chase et al. 2011; Saturno et al. 2007; Weller 2006). Archaeologists in the Maya area have long reported a spatial correlation between reservoirs and ancient settlement remains (Adams 1991; Bullard 1960; Folan et al. 1995; Morley 1937; Scarborough 1993; Smith 1950). Despite this recognition, however, it wasn t until the 1980s that scientific investigations began to focus directly on and around these features (e.g. Domínguez and Folan 1996; Dunning et al. 2003; Scarborough et al. 1995; Wahl et al. 2007). The vast majority of research on these reservoirs has examined them within the context of centralized, elite-controlled construction, maintenance, access, and water distribution with data from large sites such as La Milpa (Scarborough et al. 1995) and Tikal 89

99 (Scarborough and Gallopin 1991; Scarborough 2003; Scarborough et al. 2012) forming the basis of much of our knowledge on their construction and use-histories. Recently, however, as part of a general shift toward understanding the rich diversity that characterized the social and environmental activities, economic independence, and daily practices of non-elites or commoners has taken place in Maya studies, a more complex and dynamic picture of water management activities at the household scale has begun to emerge. This article discusses the results of decentralized, domestic-scale water management investigations from two central lowland sites: the hinterland agricultural community of Medicinal Trail in northwestern Belize and the urban center of Yaxnohcah in southern Campeche, Mexico. Research included a single reservoir and additional water features associated with two residential courtyard groups investigated at Medicinal Trail and a series of four reservoirs associated with residential structures throughout Yaxnohcah. An interdisciplinary approach consisting of landscape survey, excavation, and material analysis of collected cultural material was undertaken to understand the household water management practices in which the inhabitants of these two sites were engaged, especially their reliance on small household tanks operating separately from the larger, centralized, elite-controlled reservoirs. STUDY AREA The Maya lowlands, divided into northern, central, and southern segments as defined by Hammond and Ashmore (1981), is both a cultural and physical region characterized by a series of related cultures that have occupied this area of generally low-lying, limestone terrain with a tropical wet/dry climate (Beach et al. 2006). Evidence of Maya culture and its impacts on the environment date from the Early Preclassic period ( BC) and have lasted, with 90

100 significant ebbs and flows, to the present (Brenner et al. 2003; Pohl et al. 1996). Within the region, environmental and geological factors that create a variety of habitats include fluctuations in intra- and inter-annual rainfall, soil and geomorphic processes, and slope gradients and drainage based on structural geology (Dunning et al. 1998; Beach et al. 2008). The study area in which the two investigated sites are located is defined by a slightly variable physical geography (Figure 1). Medicinal Trail is located roughly 5 km east of the major site of La Milpa in northwestern Belize, along the eastern edge of the Elevated Interior Region (or Central Karst Uplands) of the Yucatán Peninsula (Dunning et al. 2012) and within the Programme for Belize (PfB) conservation area, a 1000 km2 ecological reserve located within the greater Three Rivers Region (Scarborough and Valdez 2003). Long-term investigations undertaken within and in the immediate vicinity of the PfB boundaries have produced valuable information about occupation and land use of this region from the Middle Preclassic ( BC) to the Late Classic periods (AD ), with a particular focus on small sites or communities and their relationship to the environment (Adams and Valdez 1993; Adams et al. 2004; Beach et al. 2002; Guderjan 2007; Hammond et al. 1998; Houk 2000; Lohse et al. 2005). The hinterland site of Medicinal Trail is part of the La Lucha Uplands within the Three Rivers physiographic region, an area comprising the fractured eastern side of the central Petén Karst Plateau and heavily influenced by extensive normal faulting (Dunning and Beach 2010; Dunning et al. 1999, 2003). The region is underlain by Late Cretaceous Early Tertiary marine carbonates, primarily marl and chert-containing limestone. The well-developed contemporary karst landscape, characterized by large solution depressions (bajos), rugged ridges, and conical hills, is the result of extensive, long-term physical and chemical weathering of carbonate bedrock (Dunning et al. 1998, 2003). Similar to elsewhere in the Central lowlands, annual rainfall in the 91

101 region today varies between 1400 and 1800 mm (Belize Meteorological Service, Rio Bravo Station). The geological and climatic characteristics of the area lead to significant variation in hydrology, soils, and vegetation types. The sloping, well-drained landscapes of the La Lucha Uplands are usually mantled in thin, calcareous soils (Rendolls) that support an upland forest mixture of subtropical moist forest species (Dunning et al. 2003). Yaxnohcah is located within the Calakmul Biosphere Reserve, part of the 53.3-millionhectare Mesoamerican Biological Corridor linking protected natural areas of Mexico, Guatemala, Belize, El Salvador, Nicaragua, Honduras, Costa Rica, and Panama under the mission of preserving and protecting the biodiversity of the rainforest (Geovannini 2013). Like Medicinal Trail, the site lies within the Elevated Interior Region. Although this region was the heartland of much of both Preclassic and Classic Maya civilization, it is also an area with severe environmental challenges. In particular, the lack of perennial surface water during the lengthy Winter Spring dry season, accentuated by a deep and largely inaccessible karst groundwater system, posed a significant problem for year-round occupation. Across the region, large bajos occupy as much as 40-60% of the land surface. At present, these large depressions are characterized by severe moisture fluctuation ranging from inundation in the wet season to complete desiccation in the dry. Yaxnohcah is situated on the southern flank of one of the largest bajos in the Maya lowlands: Bajo El Laberinto. Previous investigations within the bajo interior revealed highly gypsic or acidic soils that would have been unproductive for agricultural uses (Gunn et al. 2002). However, the margins of the bajo, where soils have developed in calcareous colluvium derived from the adjacent uplands, would potentially have been highly productive sites for cultivation (Dunning et al. 2006; Geovannini Acuña 2008). Shallow soils (Lithosols) predominate in the hills and upper elevations of the site, while Vertisols and other hyromorphic 92

102 soils are found in the depressions and lower-lying areas. Vegetation consists of a mixture of bajo, transitional, and upland semi-deciduous subtropical forest. Some of the principal species observed in the upland forest include: chicozapote (Manilkara zapota), ramón (Brosimum alicastrum), chakah (Bursera simaruba), sapotillo (Pouteria reticulate), and pimiento (Pimenta dioica). In the bajo, dominant species consist of: palo tinto (Haematoxylum campechanum), chaká (Bursera simaruba), guayabillo (Eugenia capuli), and zapote (Manikara zapota), in addition to a variety of bromeliads, orchids, and creeping vines (Geovannini 2013). In addition, the Calakmul Reserve is home to over 1500 species of animals, including jaguar, puma, ocelot, and many types of mammals, birds, reptiles, and fish. THE HYDROLOGIC LANDSCAPE OF MEDICINAL TRAIL Medicinal Trail (16Q ) extends from the Turtle Pond Aguada that defines its western margin to the escarpment on its east (Figure 2). The site name is derived from the PfB s tourist attraction called the Medicinal Trail, a proposed eco-tourism tract populated by vegetative species with a variety of native healing applications. Scarborough and Valdez (2003, 2009) define it as a small resource-specialized community a highly interdependent settlement established on the economic benefits of unique sets of goods and services differentiating one community from another focused on agricultural production and engaged in the communal management of water resources. Initial excavations at the site consisted of two separate studies that investigated mounds and terraces that cross the trail from which the site gets its name (Farnand 2002; Ferries 2002). Two seasons later, excavations took place at Turtle Pond, a large aguada at the base of a slope on the western edge of the site (Chmilar 2005). The extensive capacity of the tank, 785 m3, suggests a function exceeding the household level, 93

103 possibly even providing access beyond the boundaries of the community itself (Scarborough and Valdez 2009). Subsequent field seasons have revealed at least two closely associated formal courtyard groups, a number of informal mound clusters, and multiple landscape modifications including agricultural terraces and depressions and linear features with apparent residential scale water management functions (Brewer et al. 2013; Gill 2009; Hyde 2011; Lowe 2008; Percy 2009). Medicinal Trail Reservoir (Operation 10) The Medicinal Trail Reservoir designated Depression A-3 is located adjacent to Group A, the largest settlement group identified at the site (Figure 3). Defined as a First Tier Household, a formal courtyard group composed of multiple structures with complex basal platforms, Group A consists of six mounds organized around three contiguous courtyards with the largest construction placed on the east side of the group (Hageman and Lohse 2003; Hyde and Atwood 2008; Hyde and Valdez 2007). The depression itself is situated 40 m northeast of the center of Group A and fewer than 10 m south of Structure A-7, and is roughly ellipsoidal in shape. Initial support for Depression A-3 serving as a water tank was based on its surface area, estimated capacity, and location adjacent to a probable residential structure (Structure 7). The shape, potential volume (accounting for sedimentation and soil formation over time), and calculated surface area fit the profile of other topographical depressions investigated as water resources in the central Maya lowlands (Weiss-Krejci and Sabbas 2002). In addition, the depression s proximity to several agricultural terraces and additional house mounds, coupled with the absence of any permanent water sources in the immediate vicinity, suggest its function as a principle source of water for neighboring residents. 94

104 Numerous small aguadas or small reservoirs in the immediate area of household plots or house mounds throughout the lowlands have been noted (Bullard 1960; Matheny 1976; Scarborough 1993, 1998). Bullard (1960:364), in connection with his survey in the northeastern Petén, initially reported that there is a marked tendency for these house ruins to occur in the vicinities of present-day, and possibly former, water sources It is my impression that the maximum distance from a water source was probably less than two kilometers. Scarborough (1993) also points out the importance of households situating themselves as close to a water source as possible, highlighting the competitive aspects of neighboring households and a finite water supply. This spatial relationship between habitation structures and water features has recently been confirmed in the central lowlands at sites in proximity to Medicinal Trail, such as La Milpa, Dos Barbaras, and Wari Camp (Scarborough et al. 1995; Weiss-Krejci and Sabbas 2002). The structures nearest the Medicinal Trail Reservoir fall well within this 2-km radius. The primary goal of the Depression A-3 excavations was to clarify the role of this feature as a potential water resource and demonstrate its use history. A secondary goal was to gain a better general understanding of the role that household tanks played in the daily lives of ancient Maya people, particularly those within a decentralized or hinterland context outside of large urban centers. The earliest occupation evidence in the vicinity of the Medicinal Trail Reservoir dates to the Late Preclassic (400 BC AD 250) and derives from Hyde s house mound and courtyard excavations in Group A (Hyde 2009). Excavations began under the assumption that the reservoir would also have been in operation during this period. The depression was excavated to sterile sascab (calcareous marl) and all recovered cultural material was assessed and catalogued by the project ceramicist and lithicist (Hyde 2007). Ceramic material was assigned a preliminary date (if possible) based on the chronologic ceramic 95

105 typology being developed for Medicinal Trail. These data would prove useful in establishing a date range for the active lifespan of the feature, as well as potentially drawing a parallel between its use and the occupation of nearby residential complexes. Total reservoir capacity for Depression A-3 (and all other reservoirs investigated in this study), based on their combined current and maximum excavated depths, was estimated by taking the formula for an elliptical cone, H(⅓) (A/2)(B/2), where H is the height and A and B are the length and width, respectively, and modified it to accommodate the fact that smaller reservoirs tend to be more spherical in shape. The derived formula, H(½) (A/2)(B/2), employs a constant (½) which is halfway between the constants of the formulas for a cone (⅓) and a hemisphere (⅔), and was used to calculate the volume of the reservoirs studied at Medicinal Trail and Yaxnohcah (Brewer 2007). The result was a volume of over 153 m3 (153,000 liters or over 40,000 gallons), which would have provided over 80,000 person/day servings maximum (Table 1). McAnany s (1990) extremely conservative standard of 1.9 liters (½ gallon, based on eight 8 oz. servings/day) of water per person per day has been debated, with a more recent estimate of 17 liters (4.5 gallons, person/day) applied to the site of Xculoc (Becquelin and Michelet 1994). This figure, based on population projections derived from a combination of residential structure size and estimated capacities of subterranean cisterns (chultuns), may be liberal, with the actual ancient Maya dry season water use likely falling somewhere in between. As a compromise, an estimate halfway between these two figures (9.5 l person/day) was used to calculate the maximum person/day servings to project a localized support population for the Medicinal Trail tank as a source of potable water. A full reservoir, then, with no evaporation could provide over 43 people with enough drinking water each day for one year. This figure would undoubtedly be reduced when taking evaporation rates and additional agricultural and domestic water uses, such as cooking and 96

106 washing, into account. Nonetheless, the reservoir would have served as a reliable water source for residents in this area of the community. The presence of a degraded plaster floor located between a layer of clayey reservoir sediment and a lower clay-rich (Ab) horizon atop sascab is another positive indicator that the depression functioned as a water storage feature. Though sometimes difficult to differentiate from in-washing clay sediment, the degraded remains of this floor were encountered in multiple units throughout the depression roughly cm below the surface and interpreted in the field as the remains of a relatively impervious sealant (Figure 4). The limestone bedrock underlying the depression is highly permeable and modification either natural or artificial would have been necessary to retard leakage. Weiss-Krejci and Sabbas (2002) report the presence of a similar hard gray layer, between 10 and 30 cm thick, overlaying white medium-hard, smooth bedrock at four excavated depressions near the neighboring Wari Camp and La Milpa East sites that they interpret as small water reservoirs. Similarly, the large, extensively excavated La Milpa Aguada was subjected to extensive quarrying and contained a huge berm or dam on the western side that created or enhanced its capacity. A small stone-lined well (bukte) that functioned to extract water from surrounding saturated clay, originally Late Classic before being reopened in the Terminal Classic/Early Postclassic, was also discovered within the bottom of the reservoir. Core samples collected from the base of the aguada revealed a kaolin clay lining of either natural or unnatural origin (Scarborough et al. 1995). Further evidence of the Maya practicing these simple, yet practical modifications has also been noted in multiple aguadas at Calakmul (Folan et al. 1995), Uaxactún (Smith 1950:61, Figure 99b), and near Dos Hombres (Trachman 2007). Within Depression A-3, a layer of in-washing clay sediment overlays the sascab base of the 97

107 feature. Atop this, the highly-degraded floor composed of compacted clay with sascab rubble intermixed lies beneath additional reservoir sediment and modern soil. Bedrock at the base of each excavation unit was examined for evidence of quarrying activities and the possible presence of a water-retention or otherwise culturally-modified wall or a channel that would have permitted the flow of water into and out of the reservoir. Small depressions elsewhere in the Maya lowlands have been interpreted as rock quarries, sascaberas (sascab mines), and clay mines (Folan 1982; Tourtellot and Rose 1993). The notion that small water tanks originating as natural depressions also served as quarries for building material before being modified to store water bears further examination and is also being considered in water management investigations elsewhere in the Maya lowlands (Weiss-Krejci and Sabbas 2002; Brewer et al. forthcoming). Most of the bedrock within Depression A-3 appeared to be naturally decomposing and exhibited no conclusive evidence of quarrying except for visible cut marks and flattened edges on the limestone revealed at the base of Op. 10J. This discovery, combined with a lithic inventory consisting of multiple tools (N=57) including at least one intact mid-sized biface, supports the notion that quarrying activities may well have been taking place within the depression at some time during its use history. Based in part on experimental archaeological research conducted at the site of Nakbe, a variety of stone tools composed of chert, including bifacial axes, picks, and hammer stones, have been demonstrated to be excellent tools for mining, cutting, and shaping limestone (Hansen 1998). The stone tools recovered from Depression A-3 were generally poor to very poor in quality and, apart from a single chalcedony scraper, composed entirely of chert. Biface tools comprise over half of the tool assemblage (N=30), more than half of which are forms related to agriculture (oval bifaces and bifacial celts). Informal and expedient tools, including scrapers and utilized flakes, account for the remaining 98

108 majority (N=13) of the remaining tool assemblage. The overall lithic assemblage from the depression is best characterized as being dominated by low quality chert material used to produce bifaces for local consumption, agriculture, and limestone quarrying (Hyde 2007). Recovered ceramic material included a wide variety of rims and body sherds, basal flanges, and a single eroded vessel base with a broken mammiform foot. The majority of sherds from inside the depression, deriving from all operations except seven of the posthole units initiated around the perimeter of the feature, were eroded and only a portion could be stylistically dated. It is likely that the Maya employed ceramic materials in the course of utilizing the reservoir but due to the continuous exposure to water and pressure, most of this material is either too small, too heavily eroded, or has disintegrated entirely. With the exception of a few rim forms and ring bases that belong to the Chicanel phase (Late Preclassic period; 350 BC AD 250), the majority of recovered ceramic material dates between the Tzakol (Early Classic; AD ) and Tepeu I/II (Late Classic; AD ) complexes. This range of forms, phases, and dates correlates strongly to other ceramic artifacts recovered near the depression (Hyde 2011). Ceramics recovered from Group A clearly indicate an occupation from at least the Late Preclassic through the Terminal Classic, with a likely Middle Preclassic presence. The analysis of recovered soil samples provides additional support for the depression functioning as a reservoir. In an attempt to reconstruct the types of vegetation immediately surrounding the depression through organic or botanical remains, a 1.5-liter soil sample was collected from Op. 10C at a depth of 67 cm below the surface. The sample was floated for micro and macro plant remains using a Flote-Tech Model A-1 machine in the Anthropology lab at Northern Illinois University. After the samples were air dried, they were passed through a standard series of graduated sieves for better viewing, mm mesh. Overall, the sample 99

109 consisted of small limestone chunks and hard clays, with several small pieces of heavily eroded wood charcoal suspended in the matrix. A single palm seed fragment, tentatively identified as Arecaceae sp., was recovered. Palms were used and tended by the ancient Maya in and around settlements, and continue to be grown and harvested by their modern descendants (Darch 1983). In addition, a total of nine seeds determined to be cf. Onagraceae, possibly of the Oenothera genus were identified from the sample (Figure 5). These are seeds from a plant currently being investigated as part of the evening primrose family, and samples have been recovered from several other sites in the PfB area including Blue Creek, the Barba Group, and Guijarral (Hageman and Goldstein 2009). Although its exact significance or relationship to human activity in this time period is unclear, it has been found in conjunction with both domestic and feasting loci and its presence appears to be an indicator of human presence and activity (Goldstein and Brandt 2007; Goldstein and Hageman 2010). Ultimately, Depression A-3 was determined to have originated as a quarry before functioning as a domestic-scale reservoir based on a wealth of recovered physical and cultural data. The depression s surface area and calculated capacity, the presence of a degraded clay sealant overlaying the bedrock, and its association with a large formal household group all support its role as a water storage feature. Analysis of recovered ceramic material revealed that the reservoir was in use contemporaneously with occupation of the adjacent structures, lending further support for its importance to the community s inhabitants. Environmental indicators in the form of recovered palm seeds provide additional evidence of human activity around the depression. Additional Household Water Management Infrastructure at Medicinal Trail 100

110 Initial excavations at Medicinal Trail consisted of two studies that investigated mounds (Ferries 2002) and terraces (Farnand 2002) that cross the main trail through the site. Prehistoric terraces are common features of Maya archaeological landscapes and have often been used to define agricultural production areas. However, these features do not invariably delineate arable land and other uses, such as providing foundations for structures or managing the flow of water, have also been hypothesized. Terraces at Medicinal Trail were studied with the goal of understanding their formation processes and possible role in water management. Farnand determined that each of the terraces investigated were used for a unique purpose, including agricultural, structural, and possibly even territorial functions. Of the four alignments studied, one (Alignment 1) was determined to support a large terrace extending upslope into land that would be optimal for cultivation. A second terrace (Alignment 3) is believed to have served either as an agricultural terrace or a property boundary marker. Alignment 4 is a structural terrace directly joining another contour terrace upslope. The fourth terrace (Alignment 2), however, may well have served in a water management capacity. Like Alignment 1, it supports a large terrace that extends upslope. However, if subjected to higher erosional rates, it is possible that it would not have been able to sustain agricultural activity and may have instead aided in erosion prevention or water diversion (Farnand 2002). If so, this feature might have functioned as a check dam or weir terrace (Beach et al. 2002; Dunning and Beach 1994) acting to restrict the flow of water and create land for cultivation reasonable assumptions for this agrarian community. More recently, excavations took place within two small depressions on the west side of Group A, roughly 15 m west-northwest of the northern courtyard (Depression A-1) and 15 m west of the middle courtyard (Depression A-2) of the group seeking evidence of water 101

111 management modifications (Lowe 2008). After testing both depressions to determine their nature and function, Lowe concluded that neither feature served as a reservoir. Both were small and shallow, with total surface areas of 38 m 2 (Dep. A-1) and 40 m 2 (Dep. A-2) and catchment areas that were difficult to identify or assess. Both features also lacked a distinct clay or other seal that would have aided in water retention. Ultimately, Dep. A-1 was determined to be an unmodified natural feature of the landscape possibly the result of a tree fall and Dep. A-2 appears to be the result of a natural depression being modified and expanded into a small limestone quarry. Although Lowe dismissed Dep. A-2 as a reservoir given the lack of physical or cultural evidence, if it did retain water in some capacity following quarrying activities this small tank would have offered a convenient supplemental, probably non-potable, water source for the inhabitants of Group A. Three studies associated with Group C have been conducted in attempts to identify additional structures and features and their potential role in water management activities in this area of the Medicinal Trail site (Gill 2009; Dedrick 2009; Percy 2009). Group C is classified as a small Second Tier courtyard group and is located 50 meters south of the Tapir Group. Households in Second Tier groups are characterized by two or three structures arranged around a semi-formal courtyard and lacking eastern ceremonial structures (Brewer et al. 2013). Group C consists of three courtyard groups dispersed around a series of landscape modifications, including terraces, berms, and depressions, likely related to agricultural activities (Figure 6). The topography surrounding the group is steep relative to the extensive, low lying plain to the east and a series of interconnected berms appear to function to direct water away from the larger Tapir Group household toward four reservoirs. The flat plain below groups C and D contains 102

112 multiple terraces and reservoirs that, if coordinated, would have collected rainwater from upslope and distributed into adjacent agricultural fields. Operation 15 focused on determining the function of three landscape features posited to relate to water management activities: two berms and an associated depression (Gill 2009). The relationship between water management features used to capture and redirect water for consumption and agriculture has been well documented throughout the Maya area and berms, in particular, have been found at a variety of nearby sites within the Rio Bravo Conservation Area including La Milpa and Guijarral (Dunning and Beach 2003; Hughbanks 1995; Scarborough 2003). The slope, alignment, and composition limestone and chert construction cobbles ranging in size from 5 cm to 45 cm in diameter of Feature 15-A present both internal and external features of a berm. The pitched, roof-like construction may have been used to divert water along its east and west margins (Gill 2009). Similarly, Feature 15-B appears to have been used to slow and direct the flow of water downslope toward the southeastern depression. Ground survey of this depression revealed several cut marks in the limestone along its southern boundary and a large, poorly preserved limestone boulder was removed from its western half (Op. 15F). Taken together, these findings suggest that the depression was quarried for construction material before serving in some capacity as a reservoir. Analysis of the ceramic material recovered in the Op. 15 excavations reveal a Late Classic construction and use of these features, which correlates with occupation dates documented for this area of the site (Dedrick 2009; Gill 2009). At the north end of Group C is a semi-formal courtyard group of three mounds with a possible chultun and two depressions separated by a long berm to the east and southeast (Dedrick 2009). At its southern end, this nearly 40 m long berm hooks to the east and appears to intersect with a small separate berm investigated as part of the Op. 15 excavations. The depression 103

113 studied at Op. 15 is located southeast of this intersection and a ditch that appears to be a drainage feature begins at this feature and winds its way south to an additional depression (Brewer et al. 2013). The primary objective of Op. 13 was to examine three structures that share a common basal platform, their immediate surroundings, and the role this segment of the community may have played in the economic activities of Medicinal Trail. In northwestern Belize, multiple hilltop-focused agricultural sites are composed of large, economically prosperous household groups surrounded by smaller household groups associated with visible agricultural features (Lohse 2004). The nature of these smaller households in the agricultural production of hilltop sites lacks in-depth study, and the similar configuration of Medicinal Trail offers an analogous dataset to consider. The water management component of the Op. 13 investigations was the series of excavations conducted at an apparent chultun within the center of Group C (Op. 14). Limited in scope, the primary purpose of this operation was to establish a basic soil stratigraphy and examine the bedrock morphology within the feature. Although no obvious modifications to the bedrock were observed, cultural material in the form of lithic flakes and ceramic sherds were collected. No dates were assigned to the ceramic material, and it is hypothesized that this depression functioned in tandem with the adjacent berms to collect, and possibly redistribute, water for agricultural purposes (Percy 2009). As with other potential water management features surveyed and cursorily tested throughout the site, the hydrologic infrastructure associated with Group C requires additional excavation and analysis. HOUSEHOLD WATER MANAGEMENT AT YAXNOHCAH Yaxnohcah (16Q ) lies between the major sites of Calakmul and El Mirador within the Central Karst Uplands of the Yucatán Peninsula. The site was initially 104

114 reported by Karl Ruppert and John Dennison (1943:45) following a brief visit in 1933 and given the name Monterey after a nearby aguada. The site was revisited in 2004 by Iván Šprajc and colleagues who investigated the area in more detail, mapped six groups of monumental architecture, and renamed the site Yaxnohcah meaning first big city (Šprajc 2008). In addition to the large civic-ceremonial architectural complexes (designated Groups A F) and two expansive elite residential complexes included in this survey, the site is also defined by two expansive bajos: Bajo Tomatal to the south and Bajo Laberinto to the north and northeast (Figure 7). Today, most bajos are characterized by severe moisture fluctuation ranging from inundation in the wet season to complete desiccation in the dry. The role that bajos played in the rise and fall of Maya civilization in this region has been a matter of vigorous debate and, despite considerable recent research, still eludes consensus. Scholarly views of ancient land use in bajos have ranged from very limited agriculture to widespread intensive wetland field systems. The reality likely lies somewhere in between these opposing views and further investigation is certainly required (cf., Adams 1991; Culbert et al. 1990; Dunning et al. 2006; Hansen et al. 2002; Pope and Dahlin 1989; Wahl et al. 2006). The presence of aguadas within and along the margins of bajos suggests that these reservoirs may have facilitated agriculture. At Yaxnohcah, two flanking bajo-margin aguadas located next to major architectural groups, the Brisa and Fidelia reservoirs, have undergone archaeological testing to gather data on their chronology and construction, with additional large aguadas surveyed and scheduled for similar study in upcoming seasons (Dunning et al 2016). At the household level, small reservoirs located adjacent to many of the residential groups identified at Yaxnohcah would have provided an additional accessible water supply for the community. During the past two field seasons of the Yaxnohcah Archaeological Project 105

115 (YAP), five of these residential tanks have been excavated to determine the chronology of their construction and the nature of their role in water management at the household level (Brewer 2016, Brewer and Carr forthcoming; Brewer et al. forthcoming). Prospective reservoirs were initially identified in the analysis of lidar imagery obtained by YAP (Brewer et al. forthcoming; Reese-Taylor et al. 2016a, 2016b) and then verified through ground survey. The goal was to examine and evaluate these small depressions as potential water features in areas where the project had gathered, or was in the process of accumulating, chronological and cultural data from residential structures (Peuramaki-Brown et al. 2016). Wo Group Residential Reservoir (Operations 16B, 16C, 16D) The first reservoir tested was located 45 m southeast of the Wo Group residential complex then being investigated (Peuramaki-Brown et al. 2016). Three excavations (16B, 16C, and 16D) were opened in this irregularly-shaped, shallow depression resulting in a contiguous 1.5 m x 1.5 m unit positioned on the eastern margin lip of the feature. The goal of these excavations was to investigate the possible presence of a water retention wall, channel, or other modification, targeting places with exposed limestone visible on the surface of the rim and with suggestive surrounding topography. Based upon the exposed limestone at the base of the unit which displays evidence of cut marks on several blocks this depression appears to have initially functioned as a quarry for building material (Figure 8). If limestone blocks were being mined for construction in the immediate vicinity, their most likely destination would be the adjacent Wo Group. Indeed, initial comparative analyses of ceramic material from the depression and Op. 17F at the structure group itself indicate contemporaneity in Early/Middle Preclassic forms (Walker 2016). This 106

116 apparent temporal (and type) relationship between operations, combined with the accessibility of the depression s location, support the notion that the feature could well have functioned as a quarry for Preclassic period building material at the Wo Group. The lack of either a channel or other culturally modified feature in this area of the depression or evidence of a clay sealant call into question the notion that this feature served as a residential water tank. However, its surface area, depth, and location bordering a residential complex in an area lacking any other visible water features support the idea that this depression may have been utilized as an opportunistic water source, or quarry reservoir, perhaps for non-potable purposes. In addition, the catchment area bordering the south, west, and southeast sides of the depression would have been positioned to either receive outflow from or shed runoff into the reservoir, probably for agricultural purposes. Similar depressions originating as quarries before being modified and used as reservoirs have been identified elsewhere across the Maya lowlands (Folan 1982; Weiss-Krejci and Sabbas 2002). Suburban Alba North Residential Reservoir (Operations 16E, 16F, 16G, 16H, 16I, 16K) The second small depression was located 130 m northeast of Structure A-1 of the Alba Group, near a series of nearby house mounds that had not yet been excavated. Three separate operations were initiated across the center and along the north rim of the feature. Sascab exposed at the base of the depression was flat and smooth and was overlaid by a thick layer of sediment (Figure 9). The depression likely originated as a quarry associated with the construction of the nearby massive Alba Group monumental architecture, before becoming partly filled with inwashing sediment (C3 and C4 horizons) and sealed with a now highly degraded plaster floor (C2 horizon) and further filled by accumulating reservoir sediments (C1 and AC horizons). Although 107

117 not representative of the type of well-constructed and maintained watertight seals overlying the bedrock in the larger central precinct reservoirs found within many Maya urban centers, this plaster surface would nonetheless have served to greatly increase water retention within the reservoir. The profiles of Operations 16E and 16F revealed a similar deposition of reservoir sediments, but evidence of this former floor was visible only in 16F, located in the center of the depression. Contiguous operations 16G, 16H, 16I, and 16K, located on the depression s north rim, presented another interesting aspect of the tank. Large rocks exposed throughout this unit appear to be the result of deliberate cultural activity, based on their placement above the C soil horizon and their alignment (Figure 10). In addition, the collapse of these stones also appears to have dislodged other unnaturally placed stones. The type and nature of the feature that these stones would have comprised remains unknown, although it is evident that some type of construction activity was taking place around the perimeter of the reservoir. Excavation findings combined with a high density of ceramic material (N=610) and a proximate spatial association with the Alba Complex support the idea that this depression functioned as a reservoir. The reservoir s surface area and calculated volume combined with the topography of the surrounding landscape which could have shed water downslope into the reservoir from a large catchment surrounding it on the west, south, and east sides further support the feature s water storage capabilities. Although the neighboring residential groups have yet to be excavated it is likely that ceramics recovered in the depression will coincide with dates associated with these structures, supporting the idea that the reservoir was in use concurrent with occupation in this area of the site. 108

118 Suburban Fidelia North Residential Reservoir (Operation 16J) The only depression not located on the Alba-Fidelia transect, this small tank was located 130 m north of Structure F-1 of the Fidelia Complex. An L-shaped mound group is located immediately adjacent to the northern edge of the reservoir and the working hypothesis, based on its layout and recovered ceramic dates from nearby test excavations, was that this location was the likely setting for an Early Classic period farmstead. Another apparent reservoir that contains a possible sluice gate or dam feature is located 80 m to the south. Earlier investigations in this area included a 2 x 2 m excavation unit placed between these two depressions, from which DNA samples were recovered from soil containing both cotton and maize indicators (Anaya 2013). A water retention feature in the south depression could have served to allow and regulate the release of water into this field separating the two tanks, with the agricultural water runoff then funneling into the Suburban Fidelia North reservoir. This collected runoff could have then been recycled for non-potable uses such as additional pot irrigation. The current shallow profile of the small reservoir is the result of over a thousand years of sedimentation and soil formation following its use and abandonment. A series of exposed stones that appear to be aligned possibly indicating the remains of a cultural feature are located along the western margin of the depression, which also forms its highest present edge. Within the depression, evidence of a decomposing floor and superadjacent reservoir sediments like those examined in the Suburban Alba North reservoir are visible in the profile (Figure 11). Nearly 150 cm below the surface, the C3 horizon consists of weathered remains of a floor constructed of hard-tamped clay mixed with sandy sascab, a technique still used today to line irrigation ditches and other hydrologic features in the Yucatán (Dunning et al. 2013b). Above this surface, a thick layer of clayey reservoir sediments (C2 horizon) accumulated over time, with broken ceramics 109

119 and other artifacts intermixed. Episodic seasonal desiccation of these sediments appears evident in the development of slickensides, a product of shrinking and swelling clay, which also caused distortion in the underlying floor. During the final years of its use, or post-abandonment, the depression then began to fill with colluvial sediment (C horizon) derived from the disintegration of the adjacent residential complex, out of which the modern soil that has developed over a millennium. Two small charcoal fragments were recovered from the southwest corner of the unit at a depth approximately 1.55 m below the surface and submitted for C14 dating. This analysis returned a date of A.D (+/- 30), correlating strongly with the presence of Middle Classic ceramic material also present at this depth (Walker 2016). Similar to the Suburban Alba North reservoir, dates attached to recovered ceramics coincided with dates associated with the neighboring residential complex, indicating that this reservoir was in use during the occupation of this area of the site. This relatively small, but originally deep, depression appears to have functioned in a water management capacity based on multiple spatial, physical, and cultural characteristics. The depression s position immediately adjacent to an agricultural field (with another probable reservoir located on the west side of the field) and a structure group; its depth combined with the presence of a weathered plaster floor and evidence of intermittent desiccation/rehydration cycles in the stratigraphy; and high densities of ceramic material (N=208) throughout the unit suggestive of cultural activity all support the notion that this depression served as a household reservoir, possibly to collect water from, or distribute water to, the bordering agricultural field. Yax-3 Hinterland Aguada (Operation 22A) 110

120 The fourth depression excavated at Yaxnohcah is located roughly 500 m northeast of the Yax-3 group s center. The size (30 m N/S x 30 m E/W) and roughly circular shape of this feature resembles a natural karst sinkhole with a clay plug in its bottom (Nick Dunning, personal communication 2016). Unlike the other small tanks tested, this depression contained vegetation indicative of the presence of water for at least part of the year (Figure 12). Dense stands of saw grass and sedges, including Cyperus and Rhynchospora species and the cattail Typha dominguensis species were documented. These species have been associated with seasonallyinundated aguadas elsewhere in the Maya area, including the sites of Tikal (Dunning et al. 2015) and Pulltrouser Swamp (Darch 1983). The fact that the depression was currently completely dry is not surprising considering investigations took place in mid-may, near the end of the annual dry season. Its profile as an active aguada was further supported by the presence of modern artifacts noted around the perimeter including a chiclero (gum tapper) boiling pot, multiple glass bottles, and a shoe suggesting recent historical use of the feature. Two small metal fragments were collected 70 cm below ground surface, also evident of recent activity. The nearest ancient structure to the aguada, located 30 m west, has not been investigated but falls well within an accessible radius to take advantage of this reservoir. Additional residential groups lie 100 m northeast and 125 m southwest. The depression is oval-shaped and currently shallow, due to continual sedimentation and soil formation. A rise or rim is evident along the south edge, increasing significantly from west to east around the aguada s perimeter. This rim is also visible, but less defined, on the north side of the feature. A single 1 m x 1 m unit was placed in the center and excavations reached a total depth of nearly 2 m below present ground surface. Although no surface water was present, massive surface cracks visible in the stratigraphic profile are indicative of recent soil 111

121 dehydration. These cracks continue over 50 cm into the profile through an increasingly sticky and moist clay matrix. The modern soil is a Vertisol exhibiting significant shrink-swell behavior and subsequent clay movement (argilliturbation) over time. Aguadas have proven to be important sources of paleoecological data, with the sediments within often deposited in standing water with anaerobic conditions leading to the preservation of pollen and other microfossils (Akpinar et al. 2012; Beach et al. 2015b). In an effort to identify ancient vegetation present in the localized area surrounding the aguada, 15 pollen samples were collected from the south wall of the unit at 10 cm intervals between depths of 20 and 150 cm below ground surface. Analysis of these samples is ongoing. A total of 94 ceramic sherds were recovered, tentatively dated to the Classic period (Debra Walker, personal communication 2016). Although the density of ceramic material was relatively consistent throughout most the excavation unit, no artifacts were encountered in the dense, sticky clay (approximately 20 cm deep) overlying the sascab base of the aguada. Combined with a lack of cultural material at this depth, the composition of this thick bajo clay suggests that it acts as a natural aid in water retention. A pronounced topographical depression represents a natural place to collect water and, as has been reported elsewhere in the Maya lowlands, aguadas may begin to collect water due to the natural consequences of sedimentation by clays (Beach et al. 2015b; Dunning et al. 2002). Whether deliberately deposited or not, the clay lining within the Yax-3 hinterland aguada would have increased the reservoir s ability to store water and supplement ground and surface water, or provide water in dry periods. Although lacking evidence of a former floor or sealed surface, the present vegetation and significant surface cracks are indicative of a current water feature. Based on associated ceramic dates, this feature appears to have originated naturally and, after partially filling with in-washing clay 112

122 reservoir sediments, served as a reservoir beginning in the Classic period. It is possible that the reservoir was in use earlier, with antecedent sediment and associated ceramics having been removed by later (Classic period) dredging. Many aguadas have stratigraphic disjunctions as a result of periodic ancient dredging (e.g. Dunning et al. 2015). Yax-3 Group Reservoir (Operation 22C) The final small water feature investigated to-date at Yaxnohcah is located within the Yax-3 residential group currently being investigated by YAP (Anaya et al 2016). Massive exposed limestone blocks visible around the feature s rim, particularly along its south and west boundaries, and the presence of multiple structures within 75 m of the reservoir (including a large structure bordering its eastern edge) indicate that this depression likely originated as a quarry associated with the construction of this residential complex before being modified and used as a small reservoir. Unfortunately, excavation could not be carried to bedrock in this depression because of time constraints and safety concerns. Excavation of the depression produced a wealth of cultural and physical data, including a significant density of ceramic material. This consisted of multiple shallow vessel rim sherds, including one large Polvero Black bowl rim fragment with slip still largely intact. The ceramic mix from this operation tentatively dates from the Late Preclassic to Late Classic period (Debra Walker, personal communication 2016). A charcoal sample recovered from the south wall (125 cm below ground surface) returned a 2-sigma calibrated age range of BC (+/- 20), lending further support for use of the reservoir during the Late Preclassic. The weathered remains of two distinct surfaces visible in the profile represent multiple periods of use and refurbishment of the reservoir that appears to coincide with this range of ceramic dates (Figure 113

123 13). Atop a layer of in-washed sediment, the 5C horizon consists of a now highly weathered plaster floor beneath a thick dark layer of reservoir sediments accumulated along with cultural material (4C horizon). These sediments also contained manganese and iron oxide concretions, indicative of the presence of abundant soil water at this depth of the depression. Above this is a degraded floor composed of a mixture of very dark hard-tamped clay and sandy sascab similar to the former surface uncovered in Op. 16J. Another layer of reservoir sediments (2C horizon) and colluvium sediment resulting from the disintegration of the adjacent residential complex are located beneath the well-developed modern soil. Taken together, the physical and cultural data recovered from this operation indicates that the depression likely originated as a limestone quarry during the Late Preclassic, or perhaps earlier. After subsequently filling with a layer of in-washing sediment, the tank was sealed with a plaster floor for use as a reservoir during this period. Following the accumulation of additional reservoir sediments, perhaps indicating a period of disuse or abandonment in this area of the site, the tank was again sealed with a clay-and-sascab surface to improve its water retention. The presence of a similarly constructed floor in the Operation 16J tank north of Fidelia also dating to the Classic period suggests this technique was employed as a low-cost (relative to plaster) solution for seasonal water storage throughout Yaxnohcah during this time. DISCUSSION Investigations of residential-scale water management infrastructure at Medicinal Trail and Yaxnohcah have revealed that these small reservoirs and associated landscape modifications served various purposes within their communities. These studies suggest that many small depressions likely originated as quarries before being modified for water storage and agricultural 114

124 uses. Small household reservoirs, modified with a clay or plaster lining and combined with the use of terraces, berms, and silting tanks, represented a sustainable engineering solution for the ancient Maya to help store water in a seasonally water-deficient environment. In a region largely devoid of significant permanent water sources, small reservoirs were vital components in the daily lives of their community s residents and undoubtedly served a variety of potable and non-potable needs. For researchers seeking to understand the lifeways and cultural activities of an area s inhabitants, these features have proven to be important sources of cultural and environmental data. The study of aguadas provides archaeologists with the added benefit of the paleoenvironmental data they preserve, including ancient pollen, charcoal, and sedimentary evidence of past climatic drying. Pollen sequences gathered from aguadas reflect local vegetation, as these small bodies of water can capture the localized pollen of less-common plant species whose pollen does not travel far (Akpinar-Ferrand et al. 2012). As a result, the pollen extracted from aguadas particularly when combined with supporting material data could also prove valuable in revealing the function of these hydrologic features within ancient Maya agriculture. Although the majority of completely dry residential tanks unfortunately lack useful pollen indicators, the evidence they provide in terms of anthropogenic modifications to enhance their water holding capabilities, their physical profile (surface area, depth, and volume), and their spatial and temporal integration into communities supports the theory that they were employed similarly to aguadas to fulfill the water needs of the ancient Maya. A series of water management studies conducted at Medicinal Trail revealed a multicomponent system of water features incorporated into a dispersed hinterland community engaged in agricultural production. Located within a few kilometers of the major site of La Milpa, this community consists of multiple formal courtyard groups, numerous informal mound clusters, 115

125 and multiple landscape modifications including depressions, terraces, and berms (Hyde 2011). In their discussion of dualistic economies, emerging nation states, and the ancient economy of the lowland Maya, Scarborough and Valdez (2009) propose the formation of resource-specialized communities, of which Medicinal Trail might be classified. These settlements were linked in a largely interdependent matrix that permitted a certain amount of self-sustaining separation from the large monumental centers with which they were spatially associated and about which we have the greatest knowledge. Similar to minor sites found elsewhere in the Three Rivers Region such as Chawak But o ob (Walling et al. 2005, 2006) and Grupo Agua Lluvia (Trachman 2007), Medicinal Trail is characterized as a small community of relatively diminutive civic architecture, organized around a defined center, and engaged in low-cost water management engineering. Located on an escarpment edge, the site contains the centrally located Medicinal Trail Reservoir and a series of landscape modifications geared toward domestic water collection and storage. The capacity of the Medicinal Trail Reservoir alone, 153 m 3, would have supplied the households around Group A with a significant amount of water throughout much of the year. Combined with the additional supply provided by the other, smaller reservoirs, Medicinal Trail and its water supply would likely have functioned as a centralizing node for resource access among intercommunity members during the dry season and perhaps beyond. Despite its importance as a specialized resource community, however, and like other resource specialized communities in the region, Medicinal Trail never expanded beyond a hinterland companion to neighboring La Milpa. The political, economic, and social gravity of this major center would have attracted tribute, trade, and ceremonial interactions with the residents of Medicinal Trail some of which would likely have involved water requirements. As Lucero (2006) details, traditional rituals performed on a grand scale including prayers for 116

126 abundant water and various dedication, ancestor veneration, termination, and other household and community rites were a primary means that ruling elites used to bring together Maya farmers and other commoners during the Classic period. This elite-commoner relationship, centered in part on the importance of water collection and management, would have served as an important link between the agricultural community of Medicinal Trail and the urban center of La Milpa. Nonetheless, this investment in water management at the household scale permitted the social and economic interrelationships necessary for the community to organize and operate alongside other specialized communities largely independent from a large urban center. In contrast, Yaxnohcah emerged within a network of Preclassic cities, including Calakmul, Nakbe, and El Mirador, as a massive archaic city-state in the Central Karstic Uplands. The site grew into a large urban center early in the Preclassic period, survived the Preclassic collapse that engulfed some of its neighbors, and became a Classic period center of some significance before abandonment in the Terminal Classic (Anaya et al. 2016). The evolution and sustainability of Yaxnohcah was based on significant hydraulic, agricultural, settlement, and landscape modification strategies that allowed the persistent occupation of the site amid changing environmental and political economic conditions. The massive Reservario Brisa and Aguada Fidelia, likely constructed during the Middle and Late Preclassic periods, respectively, illustrate the sophistication of an urbanized reservoir-based water system that operated into the Classic period (Dunning et al. 2016). The system was established and modified over time to compensate for changes in population density and distribution around the urban core, fluctuations in water availability, and developments in political structure and control. This deliberate built environment adopted a comprehensive land-use strategy of sizable reservoirs designed to receive large quantities of surface water runoff flowing seasonally from the Fidelia 117

127 and Brisa Group plazas, pyramids, and palaces that typified Yaxnohcah during the Classic period. At the household, or decentralized, scale, hydrologic modifications in the form of creating and maintaining small water tanks took place in tandem with these community-wide, centralized investments in water management. The creation and use of smaller aguadas and household tanks by a growing hinterland population fanning out from the urban core, combined with the continued intensive exploitation of bajo areas as cultivated and managed wetland forests and agricultural zones, characterized residential scale water management activities at Yaxnohcah from the Middle Preclassic into the Classic period. Chultuns, identified through ground survey at multiple locations within the site, may also have served to store water for the residents of Yaxnohcah. Their role in water management has been documented throughout the Maya area, particularly in the Puuc Hills and northern Yucatán Peninsula (Gill 2000; Matheny 1982; McAnany 1990; Puleston 1971; Scarborough 1991, 2003; Scarborough and Gallopin 1991) CONCLUSION Small depressions are a recurrent landscape feature throughout the central Maya lowlands and systematic studies conducted over the past few decades have demonstrated that they served a variety of cultural functions, including water reservoirs. Most these small tanks are thought to have originated naturally within the topography as karst sinkholes and were then modified by the Maya to enhance water storage. In order to determine their origin and function a combination of approaches from surface observation, excavation, and artifact and soil analysis is required. Excavation data from a series of residential scale water features at the sites of Medicinal Trail and Yaxnohcah in the central Maya lowlands demonstrates that at least some of these features originated as limestone quarries prior to serving as household or small communal reservoirs. 118

128 Occupying an area largely devoid of lakes and rivers, the lowland Maya of the Preclassic and Classic periods met the challenges of the annual 4- to 5-month dry season by constructing water catchment systems to provide water that would last until the rains began again. In hinterland communities such as Medicinal Trail, these small-scale systems were built and maintained at the household level among families living in farmsteads dispersed throughout the jungle. Data from Yaxnohcah shows that within large urban centers, similar hydrologic infrastructure developed at the household level, indicating at least some degree of autonomy in individuals and families meeting their daily water needs throughout the year. Reservoirs, check dams, canals, terraces, and other water or agricultural subsistence technologies reflect local adaptive strategies as well as the number of people they can service. Small-scale systems are built and maintained at the household or community level and their location and scale typically reflect their role in social, economic, and political life. Additional studies of small water reservoirs in the Maya lowlands are necessary to further clarify the nature and extent of these water management adaptations. These findings will help illuminate possibly changing patterns of water management at the household scale, such as a reliance on decentralized instead of centralized sources, and how these relationships varied across space and time both adjacent to the urban core and within the more distant hinterlands. 119

129 SPANISH SUMMARY Este artículo discute los resultados de investigaciones descentralizadas de manejo de agua a escala nacional de dos sitios centrales de tierras bajas: la comunidad agrícola del interior de Medicinal Trail en el noroeste de Belice y el centro urbano de Yaxnohcah en el sur de Campeche, México. La investigación incluyó un solo reservorio y características adicionales del agua asociadas con dos grupos del patio residencial investigados en el rastro medicinal y una serie de cuatro depósitos asociados con las estructuras residenciales a través de Yaxnohcah. Se llevó a cabo un enfoque interdisciplinario consistente en encuestas de paisaje, excavación y análisis material de material cultural recolectado con el fin de comprender las prácticas de manejo de agua en las viviendas de los habitantes de estos dos sitios, especialmente su dependencia de tanques pequeños Reservorios más grandes, centralizados y controlados por elite. Las investigaciones de la infraestructura de gestión de agua a escala residencial en el Camino Medicinal y Yaxnohcah han revelado que estos pequeños embalses y las modificaciones de paisaje asociadas servían para diversos propósitos dentro de sus comunidades. Estos estudios sugieren que muchas pequeñas depresiones probablemente se originaron como canteras antes de ser modificadas para el almacenamiento de agua y usos agrícolas. Los pequeños embalses domésticos, modificados con un revestimiento de arcilla o yeso y combinados con el uso de terrazas, bermas y tanques de sedimentación, representaron una solución de ingeniería sostenible para los antiguos mayas para ayudar a almacenar el agua en un ambiente estacionalmente deficiente en agua. Las depresiones pequeñas son una característica recurrente del paisaje a través de las tierras bajas mayas centrales y los estudios sistemáticos realizados durante las últimas décadas 120

130 han demostrado que sirvieron a una variedad de funciones culturales, incluyendo reservorios de agua. Se cree que la mayoría de estos pequeños tanques se originaron naturalmente dentro de la topografía como sumideros cársticos y luego fueron modificados por los mayas para mejorar el almacenamiento de agua. Para determinar su origen y función se requiere una combinación de enfoques de observación superficial, excavación y análisis de artefactos y suelos. Datos de excavación de una serie de características de agua de escala residencial en los sitios de Medicinal Trail y Yaxnohcah en las tierras bajas mayas centrales demuestra que al menos algunas de estas características se originaron como canteras de piedra caliza antes de servir como residenciales o pequeños reservorios comunales. Los mayas de las tierras bajas de los períodos Preclásico y Clásico, que ocupaban un área en gran parte desprovista de lagos y ríos, resistieron los retos de la estación seca anual de 4 a 5 meses mediante la construcción de sistemas de captación de agua para proporcionar agua que duraría hasta que comenzaran las lluvias. En las comunidades del interior como el Camino Medicinal, estos sistemas a pequeña escala fueron construidos y mantenidos a nivel familiar entre familias que viven en granjas dispersas por toda la selva. Los datos de Yaxnohcah muestran que dentro de los grandes centros urbanos, una infraestructura hidrológica similar se desarrolló a nivel del hogar, indicando al menos cierto grado de autonomía en individuos y familias que satisfacen sus necesidades diarias de agua a lo largo del año. 121

131 ACKNOWLEDGEMENTS I would like to thank the anonymous reviewers, Nick Dunning, Chris Carr, and Vernon Scarborough for their valuable input and suggestions. Special thanks go to the local field workers who assisted with the archaeological excavations detailed here, particularly Javier Cobos, Augustín Diaz, and Reyes Pérez Martínez of YAP. The work reported in this article was funded by a National Science Foundation grant (PI: Nicholas P. Dunning, Co-PI: Jeffrey L Brewer; NSF Award Number ) and the Taft Research Center at the University of Cincinnati. The work was accomplished as part of the Programme for Belize Archaeological Project directed by Fred Valdez, Jr. and the Proyecto Arqueológico de Yaxnohcah directed by Kathryn Reese-Taylor and Armando Anaya Hernández. 122

132 REFERENCES Adams, R. E. W Nucleation of Population and Water Storage Among the Ancient Maya. Science 251(4994):632. Adams, R. E. W. and Richard C. Jones 1981 Spatial Patterns and Regional Development Among Classic Maya Cities. American Antiquity 46: Adams, Richard E. W. and Fred Valdez Jr The Programme for Belize (PfB) Archaeological Project: Report of Field Activities, University of Texas at San Antionio. Adams, Richard E. W., Vernon Scarborough, Laura Levi, Stanley Walling, Nicholas Dunning, Brandon Lewis, Leslie Shaw, Eleanor King, Lauren Sullivan, Kathryn Reese-Taylor, and Fred Valdez, Jr Programme for Belize Archaeological Project: A History of Archaeological Research. In Research Reports in Belizean Archaeology, Vol. 1, edited by Jaime Awe, John Morris, and Sherilyne Jones, pp Institute of Archaeology, National Institute of Culture and History, Belmopan, Belize. Akpinar-Ferrand, Ezgi, Nicholas P. Dunning, David L. Lentz, and John G. Jones 2012 Use of Aguadas as Water Management Sources in Two Southern Maya Lowland Sites. Ancient Mesoamerica 23: Anaya Hernández, Armando 2013 Pozos de Sondeo. In Proyecto Arqueológico Yaxnohcah, 2011 Informe de la Primera Temporada de Investigaciones, edited by Kathryn Reese-Taylor and Armando Anaya Hernández, pp Mesoweb. Anaya Hernández, Armando, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor 2016 Introducción. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor, pp Mesoweb. Beach, Timothy and Nicholas Dunning 1997 An Ancient Maya Reservoir and Dam at Tamarindito, el Peten, Guatemala. Latin American Antiquity 8(1): Beach, Timothy, Sheryl Luzzadder-Beach, Nicholas Dunning, Jon Hageman, and Jon Lohse 2002 Upland Agriculture in the Maya Lowlands: Ancient Maya Soil Conservation in Northwestern Belize. Geographical Review 92(3): Beach, Timothy, Nicholas Dunning, Sheryl Luzzadder-Beach, D. E. Cook, and Jon Lohse 2006 Impacts of the Ancient Maya on Soils and Soil Erosion in the Central Maya 123

133 Lowlands. Catena 65: Human and Natural Impacts on Fluvial and Karst Depressions of the Maya Lowlands. Geomorphology 101: Beach, Tim, Sheryl Luzzadder-Beach, Duncan Cook, Nicholas Dunning, Douglas J. Kennett, Samantha Krause, Richard Terry, Debora Trein, and Fred Valdez 2015a Ancient Maya impacts on the Earth s surface: An Early Anthropocene analog? Quaternary Science Reviews 124:1-30. Beach, Timothy, Sheryl Luzzadder-Beach, Jonathan Flood, Stephen Houston, Thomas G. Garrison, Edwin Román, Steve Bozarth, and James Doyle 2015b A Neighborly View: Water and Environmental History of the El Zotz Region. In Tikal: Paleoecology of an Ancient Maya City, edited by David L. Lentz, Nicholas P. Dunning, and Vernon L. Scarborough, pp Cambridge University Press, Cambridge. Becquelin, Pierre and Dominique Michelet 1994 Demografia en la zona Puuc: el recurso del método. Latin American Antiquity 5(4): Brenner, Mark, David A. Hodell, Jason H. Curtis, Michael F. Rosenmeier, Flavio S. Anselmetti, and Daniel Ariztegui 2003 Paleolimnological Approaches for Inferring Past Climate Change in the Maya Region: Recent Advances and Methodological Limitations. In The Lowland Maya Area, edited by Arturo Gomez-Pompa, Michael Allen, Scott L. Fedick, and Juan J. Jimenez-Osomio, pp Food Products Press, New York. Brewer, Jeffrey L 2007 Understanding the Role of a Small Depression in Ancient Maya Water Management at the Medicinal Trail Site, Northwestern Belize. Master's thesis, Department of Anthropology, Division of Graduate Studies and Research of the University of Cincinnati Investigaciones en Aguadas de Escala Residenciales. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan PeuramakiBrown, and Kathryn Reese-Taylor, pp Mesoweb. Brewer, Jeffrey L and Christopher Carr Forthcoming Operation 22: Continuing Household Water Management Investigations at Yaxnohcah. In Proyecto Arqueológico Yaxnohcah, Informe de la 2016 Temporada de Investigaciones, edited by Armando Anaya Hernández, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor. Mesoweb. Brewer, Jeff, David M. Hyde, and Michael Stowe 124

134 2013 Mapping Medicinal Trail: A Summary From 2004 to In Research Report from the Programme for Belize Archaeological Project, Volume 7, edited by Fred Valdez, Jr., pp Occasional Papers, Number 14. Mesoamerican Archaeological Research Laboratory, University of Texas at Austin. Brewer, Jeffrey L, Christopher Carr, Nicholas P. Dunning, Debra S. Walker, Armando Anaya Hernández, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor Forthcoming Employing Airborne Lidar and Archaeological Testing to Determine the Role of Small Depressions in Water Management at the Ancient Maya site of Yaxnohcah, Campeche, Mexico. Journal of Archaeological Science:Reports. Bullard, Jr., William R Maya Settlement Pattern in Northeastern Peten, Guatemala. American Antiquity 25(3): Chase, Adrian S. Z Beyond elite control: residential reservoirs at Caracol, Belize. WIREs Water 3: Chase, Arlen F., Diane Z. Chase, John F. Weishampel, Jason B. Drake, Ramesh L. Shrestha, K. Clint Slatton, Jaime J. Awe, and William E. Carter Airborne LiDAR, Archaeology, and the Ancient Maya Landscape at Caracol, Belize. Journal of Archaeological Science 38: Chmilar, Jennifer A Ancient Maya Water Management: Archaeological Investigations at Turtle Pond, Northwestern Belize. Master's thesis, Department of Anthropology, Division of Graduate Studies and Research of the University of Cincinnati. Culbert, T. Patrick, Laura J. Levi, and Luis Cruz 1990 Lowland Maya Wetland Agriculture: The Rio Azul Agronomy Program. In Vision and Revision in Maya Studies, edited by Flora S. Clancy and Peter D. Harrison, pp University of New Mexico Press, Albuquerque. Darch, Janice P Vegetation Associations at Pulltrouser Swamp. In Pulltrouser Swamp: Ancient Maya Habitat, Agriculture, and Settlement in Northern Belize, edited by B. L. Turner, II and Peter D. Harrison, pp University of Texas Press, Austin. Domínguez Carrasco, María del Rosario and William J. Folan 1996 Calakmul, México: Aguadas, Bajos, Precipitación y Asentamiento en el Petén Campechano. In IX Simposio de Investigaciones Arqueológicas en Guatemala, edited by Juan Pedro Laporte and Hector L. Escobedo, pp Museo Nacional de Antropología y Etnología, Guatemala. Dunning, Nicholas P. and Timothy Beach 1994 Soil Erosion, Slope Management, and Ancient Terracing in the Maya 125

135 lowlands. Latin American Antiquity 5(1): Fruit of the Luum: Lowland Maya Soil Knowledge and Agricultural Practices. Mono y Conejo 2: Farms and Forests: Spatial and Temporal Perspectives on Ancient Maya Landscapes. In Landscapes and Societies: Selected Cases, edited by I. Peter Martini and Ward Chesworth, pp Springer, Dordrecht, Netherlands. Dunning, Nicholas P., Timothy P. Beach, Patrice Farrell, and Sheryl Luzzadder-Beach 1998 Prehispanic Agrosystems and Adaptive Regions in the Maya Lowlands. Culture and Agriculture 20: Dunning, Nicholas, Vernon Scarborough, Fred Valdez, Jr, Sheryl Luzzadder-Beach, Timothy Beach, and John G. Jones 1999 Temple Mountains, Sacred Lakes, and Fertile Fields: Ancient Maya Landscapes in Northwestern Belize. Antiquity 73: Dunning, Nicholas P., Sheryl Luzzadder-Beach, Timothy Beach, John Jones, Vernon Scarborough, and Patrick Culbert 2002 Arising from the Bajos: The Evolution of a Neotropical Landscape and the Rise of Maya Civilization. Annals of the Association of American Geographers 92: Dunning, Nicholas P., Timothy Beach, and Sheryl Luzzadder-Beach 2006 Environmental Variability Among Bajos in the Southern Maya Lowlands and its Implications for Ancient Maya Civilization and Archaeology. In Precolumbian Water Management, edited by Lisa J. Lucero and Barbara W. Fash, pp The University of Arizona Press, Tucson. Dunning, Nicholas P., Ezgi Akpinar, Chris Carr, Robert Griffin, John G. Jones, David L. Lentz, Andrew Miller, and Jasmin Prater 2007 Investigaciones geoarqueológicas y el medioambiente en los alrededores de San Bartolo, Peten. In Proyecto Arqueológico San Bartolo: Informe Preliminar No. 6, Sexta Temporada, edited by M. Urquizú and W. Saturno, Dirección de Patrimonio Cultural y Natural de Guatemala, Guatemala City. Dunning, Nicholas P., David Wahl, Timothy P. Beach, John G. Jones, Sheryl Luzzadder-Beach, and Carmen McCane 2012 Environmental Instability and Human Response in the Late Preclassic EastCentral Yucatan Peninsula. In The Great Maya Droughts in Cultural Context, edited by Gyles Iannone, University of Colorado Press, Boulder. Dunning, Nicholas P., Timothy Beach, Liwy Grazioso Sierra, John G. Jones, David L. Lentz, Sheryl Luzzadder-Beach, Vernon L. Scarborough, and Michael P. Smyth 2013a Historia de dos colapsos: Variabilidad ambiental e interupción cultural en las 126

136 Tierras Bajas mayas. Dialogo Andino 41: Dunning, Nicholas, Eric Weaver, Michael Smyth, and David Ortegon Zapata 2013b Xcoch: Home of Maya Rain Gods and Water Managers. In The Archaeology of Yucatan: New Directions and Data, edited by T. W. Stanton. BAR International Series, Oxford. Dunning, Nicholas P., Robert E. Griffin, John G. Jones, Richard E. Terry, Zachary Larsen, and Christopher Carr 2015 Life on the Edge: Tikal in a Bajo Landscape. In Tikal: Paleoecology of an Ancient Maya City, edited by David L. Lentz, Nicholas P. Dunning, and Vernon L. Scarborough, pp Cambridge University Press, Cambridge. Dunning, Nicholas P., Armando Anaya Hernández, and Helga Geovannini 2016 Operaciones 13 y 19: Investigaciones en los Reservorios. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by Armando Anaya Hernández, Meaghan Peuramaki-Brown, and Kathryn Reese-Taylor, pp Mesoweb. Farnand, Danica M Agricultural Formation Histories of Prehistoric Terraces of the Medicinal Trail Site, Northwestern Belize. Master's thesis, Department of Anthropology, Division of Graduate Studies and Research of the University of Cincinnati. Ferries, Laura C Site Formation and Occupation History of the Medicinal Trail House Mound Group at the Programme for Belize Archaeological Project, Belize. Master's thesis, Department of Anthropology, Division of Research and Advanced Studies of the University of Cincinnati. Folan, William J Mining and Quarrying Techniques of the Lowland Maya. Anthropology 6: Folan, William J., Joyce Marcus, Sophia Pincemin, María del Rosario Domínguez Carrasco, Laraine Fletcher, and Abel Morales Lopez 1995 Calakmul: New Data from an Ancient Maya Capital in Campeche, Mexico. Latin American Antiquity 6(4): Ford, Anabel 1996 Critical Resource Control and the Rise of the Ancient Maya. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use, edited by Scott L. Fedick, pp University of Utah Press, Salt Lake City. Garrison, Thomas G Remote Sensing Ancient Maya Rural Populations Using QuickBird Satellite 127

137 Imagery. International Journal of Remote Sensing 31(1): Geovannini Acuña, Helga 2008 Rain Harvesting in the Rainforest: The Ancient Maya Agricultural Landscape of Calakmul, Campeche, Mexico. BAR International Series, Oxford El Paisaje Natural del Sitio Arqueológico de Yaxnohcah. In Proyecto Arqueológico Yaxnohcah, 2013 Informe de la Segunda Temporada de Investigaciones, edited by Kathryn Reese-Taylor and Armando Anaya Hernández, pp Mesoweb. Gill, Erin E Operation 15: Berm Structures and Water Management at Medicinal Trail (RB 62), Belize. In Research Reports from the Programme for Belize Archaeological Project, Volume 3, edited by Rissa M. Trachman and Fred Valdez, Jr., pp Occasional Papers, Number 10. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Goldstein, David and Gardner Brandt 2007 Report on Flotation Sample #367 from 2006 PFB Fieldwork. Manuscript on file, Department of Geography, University of Cincinnati, Cincinnati, Ohio. Goldstein, David J. and Jon B. Hageman 2010 Power Plants: Paleobotanical Evidence of Rural Feasting in Late Classic Belize. In Food and Feasting in Mesoamerican Civilization: Interdisciplinary Approaches to Food, Culture, and Markets in Ancient Mesoamerica, edited by John E. Staller and Michael Carrasco, pp Springer, New York. Guderjan, Thomas H The Nature of an Ancient Maya City: Resources, Interaction, and Power at Blue Creek, Belize. The University of Alabama Press, Tuscaloosa. Gunn, Joel D., William J. Folan, Maria del Rosario Domínguez-Carrazco, and Betty J. Faust 2002 Bajo Sediments and the Hydraulic Landscape of Calakmul, Mexico. Ancient Mesoamerica 13: Hageman, Jon B. and Jon C. Lohse 2003 Heterarchy, Corporate Groups, and Late Classic Resource Management in Northwestern Belize. In Heterarchy, Political Economy, and the Ancient Maya, edited by Vernon L. Scarborough, Fred Valdez Jr., and Nicholas Dunning, pp The University of Arizona Press, Tucson. Hageman, Jon B. and David J. Goldstein 2009 An Integrated Assessment of Archaeobotanical Recovery Methods in the Neotropical Rainforest of Northern Belize: Flotation and Dry Screening. Journal of Archaeological Science 36:

138 Hammond, Norman and Wendy Ashmore 1981 Lowland Maya Settlement: Geographical and Chronological Frameworks. In Lowland Maya Settlement Patterns, edited by Wendy Ashmore, pp University of New Mexico Press, Albuquerque. Hammond, Norman, Gair Tourtellot, Sara Donaghey, and Amanda Clarke 1998 No Slow Dusk: Maya Urban Development and Decline at La Milpa, Belize. Antiquity 72(278): Hansen, Richard D Continuity and Disjunction: The Pre-Classic Antecedents of Classic Maya Architecture. In Function and Meaning in Classic Maya Architecture, edited by Stephen D. Houston, pp Dumbarton Oaks, Washington, D.C. Hansen, Richard D., Steven Bozarth, John Jacob, David Wahl, and Thomas Schreiner 2002 Climatic and Environmental Variability in the Rise of Maya Civilization: A Preliminary Perspective from Northern Peten. Ancient Mesoamerica 13: Harrison, Peter D Aspects of Water Management in the Southern Maya Lowlands. In Economic Aspects of Water Management in the Prehispanic New World, edited by Vernon L. Scarborough and Barry L. Isaac, pp Research in Economic Anthropology, supplement No. 7, JAI Press, Greenwich, Connecticut. Hester, Thomas R. and Harry J. Shafer 1984 Exploitation of Chert Resources by the Ancient Maya of Northern Belize. World Archaeology 16(2): Houk, Brett A. (editor) 2000 The 1998 and 1999 Seasons of the Chan Chich Archaeological Project. Papers of the Chan Chich Archaeological Project, No. 4. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Hughbanks, Paul J Research at Guijarral (RB-18), In The Programme for Belize Archaeological Project 1994 Interim Report, edited by Richard E. W. Adams and Fred Valdez, Jr., pp The Center for Archaeology and Tropical Studies, The University of Texas at San Antonio. Hyde, David M Report of the Stone Tool Assemblage at the Medicinal Trail Site, Operation 10. Manuscript on file, Mesoamerican Archaeological Research Laboratory, University of Texas at Austin The 2008 Field Season at Group A of the Medicinal Trail Site, Northwestern 129

139 Belize. In Research Reports from the Programme for Belize Archaeological Project, Volume 3, edited by Rissa M. Trachman and Fred Valdez, Jr., pp Occasional Papers, Number 10. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin Power Dynamics at a Commoner Hinterland Community in the Maya Lowlands: The Medicinal Trail Site, Northwestern Belize. Ph.D. dissertation, Department of Anthropology, The University of Texas at Austin. Hyde, David M. and Fred Valdez, Jr Overview of Activities at the Medicinal Trail Site for the 2006 Season. In Research Reports from the Programme for Belize Archaeological Project, edited by Fred Valdez, Jr., pp Occasional Papers, Number 8. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Hyde, David M. and Kirsten Atwood 2008 Excavations at Group A of the Medicinal Trail: Results from the 2007 Investigations. In Research Reports from the Programme for Belize Archaeological Project, Vol. 2, edited by Fred Valdez, Jr., pp Occasional Papers Number 9. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Isendahl, Christian 2011 The Weight of Water: A New Look at Pre-Hispanic Puuc Maya Water Reservoirs. Ancient Mesoamerica 22: Johnston, Kevin J Lowland Maya Water Management Practices: The Household Exploitation of Rural Wells. Geoarchaeology 19(3): Johnston, Kevin J., Andrew J. Breckenridge, and Barbara C. Hansen 2001 Paleoecological Evidence of an Early Postclassic Occupation in the Southwestern Maya Lowlands: Laguna Las Pozas, Guatemala. Latin American Antiquity 12(2): Kunen, Julie L Ancient Maya Life in the Far West Bajo: Social and Environmental Change in the Wetlands of Belize. University of Arizona Press, Tucson. Lohse, Jon C Intra-Site Settlement Signatures and Implications for Late Classic Maya Commoner Organization at Dos Hombres, Belize. In Ancient Maya Commoners, edited by Jon C. Lohse and Fred Valdez, Jr., pp University of Texas Press, Austin. Lohse, Jon C. and Patrick N. Findlay 130

140 2000 A Classic Maya House-Lot Drainage System in Northwestern Belize. Latin American Antiquity 11(2): Lohse, Jon C., Jimmy Barrera, and Antonio Padilla Season Excavations in the Gran Cacao Ballcourt, Northwestern Belize. In 2004 Season Summaries of the Blue Creek Political Ecology Project, edited by Jon C. Lohse and Kerry L. Sagebiel, pp Maya Research Program, Ft. Worth, and the University of Texas at Austin. Lowe, John D Excavations of Two Small Depressions at Group A, The Medicinal Trail Site RB62, Belize. Unpublished Master's thesis, Department of Anthropology, The University of Texas at Austin. Lucero, Lisa J Water Control and Maya Politics in the Southern Maya Lowlands. In Complex Polities in the Ancient Tropical World, edited by Elisabeth A. Bacus and Lisa J. Lucero, pp Archaeological Papers of the American Anthropological Association Number 9, American Anthropological Association, Arlington, Virginia The Collapse of the Classic Maya: A Case for the Role of Water Control. American Anthropologist 104(3): Water and Ritual: The Rise and Fall of Classic Maya Rulers. University of Texas Press, Austin. Lucero, Lisa J., Joel D. Gunn, and Vernon L. Scarborough 2011 Climate Change and Classic Maya Water Management. Water 3: McAnany, Patricia C Water Storage in the Puuc Region of the Northern Maya: A Key to Population Estimates and Architectural Variability. In Precolumbian Population History in the Maya Lowlands, edited by P.T. Culbert and D. Rice, University of New Mexico Press, Albuquerque. Matheny, Ray T Maya Lowland Hydraulic Systems. Science 193(4254): Northern Maya Lowland Water-Control Systems. In Pre-Hispanic Maya Agriculture, edited by Peter D. Harrison and B.L. Turner II, pp University of New Mexico Press, Albuquerque Ancient Lowland and Highland Maya Water and Soil Conservation Strategies. In Maya Subsistence: Studies in Honor of Dennis E. Puleston, edited by Kent V. Flannery, pp Academic Press, New York. 131

141 Morley, Sylvanus 1937 Inscriptions of the Peten. Vol. 1. Carnegie Institution of Washington, Publication 437. Washington, D.C. Percy, Madelyn 2009 Excavation at the Medicinal Trail Site (2008): Operation 14. In Research Reports from the Programme for Belize Archaeological Project, Volume 3, edited by Rissa M. Trachman and Fred Valdez, Jr., pp Occasional Papers, Number 10. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Peuramaki-Brown, Meaghan, Shawn G. Morton, and Ana Castro 2016 Operacíon 17 en el Grupo Wo, Excavaciones de Prueba en un Lote Doméstico Peri-Urbano. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, pp Mesoweb. Pohl, Mary D., Kevin O. Pope, John G. Jones, John S. Jacob, Dolores R. Piperno, Susan D. defrance, David L. Lentz, John A. Gifford, Marie E. Danforth, and J. Kathryn Josserand Early Agriculture in the Maya Lowlands. Latin American Antiquity 7(4): Pope, Kevin O. and Bruce H. Dahlin 1989 Ancient Maya Wetland Agriculture: New Insights from Ecological and Remote Sensing Research. Journal of Field Archaeology 16(1) Puleston, Dennis E An Experimental Approach to the Function of Classic Maya Chultuns. American Antiquity 36(3): Ruppert, Karl and John H. Dennison 1943 Archaeological Reconnaissance in Campeche, Quintana Roo, and Petén. Carnegie Institution of Washington, Publication 543. Washington, D.C. Saturno, William, Thomas L. Sever, Daniel E. Irwin, Burgess F. Howell, and Thomas G. Garrison 2007 Putting Us on the Map: Remote Sensing Investigation of the Ancient Maya Landscape. In Remote Sensing in Archaeology: Interdisciplinary Contributions to Archaeology, edited by J.R. Wiseman and F. El-Baz, pp Springer, New York. Scarborough, Vernon L Water Management Adaptations in Nonindustrial Complex Societies: An Archaeological Perspective. In Archaeological Method and Theory, Vol. 3, edited by Michael B. Schiffer, pp University of Arizona Press, Tucson. 132

142 1993 Water Management in the Southern Maya Lowlands: An Accretive Model for the Engineered Landscape. In Economic Aspects of Water Management in the Prehispanic New World, Vol. Research in Economic Anthropology, supplement No. 7, edited by V.L. Scarborough and B.L. Isaac, pp JAI Press, Greenwich, Connecticut Maya Water Management. National Geographic Research and Exploration Reservoirs and Watersheds in the Central Maya Lowlands. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use, edited by S.L. Fedick, pp University of Utah Press, Salt Lake City The Flow of Power: Ancient Water Systems and Landscapes. SAR Press, Santa Fe. Scarborough, Vernon L. and Gary G. Gallopin 1991 A Water Storage Adaptation in the Maya Lowlands. Science 251(4994): Scarborough, Vernon L. and Fred Valdez, Jr The Engineered Environment and Political Economy of the Three Rivers Region. In Heterarchy, Political Economy, and the Ancient Maya: The Three Rivers Region of the East-Central Yucatán Peninsula, edited by Vernon L. Scarborough, Fred Valdez Jr., and Nicholas Dunning, pp University of Arizona Press, Tucson An Alternative Order: The Dualistic Economies of the Ancient Maya. Latin American Antiquity 20(1): Scarborough, Vernon L., Matthew E. Becher, Jeffrey L. Baker, Garry Harris, and Fred Valdez Jr Water and Land at the Ancient Maya Community of La Milpa. Latin American Antiquity 6(2): Scarborough, Vernon L., Nicholas P. Dunning, Kenneth B. Tankersley, Christopher Carr, Eric Weaver, Liwy Grazioso, Brian Lane, John G. Jones, Palma Buttles, Fred Valdez, and David L. Lentz 2012 Water and Sustainable Land Use at the Ancient Tropical City of Tikal, Guatemala. Proceedings of the National Academy of Sciences 109(31) Siemens, Alfred H Karst and the Pre-Hispanic Maya in the Southern Lowlands. In Pre-Hispanic Maya Agriculture, edited by Peter D. Harrison and B.L. Turner II, pp University of New Mexico Press, Albuquerque. 133

143 Smith, A. L Uaxactun, Guatemala: Excavations of Carnegie Institution of Washington, Publication 588. Washington, D.C. Šprajc, Ivan 2008 Reconocimiento arqueológico en el Sureste del Estado de Campeche, Mexico: Paris Monographs in American Archaeology 19. BAR International Series, 1742, Archaeopress, Oxford. Stowe, Michael and David M. Hyde 2014 Report of the 2013 Mapping Season Within the Medicinal Trail Locality. In Research Reports from the Programme for Belize Archaeological Project, Volume 8, edited by Fred Valdez, Jr., pp Occasional Papers, Number 15. Mesoamerican Archaeological Research Laboratory, University of Texas at Austin. Tourtellot, Gair, III and John Rose 1993 More Light on La Milpa Mapping: Interim Report on the 1993 Season. Manuscript on file, Department of Archaeology, Boston University. Trachman, Clarissa M Excavated Households, Excavated Lives: Social Reproduction, Identity, and Everyday Life for the Ancient Maya in Northwestern Belize. Ph.D. dissertation, Department of Anthropology, The University of Texas at Austin. Wahl, David, Roger Byrne, Thomas Schreiner, and Richard Hansen 2006 Holocene Vegetation Change in the Northern Peten and its Implications for Maya Prehistory. Quaternary Research 65(3) Wahl, David, Thomas Schreiner, Roger Byrne, and Richard Hansen 2007 A Paleoecological Record from a Late Classic Maya Reservoir in the North Petén. Latin American Antiquity 18(2): Walker, Debra S Apuntes Sobre La Secuencia Cerámica de Yaxnohcah. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, pp Mesoweb. Walling, Stanley L., Peter Davis, Sandra Dias, and Melissa DeVito 2005 Report of the 2004 Rio Bravo Archaeology Project: Site R.B. 47, Chawak But o ob. In Programme for Belize Archaeological Project: Report of the Activities of the 2004 Field Season, edited by Fred Valdez, Jr., pp Occasional Papers, Number 4. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Walling, Stanley L., Peter F. Davis, Jonathan Hanna, Leah Matthews, Nahum Prasarn, and Christine Taylor 134

144 2006 Residential Terracing, Water Management, Matrix Analysis, and Suburban Ceremonialism at Chawak But o ob, Belize. In Programme for Belize Archaeological Project Report of Activities for the 2005 Field Season. Occasional Papers, Number 6. Mesoamerican Archaeological Research Laboratory, The University of Texas at Austin. Weiss-Krejci, Estella 2013 Ancient Maya Rain Water Reservoirs in Northwestern Belize. Contributions in New World Archaeology 5: Weiss-Krejci, Estella and Thomas Sabbas 2002 The Potential Role of Small Depressions as Water Storage Features in the Central Maya Lowlands. Latin American Antiquity 13(3): Weller, Erin T Satellites, Survey, and Settlement: The Late Classic Maya Utilization of Bajos (Seasonal Swamps) at Tikal and Yaxha, Guatemala. In From Space to Place: 2nd Annual Conference on Remote Sensing in Archaeology, edited by Stefano Campana and Maurizio Forte, pp Archaeopress, Oxford. 135

145 Table 1. Operation Op. 10A Residential Group (Site) Distance to nearest structure (m) 7 Op. 16B, C, D Group A (Medicinal Trail) Wo (Yaxnohcah) Op. 16E, F, G, H, I, K Alba (Yaxnohcah) 60 Op. 16J Fidelia (Yaxnohcah) 40 Op. 22A Yax-3 (Yaxnohcah) 30 Op. 22C Yax-3 (Yaxnohcah) 5 60 Chronology Surface Area (m2) Present Depth (m) Excavated depth below ground surface (m) Observations and Materials Postulated Function Late Preclassic Late Classic (350 BC AD 850) Early/Middle Preclassic Late Classic (900 BC AD 800) Middle Preclassic Late Classic (700 BC AD 800) Middle Classic Postclassic (600 BC AD 1000) Classic period (AD ) Degraded plaster floor; cut marks on exposed limestone Limestone quarry reservoir Cut marks on exposed limestone; large catchment area Limestone quarry ( reservoir?) 1, Reservoir Degraded plaster floor; placed stones (cultural feature) Degraded clay/sascab floor; slickensides ,142 Reservoir Aguada vegetation and modern artifacts; surface cracks and argilloturbation Adjacent structure and exposed limestone blocks; multiple degraded plaster floors Late Preclassic Late Classic (350 BC AD 800) Table 1. Data summary and evaluation of all reservoirs investigated by the author. 136 Total Depth (m) Total Capacity (m3) Reservoir (agricultural?) Limestone quarry reservoir

146 Figure 1. Figure 1. Map of the Maya region showing Medicinal Trail and Yaxnohcah. Both sites are located in the Elevated Interior Region of the Yucatán Peninsula. 137

147 Figure 2. Figure 2. Medicinal Trail site map showing structure groups discussed in the text. Contour interval is 50 cm. Map adapted from Brewer et al

148 Figure 3. Figure 3. Map of Group A at Medicinal Trail. The Medicinal Trail Depression (Op. 10) is shown in the upper right. Depressions A-1 and A-2 were investigated by Lowe (2008). Contour interval is 15 cm. Map adapted from Stowe and Hyde

149 Figure 4. Figure 4. South profile drawing of Operations 10Q, 10R, 10Y, and 10Z from Depression A-3 at Medicinal Trail showing the remnants of a heavily weathered floor. Figure 5. Figure 5. Image of cf. Onagraceae seeds recovered from the Medicinal Trail Reservoir. 140

150 Figure 6. Figure 6. Plan map of Group C at Medicinal Trail detailing Operations 14 and 15. Dashed form lines show general contours and arrows indicate probable flow of water along berms toward the Op. 14 chultun and Op. 15 depression. Map adapted from Hyde

151 Figure 7. Figure 7. Overview of Yaxnohcah showing the location of the five operations discussed in the text. 142

152 Figure 8. Figure 8. Op. 16B, 16C, and 16D south profile drawing showing location of cut stones. Figure 9. Figure 9. Operation 16F west profile showing remnants of a highly weathered plaster floor. 143

153 Figure 10. Figure 10. Base of Operations 16G, 16H, 16I, and 16K showing large placed stones. 144

154 Figure 11. Figure 11. Op. 16J north profile drawing showing location of colluvium deposits. 145

155 Figure 12. Figure 12. Pre-excavation image of the Yax-3 Hinterland Aguada. 146

156 Figure 13. Figure 13. North profile of the Yax-3 Group Reservoir showing evidence of multiple highly weathered floors. 147

157 5. Research Article 3 HARVESTING HA: ANCIENT WATER COLLECTION AND STORAGE IN THE ELEVATED INTERIOR REGION OF THE MAYA LOWLANDS Nicholas Dunning, Jeffrey Brewer, Christopher Carr, Armando Anaya Hernández, Timothy Beach, Jennifer Chmilar, Liwy Grazioso Sierra, Robert Griffin, David Lentz, Sheryl Luzzadder-Beach, Kathryn Reese-Taylor, William Saturno, Vernon Scarborough, Michael Smyth, and Fred Valdez (Chapter from edited volume - A Path to Sustainability: The Past and Future Role of Water Management, edited by Jean T. Larmon, Lisa J. Lucero, and Fred Valdez. University Press of Colorado.) DO NOT CITE IN ANY CONTEXT WITHOUT PERMISSION OF THE AUTHOR(S) 148

158 ABSTRACT The Elevated Interior Region (EIR) of the Maya Lowlands posed especially difficult challenges for year-round ancient human occupation and urbanization. Accessible surface and groundwater sources are rare and a 5-month dry season necessitated the annual collection and storage of rainwater in order to concentrate human population. Here we review ancient Maya water storage adaptation in the EIR including urban and hinterland reservoirs as well as residential scale tanks and cisterns. Large reservoirs were devised as early as the Middle Preclassic period and continued to be an important adaptation for urban centers in the EIR throughout their occupation. Residential scale tanks and cisterns may also have early origins, though these have been less investigated. Considerable inter-regional variation existed in water management strategies. 149

159 Among the wonders unfolded by the discovery of these ruined cities, what made the strongest impression on our minds was the fact that their immense population existed in a region so scantily supplied with water. - J.L. Stephens (1841) The important role of water in Maya thought and life has long been recognized, as has the great lengths gone to by the ancient Maya to secure an adequate water supply (e.g., Stephens 1841; Lundell 1937; Morley 1938; Matheny 1978; Adams 1991; Scarborough 1998; Lucero 2002; Dunning 2003; Winemiller 2003; Wyatt 2014; Luzzadder-Beach et al. 2016). Water management in the Maya lowlands emphasized collection over diversion, source over allocation (Scarborough and Gallopin 1991:658). In elevated, interior parts of the Maya Lowlands elaborate rain collection and storage systems were devised at multiple scales and social levels, because groundwater access and perennial surface streams were rare indeed. The word ha (or ja ) and its cognates designates water across the Mayan language family and also appears in Maya glyphic script of the Classic period (Kettunen and Helmke 2013). In Classic period texts ha al refers specifically to rainfall, and nahb references bodies of water, most commonly pools of water (aguadas/reservoirs). Water and water bodies, including those in caves, form fundamental components of Maya cosmology, reflective of the tropical karst environment of the Maya Lowlands (Dunning 2003). Water-related landscape features became closely associated with the sacred places of origin (ch en) referenced by Maya lineages and dynasties in Classic inscriptions (Tokovinine 2013). This fundamental link between Maya culture, society, place-based identity, and water reflects the critical nature and seasonal scarcity of this resource across much of the Maya Lowlands. Almost all ancient Maya agriculture was 150

160 rainfall dependent and in many areas water for consumption and other needs had to be harvested and stored in the rainy season and meted out through the dry. This relationship was most pronounced in the Elevated Interior Region (EIR) where Maya water collection and storage technology evolved over many centuries. The Elevated Interior Region (EIR) of the Maya Lowlands is a rugged karst landscape of hills, escarpments and valleys created by block faulting and variable limestone dissolution (Figure 1) (Dunning et al. 2012). Drainage is complex, with some seasonally flowing rivers, but largely dominated by subsurface drainage and flow typical of karst geomorphology. Depth to permanent groundwater in the EIR is typically tens to hundreds of meters below surface, though a few accessible perched aquifers occur in places. About 90% of rain falls between late May and mid-december. Average annual rainfall totals vary from about 1,145 mm in the Puuc Hills in the north to some 1,600 mm in the south around Tikal (Isendahl 2011; Rosenmeier et al. 2002). During the rainy season, the surface drainage systems come alive and course with runoff, only to later dry into scattered, shrinking pools along river channels. The EIR contains numerous natural depressions (bajos) created by faulting and limestone dissolution, and ranging in size from one or two to hundreds of square kilometers in size. Today, most bajos are characterized by seasonal swamp forest ecosytems. Limited paleoenvironmental data indicate that some of these depressions held perennial wetlands or shallow lakes at the outset of Maya occupation in the Preclassic, but gradually desiccated due to some combination of natural climate change and anthropogenic sedimentation (Dunning et al 2002; Dunning et al. 2006a). The highly seasonal distribution of rainfall in the Maya Lowlands and a lack of, or declining quantity of natural perennial water sources in the EIR made collection and storage of rain water essential for permanent settlement of the region (Dunning et al. 2012). Additionally, 151

161 the climate record of the Maya Lowlands is punctuated by periods of more frequent and intense droughts that also posed challenges for water management (e.g., Lucero et al. 2011; LuzzadderBeach et al. 2016). Arguably, in the EIR the need to collect and store sufficient water to survive the dry season became a significant urbanizing force (Dunning et al 1999; Scarborough 1993). While natural sources could provide water for small number of people at select locations, concentrating larger numbers in response to the availability of other resources (e.g., agricultural land) required either cooperative or coerced investment in water storage at multiple scales. The Maya devised several types of features in which to collect and store water, including reservoirs, smaller scale tanks, underground cisterns (chultuns), and canals. The exact nature and origin of reservoirs and other water storage features, including actual dimensions, and construction and use history, is typically impossible to determine without excavation. Unfortunately, relatively few such features have been excavated across the EIR, thus constraining our ability to adequately understand the development history of these vital components of Maya civilization. Nonetheless, enough data has been amassed at his point to begin to discern a number of chronological and geographical patterns which are the subject of this paper. Overall, there is a demonstrable increase in the diversity and sophistication of water collection and storage between the Middle Preclassic and Terminal Classic periods (ca., 800 BC 900 AD). Natural Models for Water Collection and Storage The Maya Lowlands environment provided a number of natural features that might have served as models for the creation of water collection systems: sartenejas, natural aguadas, bajos, and deep pools within seasonal rivers. Especially in the northern lowlands processes of limestone 152

162 weathering and soil formation in the seasonally wet/dry climate combine to create a thick, casehardened caprock layer anywhere from a few centimeters to two or more meters below the surface (Dunning 1992). Where this caprock is exposed at the surface, solution hollows often form, some of which hold pools of rainwater known locally as sartenejas. Sartenejas can hold volumes of water ranging from a few cubic to many tens of cubic meters of water, though few of these pools are sufficiently large to persist through the entirety of the dry season The term aguada ( pond ) is widely used across the Maya Lowlands for self-contained bodies of water that typically hold water for much or all of the year, but, depending on year-toyear variation in precipitation, may desiccate for some period. Aguadas can form naturally, typically where a karst sinkhole has gradually become plugged with clay-rich sediment allowing water to accumulate (Siemens 1978; Dunning et al. 2015). Examples of natural aguadas can be found throughout the Maya Lowlands. Many of today s aguadas were entirely or partly anthropogenic in origin, that is, are the product of ancient reservoir construction. The word bajo ( low-lying area ) is widely used in the Maya Lowlands to designate natural depressions of varying size (as little as square kilometer, to many hundreds of square kilometers). Bajos typically are floored with deposits of clay-rich sediments. Topographic lows within bajos often fill with water during parts of the rainy season an aspect of the environment easily observable by early Maya settlers. The rivers that originate in the Elevated Interior Region and flow towards the coast have highly seasonal discharge patterns, largely desiccating in the dry season, but with pools persisting within deep solution pockets within the channel. Along the Rio Holmul drainage, such pools were the loci of many Preclassic settlements (Fialko 2000, 2005). 153

163 Many early Maya settlements were established at natural water sources, including springs and water-bearing caves in addition to the four natural water collection features mentioned above. However, sartenejas, aguadas, bajos, and river channel pools would have provided natural models for early settlers to observe the manner in which rainwater and seasonal streams could be manipulated to enhance water availability. Water Capture There are a few examples of places where water emerging from natural springs was captured and stored locally or shunted off for use elsewhere; for example, at Chan and Itzan, sites outside of the EIR, where springs are much more common (Johnston 2004). Local, perched aquifers can result in springs in the EIR, though these are quite rare. One example was located at Tikal, in the so-called silting tank of the Temple Reservoir. Here a basin was constructed to capture spring water, perhaps a foundational part of the Preclassic colonization of Tikal (Scarborough and Grazioso 2015). Ironically, as Tikal grew and the recharge surfaces for the spring were covered by buildings and pavements, the usefulness of the spring would have waned. For the EIR as a whole, water was only generally available in the dry season where it had collected during the preceding rainy season, either naturally (see above) or by human manipulation of the landscape, though noting that hydrological conditions within bajos appears to have changed over time making water progressively less available during the course of Maya civilization (see above). During the rainy season, rainfall is often intense, most typically in the form of thunderstorms, but also, with some frequency, tropical storms and hurricanes (e,g, Dunning and Houston 2011; Frappier et al. 2014; Medina Elizade et al. 2016). These intense rainfalls have the capacity to generate considerable surface runoff even under forested 154

164 conditions. In fact, ancient Maya water systems needed to be able deal with excess runoff in addition to their primary function of collecting and capturing water. However, hurricanes and tropical storms also likely provided important bumps in precipitation that allowed reservoirs to fill completely. As the ancient Maya cleared forest and urbanized the landscape, they changed its hydrologic characteristics. In the case of urban watersheds, these changes became progressively more intentional, with plastered structures and plazas canted specifically to shed water directly into storage features or into channels directing the water into reservoirs some distance away. This system of enhanced collection and capture was established as early as 800 BC at Yaxnohcah (Dunning et al. 2016; Dunning et al. 2017a), and became a fundamental part of urban design across the EIR. For example, at the Preclassic Puuc region site of Xcoch, water was collected off of elevated paved building and plazas and shunted into plastered channels carrying it into the La Gondola reservoir (Dunning et al. 2014a; Smyth et al. 2017). Apparently, that system was not always able to fill the reservoir adequately and later a large elevated platform was constructed adjacent to the reservoir in order to generate additional runoff. Outside of site cores, runoff was collected off of many non-paved surfaces in both urban and rural settings and funneled into reservoirs. At the household scale, small paved surfaces within residential compounds were often used to collect water into local open air tanks or chultuns. Reservoirs Reservoirs were both adapted from natural features or completely constructed across the EIR beginning at least as early as the Middle Preclassic (ca. 800 BC). The origins of individual reservoirs is usually difficult to determine with excavation, and, because relatively few have 155

165 been probed, their exact nature and chronology often remain obscure. Excavations at several sites have revealed that reservoirs were built by modifying hydrologic ponding points such as natural aguadas and parts of bajos, by sealing depressions created by quarrying, as well as impoundments created within stream channels. Examination of regional archaeological surveys in various parts of the EIR reveal the ubiquity of reservoirs embedded with the ancient settled landscape and their direct spatial association with sites of any size (e.g., Bullard 1960; Šprajc 2008). Examination of satellite imagery from areas with the EIR also quickly indicates that the Maya constructed thousands of reservoirs, only a fraction of which are visible after a thousand years of abandonment. Scarborough (1993) proposed a model of Maya watershed manipulation in which he contrasted concave systems based on passive collection of runoff at the low point of local drainages with convex system that involved water capture at higher points with a local watershed, further proposing that the concave systems were more characteristic of the Maya Preclassic, whereas convex system were more typical of the Classic. While we do not refute the overall veracity of this model, we propose some refinements and changes in terminology. From a geomorphological standpoint, all stream channels are concave; that is, the floor of the channel drops in a general concave arc from upper watershed to lower, though the flow of this line is usually interrupted by grade changes for various reasons, most typical changes in bedrock hardness. We suggest that depression filling reservoirs be used to denote those features situated at a low point in a local drainage system (replacing Scarborough s concave ), and stream damming reservoir be used to indicate those located higher in the drainage and created by damming an incised valley (replacing Scarborough s convex ) (Figure 2). Natural aguadas and low, seasonally flooded areas within bajos provided early models for depression-filling 156

166 reservoirs, and, in fact, many such places were developed as reservoirs. Deep pools within descending streams may have had a similar role for modeling the potential of stream damming. Two useful characteristics of reservoirs that facilitate comparison are their relative location and capacity, both of which speak to their origins and control. The location of reservoirs within ancient settlements has been presumed to offer information on their control. For example, centrally located urban reservoirs that are often intimately associated with monumental architecture have been used as evidence supporting the role of water management by ancient Maya political rulers (Scarborough and Gallopin 1991; Ford 1996; Scarborough 1998; Lucero 1999, 2002, 2006;). In contrast, reservoirs and smaller scale features in urban peripheries or associated with smaller settlements, are suggestive of more decentralized water management (Weiss-Krejci and Sabbas 2002). However, this picture is complicated by the presence of reservoirs of many sizes located in varied parts of urban centers, urban fringes, smaller settlements, and rural areas. We suggest that the terms central, community, and household be applied to water collection and storage features. Central Reservoirs include spatial association with the monumental site core and elite control. Community Reservoirs are characterized by association with locations outside of the site cores, but are of a scale that required participation by many households for creation and maintenance. Household features (such as tanks and chultuns) are smaller features that are closely associated with an individual or small cluster of residential compounds. These terms are intentionally both spatial and social. Another approach to comparing reservoirs is by their capacity. However, there are limitations to this approach. Available data are limited by the small quantity of excavated features which makes accurate measurement of volumetric capacity impossible because most reservoirs have experienced significant sedimentation, often both during their active use as well 157

167 as post-abandonment. Even surface area calculations can be difficult. While some reservoirs are rectangular, circular, or elliptical in shape, many others have very irregular outlines; and the original surface outline is often obscured by sedimentation and post-abandonment vegetative succession. Additionally, even where capacity is known after excavation, it does not necessary reflect the importance of a reservoir to a particular place. For example, even a small reservoir may have played a vital ( central ) role in a relatively small site. Individual reservoir size may also not be especially important in places where there were multiple reservoirs or systems of reservoirs. Documented Community and Central reservoirs ranged in capacity from less than 1,000 to some 90,000 m3. Open-air, water-holding ponds smaller than 500 m3 are here considered as household tanks essentially small capacity reservoirs typically closely spatially associated with ancient residential groups. These capacity ranges are simply estimates and the spatial settlement associations of reservoirs and tanks is critical to their interpretation; for example, tanks are typically closely associated with specific residential groups. These household-scale features will be discussed separately below. Table 1 lists a sample of ancient Maya Central and Community Reservoirs for which there is sufficient excavation data to assess origin, use history, and capacity. This list is meant to be illustrative and is not comprehensive. These data help illustrate the history of reservoir creation in the Maya Lowlands, especially the EIR. Preclassic Period Reservoirs The origin of ancient Maya reservoirs is typically impossible to ascertain without excavation. Given that relatively few reservoirs have been excavated at this time, our understanding of the origins of this form of water collection and storage in the Preclassic is very 158

168 limited. Two of the oldest reservoirs known are at San Bartolo and Yaxnohcah, both with clear origins in the Middle Preclassic. San Bartolo is a moderately sized site that is best known for its well-preserved Preclassic murals and inscriptions (Saturno et al. 2006). The site includes several groups of monumental architecture, one of which sits immediately adjacent to an aguada that gave the contemporary name to the site. Excavations in San Bartolo Aguada in 2005 revealed that it began as a limestone quarry (likely to provide material for the adjacent elevated plaza and pyramid) which was converted to a Central Reservoir by lining the floor with a thick coat of plaster (Figure 3) (Dunning et al. 2006b). Charcoal embedded in the plaster floor produced a radiocarbon date with a calibrated radiocarbon date of BC. Preclassic ceramics were recovered in abundance in the overlying reservoir sediments. San Bartolo was abandoned at the end of the Preclassic period (c.a. 150 AD), but was reoccupied by in the Late Classic, at which point the reservoir was partially dredged and a thin plaster floor was added atop remaining sediment. Yaxnohcah is a sprawling site with numerous groups of monumental architecture featuring characteristic Preclassic triadic pyramid complexes (Šprajc 2008; Reese-Taylor and Anaya Hernandez 2013). Although several reservoirs were known at the site, an airborne LiDAR survey in 2014 revealed more, including Brisa Reservoir, the largest at Yaxnohcah (Anaya Hernandez et al. 2016; Reese-Taylor and Anaya Hernandez 2017). Excavations in Brisa Reservoir have revealed that it originated around 800 BC in the Middle Preclassic (Dunning et al. 2017a). The reservoir is enormous, with a surface area of over 28,000 m 2 and capacity of over 84,000 m 3, but relatively primitive. It was created by walling off a section of the Bajo Tomatal adjacent to an angled quarried limestone scarp lying below the Brisa Group of monumental architecture that also appears to have originated in the Middle Preclassic (Figure 3). No floor was 159

169 added to the reservoir, which relied on the slow permeability of the underlying bajo clay soil to retain water. In both the case of San Bartolo and Yaxnohcah, the early reservoirs were at least partially filled by runoff directed from superadjacent monumental architecture, indicating that the relationship between central reservoirs and elite-directed activities in the site core existed from an early age in Maya civilization. Due to lack of excavations, it is unclear how many other central reservoirs date to the Middle Preclassic. The sprawling Xpotoit Aguada at Yaxhom appears to have originated in the Middle Preclassic, though evidence to support this dating is preliminary (Ringle 2011). Many additional reservoirs are known from the ensuing Late Preclassic period from throughout the EIR. At Yaxnohcah, excavations have revealed that the Fidelia Aguada was created in the Late Preclassic period (Dunning et al. 2016). A section of bajo floor adjacent to the Fidelia Group was walled off by a large berm creating an impoundment with a capacity of about 29,000 m 3. Parts of the reservoir appear to have originated by rock quarrying. Where revealed by excavation, the floor the reservoir was formed by bedrock, with lower pockets filled in with cobbles and clay. Excavation in Aguada La Gondola at Xcoch revealed a succession of three floors, the lowest of which was associated with Late Preclassic ceramics and a radiocarbon date of 89 BC AD 1 (Dunning et al. 2014a; Smyth et al. 2017). La Gondola was refloored twice, after a period of apparent abandonment in the Early Classic, and again in the Late Classic. At El Mirador, often considered the preeminent Late Preclassic Maya center, an elaborate system of reservoirs was constructed, including a number located within low points immediately adjacent to acropolis complexes, as well as several very large depression-filling, bajo margin components located at the end of drainages descending the tall escarpment on which the site core is situated 160

170 (Matheny et al. 1980; Morales-Aguilar 2009). Excavation in the large Aguada Limon on the bajo margin, revealed a plaster floor and two distinct episodes of sedimentation; construction was dated to the Late Preclassic by both ceramics and a radiocarbon sample (Dahlin et al. 1980). Other Late Preclassic reservoirs are known from EIR sites including at Tikal, and the Xultun San Bartolo area (Akpinar-Ferrand et al. 2012; Scarborough et al. 2012). It is likely that a significant number of the hundreds of reservoirs in the EIR had Preclassic origins, but how many cannot be determined without excavation. Classic Period Reservoirs and their Attributes During the Classic period the number and diversity of reservoirs expanded, including the adoption of features designed to better control inflow and release of water, expand capacity, enhance access to water, and improve potability. We review several key attributes below. Reservoir seals Reservoirs in the EIR exhibit a wide variety of strategies for retaining water by reducing loss via infiltration in their bottoms and sides. The default option was to simply rely on the natural impermeability of basal substrate. In a few instances, bedrock floors were exposed at the base of reservoirs during archaeological excavations. When not fractured, dense limestone and marl can have low permeability, but fractures would need to have been filled with clay or another dense material to provide a reasonable basal seal. Reservoirs located along bajo margins or other, localized depressions could have taken advantage of the low permeability of clay soils and sediments typically found in such locations and examples of natural clay floors are not unusual. At Uxul, Seefeld (2013) found that the Aguada Occidental originated as a natural aguada in a small clay-bottomed depression on the margin of a larger bajo, but was later 161

171 modified to enhance water storage. One problem with reliance on clay seals, natural or anthropogenic, is the smectite-rich clays characteristic of the Maya Lowlands have a propensity to shrink and crack when desiccated, potentially compromising water retention until fully rehydrated. The ubiquity of clayey soils and sediments in the Maya Lowlands can make it difficult to distinguish whether or not clay resulted from natural sedimentation, or was intentionally introduced by the Maya. A case in point is the Aguada de Terminos on the outskirt of Tikal (Dunning et al. 2015). This reservoir appears to have originated as a quarry targeting chert and limestone in the Late Preclassic, but began to function as a reservoir after the accumulation of clayey sediment or its introduction. In either case, the depression was converted to a substantial reservoir that continues to effectively hold water today. At some point in the Preclassic the ancient Maya discovered tapial, sometimes referred to as rammed earth. In the Maya case, clay-rich soil was mixed with natural limestone marl (sascab or sahcab) which was pressure packed into a hard, dense layer. This technique was used to create the floors of several reservoirs excavated at Xcoch, in the Puuc region, where it is still used today to line modern irrigation ditches (Dunning et al. 2014a). Examples of tapial seals can also be found in the southern parts of the EIR, including within at least one small tank investigated at Yaxnohcah (Brewer 2016). In some instances, limestone slabs were used as part of the flooring of reservoirs, with plaster or clay tuck-pointed between or overlaid on the stones to complete the seal. Examples of such flooring in the EIR are documented for the Palace Reservoir at Tikal (Scarborough and Grazioso 2015); the large reservoir at El Zotz (Beach et al. 2015a); the Aguada Oriental at Uxul 162

172 (Seefeld 2013), Aguadas 4 and 6 at Calakmul (Domínguez Carrasco and Folan 1996; Geovannini Acuña 2008), and Aguada Zacatal near Nakbe (Wahl et al. 2007). Examples of plaster lining in reservoirs are relatively rare, probably reflecting the high cost of producing plaster and its need for other purposes, most notably on monumental architecture though, as noted, such architectural complexes provided important runoff used to fill reservoirs and other water storage features. Another rare treatment of reservoirs was the apparent application of a layer of ceramic fragments from broken vessels compressed into clay. Examples have been found at El Zotz (Beach et al. 2015a) and Uxul. At Uxul, Seefeld (2013) noted that the ceramic layer on one floor of the Aguada Orriental was discontinuous, raising the question of whether this treatment was meant to enhance water retention or served another function. We speculate that such tiling may have been decorative? Dams and Berms Dams and berms are essentially variations of the same idea: walls constructed to restrict the movement of water. Dams can be distinguished as walls put in place across stream channels, whereas berms are walls constructed to impound water accumulated in depressions. Examples of dams can be found widely across the Maya Lowlands wherever the Maya sought to manipulate the flow of moving water. In the EIR, dams were often used to create reservoirs within incised drainages. For example, the multiple dams at constructed in the elevated site core of Tikal are of the stream damming type (Scarborough et al. 2012; Grazioso Sierra and Scarborough 2013; Scarborough and Grazioso 2015). In many cases, the tops of dams doubled as causeways facilitating passage across the valleys. The Palace Reservoir dam is the largest known thus far in Mesoamerica. When complete the dam was about 80-m long, 60-m wide at the base, and over

173 m in height, creating a reservoir with a capacity of about 74,630 m3 of water. The dam was constructed with a fill of limestone rubble packed with clay soil obtained from bajos, and an armoring of limestone slabs. Buried deep within the dam is a much smaller Early Classic construction. Examples of depression filling reservoirs that employed berms are found ubiquitously across the EIR. Berms sometimes form partial enclosures, capturing runoff water at the base of a drainage. In other examples berms essentially encircle the entire reservoir except for restricted inlet and outlet points. Berms were constructed of a variety of materials, most typically expediently from readily available materials such as quarry rubble, bajo clay, and soil/sediment excavated while deepening the reservoir pool. Examples exist of more refined berm construction, most notably stone-lined inner walls sealed with clay, tapial, or plaster. While more elaborate berm construction is more often seen in urban reservoirs, it can sometimes also be found even in hinterland or rural areas, such as Aguada Tintal outside of San Bartolo (Dunning et al. 2008). Berms around larger reservoirs required large investments in labor. For example, at Xuch in the Puuc region Isendahl (2011) found that the massive berms ringing the huge central aguada represented the single largest construction investment at the site Berms essentially served two functions in reservoir construction and management. One was to limit the sourcing of water allowed to enter a reservoir, and to exclude other water from entering. This is most clearly evident in bajo-margin reservoirs flanked on one or more sides by low-lying, seasonally inundated terrain where inflow from surrounding areas needed to be prevented in the rainy season (e.g., Brisa Reservoir at Yaxnohcah; Figure 3). However, even in some reservoirs situated within stream channels berms could function in a similar way. For example, the Corriental Reservoir at Tikal was created at the confluence of several seasonal 164

174 streams where a natural wide spot was probably expanded and deepened by quarrying and berms were used to control inflow and outflow of water which could be either allowed into the reservoir or shunted around the outside of the berm (Scarborough et al. 2012). Another function of berms was to enhance water storage capacity by allowing the water pool to elevate within the reservoir. At Xcoch s Aguada La Gondola, excavations revealed that the surrounding berm was increased in height and width three times over its use history in association with the addition of new floors within the reservoir (Dunning et al. 2014a). Dredging, Enhanced Capacity, and Access to Water In order to maintain their water storage capacity, reservoirs were sometimes dredged of sediment to maintain pool depth, though the archaeological record offers few glimpses on how regularly this occurred. Excavations and sediment coring often produce evidence of severe disjunctions in the sediment record within reservoirs. A common finding is some depth of Preclassic age sediment in the lowest cultural level, separated abruptly from accumulations of Late or Terminal Classic and later sediments; sometimes these disjunctions are clearly visible in the sediment strata, but not always. Good examples of such disrupted sediments can be seen in the Aguada de Terminos at Tikal (Dunning et al. 2015), Aguada Fidelia at Yaxnohcah (Dunning et al. 2016), and Aguada Tintal near San Bartolo (Dunning et al. 2014b) (Figure 4). Such dredging has proved to be a problem for studies attempting to use paleoenvironmental proxies such as pollen from reservoirs because sediment generated during much of the Classic Period is often absent (Dunning et al. 2003; Akpinar-Ferrand et al. 2012). How frequently dredging occurred during the Classic is difficult to ascertain. In the case of the Palace Reservoir at Tikal a partial inward collapse of the reservoir dam in the Terminal Classic happened to preserve beautifully varved, visually distinguishable wet and dry season 165

175 sediments representing 5.5 years of accumulation between a cleaning and dam collapse. This sample implies fairly frequent sediment removal even in the waning days of Tikal s political and economic power, though it is unwise to draw broad conclusions from one sample. Furthermore, this reservoir is located immediate adjacent to Tikal s royal palace and may well have been subject to more rigorous maintenance than was typical for Maya reservoirs in general. Coring in the berms of two reservoirs in the Puuc region indicates one, two, or more episodes of dredging with some of the excavated sediments heaped atop the berms (Dunning 1992: 23). There are also many examples of reservoirs which were evidently not subject to dredging, or to less thorough sediment extraction. As noted above, the central reservoir at San Bartolo was only partially dredged before a new floor was added when the site was reoccupied after a period of abandonment (Figure 3). Excavations in the huge reservoir at El Zotz revealed an original Early Classic floor buried by accumulated sediment, a Late Classic floor, and a further accumulation of sediment (Beach et al. 2015a). Aguada La Gondola at Xcoch has three floors separated by thick zones of sediment accumulation (Dunning et al. 2014a). In this case, loss of pool depth due to sedimentation was partially offset by adding elevation to the surrounding berm. The decision to simply add a new floor atop accumulated sediment within reservoirs with no or only partial dredging may have been made for expediency, because dredging would have been a very labor intensive undertaking and would have potentially kept a reservoir out of commission for a longer period. Partial rather than thorough dredging may have also been motivated by a desire to avoid potentially damaging underlying seals whose exact depth and composition may no longer have been in the collective memory of later site occupants. In some instances, the Maya sought to prolong the dry season yield of water in reservoirs by excavating filtration wells in their floors. Known as buk té ob in Yukatek Maya (singular 166

176 buk té), these wells feature dry-laid stone walls surrounding openings of varying depth and width. In clay-bottomed reservoirs, or those with significant accumulated sediment into which the well was sunk, throughflow of water from surrounding saturate sediments could be accessed to obtain a final water yield from the drying reservoir. Buk té ob were first reported by John Lloyd Stephens (1841) who was told by a hacienda owner about his efforts to clean out and reuse these features. The practice of resurrecting ancient reservoirs, including their buk té ob occurred elsewhere in the Puuc and Chenes regions during periods of economic expansion in the 19 th century (for example, re-opened bukte ob are still clearly visible in a reservoir at Ichpich; Dunning 1992: Fig. 2-11). Huchim Herrara and Sanchez (1990) excavated a buk té in the floor of the ChenChan reservoir at Uxmal that included a plaster floor at the base of the filtration wall. In the La Milpa Aguada, excavations exposed a buk té visible as a stone ring on the reservoir floor after complete removal of grassy vegetation. This well proved to be the second of two, with the one visible at the surface having been excavated into an earlier, lower well after a period of apparent neglect and sedimentation (Figure 5). How widespread such wells are within Maya reservoirs is difficult to assess since the majority are likely to be essentially invisible due to sedimentation and vegetation growth, though some aguadas contain small depressions in their floors that may indicate the presence of wells. Stephens (1841) was also told of, but did not observe, features resembling chultuns reportedly uncovered in the floor of an aguada. Especially in wider and deeper reservoirs, the ancient Maya would have been faced with the problem of gaining access to water as levels dropped during the course of a dry season. Features observed in a number of reservoirs appear to have been put in place to facilitate access to water in the shrinking pool. In Aguada La Gondola at Xcoch inclined stone-faced walkways were added at different heights on each side of the reservoir s interior walls so that users could 167

177 have followed the water down as it descended over time (Dunning et al. 2014a). Other reservoirs included stone piers that projected into the pool, such as Aguada de Carlos near Tikal (Dunning et al. 2015). In other instances, stone-based platforms or islands were built within reservoirs, perhaps to facilitate access to water, such as at Aguada Lagunita Elusiva near La Milpa (Weiss- Krejci 2013). Ingress Controls and Potability Features If reservoir water was to be used for human consumption, measures needed to be taken to maintain potability. As noted above, the collection surfaces used to supply water to reservoirs varied greatly, ranging from plastered surfaces of varying sizes, that could presumably be kept relatively clean, to open slopes where residential and agricultural land use were likely intertwined. Another part of the collection and storage system in which cleanliness could be enhanced was via the control of water flow into the reservoir. Many reservoirs include restricted entry points that could potentially have been closed off to limit incoming flow; see, for example, the Perdido Reservoir at Tikal (Scarborough et al. 2012; Scarborough and Grazioso 2015). Such control would have served multiple purposes, including limiting inflow during times of excessive rainfall and potentially destructive volumes or velocities of runoff, as well as to exclude water that may have appeared dirtier than desired. Water in channels leading into reservoirs was sometimes interrupted in its journal by the construction of silting or settling tanks such as at Kinal (Scarborough et al. 1994), Oxpemul (Volta et al. 2013), and Aguada Los Loros near San Bartolo (Akpinar-Ferrand et al. 2012). Such features functioned to quickly slow water flow, allowing at least some particulate matter to precipitate before the water entered to main reservoir. At the Aguada Oriental at Uxul, channelized water was run through a thick, dry-laid stone filter wall before it could enter the 168

178 reservoir (Seefeld 2013). The remains of this wall were not visible prior to excavation, and such features may be more commonplace than realized. Excavations in the Corriental Reservoir at Tikal produced findings suggestive of an even more sophisticated filtration system (Scarborough et al. 2012). Lenses of quartz sand were found stratified between typical clayey sediments within the reservoir. The nearest know source for such sand is some 39 km away in the Bajo de Azúcar; though closer sources are possible, none are present in the reservoir watershed. This suggests that this exotic material was intentionally introduced and may well be the blown-out remains of former sand filtration boxes constructed in the reservoir s several intake points. The contributing watershed included residential areas interspersed with garden/orchard zones, thus the ancient residents of Tikal may have sought a more thorough means of cleansing water entering the reservoir. A few, less clear sand lenses were also found in the sediments in the interconnected Temple and Palace reservoirs suggesting that this filtration practice may have been more widespread at Tikal. Robert Rands (1953) noted the prevalence of water lilies in royal iconography from the Classic period, leading to suggestions that these plants may have played a role in maintaining water purity in reservoirs and as a mark or royal religious and political power (Ford 1996; Lucero 2006; Schele and Friedel 1990). While water lilies (Nymphaea spp.) are rarely found in ancient reservoirs that still hold water today, pollen recovered from sediments in some reservoirs contain lily pollen supporting the idea that these plants were cultivated in reservoirs (Hansen et al. 2002). Lilies would have the beneficial properties of slowing evaporation off of water surfaces as well as biologically filtering some toxins that may have been present in reservoirs (Ford 1996; Gill 2000; Harrison 1993; Lucero 1999; Pohl et al. 1996). Outflow Control 169

179 Many reservoirs include features that allowed for the release of water. Typically, these egresses are manifest today simply as gaps in the berms or dams used to create the reservoir. Examples include the Middle Preclassic Brisa reservoir at Yaxnohcah and the Early Classic Perdido reservoir at Tikal (Scarborough et al. 2012; Dunning et al. 2017a). Excavation revealed that these gaps are the badly blown out remains of spillways. Presumably, features such as wooden gates may have existed in these spillways in the past allowing for more regulated release of water. Perdido Reservoir, along with the nearby Pital Reservoir, was positioned to release captured water into a sub-adjacent pocket bajo, where there is isotopic evidence for intensive maize cultivation (Dunning et al. 2015). The egress for Corriental Reservoir at Tikal was apparently more complex, at one point being transformed from a simple spillway to a switching station that could allow water to either enter or be released from the reservoir (Scarborough and Grazioso 2015). The Palace Reservoir and dam featured even more sophisticated outflow control; the dam included a series of stacked sluice channels that would have allowed water to be released at varying levels (Scarborough et al. 2012). Religious and Ritual Aspects and Connections Water played a central role in both Maya pragmatic and cosmological understanding of the world, thus, divine and elite political control of this substance became intertwined as Maya Civilization evolved (e.g., Scarborough 1998; Dunning 2003; Lucero 2006; Houston 2010). This complex relationship is evident in the incorporation of water collection and storage features in the elite sponsored monumental architecture within site cores from the Middle Preclassic onward; for example, the shedding of water from the Brisa Group in the Brisa reservoir at Yaxnohcah (Dunning et al. 2017a). While the precise meaning and function of many Maya 170

180 pyramid temples remains unclear, at least some of these building were apparently manifestations of the pan-mesoamerican concept of a water mountain (Lucero 2006). Conceptually tied to mountain caves and springs, the architectural creation of such symbolically potent places was another symbol of the cosmologically endowed authority of rulers. The symbolic recreation of water mountains as plastered pyramids and associated plazas and structures allowed these features to literally produce water as rain was shed that could be collected in reservoirs (Scarborough 1998). In Maya cosmology, water-bearing caves are believed to be the home of the rain deity Chaak (Brady 2010). Caves that penetrate to the permanent groundwater table are extremely rare in the EIR. One such cave is located at Xcoch, where a tortuous series of narrow passageways, drops, and wider chambers eventually reach a small pool of water which was a focus of rainrelated rituals from at least as early as 800 BC and continuing into the modern era, including as a source of zuhuy ha, sacred water used in rain calling ceremonies to this day (Stephens 1841; Smyth and Ortegon Zapata 2008; Dunning et al. 2014a; Weaver et al. 2015; Smyth et al. 2017). On the surface, the Great Pyramid of Xcoch and an associated acropolis and plaza complex, grew incrementally over many centuries beginning in the Middle Preclassic, with numerous rebuilding episodes reinvigorating this sacred site as well as literally creating a water mountain shedding rain that could be harvested. Over centuries a city grew around and above the cave which remained at its ceremonial heart and the geomantic centering point for the community, its identity, its public architecture and its water management infrastructure. Plastered channels funneled rain water collected in several plazas into the city s largest reservoir, Aguada La Gondola. A causeway (sacbe) was constructed connecting an altar complex at the cave entrance to a temple group on the eastern lip of the reservoir. Rare Chac Polychrome and 171

181 related Dos Arroyos water jars are found only deep within the cave and on the Preclassic and Early Classic floors of the reservoir. These specialized jars were likely used to bring zuhuy ha to the reservoir to symbolically initiate the refilling of the reservoir by the perceived combined actions of a shaman/ruler and the rain gods being called forth from the cave and sky. The repetition of the urban pattern of pyramid, plazas, and reservoirs in the core areas of Maya cities and towns in the EIR is indicative of the pervasiveness of a belief system that included the close relationship between rulers and rain deities that persisted from the Middle Preclassic through Terminal Classic periods. Whether natural or created, water sources at the heart of Maya communities were intrinsically tied to place identity and the authority of dynastic rulership (Tokovinine 2013; Dunning and Weaver 2015). Many of these water catchment systems continued to evolve as central architectural complexes grew (e.g., a system documented at Nakum: Kozskul and Žralka 2013). Canals The ancient Maya excavated canals for many reasons including facilitating transportation, draining perennially or seasonally inundated terrain to facilitate cultivation, aquaculture, and water storage. The multiple functions of canals and their history will not be reviewed here. However, canals clearly had the potential to store large volumes of water and cannot be ignored as at least a backup source for human consumption, as well as to facilitate agricultural production and aquaculture. Most documented canal systems are situated at considerable distances from urban population centers, but there are examples of systems that may have functioned as either vital or secondary urban water supply; examples include Edzna, Dzibanche, and Baking Pot (Matheny et al. 1983; Dunning and Beach 2010; Ebert et al. 2016). 172

182 While the majority of the documented canal systems lie outside of the EIR in lower-lying areas with more abundant perennial water, examples of canal systems can be found in bajos within the EIR (e.g., the Bajo de Azúcar: Dunning et al. 2017b). Community and Household Water Collection and Storage The majority of the reservoirs discussed previously, though not all, are examples of Central Reservoirs intimately associated with core monumental architecture. Central Reservoirs have been subject to far more archaeological investigation than Community Reservoir located in outlying residential areas of urban centers, hinterland, and rural areas. Community Reservoirs tend to be smaller in size, though many exceptions exist. These reservoirs are typically embedded within the residential landscape suggesting that their construction, use, and maintenance was directed by spatially adjacent residential groups. Many residential groups also collected water at the household level. What we are calling Community Reservoirs and Household Tanks are open-air ponds that grade into one another in terms of size and are best distinguished subjectively by their association with multiple or individual residential groups. Community Reservoirs Good examples of Community Reservoirs have been documented at Turtle Pond and Aguada Lagunita Elusiva, both on the urban fringe of La Milpa. Turtle Pond Aguada originated as a natural aguada within a pocket bajo amongst the far eastern settlement zone of La Milpa (Chmilar 2005). Modifications included construction of a low berm, an inlet channel, and deepening resulting in a capacity of some 785 m3, probably serving as a water supply for residential groups spread along a nearby ridge of higher ground. Aguada Lagunita Elusiva lies even further out on the eastern urban fringe of La Milpa but was significantly larger, with a 173

183 capacity of at least 7,800 m3 (Weiss-Krejci 2013) despite lying in an area of lesser population density. The area around this reservoir includes field walls indicative of intensive cultivation, an idea also supported by abundant maize pollen recovered in its sediments (Dunning and Beach 2010). There is no direct evidence that the reservoir was used for pot or another form of irrigation, but that function remains a possibility Some other Community Reservoirs were significantly larger in size. The Xcoch South Aguada 1 had a capacity of about 12,000 m3, seemingly far larger than needed to supply the household needs in nearby areas on the urban fringe of Xcoch. (Dunning et al. 2014a). The remains of two canals, now badly marled by modern agriculture, linking the reservoir with nearby farmland suggests than some capacity may have been used for agricultural production. Similarly, the Aguada Tintal located in the hinterland of San Bartolo was oversized for its lowpopulation-density setting, but could have been used to facilitate nearby cultivation (Dunning et al. 2008). In the densely settled, but distinctly non-urban landscape of the Rio Bec region, reservoirs of varying size are a ubiquitous part of the landscape, perhaps serving both domestic and agricultural functions (Lemonnier and Vanniere 2013). Household Tanks and Cisterns Considerable inter-regional variation existed in domestic scale water management strategies with the EIR from the Puuc Hills south to the Tikal area. In general, in the more northerly parts of the EIR chultuns were most common, whereas household tanks predominate in the south, though examples of each type of feature can be found in both the north and south. Residential scale tanks and chultuns have been less investigated than their larger, more visible, counterparts, but data from multiple sites within the region indicates that significant temporal variability was likely also found in household water features. The picture emerging from ongoing 174

184 investigations across the Maya lowlands details the frequency and importance of household tanks and cisterns in providing for the daily water needs of the region s inhabitants. These features have been investigated at sites across multiple settlement scales within the EIR, from minor agricultural communities to highly developed urban centers. The regional surficial geology of the Puuc region was notably favorable for chultun construction, with abundant surface exposures of hard caprock with an underlying zone of softer sascab (Dunning 1992). After punching openings through the caprock, Maya people then excavated large, typically bell-shaped chambers made water-tight with layers of plaster (Figure 6a and b). Surfaces surrounding the opening were canted into the mouth and plastered to funnel rainfall into the cistern. Chultuns had capacities ranging from about 5 to as much as 90 m3. Chultuns are found in the majority of residential patio groups within Puuc communities, and their presence (or absence) can be used to help distinguish architectural groups with nonresidential functions (Dunning 1992). Chultuns are found in a relatively consistent ratio in relation to the number of presumed residential structures in ancient Puuc communities (Dunning et al. 2014a). Although highly dependent upon the extent and accuracy of mapping between sites, this spatial relationship is evident at other Puuc sites such as Sayil, Xuch, and Xculoc, among others, where significant numbers of both residential structures and chultuns have been documented. Despite the presence of sizable reservoirs at some of these sites, including Xcoch, the ratio of residential structures per chultun indicates that residential compounds were likely expected to satisfy their own water needs. However, chultun construction many have been restricted. Many rural hamlet sites in the Puuc lack chultuns suggesting that these smallest of settlements represent seasonally-occupied hamlets/farmsteads (Dunning 2004). Or was there an indentured rural population dependent on nearby minor or major centers for water? 175

185 Very few household tanks have been noted at Puuc region sites, though this may reflect a regional bias whereby archaeologists have not really looked for these features given the ubiquity of chultuns. Several household tanks were identified and two excavated at Xcoch (Dunning et al. 2014a; Smyth et al. 2017). Both excavated examples proved to be Preclassic constructions, a finding consistent with the early dates from associated residential groups. Although the sample size is miniscule, data so far suggest that tanks may have been an early strategy in the Puuc that was later largely replaced by a preference for chultuns. Chultuns had the distinct advantage of being sealable, closed by stone of wooden lids to minimize evaporative loss and contamination. Further south within the EIR, the evolution and sustainability of Yaxnohcah was dependent on significant hydraulic, agricultural, settlement, and landscape modification strategies that allowed the persistent occupation of the site from early in the Preclassic period, through the Preclassic collapse that engulfed some of its neighbors, and well into the Classic period (Anaya et al. 2016). At the household, or decentralized, scale, hydrologic modifications in the form of creating and maintaining small residential tanks took place in tandem with the community-wide, centralized investments in water management exhibited by the Middle and Late Preclassic Brisa and Fidelia Reservoirs. Small reservoirs located adjacent to many of the household groups identified at Yaxnohcah would have provided an additional accessible water supply for the community (Figure 7a, b, c). Ongoing investigations through the Yaxnohcah Archaeological Project (YAP) have identified five of these residential tanks, with excavations focused on understanding the chronology of their construction and the nature of their role in water management at the household level (Brewer 2016; Brewer forthcoming; Brewer and Carr forthcoming; Brewer et al. forthcoming). Each was determined to be a water storage feature based on a combination of 176

186 physical and spatial characteristics (Figure 7a, b, c). Four of the depressions contained either a degraded plaster surface or thick layer of bajo clay overlying their base, which would have functioned as a water retention sealant. The fifth fit the profile of a residential tank in terms of surface area, depth, and location adjacent to a residential complex in an area of the site lacking any visible water features and was bordered by an extensive catchment area along its south, west, and southeast sides that would have been positioned to either receive outflow from or shed runoff into the reservoir, probably for agricultural purposes. Three of the five tanks also exhibited evidence of originating as limestone quarries, with either (or both) exposed limestone blocks visible around their rims or cut stone blocks forming their base. Similar depressions originating as quarries before being modified and used as small reservoirs have been identified elsewhere across the EIR (Folan 1982; Weiss-Krejci and Sabbas 2002). Analysis of recovered ceramic material and charcoal samples submitted for C14 dating was used to assess the chronology of the household tanks at Yaxnohcah. Dates for these features ranged from the Middle Preclassic to the Classic period, with each matching dates assigned to the nearest corresponding residential group (Peuramaki-Brown et al. 2016; Walker 2016). The creation and use of smaller aguadas and household tanks by a growing hinterland population fanning out from the urban core, combined with the continued exploitation of bajo areas as cultivated and managed wetland forests and agricultural zones, characterized residential scale water management activities at Yaxnohcah from the Middle Preclassic into the Classic period (Brewer forthcoming; Brewer et al. forthcoming). A series of household water management studies conducted at Medicinal Trail have revealed a multi-component system of water features incorporated into a dispersed hinterland community engaged in agricultural production (Brewer 2007; Brewer et al. 2013; Chmilar 2005; 177

187 Gill 2009; Hyde 2011; Lowe 2008; Percy 2009). Located on an escarpment edge, the site consists of multiple formal courtyard groups, numerous informal mound clusters, a series of landscape modifications geared toward domestic water collection and storage, including depressions, terraces, and berms, and the centrally located Medicinal Trail Reservoir (Hyde 2011). Extensive investigation of the reservoir revealed the presence of a degraded plaster floor located between a layer of clayey reservoir sediment and a lower clay-rich (Ab) horizon atop sascab (calcareous marl). Combined with the tank s calculated capacity (153 m3) and its position adjacent to a residential structure (Structure A-7), the former plastered surface is a positive indicator that the reservoir functioned as a water storage feature. Weiss-Krejci and Sabbas (2002) report the presence of a similar hard gray layer, between 10 and 30 cm thick, overlaying white medium-hard, smooth bedrock at four excavated depressions near the neighboring Wari Camp and La Milpa East sites that they also interpret as small water reservoirs. Visible cut marks and flattened edges on the limestone base of the depression and a recovered lithic inventory consisting of multiple tools (N=57), including at least one intact mid-sized biface, supports the notion that the reservoir may well have originated as a limestone quarry. The capacity of the Medicinal Trail Reservoir alone would have supplied the households in its immediate vicinity with a significant amount of water throughout much of the year. Combined with the additional supply provided by the other, smaller reservoirs, Medicinal Trail and its water supply would likely have functioned as a centralizing force for resource access among intercommunity members during the dry season and perhaps beyond. Chultuns are also a common feature in the southern portion of the EIR, but the majority appear to have been used for purposes other than water storage as evident by their lack of interior plaster sealant and paved catchment areas around their openings (Dunning et al. 2014a; 178

188 McAnany 1990; Scarborough 2003). However, a few examples of water-storage chultuns have been noted around La Milpa and Yaxnohcah. Though none of these features has yet been excavated, there location within drainages or associated of catchment rings strongly supports their hydrologic function. Discussion The above discussion is limited to the capture and storage of water in tanks of variable size in the EIR. Reservoirs, tanks, and chultuns were also used to a lesser extent in ancient Maya communities outside of the EIR. A few areas outside of the EIR also faced severe seasonal water availability challenges and were dependent on rainfall capture, most notably the Vaca Plateau in east-central Belize where the huge urban center of Caracol was dependent of reservoirs that were constructed at central and community scales (Chase 2016) We recognize that the Maya managed water in other forms, most notably to manage soil moisture levels on hillslopes via terracing and in wetland setting via ditching (e.g. Beach et al. 2002; Beach et al. 2009; Luzzadder-Beach et al. 2012; Wyatt 2012, 2014; Beach et al. 2015a; Beach et al. 2015b). Due to severe constraints posed by the natural environment in the EIR, the availability of water was a limiting factor for the human occupation of this region which, nevertheless developed into the well-populated heartland of Maya Civilization. Water was most fundamentally needed for human consumption and sustenance, but other needs were also critical, including bathing, cooking, production of ceramics and plaster, and, where possible, localized irrigation. Water storage capacity has been proposed as a means of estimating ancient urban population numbers (e.g., McAnany 1990). However, this method is problematic, because normal human water needs go far beyond bare-minimum survival needs (see Becquelin and 179

189 Michelet 1994). At best, water can be used as a constraining factor for archaeologically-derived population estimates. Rainwater capture and storage are evident as a fundamental aspect of urban development and design from the outset of population nucleation in the EIR at the beginning of the Middle Preclassic period (c.a., 800 BC). Urban design elements were of both a pragmatic (water collection and storage) as well as symbolic nature: reflecting the basic importance of water in the evolving cosmology of the ancient Maya and the increasingly intertwined nature of religious, political, and economic power in Preclassic and Classic period Maya culture and society. The importance of water for urban Maya society can be seen in the enormous amount of labor invested in its harvest. In addition to examples discussed in the preceding text, there is an archaeological legacy of literally thousands of ancient reservoirs in the EIR, some of truly monumental proportions. For example, the main reservoir at Chactun is some 220 x 170 m (Šprajc 2015), Aguada Maya, near the site of Pozo Maya, is some 280 x 230 m in size (Culbert et al. 1996). The actual capacity and age of these reservoirs is not known due to lack of excavation. Many larger reservoirs, especially those that continue to retain water for part of the year and are not overly obscured by tree canopy, are readily available on multispectral satellite imagery such as that regularly posted on Google Earth, but many are not. The introduction of lidar for mapping in Maya archaeology has the capability of revealing the full extent of reservoir construction (Chase et al. 2012). For example, at Yaxnohcah the acquisition of 25 km 2 of lidar centered over the site revealed 8previously unknown reservoirs in addition to the 5 already known (Anaya Hernandez et al. 2016; Dunning et al. 2016; Reese-Taylor et al. 2016). Lidar also has the capacity to reveal even small closed depressions, many of which served as residential tanks, though excavation is needed to verify their nature and function (Brewer et al. forthcoming). 180

190 Evidence collected from geoarchaeological excavations in several EIR bajos suggests that some of these depressions were likely wetter, with some even holding shallow lakes, in the Preclassic, but progressively dried into and through the Classic, exacerbating the water challenges for Maya occupation (Dunning et al. 2002; Hansen et al. 2002; Dunning et al. 2006a). The Maya Lowlands was also beset with repeated episodes of severe or prolonged droughts, most notably in the Terminal Preclassic and Terminal Classic periods (e.g., Haug et al. 2003; Hodell et al. 2005; Median Elizade et al. 2010; Aimers and Hodell 2011; Medina Elizade and Rohling 2012; Kennett et al. 2012; Dunning et al. 2014b; Douglas et al. 2015). Over the centuries, Maya farmers and rulers alike became acutely aware of the propensity for drought in their environment and sought to intercede via religious ritual as well as the pragmatic response of increasing water collection and storage capacity (e.g., Lucero et al. 2011). In the Late Preclassic period, San Bartolo and Xultun were rival centers only 8 km apart. San Bartolo was abandoned at the end of the Preclassic, while nearby Xultun arose to become one of the most preeminent and long-lived centers of the northeast Peten region. Only one small reservoir is known at San Bartolo, whereas there are at least six reservoirs at Xultun, suggesting that perhaps its rulers or populace responded more effectively to increasing water scarcity (Akpinar-Ferrand et al. 2012). Tikal provides another example of a Preclassic urban center that adapted an elaborate system of reservoirs and emerged as a regional power in the Classic period. Analysis suggests that Classic Tikal was hydrologically overbuilt with its reservoirs potentially supplying more water than needed at least as long as rainfall remained reasonably dependable (Scarborough et al. 2012; Lentz et al. 2014). Drought was not the only factor that could drive the Maya to increase investment in water storage. During a period of escalating regional warfare, the city of Tamarindito (outside of the EIR) chose to invest in a reservoir near the site center atop a 181

191 defensible bluff rather than continue to rely only on several natural springs lying below more easily defended terrain (Beach and Dunning 1997). Over time in the Classic period, water storage features became increasingly complex, diverse, and redundant perhaps reflecting lessons learned in the Terminal Preclassic droughts, perhaps indicative of changing political, economic, or social organization. In addition to large central reservoirs, water storage was also developed at community and household levels. At Xcoch, spatial analysis indicates a positive correlation between chultun and reservoir locations a relationship that suggests that the reservoirs may have functioned in part to allow chultuns to be refilled during the dry season. That is, the old Preclassic reservoirs, refurbished in the Classic period, became a back-up system for household chultuns, perhaps giving some Puuc communities greater resilience to withstand drought (Dunning et al. 2014a). Nevertheless, in the Terminal Classic, droughts may have become too severe to survive, especially give the precariously high populations concentrated in many parts of the EIR (Isendahl et al. 2014). There are more reservoirs located within the urban zone of Uxmal than at any other Puuc site, a fact that likely reflects a natural abundance of karst depressions as well as a huge investment in labor in their modification for water storage (Figure 8). Its locational advantages may have facilitated Uxmal s rise to regional prominence in the late 9 th century AD (Dunning 1992). The apogee of monumental construction at Uxmal between 870 and 910 AD is marked by continued growth at the allied sites of Nohpat and Kabah, but a suppression of construction at many surrounding sites, suggesting that Uxmal was siphoning away labor formerly bound to these other communities during a time of increasing environmental stress related to drought and declining agricultural stress (Isendahl et al. 2015). Evidence suggests that Uxmal was a predatory state that expanded by military conquest, but that itself met a violent end after 910 AD (Carmean et al. 182

192 2004). Uxmal, like almost all of the cities and towns in the EIR was effectively abandoned by 1000 AD. The question arises why most of the EIR remained largely abandoned throughout the Postclassic? Among other reasons that may have made Postclassic resettlement of the Elevated Interior Region by large numbers of Maya difficult was the huge time and labor required to revitalize infrastructure especially for vital water collection and storage systems that had grown incrementally over many generations and lay neglected and increasingly dysfunctional with each passing generation. The Intergovernmental Panel on Climate Change has recommended expanding rainwater harvesting, improving water storage, conservation, and re-use among other strategies to offset negative impacts of climate change in Latin America (Intergovernmental Panel on Climate Change 2007). These are precisely many of the techniques that the ancient Maya honed while seeking sustainable occupation of the EIR, and which could inform present-day population expansion in the region (van der Leeuw 2008; Lucero et al. 2011). Today, the expansion of new towns and villages in many parts of the EIR relies on the pumping of groundwater from deep aquifers, both for domestic purposes and, increasingly, for irrigation agriculture. In some areas tapped freshwater aquifers are already dropping and gypsic or saline aquifers are rising, threatening the long-term sustainability of this development strategy. Resurrecting the rain harvesting adaptations pioneered by countless generations of Maya people offers an alternative model either on its own, or in combination with groundwater-based development. In this way, the ancient Maya could perhaps aid their descendants, and other, non-maya immigrants to enjoy a better future. 183

193 REFERENCES CITED Adams, R.E.W Nucleation of Population and Water Storage Among the Ancient Maya. Science 251:632. Aimers, J.J., and D. Hodell 2011 Societal Collapse: Drought and the Maya. Nature 479: Akpinar-Ferrand, E., N.P. Dunning, D.L. Lentz, and J.G. Jones 2012 Use of Aguadas as Water Management Sources in Two Southern Maya Lowland Sites. Ancient Mesoamerica 23(1): Anaya Hernández, A., M. Peuramaki-Brown, and K. Reese-Taylor 2016 Introducción. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor, pp Mesoweb. Beach, T., S. Luzzadder-Beach, D. Cook, N. Dunning, D. Kennett, S. Krause, F. Valdez, D. Trein, and R. Terry 2015a Ancient Maya Impacts on the Earth's Surface: An Early Holocene Analogue. Quaternary Science Reviews 124:1-30. Beach, T., S. Luzzadder-Beach, S. Krause, N. Brokaw, N.P. Dunning, J. Flood, T. Guderjan, F. Valdez, and S. Walling 2015b "Mayacene" Floodplain and Wetland Formation in the Rio Bravo Watershed, Northwestern Belize. The Holocene 25: Beach, T., S. Luzzadder-Beach, N. Dunning, J. Jones, J. Lohse, T. Guderjan, S. Bozarth, S. Millspaugh, and T. Bhattacharya 2009 A Review of Human and Natural Changes in Maya Lowlands Wetlands Over the Holocene. Quaternary Science Reviews 28: Beach, T., S. Luzzadder-Beach, N.P. Dunning, J. Hageman, and J. Lohse 2002 Upland Agriculture in the Maya Lowlands: Ancient Maya Soil Conservation in Northwestern Belize. Geographical Review 92: Beach, T., and N. Dunning 1997 An Ancient Maya Reservoir and Dam at Tamarindito, El Peten, Guatemala. Latin American Antiquity 8(1): Becquelin, P., and D. Michelet 1994 Demografía en la Zona Puuc: el Recurso del Método. Latin American Antiquity 5(4):

194 Brady, J Offerings to the Rain Gods: The Archaeology of Maya Caves. In Fiery Pool: The Maya and the Mythic Sea, edited by D. Finamore and S.D. Houston, pp Yale University Press, New Haven, Connecticut. Brewer, J.L Understanding the Role of Small Depressions in Ancient Maya Water Management at the Medicinal Trail Site, Northwestern Belize. Unpublished Master s thesis, Department of Anthropology, University of Cincinnati Investigaciones en Aguadas de Escala Residenciales. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor, pp Mesoweb. Forthcoming (in review) Householders as Water Managers: A Comparison of DomesticScale Water Management Practices from Two Central Maya Lowland Sites. Ancient Mesoamerica. Brewer, J.L. and C. Carr 2017 Forthcoming (in press) Operation 22: Continuing Household Water Management Investigations at Yaxnohcah. In Proyecto Arqueológico Yaxnohcah, Informe de la 2016 Temporada de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor. Mesoweb. Brewer, J., D.M. Hyde, and M. Stowe 2013 Mapping Medicinal Trail: A Summary from 2004 to In Research Reports from the Programme for Belize Archaeological Project, Volume 7. Occasional Papers, Number 14, edited by J. Valdez F., pp Mesoamerican Archaeological Research Laboratory, University of Texas at Austin. Brewer, J.L., C. Carr, N.P. Dunning, D.S. Walker, A. Anaya Hernández, M. Peuramaki-Brown, K. Reese-Taylor Forthcoming (in press) Employing Airborne Lidar and Archaeological Testing to Determine the Role of Small Depressions in Water Management at the Ancient Maya Site of Yaxnohcah, Campeche, Mexico. Journal of Archaeological Science: Reports. Bullard, Jr., W.R Maya Settlement Pattern in Northeastern Peten, Guatemala. American Antiquity 25(3): Carmean, K., N. Dunning, and J.K. Kowalski 2004 High Times in the Hill Country: A Perspective from the Terminal Classic Puuc Region. In The Terminal Classic in the Maya Lowlands: Collapse, Transition, and Transformation, edited by A.A. Demarest, P.M. Rice, and D.S. Rice, pp University Press of Colorado, Boulder. 185

195 Chase, A.S.Z Beyond Elite Control: Residential Reservoirs at Caracol, Belize. WIREs Water 3: Chase, A.F., D.Z. Chase, C. Fisher, S. Leisz, and J. Weishampel 2012 The Geospatial Revolution in Mesoamerican Archaeology. Proceedings of the National Academy of Sciences 109: Chmilar, J Ancient Water Management: Archaeological Investigations at Turtle Pond, Northwestern Belize. Unpublished Master s thesis, Department of Anthropology, University of Cincinnati. Culbert, T.P., V. Fialko, B. McKee, L. Grazioso, and J. Kunen 1996 Investigación Arqueológica en el Bajo La Justa: la Temporada de In X Simposio de Investigaciones Arqueológicas en Guatemala, edited by J.P. Laporte and H.L. Escobedo, pp Museo Nacional de Arqueología y Etnología de Guatemala, Guatemala City. Dahlin, B.H., J. Foss, and M.E. Chambers 1980 Project Akalches: Reconstructing the Natural and Cultural History of a Seasonal Swamp: Preliminary Results. In El Mirador, El Petén, Guatemala: An Interim Report. Papers of the New World Archaeological Foundation No. 45, edited by R.T. Matheny, pp Brigham Young University, Provo, Utah. Domínguez Carrazco, M., and W.J. Folan 1996 Calakmul, México: Aguadas, Bajos, Precipitación y Asentamineto en el Peten Campechano. In IX Simposio de Investigaciónes Arqueológicas en Guatemala, edited by J.P. Laporte and H.L. Escobedo, pp Museo Nacional de Antropología y Historia, Guatemala City. Douglas, P.M., M. Pagani, M.A. Canuto, M. Brenner, D.A. Hodell, T.I. Eglinton, and J.H. Curtis 2015 Drought, Agricultural Adaptation, and Sociopolitical Collapse in the Maya Lowlands. Proceedings of the National Academy of Sciences 112: Dunning, N Lords of the Hills: Ancient Maya Settlement in the Puuc Region, Mexico. In Monographs in World Archaeology No. 15. Prehistory Press, Madison, Wisconsin Birth and Death of Waters: Environmental Change, Adaptation, and Symbolism in the Southern Maya Lowlands. In Espacios Mayas: Representaciones, Usos, Creencias, edited by A. Breton, A. Monod-Becquelin and M.H. Ruz, pp Universidad Autonoma de México, Mexico City. Dunning, N.P Down on the Farm: Classic Maya Houselots as Farmsteads. In Ancient Maya 186

196 Commoners, edited by J. Lohse and F. Valdez, pp University of Texas Press, Austin. Dunning, N., and T. Beach 2010 Farms and Forests: Spatial and Temporal Perspectives on Ancient Maya Landscapes. In Landscapes and Societies, edited by I.P. Martini and W. Chesworth, pp Springer-Verlag, Berlin. Dunning, N.P., and S.D. Houston 2011 Hurricanes as a Disruptive Force in the Maya Lowlands. In Ecology, Power, and Religion in Maya Landscapes, edited by C. Isendahl and B.L. Persson, pp Verlag Anton Sauerwein, Möckmühl, Germany. Dunning, N. and E. Weaver 2015 Final Thoughts: Space, Place, Ritual, and Identity in Ancient Mesoamerica. In Memory Traces: Sacred Space at Five Mesoamerican Sites, edited by L. Amrhein and C. Kristen-Graham, pp University of Colorado Press, Boulder. Dunning, N., V. Scarborough, J. Valdez F., S. Luzzadder-Beach, T. Beach, and J.G. Jones 1999 Temple Mountains, Sacred Lakes, and Fertile Fields: Ancient Maya Landscapes in Northwestern Belize. Antiquity 73: Dunning, N.P., S. Luzzadder-Beach, T. Beach, J.G. Jones, V.L. Scarborough, and T.P. Culbert 2002 Arising from the Bajos: The Evolution of a Neotropical Landscape and the Rise of Maya Civilization. Annals of the Association of American Geographers 92: Dunning, N., J.G. Jones, T. Beach, and S. Luzzadder-Beach 2003 Physiography, Habitats, and Landscapes of the Three Rivers Region. In Heterarchy, Political Economy, and the Ancient Maya: The Three Rivers Region of the East-Central Yucatan Peninsula, edited by V. Scarborough, F. Valdez and N.P. Dunning, pp University of Arizona, Tempe. Dunning, N., T. Beach, and S. Luzzadder-Beach 2006a Environmental Variability Among Bajos in the Southern Maya Lowlands and its Implications for Ancient Maya Civilization and Archaeology. In Precolumbian Water Management: Ideology, Ritual, and Power, edited by L. Lucero and B. Fash, pp University of Arizona Press, Tucson. Dunning, N.P., J. Jones, J. Chmilar, and J. Blevins 2006b Investigaciones Ambientales y Geoarqueologicas in las Alrededoras de San Bartolo. In Proyecto Arqueológico Regional San Bartolo, Informe Preliminar #3, Tercera Temporada 2005, edited by W. Saturno and M. Urquizú. Instituto Nacional de Antropología y Historia, Guatemala City. 187

197 Dunning, N.P., E. Akpinar, C. Carr, R. Griffin, J.G. Jones, D. Lentz, A. Miller, and J. Prater 2008 Investigaciones Geoarqueológicas y del Medioambiente en los Alrededores de San Bartolo, Peten: In Proyecto Arqueológico San Bartolo: Informe Preliminar No. 6, Sexta Temporada 2007, edited by M. Urquizú and W. Saturno. Instituto Nacional de Antropología y Historia, Guatemala City. Dunning, N.P., T. Beach, and S. Luzzadder-Beach 2012 Kax and Kol: Collapse and Resilience in Lowland Maya Civilization. Proceedings of the National Academy of Sciences 109: Dunning, N., E. Weaver, M. Smyth, and D. Ortegón 2014a Xcoch: Home of Ancient Maya Rain Gods and Water Managers. In The Archaeology of Yucatan: New Directions and Data, edited by T. Stanton, pp BAR International Series, Oxford. Dunning, N., D. Wahl, T. Beach, J. Jones, S. Luzzadder-Beach, and C. McCormick 2014b The End of the Beginning: Drought, Environmental Change, and the Preclassic to Classic Transition in the East-Central Yucatan Peninsula. In The Great Maya Droughts in Cultural Context, edited by G. Iannone, pp University of Colorado Press, Boulder. Dunning, N., R. Griffin, J. Jones, R. Terry, Z. Larsen, and C. Carr 2015 Life on the Edge: Tikal in a Bajo Landscape. In Tikal: Paleoecology of an Ancient Maya City, edited by D.L. Lentz, N. Dunning and V. Scarborough, Cambridge University Press, Cambridge. Dunning, N.P., A. Anaya Hernández, and H. Geovannini 2016 Operaciones 13 y 19: Investigaciones en los Reservorios. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor, pp Mesoweb. Dunning, N.P., A. Anaya Hernández, A. Haggard, and C. Carr 2017a (Forthcoming) Investigaciones en el Reservario Brisa. In Proyecto Arqueológico Yaxnohcah, Informe de 2016 Temporada de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor. Mesoweb. Dunning, N., R. Griffin, T. Sever, W. Saturno, and J. Jones 2017b The Nature and Origin of Linear Features in the Bajo de Azúcar, Guatemala: Implications for Ancient Maya Adaptations to a Changing Environment. Geoarchaeology 32: Ebert, C., J. Awe, and J.A. Hogarth 2016 Classic Period Maya Water Management and Ecological Adaptation in the Belize River Valley. Research Reports in Belizean Archaeology 13: Fialko, V Recursos Hidraulicos en Tikal y sus Periferias. In XIII Simposio de Investigaciones 188

198 Arqueológicas en Guatemala, edited by J.P. Laporte, H. Escobedo, B. Arroyo and A.C. Suasnavar, pp Museo Nacional de Arqueología y Etnología, Guatemala City Diez Años de Investigaciones Arqueológicas en la Cuenca del Río Holmul, Region Norteste del Petén. In XVIII Simposio de Investigaciones Arqueológicas en Guatemala, edited by J.P. Laporte, B. Arroyo and H. Mejía, pp Museo Nacional de Arqueología y Etnología, Guatemala. Folan, W.J Mining and Quarrying Techniques of the Lowland Maya. Anthropology 6: Ford, A Critical Resource Control and the Rise of the Ancient Maya. In The Managed Mosaic: Ancient Maya Agriculture and Resource Use, edited by S.L. Fedick, pp University of Utah, Salt Lake City. Frappier, A.B., J. Pyburn, A.D. Pinkey-Drobnis, X. Wang, D.R. Corbett, and B.H. Dahlin 2014 Two Millennia of Tropical Cyclone-Induced Mud Layers in a Northern Yucatán Stalagmite: Multiple Overlapping Climatic Hazards During the Maya Terminal Classic Megadroughts. Geophysical Research Letters 41: Geovannini-Acuña, H Rain Harvesting in the Rainforest: The Ancient Maya Agricultural Landscape of Calakmul, Campeche, Mexico. BAR International Series, Oxford. Gill, E Operation 15: Berm Structures and Water Management at Medicinal Trail (RB 62), Belize. In Research Reports from the Programme for Belize Archaeological Project, Volume 3. Occasional Papers, Number 10, edited by R. Trachman and J. Valdez F., pp Mesoamerican Archaeological Research Laboratory, University of Texas at Austin. Gill, R.B The Great Maya Droughts: Water, Life, and Death. University of New Mexico Press, Albuquerque. Grazioso Sierra, L., and V. Scarborough 2013 Control de Agua por los Antiguos Mayas: El Sistema Hidráulico de Tikal. Contributions to New World Archaeology 5: Hansen, R.D., S. Bozarth, J. Jacob, D. Wahl, and T. Schreiner 2002 Climatic and Environmental Variability in the Rise of Maya Civilization: A Preliminary Perspective from Northern Peten. Ancient Mesoamerica 13: Harrison, P.D Aspects of Water Management in the Southern Maya Lowlands. In Economic Aspects of Water Management in the Prehispanic New World, edited by V.L. Scarborough and B.L. 189

199 Isaac, pp Research in Economic Anthropology, supplement No. 7, JAI Press, Greenwich, Connecticut. Haug, G.H., D. Gunther, Peterson L.C., D.M. Sigman, K.A. Hughen, and B. Aeschliman 2003 Climate and the Collapse of Maya Civilization. Science 299: Hodell, D.A., M. Brenner, and J.H. Curtis 2005 Terminal Classic Drought in the Northern Maya Lowlands Inferred from Multiple Sediment Cores in Lake Chichancanab (Mexico). Quaternary Science Review 24: Houston, S.D Living Waters and Wondrous Beasts. In Fiery Pool: The Maya and the Mythic Sea, edited by D. Finamore and S.D. Houston, pp Yale University Press, New Haven, Connecticut. Huchim Herrera, J. and I. Sánchez 1990 El Sistema Hidraúlico de Uxmal. Boletin Academico de la Facultad Ingeneria (Universidad Autonoma de Yucatan) 13: Hyde, D.M Power Dynamics at a Commoner Hinterland Community in the Maya Lowlands: The Medicinal Trail Site, Northwestern Belize. Unpublished Ph.D. dissertation, Department of Anthropology, University of Texas at Austin. Intergovernmental Panel on Climate Change Summary for Policymakers of the Synthesis Report of the IPCC Fourth Assessment Report. IPCC, Geneva, Switzerland. Electronic Document, accessed March 16, Isendahl, C The Weight of Water: A New Look at Pre-Hispanic Puuc Maya Water Reservoirs. Ancient Mesoamerica 22: Isendahl, C., N. Dunning, and J. Sabloff 2014 Growth Dependency and Decline in Classic Maya Puuc Political Economies. Archaeological Papers of the American Anthropological Association 24: Isendahl, C., V. Scarborough, J. Gunn, N. Dunning, S. Fedick, G. Iannone, and L. Lucero 2015 Applied Perspectives on Pre-Columbian Maya Water Management Systems: What are the Insights for Water Security? In The Oxford Handbook on Historical Ecology and Applied Archaeology, edited by C. Isendahl and D. Stump. Oxford University Press, Oxford. Johnston, K.J Lowland Maya Water Management Practices: The Household Exploitation of Rural Wells. Geoarchaeology 19:

200 Kennett, D.J., S.F.M. Breienbach, V.V. Aquino, Y. Asmerom, J. Awe, J.U.L. Baldini, P. Bartlein, B.J. Culleton, C. Ebert, C. Jazwa, M.J.M. Macri N., V. Polyak, K.M. Prufer, H.E. Ridley, H. Sodemann, B. Winterhalder, and G.H. Haug 2012 Development and Disintegration of Maya Political Systems in Response to Climate Change. Science 338: Kettunen, H., and C. Helmke 2013 Water in Maya Imagery and Writing. Contributions to New World Archaeology 5:1738. Koszkul, W., and J. Zralka 2013 El Manejo Ritual y Práctico del Agua: El Caso del Edificio 14 de Nakum, Guatemala. Contributions to New World Archaeology 5: Lemonnier, E., and B. Vanniere 2013 Agrarian Features, Farmsteads, and Homesteads in the Río Bec Nuclear Zone (Mexico). Ancient Mesoamerica 24(2): Lentz, D.L., N.P. Dunning, V.L. Scarborough, K. Magee, K.M. Thompson, E. Weaver, C. Carr, R.E. Terry, G. Islebe, K.B. Tankersley, L. Grazioso Sierra, J.G. Jones, P. Buttles, Valdez, F., and C. Ramos Hernandez 2014 Farms, Forests, and the Edge of Sustainability at the Ancient Maya City of Tikal. Proceedings of the National Academy of Sciences 111: Lowe, J.D Excavations of Two Small Depressions at Group A, The Medicinal Trail Site RB 62, Belize. Unpublished Master s thesis, Department of Anthropology, University of Texas at Austin. Lucero, L.J Water Control and Maya Politics in the Southern Maya Lowlands. In Complex Polities in the Ancient Tropical World, edited by E.A. Bacus and L.J. Lucero, pp Archaeological Papers of the American Anthropological Association Number 9. American Anthropological Association, Arlington, Virginia The Collapse of the Classic Maya: A Case for the Role of Water Control. American Anthropologist 104(3): Water and Ritual: The Rise and Fall of Classic Maya Rulers. University of Texas Press, Austin. Lucero, L.J., J.D. Gunn, and V.L. Scarborough 2011 Climate Change and Classic Maya Water Management. Water 3:

201 Lundell, C.L The Vegetation of the Petén. In Carnegie Institution of Washington Publication 478. Carnegie Institution, Washington, D.C. Luzzadder-Beach, S., T. Beach, and N.P. Dunning 2012 Wetland Fields as Mirrors of Drought and the Maya Abandonment. Proceedings of the National Academy of Sciences 109: Luzzadder-Beach, S., T. Beach, S. Hutson, and S. Krause 2016 Sky-Earth, Lake-Sea: Climate and Water in Maya History and Landscape. Antiquity 90: Matheny, R.T Northern Maya Lowland Water-Control Systems. In Pre-Hispanic Maya Agriculture, edited by P.D. Harrison and B.L. Turner II, pp University of New Mexico Press, Albuquerque. Matheny, R.T., R.D. Hansen, and D. Gurr 1980 Preliminary Field Report, El Mirador, 1979 Season. In El Mirador, El Petén, Guatemala: An Interim Report. Papers of the New World Archaeological Foundation No. 45, edited by R.T. Matheny, pp Brigham Young University, Provo, Utah. Matheny, R.T., D. Gurr, W. Forsyth, and R. Hauck 1983 Investigations at Edzná Campeche, Mexico. Volume 1, Part 1: The Hydraulic System. In Papers of the New World Archaeological Foundation No. 46, edited by R.T. Matheny. Brigham Young University, Provo, Utah. McAnany, P.A Water Storage in the Puuc Region of the Northern Maya: A Key to Population Estimates and Architectural Variability. In Precolumbian Population History in the Maya Lowlands, edited by P.T. Culbert and D. Rice, pp University of New Mexico Press, Albuquerque. Medina-Elizade, M., and E.J. Rohling 2012 Collapse of Classic Maya Civilization Related to Modest Reduction in Precipitation. Science 335: Medina-Elizade, M., S. Burns, D.W. Lea, S. Asmerom, L. van Gunten, V. Polyak, M. Vuille, and A. Karmalkar 2010 High Resolution Stalagmite Climate Record from the Yucatan Peninsula Spanning the Maya Terminal Classic Period. Earth & Planetary Science Letters 298: Medina-Elizade, M., J.M. Martín, F. Lases-Hernández, R. Bradley, and S. Burns 2016 Testing the "Tropical Storm" Hypothesis of Yucatan Peninsula Climate Variability During the Maya Terminal Classic Period. Quaternary Research 86:

202 Morales-Aguilar, C.A El Sistema Hidráulico de El Mirador, Peten, Guatemala: Una Perspectiva General. Proyecto Arqueologico Cuenca Mirador, Foundation for Archaeological Research and Environmental Studies (FARES). Morley, S.G The Inscriptions of Petén, Volume 1. In Carnegie Institution of Washington, Publication 437. Carnegie Institution, Washington, D.C. Percy, M Excavation at the Medicinal Trail Site (2008): Operation 14. In Research Reports for the Programme for Belize Archaeological Project, Volume 3. Occasional Papers, Number 10, edited by R. Trachman and J. Valdez F., pp Mesoamerican Archaeological Research Laboratory, University of Texas at Austin. Peuramaki-Brown, M., S.G. Morton, and A. Castro 2016 Operacion 17 en el Grupo Wo': Excavaciones de Prueba en un Lote Doméstico PeriUrbano. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor, pp Mesoweb. Pohl, M.D., K.O. Pope, J.G. Jones, J.S. Jacob, D.R. Piperno, S.D. defrance, D.L. Lentz, J.A. Gifford, M.E. Danforth, and J.K. Josserand 1996 Early Agriculture in the Maya Lowlands. Latin American Antiquity 7(4): Rands, R The Water Lily in Maya Art: A Complex of Alleged Asiatic Origin. Bulletin 151: Bureau of American Ethnology, Smithsonian Institution, Washington, D.C. Reese-Taylor, K., and A. Anaya Hernández 2013 Proyecto Arqueológico Yaxnohcah, Informe de la Primera Temporada de Investigaciones. Mesoweb. Reese-Taylor, K., A. Anaya Hernández, A. Flores Esquival, K. Monteleone, A. Uriarte, C. Carr, H. Geovannini Acuña, J.C. Fernandez-Diaz, M. Peuramaki-Brown, and N.P. Dunning 2016 Boots on the Ground at Yaxnohcah: Ground-Truthing Lidar in a Complex Tropical Landscape. Advances in Archaeological Practice 4: Reese-Taylor, K., and A. Anaya Hernández (editors) 2017 (Forthcoming) Proyecto Arqueológico Yaxnohcah, Informe de 2016 Temporada de Investigaciones. Ringle, W.M The Yaxhom Valley Survey: Pioneers of the Puuc Region. Research Grant , National Geographic Society, Washington, D.C. 193

203 Rosenmeier, M.F., D.A. Hodell, M. Brenner, J.H. Curtis, J.B. Martin, F.S. Anselmetti, D. Ariztegui, and T.P. Guilderson 2002 Influence of Vegetation Change on Watershed Hydrology: Implications for Paleoclimatic Interpretation of Lacustrine Records. Journal of Paleolimnology 27(1): Saturno, W., D. Stuart, and B. Beltran 2006 Early Maya Writing at San Bartolo, Guatemala. Science 311: Scarborough, V.L Water Management in the Southern Maya Lowlands: An Accretive Model for the Engineered Landscape. In Economic Aspects of Water Management in the Prehispanic New World, edited by V.L. Scarborough and B.L. Isaac, pp Research in Economic Anthropology, supplement No. 7, JAI Press, Greenwich, Connecticut Ecology and Ritual: Water Management and the Maya. Latin American Antiquity 9(2): The Flow of Power: Ancient Water Systems and Landscapes. SAR Press, Santa Fe, New Mexico. Scarborough, V.L., and G.G. Gallopin 1991 A Water Storage Adaptation in the Maya Lowlands. Science 251: Scarborough, V.L., and L. Grazioso Sierra 2015 The Evolution of an Ancient Earthworks System at Tikal. In Tikal: Paleoecology of an Ancient Maya City, edited by D.L. Lentz, N.P. Dunning and V.L. Scarborough, pp Cambridge University Press, New York. Scarborough, V.L., R. Connelly, and S. Ross 1994 The Prehispanic Maya Reservoir System at Kinal, Peten, Guatemala. Ancient Mesoamerica 5: Scarborough, V.L., N.P. Dunning, K.B. Tankersley, C. Carr, E. Weaver, L. Grazioso, B. Lane, J.G. Jones, P. Buttles, F. Valdez, and D.L. Lentz 2012 Water and Sustainable Land Use at the Ancient Tropical City of Tikal, Guatemala. Proceedings of the National Academy of Sciences 109(31): Schele, L., and D. Friedel 1990 A Forest of Kings: The Untold Story of the Ancient Maya. Morrow, New York. Seefeld, N Public Provisions for Dry Seasons: The Hydraulic System of Uxul and its Relevance for the Survivability of the Settlement. Contributions to New World Archaeology 5:

204 Siemens, A Karst and the Pre-Hispanic Maya in the Southern Lowlands. In Pre-Hispanic Maya Agriculture, edited by P.D. Harrison and B.L. Turner II, pp University of New Mexico Press, Albuquerque. Smyth, M.P., and D. Ortegón Zapata 2008 A Preclassic Center in the Puuc Region: A Report on Xcoch, Yucatan, Mexico. Mexicon 30: Smyth, M., N.P. Dunning, E. Weaver, P. van Beynen, and D. Ortegón Zapata 2017 (in press) The Perfect Storm: Climate Change and Ancient Maya Response in the Puuc Hills Region, Yucatan, Mexico. Antiquity. Šprajc, I Reconocimiento Arqueológico en el Sureste del Estado de Campeche: Paris Monographs in American Archaeology 19. BAR International Series 1742, Archaeopress, Oxford (editor) Exploraciones Arqueológicas en Chactún, Campeche, México. Prostor, Kraj, Cas 7, Zalozba, ZRC, Ljubljana, Slovenia. Stephens, J.L Incidents of Travel in Central America, Chiapas, and Yucatan. Vol. 2, Harper & Brothers, New York. Tokovinine, A Place and Identity in Classic Maya Narratives. In Studies in Pre-Columbian Art and Archaeology 37. Dumbarton Oaks, Washington, D.C. van der Leeuw, S.E Climate and Society: Lessons from the Past 10,000 Years. Royal Swedish Academy of Science 14: Volta, B., R. González Heredia, L.F. Folan, W.J. Folan, and A. Morales López 2013 Los Rasgos Hidráulicos de Oxpemul, Campeche. Los Investigadores de la Cultura Maya 21: Wahl, D., T. Schreiner, R. Byrne, and R. Hansen 2007 A Paleoecological Record from a Late Classic Maya Reservoir in the North Petén. Latin American Antiquity 18(2): Walker, D.S Apuntes Sobre la Secuencia Cerámica de Yaxnohcah. In Proyecto Arqueológico Yaxnohcah, Informe de las 2014 y 2015 Temporadas de Investigaciones, edited by A. Anaya Hernández, M. Peuramaki-Brown and K. Reese-Taylor, pp Mesoweb. 195

205 Weaver, E., C. Carr, N. Dunning, L. Florea, and V. Scarborough 2015 Examining Landscape Modifications for Water Management at Tikal Using 3Dimensional Modeling with ArcGIS. In Tikal: Paleoecology of an Ancient Maya City, edited by D. Lentz, N. Dunning and V. Scarborough, pp Cambridge University Press, Cambridge. Weiss-Krejci, E Ancient Maya Rainwater Reservoirs in Northwestern Belize. Contributions to New World Archaeology 5: Weiss-Krejci, E., and T. Sabbas 2002 The Potential Role of Small Depressions as Water Features in the Central Maya Lowlands. Latin American Antiquity 13(3): Winemiller, T.L Water Resource Management by the Ancient Maya of Yucatan, Mexico. Unpublished Ph.D. dissertation, Louisiana State University, Department of Geography and Anthropology. Wyatt, A.R Agricultural Practices at Chan: Farming and Political Economy in an Ancient Maya Community. In Chan: An Ancient Maya Farming Community, edited by C. Robin, pp University of Florida Press, Gainesville The Scale and Organization of Ancient Maya Water Management. WIREs Water 1(5):

206 Figure 1. Map showing the situation of the Elevated Interior Region (EIR) with the Maya Lowlands and sites discussed in the text. 197

207 Figure 2. Diagram showing stream-damming and depression-filling reservoirs [note: this is a draft]. 198

208 Figure 3a. Profile of south wall of Op. SB3C-1 in Aguada San Bartolo. Unit 8 is a Middle Preclassic plaster floor; Units 4 7 are Middle and Late Preclassic sediments; 3 is a badly distorted Late Classic plaster floor; Units 1 2 are the modern soil formed on Late Classic and later sediment. Figure 3b. Lidar-derived hillshade image of a site center reservoir and associated ceremonial center (Brisa Reservoir, Yaxnohcah). 199

209 Figure 4. Cross-sectional drawing of Aguada Tintal, northeast of San Bartolo, showing disjunction between Late Preclassic and Terminal Classic and later sediment within the reservoir. Figure 5. Cross-sectional drawing of La Milpa Aguada, including buk te in reservoir floor; inset photo showing earlier, lower part of the buk te after excavation. 200

210 Figure 6a. Idealized drawing of Puuc residential group with chultun. Figure 6b. Photo of chultun with reinforced neck and sloping catchment (Yaxhom). 201

211 Figure 7a. Idealized drawing of residential tank. Figure 7b. Photo of a residential tank at Yaxnohcah prior to excavation. Figure 7c. Photo of excavation in Yax-3 Residential Group tank (Yaxnohcah). 202

212 Figure 8. Space Shuttle photo of Uxmal showing some of the many aguadas at the site; inset photo showing Aguada ChenChan in the foreground and site center in the background. Table 1. Estimated Capacity of Selected EIR Reservoirs Site Name Reservoir Name Estimated Capacity/m 3 El Zotz El Zotz Aguada 87,920 Yaxnohcah Brisa Reservoir 84,000 Xcoch Aguada La Gondola 79,200 Uxmal Aguada ChenChan 75,000 Tikal Palace Reservoir 74,630 Tikal Corriental 57,559 Yaxnohcah Fidelia 29,000 Xcoch South Aguada 1 12,000 Uxul Aguada Oriental 10,000 La Milpa Aguada Lagunita Elusiva 7,800 La Milpa Turtle Pond