Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier in the Eastern Nepal Himalaya since 1964

83 Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier in the Eastern Nepal Himalaya since 1964 Takanobu SAWAGAKI 1*, Damodar LAMSAL 2, Alton C BYERS 3 and Teiji WATANABE 1 1 Faculty of Environmental Earth Science, Hokkaido University N10, W5, Sapporo, 060-0810, Japan 2 Graduate School of Environmental Science, Hokkaido University N10, W5, Sapporo, 060-0810, Japan 3 The Mountain Institute, 100 Campus Drive, 108 LA Elkins, WV 26241, USA *e-mail: sawagaki@ees.hokudai.ac.jp Abstract We carried out multi-date morphological mappings to document the development of the glacial lake Chamlang South Tsho in the eastern Nepal Himalaya over four decades. High-resolution Corona KH-4A and Advanced Land Observing Satellite (ALOS) PRISM stereo-data taken in 1964 and 2006 were processed in the Leica Photogrammetric Suite (LPS) to generate digital terrain models (DTMs). The DTMs produced topographic maps representing elevations and morphology of the glacier surface with a maximum error of +/- 10 m. A bathymetric map was also produced based on sonar sounding data obtained in November 2009. Extensive surface lowering was found to have occurred since 1964, as high as 156.9 m in the upper glacier area. The average lowering of the glacier for the entire 42-year period from 1964 to 2006 is 37.5 m, with the average surface-lowering rate calculated at 0.9 m/year. The average surface lowering for the 45 years from the glacier surface in 1964 to the lake bottom in 2009 was 99.5 m at a rate of 2.2 m/year. The minimum and maximum surface lowering during that period were 12 m and 153.8 m, respectively. The area surrounding the largest supraglacial pond in 1964 exhibited a low surface gradient, and there had already been a large degree of ice melting that favoured further lake expansion. The larger lowering rate at the lake area supports the previously presented idea that ice calving into the pond triggered the larger and faster up-glacial lake expansion. Key words: ALOS PRISM, Chamlang South Glacier, Corona, digital terrain model, glacial lakes, Nepal Himalaya, photogrammetry, satellite images, surface morphology, topographic map 1. Introduction Glaciers in the Hindu Kush-Himalaya have been receding and melting since the end of the Little Ice Age in 1850 (Fujita et al., 2001; Bolch et al., 2008a, 2011; Shrestha & Aryal, 2010; Fujita & Nuimura, 2011). One of the consequences has been the formation and expansion of glacial lakes, particularly during the last half century, as a result of glacial retreat that is most likely in response to contemporary global warming and regional warming trends (Yamada & Sharma, 1993; Mool et al., 2001; Bajracharya et al., 2007; Watanabe et al., 2009). As a result, the potential for glacial lake outburst floods (GLOFs) is increasing throughout the Hindu Kush- Himalaya (e.g., Watanabe et al., 1994). Supraglacial ponds and lakes are likely to form on the surface of debris-covered glaciers when conditions are Global Environmental Research 16/2012: 83-94 printed in Japan favorable (Reynolds, 2000; Quincey et al., 2007), i.e., exhibiting a glacier surface gradient of less than 2, low ice velocity, and/or complete stagnation. Sakai and Fujita (2010) discussed the importance of the glacial surface lowering since the Little Ice Age. Glacier recession and associated phenomena have multiple impacts and significances, all of which need to be considered in an integrated manner (Haeberli et al., 2007; Hegglin & Huggel, 2008). Most importantly, an understanding of the recession of debris-covered glaciers is absolutely necessary in terms of hazards assessment, primarily because the glacial lakes that form on them sometimes produce devastating GLOFs (Clague & Evans, 2000; Richardson & Reynolds, 2000a, 2000b; Ives et al., 2010), which are often several times larger than normal climatic floods (Cenderelli & Wohl, 2001). GLOFs can cause widespread damage and loss of life, property, infrastructure, 2012 AIRIES

84 T. SAWAGAKI et al. and the geomorphic environment (Ives, 1986; Vuichard & Zimmermann, 1987; Cenderelli & Wohl, 2003; Narama et al., 2010). A thorough understanding of glacial recession is also critical to the estimation of the glacier mass balance for water resource management within a given region (Bolch et al., 2008b; Immerzeel et al., 2010). Mapping of glacier area change can be carried out with non-stereo orthorectified imagery; however, multidate, stereo-capable data are required in order to calculate volumetric changes of glaciers, especially for the debriscovered type. Multi-date stereo-images with long-period time differences as well as high spatial resolution are needed to produce accurate and detailed 3D topographic or terrain maps, that in turn can expand our understanding of the morphological changes of a glacier s surface. Photographic products from the Corona program (~1960 to 1972) are the oldest remote sensing stereo images publicly available with a wide geographic coverage and high spatial resolution (~1.8 to 7.6 m). Likewise, the Advanced Land Observing Satellite (ALOS) PRISM, is a relatively new remote sensing satellite program (launched in 2006) that has stereo capability capable of generating digital terrain models (DTMs) and 3D maps, and which also offers high spatial resolution stereo-data (2.5 m). Several studies have investigated volumetric changes in glaciers in the Himalaya using either Corona or ALOS data (e.g., Bolch et al., 2008a; Lamsal et al., 2011). Corona data have a complex geometry, lack metadata (ephemeris) pertaining to image acquisition, and require highly sophisticated image- processing software to extract 3D information. Despite this, Corona data provide an invaluable source for investigation in various fields such as glaciology and archaeology. Detailed information on Corona and ALOS PRISM data and their processing is described well by Lamsal et al. (2011). During the past several decades in the Mt. Everest and Makalu-Barun national parks of Nepal, 24 new glacial lakes have formed and 34 major lakes are reported to have grown substantially as a result of climate change and regional warming trends (Bajracharya et al., 2007). Recent analyses of comparative satellite imagery have suggested that at least twelve of the new or growing lakes within the Dudh Kosi watershed, nine of which are located in the remote Hongu valley of Makalu-Barun National Park, are potentially dangerous based on their rapid growth over the past several decades (Bajracharya et al., 2007; Bajracharya & Mool, 2009). However, despite the large amount of national and international media attention recently generated by these new and/or growing lakes in the Hongu valley, relatively little was known about them in 2009 because of their extreme remoteness and difficult access. In response, we carried out the first scientific field investigation of these glacial lakes in October-November of 2009 (Fig. 1). At the time, a glacial lake situated at the southwestern foot of Mt. Chamlang was considered to be Fig. 1 Location of the Chamlang South Tsho in the Hongu valley, eastern Nepal Himalaya. CST: Chamlang South Tsho, CNT: Chamlang North Tsho.

Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier 85 among the most dangerous of the nine potentially dangerous lakes within the valley (Mool et al., 2001; Bajrachayra et al., 2007; ICIMOD, 2011), prompting the research team to focus its activities on the Chamlang South Glacier while conducting preliminary analyses of the other eight lakes. Respecting its positional relationship with the peak, we call this lake Chamlang South Tsho although it has been referred to by various other names. The aim of our study was to examine changes in surface morphology of the Chamlang glacier and the lake between 1964 and 2006. 2. Study Area Chamlang South Tsho (longitude: 86 57 33 E, latitude: 27 45 24 N), has been called different names by different researchers, e.g., Chamlang Tsho in the 1:50000 Schneider map, Chamlang Pokhari by Byers et al. (2012), and most commonly West Chamlang in the 1:50000 topographic map by the Survey Department of Nepal and in Mool et al. (2001) and Bajracharya et al. (2007), which also provided an identification number of kdu_gl 466. Mt. Chamlang has two glacial lakes on the westward side of the peak: one in the northwest, with no local name but an identification number of kdu_gl 464 (Bajracharya et al., 2007), and the other in the southwest region of the peak (Fig. 1). To avoid confusion, we propose the use of the name of Chamlang South Tsho for the lake located in the southwest region. Chamlang South Tsho has developed rapidly between 1964 and 2000 on the debris-covered Chamlang South Glacier (Fig. 2). The altitude of the lake surface is 4,940 m. 3. Data and Methods 3.1 Image processing Two sets of stereo-data (ALOS PRISM and Corona) with a time span of over four decades were used. ALOS PRISM data (acquired on 4 December 2006; Path: 140, Row: 41, spatial resolution: 2.5 m) with processing level 1B1 (radiometrically calibrated data) plus Rational Polynomial Coefficient (RPC) data files were obtained from the Remote Sensing Technology Center of Japan (RESTEC). ALOS PRISM acquires data in triple mode F, N, and B (forward, nadir, and backward). With these data, <10 m geometric accuracy is achievable (Tadono et al., 2009). Images acquired from Corona camera systems KH-1, KH-2, KH-3, KH-4, KH-4A, and KH-4B were declassified in 1995 and became available in a digital format in 2003 (McDonald, 1995; Galiatsatos et al., 2008). The early systems of KH-1, KH-2, and KH-3 were equipped with a single panoramic camera, while the latter systems of KH-4, KH-4A, and KH-4B were equipped with both forward- and backward-looking cameras. The images taken by the KH-4, KH-4A, and KH-4B camera systems, which have a stereo capability, a higher spatial resolution Fig. 2 Landform features and surrounding geomorphic environments of the Chamlang South Glacier. Corona KH-4A image depicting the supraglacial ponds and glacier surface in 1964 (a) and IKONOS image acquired in 2000 (b) showing the morphological change of a large lake (Chamlang South Tsho) from 1964 to 2000. C1 C5: Abandoned channels, P: Current largest pond.

86 T. SAWAGAKI et al. (3.00-7.60, 2.70-7.60 and 1.8-7.60 m, respectively) and wide area coverage offer the oldest available stereo data. Corona panoramic filmstrips contain inherent geometric distortions, which need to be rectified to extract reliable information from its images (Slama, 1980; Altmaier & Kany, 2002). Typical geometric distortions of a Corona image gradually increase from the centre to the extreme edges of an image along the track (Slama, 1980; Altmaier & Kany, 2002; Lamsal et al., 2011). The processing of Corona stereo images for DTM generation has been well described by Altmaier and Kany (2002), Bolch et al. (2008a) and Lamsal et al. (2011). Corona KH-4A stereo-pair images (acquired on 26 November 1964, entity ID: DS1014-2118) of the study site were obtained from the USGS in a digital format scanned at 3,600 dpi (7 microns). Stereo images were processed with the aid of the Leica Photogrammetric Suite (LPS). DTMs and topographic maps were produced after rigorous editing of the triangulated irregular network (TIN) model. Stereo Mirror TM /3D Monitor and Leica 3D TopoMouse greatly facilitated and enhanced the tasks of image processing such as ground control point (GCP) collection and terrain editing. A fully editable Leica Terrain Format (LTF) and a TIN (vector DTM) model was created for each dataset, so that the editing of inherent errors took place with automatically generated DTM to produce accurate 3D terrain maps. ALOS PRISM images (F, N and B) with RPC file and three GCPs (collected from a map of the Everest area compiled by the National Geographic Society in 1988) were employed to triangulate (RMSE ~0.40 pixel) stereo images to create a stereo model and were used for DTM generation. A stereo model was generated for the entire area covered by the scenes; however, only a DTM of the study area was created. The vector DTM was edited upon viewing the stereo model as a base. Extensive editing was carried out upon viewing the stereo model until satisfactory terrain representation with mass points was achieved. Thick or thin distribution of mass points that did not truly represent ridges, ponds, and depressions were also edited to ensure that enough had been accurately represented. Densification and thinning of mass points were mainly carried out during terrain editing, so that the produced DTM would duly represent the actual terrain surface. After rigorous terrain editing a promising DTM was produced (Fig. 3); image processing tasks, including DTM editing, are well described by Lamsal et al. (2011). To correct the geometric distortions of the Corona images, a non-metric camera model was selected in the LPS platform. This requires only the focal length and pixel size of the scanned image, and the flying height of the camera platform as input for interior orientation parameters (IOPs) (Altmaier & Kany, 2002; Casana & Cothren, 2008). Acquiring reliable GCPs from identical locations is one of the most crucial tasks in Fig. 3 Location of the ground control points (GCPs) denoted by white triangles (a). The red box shows the area shown in (b). Figure 3b shows a representation of landform by mass points, triangles and contour lines in vector DTM (LTF).

Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier 87 photogrammetric operations for producing quality DTMs and topographical maps. The GCPs for the Corona images (Fig. 3) were taken from corrected ALOS stereo models, with meticulous care given to their accuracy. Afterwards, approximately 100 tie-points were automatically generated with an image matching technique for stereo data based on the provided IOPs and exterior orientation parameters (EOPs). All tie points were automatically placed at the correct locations on both images (forward and backward). Aerial triangulation was then performed (RMSE ~1.25 pixel) to produce a vector DTM. Later, substantial DTM editing was carried out to obtain a representative terrain surface as explained above. 3.2 Evaluating the accuracy of the produced DTMs and topographic maps The pond/lake boundary of Chamlang South Tsho was delineated from the surrounding moraines and glacier while viewing the DTM superimposed over the stereo model of the ALOS PRISM data. The surroundings are higher than the lake surface, and the texture of the lake is also different. The flat lake surface is easily identifiable on the stereo model. Uncertainty in delineating the precise lake boundary over the stereo model is believed to be very small, i.e., less than one meter. The accuracy of the produced DTMs and topographic maps of the Chamlang South Glacier was also assessed by differentiating the 2006 ALOS DTM from the 1964 Corona DTM in the unaltered area outside the lateral and terminal moraine boundary. As shown in Fig. 2, the Chamlang South Glacier is surrounded by lateral and terminal moraines. The glacier surface within the lateral and terminal moraines is subject to surface lowering because of glacier melt. It is assumed that the surface beyond the lateral moraines has remained unchanged. This unchanged surface (buffer of ~0.55 km 2 ) beyond the lateral/terminal moraines is used for assessing the consistency among the DTMs and the topographic maps. Figure 4a is a portion of the elevation change map beyond the lateral moraines (unchanged ground) of the Chamlang South Glacier. The glacier surface area within the lateral moraines was masked for the calculation. The difference between the Corona and ALOS DTM ranges from 11 to 13 m (Fig. 4a). To quantify this DTM error more precisely, a histogram of DTM differences was prepared (Fig. 4b). The calculations were carried out on a 3 3 m grid, and the total pixel count was 59,001. Subsequently, the pixel count for each class was converted to an error percentage. The classes with the most dominant DTM difference were 1-1 m (26%) and 1-3 m (26%), while the classes with negligible influential difference were 13 - ( 11) m (0.1%), 11 - ( 9) m (0.2%), 9 - ( 7) m (0.4%), 7-9 m (1%), and 9-11 m (0.2%). The range of DEM differences from 5 to 5 m accounts for 90.9% and the range from 1 to 3 m comprises 52.0%. As a result, it is concluded that a maximum relative DEM error of 10 m may exist in the generated DTMs and the topographic maps of the Chamlang South Glacier. The use of Corona data to generate a DTM has been extremely limited in the scientific literature. The few studies that do exist include Altmaier and Kany (2002) with Corona KH-4B; Bolch et al. (2008a, 2011) with Corona (KH-4) and ASTER; and Lamsal et al. (2011) with Corona KH-4A and ALOS PRISM. We employed a combination of Corona KH-4A (spatial resolution, 2.7-7.6 m) and ALOS PRISM (spatial resolution, 2.5 m) data as used by Lamsal et al. (2011) to generate DTMs and topographic maps. Three-dimensional topographic mappings with Corona KH-4A and ALOS PRISM stereo imageries can represent glacial surfaces, including micro-landforms, such as supraglacial ponds, depressions, moraine ridges, and ice cliffs satisfactorily well with a high accuracy. However, the DTMs produced with automatic procedures were largely unsatisfactory, which led to extensive DTM editing to remove unnecessarily thick mass points or to add mass points where their distribution was thin. Despite careful editing of DTMs, there may still be small errors (a maximum of +/ 10 m). However, representation of the terrain of the glacier surface was entirely satisfactory for the present purpose. a) b) Fig. 4 Portion of the elevation change map beyond the lateral and terminal moraines of the Chamlang South Glacier, showing unchanged areas (a), and a histogram of DEM differences (b).

88 T. SAWAGAKI et al. 3.3 Bathymetric survey Many researchers have examined the areal expansion rates of glacial lakes in the Himalaya by using photographs, maps and satellite images (Watanabe et al., 1994; Yamada, 1998; Ageta et al., 2000; Haeberli et al., 2001; Yabuki, 2003; Byers, 2007; Sakai et al., 2007; Racoviteanu et al., 2008, 2009; Byers et al., 2012). However, few studies have investigated bathymetric changes in the supraglacial or proglacial lakes and ponds of concern (cf. Benn et al., 2001; Fujita et al., 2009), partly because it is extremely difficult to gain access to most of the high altitude glacier areas in the Himalayan mountains. Recurrent measurements have also been restricted or are completely absent. Still, it is important to measure the water depth of glacial lakes in order to have a better understanding of the expansion process mechanisms involved (e.g., Röhl, 2008; Sakai et al., 2009). We carried out geomorphic observations and a bathymetric survey at and around Chamlang South Tsho between 26 and 29 October 2009, at the end of the ice-melting period. The lake water depth was confirmed continuously using sonar sounding from a boat. Fishing sonar (Fish Elite 500C) was utilized for sounding the depth. Depth measurement points were simultaneously recorded by GPS. which have already been abandoned (Fig. 2a). These channels suggest that the former melt-water drainage over- spilled when the glacier surface was much higher than today. The magnitude of the overspill was determined not to be large, based on the small size of the depositional landforms on the terminal moraines. The largest channel located to the northwest of the terminal moraine currently has surface water (a small river), the uppermost of which is subsurface flow. The water of this small river is most likely derived from the pond shown as P in Fig. 2b. Morphologies shown in Fig. 5 were converted to DEMs (Fig. 7) for the purpose of raster-based calculations to quantify glacial morphology and its change by producing elevation change maps and surface profiles. Here, we use the term lowering in a descriptive manner, without suggesting specific processes, because this is more favorable to the description of morphological changes that occurred between 1964 and 2006. 4. Results and Discussion Figure 5 shows the surface morphologies around the Chamlang South Glacier with topographic maps in 1964 and 2006 at contour intervals of 2 m (faint lines), 10 m (lighter bold lines), and 20 m (bold lines), mapped by Corona KH-9 and ALOS PRISM stereo images, respectively. As can be seen when comparing the two images, noticeable glacial surface morphology change occurred during the 42-year period from 1964 to 2006. Corona KH-4A image shows a few supraglacial ponds in 1964 (Fig. 2a). These ponds were most probably very small in 1962 when a climbing expedition team from Hokkaido University scaled the summit of Mt. Chamlang. The largest pond was located at an altitude of 4,952 m, and the glacial surface upward from the pond had a relatively steep gradient in 1964 (Fig. 5a). The steep section of the glacier had completely melted out by 2006 resulting in the formation of a large lake, Chamlang South Tsho, that had a lake water surface altitude of 4,940 m in 2006 (Fig. 5b). This suggests that some 12 m of lake surface lowering had occurred during the period between 1964 and 2006. As shown in the bathymetric map (Fig. 6), the current lake is approximately 550 m wide (north-south) and about 1,650 m long (east-west), with the bathymetric survey suggesting that the maximum water depth is 87 m. The current lake area, 8.7 10 7 m 2, is roughly six times larger than that shown on the Schneider map (surveyed between 1955 and 1974). The total water volume is estimated to be 35.6 10 6 m 3. The terminal moraine of the Chamlang South Glacier has six dissected channels (Figs. 2a and 2b), five of Fig. 5 Topographic maps of the Chamlang South Glacier in 1964 (a) and 2006 (b). Contour intervals: 2 m (faint lines), 10 m (lighter bold lines), and 20 m (bold lines). The red lines show the lateral and terminal moraines. Fig. 6 Bathymetric DEM of the lake in 2009.

Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier 89 In Fig. 7, the surface morphologies for both years are represented in eleven classes (ranging from 4,710-5,150 m at 40-m class intervals) with the same 11 ranges and color indexes used to depict comprehensive surface morphology comparisons. The actual calculation was done using numeric values of the DEM. The two most salient features on the morphological maps of the glacier, in terms of DTMs from 1964 and 2006, include the following: (i) there was an elongated lower surface or depression along the middle of the glacier in 1964 (Figs. 5a and 7a) as depicted by the light green colour in the 4,950-4,990 m class range, and (ii) the glacier surface above the elongated depression surface was relatively steep in 1964, but subsequently melted out causing substantial surface lowering in the area by 2006. Topographic maps (Fig. 5) and DEMs (Fig. 7) clearly show that the elevations of the debris-covered glacier surface have substantially changed. A map of elevation changes for the entire area (Fig. 8) was produced by subtracting the DEM values for 2006 (Fig. 7b) from those for 1964 (Fig. 7a). The values for elevation change are presented in nine classes (Fig. 8). Extensive surface lowering, e.g., as high as 156.9 m, is visible in the up-glacier area, just up from the 2006 east lakeshore, from melting of the steep glacier surface that existed in 1964. This characteristic is similar to that of the Imja Glacier Lake (Lamsal et al., 2011). The average lowering of the glacier surface (calculated for 3 3 m grids using the elevation change map, Fig. 8) for the 42-year period from 1964 to 2006 over the entire area is 37.5 m, with the average rate of surface lowering being 0.9 m/year. Figure 8 clearly demonstrates that surface lowering at the dead-ice and nearby area is smaller, whereas the lowering gradually increases in the up-glacier area and reaches its maximum in the uppermost area. This suggests that less glacier melting has occurred in the down-glacier area when compared with the up-glacier area. The elevation change map in the lake area of 2009 (Fig. 9) was produced by subtracting the 2009 lake bottom values (bathymetric DEM, Fig. 6) from the 1964 glacier surface values (topographic DEM, Fig. 10b). The surface lowering values are grouped into eight classes to depict lowering extent. The average surface lowering for the 45 years from 1964 to 2009 was 99.5 m at a rate of 2.2 m/year; the minimum and maximum surface lowering during this period were 12 m and 153.8 m, respectively. The explicit influence of a steeper surface gradient in the upper part of the up-glacier area in the 1964 map, as well as the deepest area in the 2009 bathymetric map, are manifested in the elevation change map of the lake: i.e., greater surface lowering has occurred in the area specified above (Fig. 9). The longitudinal profile L L (Fig. 11) shows the extent of the surface lowering along the glacier of 1964, and the lake area and dead-ice area in 2006 together with the steepness of the glacial surface gradient for both years. Along profile L L (Fig. 11), the maximum surface lowering in the dead-ice area for the 42-year period from Fig. 7 Digital Elevation Model (DEM) of the Chamlang South Glacier in 1964 (a), and DEM for 2006 (b). Terrain elevation values are grouped in eleven classes with the same colors at 40 m intervals for each year. Red line shows the lakeshore in 2006. Fig. 8 Elevation change map showing the surface lowering of the Chamlang South Glacier for 42 years from 1964 to 2006. Fig. 9 Surface lowering from the glacier surface in 1964 (topographic DEM) to the lake bottom in 2009 (bathymetric DEM).

90 T. SAWAGAKI et al. a) b) Fig. 10 Topographic contour map (a) and DEM (b) of the 1964 glacier surface in the 2009 lake area, extracted from Figs. 5a and 8 respectively. 1964 to 2006 was 29 m (from 4,954 m to 4,925 m), while the maximum surface lowering in the lake area for the 45-year period from 1964 to 2009 was 134 m (from 5,026 m to 4,892 m). Figure 5a clearly shows that the area around the largest supraglacial pond had a low surface gradient by 1964, and that there had already been large melting, which favoured further lake expansion: the relative height between the lake surface and the lateral moraine (DGM of Sakai & Fujita, 2010) exceeded 100 m, so the lower area of the Chamlang South Glacier in the 1960s can be categorized as a glacier of the type with glacial lake of Sakai and Fujita (2010). The area down-glacier from the pond melted less whereas the steeper up-glacial area melted progressively more, accelerating between 1964 and 2006. Considering the steeper up-glacier surface and the water-surface level of the largest supraglacial pond in 1964, it is suggested that ice calving into the pond triggered the larger and faster up-glacial lake expansion (Kirkbride & Waren, 1997; Röhl, 2008; Fujita et al., 2009; Sakai et al., 2009). One reason for low surface lowering in the dead- ice area as well as non-downward lake expansion could be due to a thicker debris cover there, as debris thickness increases progressively down-glacier, hampering ice melting (Mattson et al., 1993; Kayastha, 2000; Benn et al., 2001). Thus, ice calving is undoubtedly a main mechanism of lake expansion, especially for the further expansion of a large glacial lake once it has formed. In contrast, direct ice melting, including processes such as water-line melting, aqueous and subaqueous melting, remain largely significantly less influential than ice calving processes. Subaqueous melting is also very small in the case of the Imja glacial lake to the north (Fujita et al., 2009). If ice calving had not been the main mechanism, downward lake expansion in Chamlang South Tsho would have been larger. Profile A A (Fig. 11) in the dead-ice area demonstrates that the glacial surface, even in 1964, was significantly lower than its lateral moraines, although the glacial surface in the 1960s had been close to, or even higher than, the lateral moraines in the lower area of the Imja Glacier (Lamsal et al., 2011). Thus, the surface lowering in the lower area of the Chamlang South Glacier from 1964 to 2006 was small, suggesting that substantial ice melting in the area had already occurred by 1964. Profile B B (Fig. 11) reveals that glacial ice melting was less in the pond area, as well as the section towards the left lateral moraine than the section towards the right lateral moraine. In this particular profile section (Profile B B ), the glacier surface in 1964 was already considerably lower than the lateral moraines. Furthermore, the undulating surface structure under the lake in 2006 seems unlikely to indicate bedrock, thus suggesting that it most probably consists of glacial ice and/or debris deposits on the ice. Two other profiles C C and D D (Fig. 11) exhibit progressively larger glacial surface lowering in the up-glacial area. Few studies reporting morphological changes of glaciers (e.g., Bolch et al., 2008a) and glacial lakes (e.g., Lamsal et al., 2011) in the Nepal Himalaya exist, particularly those concerned with pre- and post-glacial lake development and expansion as presented here. Prior to the current study, no field research at the Chamlang South Glacier had been conducted, so a comparison of glacier surface lowering of the glacier and glacial lake to its own previous period is not possible. However, the glacier surface lowering at the Chamlang South Glacier can be compared with two of the most thoroughly researched glaciers in the same region, i.e., the Khumbu and Imja glaciers (Bolch et al., 2008a; Lamsal et al., 2011; Nuimura et al., 2011). The results found at the Chamlang South Glacier agree well with the conclusions of these studies. For example, Nuimura et al. (2011) calculated a surface-lowering rate of 0.6-0.8 m/year in 17 years from 1978 to 1995 in the ablation area of the Khumbu Glacier, and Bolch et al. (2011) also reported 0.5-1.2 m/year of surface lowering in the 40-year period from 1962 to 2002 for the same glacier. Lamsal et al. (2011) found surface lowering of the dead-ice and up-glacier area of the Imja Glacier for the 42-year period from 1964 to 2006 at the rate of 0.4 m/year and 1.1 m/year, respectively. Further, Fujita et al. (2009) showed that the lowering rate of the dead-ice area of the same glacier for the period from 2001 to 2007 was 0.06 to 1.03 m/year. Of particular interest is the fact that the rate of lowering in the lake area of Imja Tsho for 38 years was the same as that of the Chamlang South Tsho at 2.2 m/year. Furthermore, Fujita et al. (2009) estimate that direct ice melting at the lake bottom of Imja Tsho for the past several years is not significant. Therefore, the evidence strongly suggests that the larger lowering rate

Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier 91 Fig. 11 Cross sections (A A, B B, C C, D D ) and a longitudinal section (L L ) showing glacier topographies in 1964 and 2006, and lake bathymetry in 2009. of the Chamlang South Tsho is attributable to ice-calving processes. The applicability of repeat remote sensing data, especially DTMs, in the study of glacial lake dynamics has been described elsewhere (e.g., Kääb, 2005; Surazakov & Aizen, 2006; Fujita et al., 2008). Nevertheless, results from this study are distinguished from the previous research through the use of remote sensing data (e.g., Lamsal et al., 2011) in the representation of the glacial morphology and surrounding terrain morphology, as well as the documentation of changes over time. Clearly, coarse resolution data such as ASTER and SRTM (15/30 m and 90 m spatial resolutions, respectively) inhibit the detection of micro-landform features (Fujita et al., 2008). However, our research further demonstrated and corroborated the previous conclusions of Lamsal

92 T. SAWAGAKI et al. et al. (2011) concerning the use of high-resolution Corona and ALOS data in the LPS platform, and its production of sufficiently accurate DTMs and detailed topographic maps that represent changes in glacier surfaces, including micro-landform features such as supraglacial ponds, ice cliffs, and moraine ridges. 5. Conclusions This study employed high-resolution Corona KH-4A (spatial resolution, 2.7-7.6 m) and ALOS PRISM (spatial resolution, 2.5 m) stereo-data taken in 1964 and 2006, respectively. The data were processed in a stereoscopic photogrammetry system to generate DTMs. The produced DTMs, and subsequently derived topographic maps, represented elevations and morphology of the glacier surface in fine detail with a maximum error of about +/- 10 m. Consistency among the produced DTMs and topographic maps in the Chamlang South Glacier was assessed in the unaltered area outside the lateral and terminal moraine boundary. A maximum relative DEM error of 10 m may exist in the generated DTMs and topographic maps. This result corroborates the findings of Lamsal et al. (2011) that DTMs created from Corona and ALOS PRISM images are suitable for use in studies of the surface change of glaciers. A substantial glacier surface lowering occurred during the period from 1964 to 2006/2009. Extensive surface lowering, e.g., as high as 156.9 m, is visible in the up-glacier area. The average lowering of the glacier for the 42 years from 1964 to 2006 over the entire area is 37.5 m, with an average rate of the surface lowering 0.9 m/year. The surface lowering in the dead-ice and nearby area was smaller, whereas the lowering gradually increased in the up-glacier area and reached a maximum in the uppermost area. The average surface lowering for the 45 years from the 1964 glacier surface to the 2009 lake bottom was 99.5 m at a rate of 2.2 m/year, and the minimum and maximum surface lowering during that period was 12 m and 153.8 m, respectively. This larger lowering rate in the lake area supports the previously presented idea that ice calving into the pond triggered the larger and faster up-glacial lake expansion. The area around the largest supraglacial pond in 1964 had a low surface gradient, and there had already been a large degree of ice melting that had taken place, which favored further lake expansion and ice melting. Acknowledgements We greatly appreciate the helpful discussions and comments related to glacial lake development in the Nepal Himalaya by Dr. K. Fujita, Dr. A. Sakai, and Dr. T. Nuimura of Nagoya University. We would also like to thank Dr. T. Yamanokuchi of RESTEC, and Dr. T. Tadono of JAXA, for their valuable insights and information regarding ALOS PRISM and Corona image processing. This research was co-financed by the Global Centre of Excellence Programme Establishment of Centre for Integrated Field Environmental Science at Hokkaido University; the National Geographic Society- Waitt Grant Programme, Washington, DC; The Mountain Institute, Washington, DC; and Himalayan Research Expeditions, Kathmandu, Nepal. References Ageta, Y., S. Iwata, H. Yabuki, N. Naito, A. Sakai, C. Narama and Karma (2000) Expansion of glacier lakes in recent decades in the Bhutan Himalayas. International Association of Hydrological Sciences Publication, 264: 165-175. Altmaier, A. and C. Kany (2002) Digital surface model generation from CORONA satellite images. ISPRS Journal of Photogrammetry and Remote Sensing, 56(4): 221-235. Bajracharya, S.R., P.K. Mool and B.R. Shrestha (2007) Impact of Climate Change on Himalayan Glaciers and Glacial Lakes. ICIMOD, Kathmandu, Nepal. Bajracharya, S.R. and P. Mool (2009) Glaciers, glacial lakes and glacial lake outburst floods in the Mount Everest region, Nepal. Annals of Glaciology, 50(53): 81-86. Benn, D.I., S. Wiseman and K.A. Hands (2001) Growth and drainage of supraglacial lakes on debris-mantled Zgozumpa glacier, Khumbu Himal, Nepal. Journal of Glaciology, 47(159): 626-639. Bolch, T., M.F. Buchroithner, T. Pieczonka and A. Kunert (2008a) Planimetric and volumetric glacier changes in the Khumbu Himal, Nepal, since 1962 using Corona, Landsat TM and ASTER data. Journal of Glaciology, 54(187): 592-600. Bolch, T., M.F. Buchroithner, J. Peters, M. Baessler and S.R. Bajracharya (2008b) Identification of glacier motion and potentially dangerous glacier lakes at Mt. Everest area/nepal using spaceborne imagery. Natural Hazards and Earth System Sciences, 8: 1329-1340. Bolch, T., T. Pieczonka and D.I. Benn (2011) Multi-decadal mass loss of glaciers in the Everest area (Nepal Himalaya) derived from stereo imagery. The Cryosphere, 5: 349-358. Byers, A., D. Mckinney, M. Somos and T. Watanabe (2012) Glacial Lakes of the Hinku and Hongu Valleys, Makalu-Barun National Park and Buffer Zone, Nepal. Andean-Asian mountain Global knowledge exchange on glaciers, glacial lakes, water and hazard management, Proceedings, The Mountain Institute, Washington, DC. Byers, A. (2007) An assessment of contemporary glacier fluctuations in Nepal s Khumbu Himal using repeat photography. Himalayan Journal of Sciences, 4: 21-26. Casana, J. and J. Cothren (2008) Stereo analysis, DEM extraction and orthorectification of CORONA satellite imagery: archaeological applications from the Near East. Antiquity, 82(317): 732-749. Cenderelli, D.A. and E.E. Wohl (2001) Peak discharge estimates of glacial lake outburst floods and normal climatic floods in the Mount Everest region Nepal. Geomorphology, 40: 57-90. Cenderelli, D.A. and E.E. Wohl (2003) Flow hydraulics and geomorphologic effects of glacial lake outburst floods in the Mount Everest region, Nepal. Earth Surface Process and Landforms, 28: 385-407. Clague, J.J. and S.G. Evans (2000) A review of catastrophic drainage of moraine-dammed lakes in British Columbia. Quaternary Science Reviews, 19: 1763-1783. Fujita, K., T. Kadota, B. Rana, R.B. Shrestha and Y. Ageta (2001) Shrinkage of Glacier AX010 in Shorong region, Nepal Himalayas in the 1990s. Bulletin of Glaciological Research, 18: 51-54. Fujita, K. and T. Nuimura (2011) Spatially heterogeneous wastage of Himalayan glaciers, Proceedings of the National Academy of Sciences of USA, 108: 14011-14014.

Changes in Surface Morphology and Glacial Lake Development of Chamlang South Glacier 93 Fujita, K., A. Sakai, T. Nuimura, S. Yamaguchi and R.R. Sharma (2009) Recent changes in Imja Glacial Lake and its damming moraine in the Nepal Himalaya revealed by in-situ surveys and multi-temporal ASTER imagery. Environmental Research Letters, 4: 045205. Fujita, K., R. Suzuki, T. Nuimura and A. Sakai (2008) Performance of ASTER and SRTM DEMs, and their potential for assessing glacier lakes in the Lunana region, Bhutan Himalaya. Journal of Glaciology, 54: 220-228. Galiatsatos, N., D.N.M. Donoghue and G. Philip (2008) High resolution elevation data derived from stereoscopic CORONA imagery with minimal ground control: an approach using Ikonos and SRTM data. Photogrammetric Engineering and Remote Sensing, 74: 1093-1106. Haeberli, W., A. Ka a b, D.V. Mu hll and P. Teysseire (2001) Prevention of outburst floods from periglacial lakes at Grubengletscher, Valais, Swiss Alps. Journal of Glaciology, 47: 111-122. Haeberli, W., M. Hoelzle, F. Paul and M. Zemp (2007) Integrated monitoring of mountain glaciers as key indicators of global climate change: the European Alps. Annals of Glaciology, 46: 150-160. Hegglin, E. and C. Huggel (2008) An integrated assessment of vulnerability to glacial hazards: a case study in the Cordillera Blanca, Peru. Mountain Research and Development, 28: 299-309. ICIMOD (2011) Glacial Lakes and Glacial Lake Outburst Floods in Nepal. ICIMOD, Kathmandu, Nepal. Immerzeel, W.W., L.P.H. Van Beek and M.F.P. Bierkens (2010) Climate change will affect the Asian water towers. Science, 328: 1382-1385. Ives, J.D. (1986) Glacial Outburst Floods and Risk Engineering in the Himalaya. ICIMOD Occasional Paper 5. ICIMOD, Kathmandu, Nepal. Ives, J.D., R.B. Shrestha and P.K. Mool (2010) Formation of Glacial Lakes in the Hindu Kush-Himalayas and GLOF Risk Assessment. ICIMOD, Kathmandu, Nepal. Kääb, A. (2005) Combination of SRTM3 and repeat ASTER data for deriving alpine glacier flow velocities in the Bhutan Himalaya. Remote Sensing of Environment, 94: 463-474. Kayastha, R.B., Y. Takeuchi, M. Nakawo and Y. Ageta (2000) Practical prediction of ice melting beneath various thickness of debris cover on Khumbu Glacier, Nepal, using positive degree-day factor. IAHS Publication, No. 264 (Symposium at Seattle 2000 - Debris-covered Glaciers): 71-81. Kirkbride, M.P. and C.R. Warren (1997) Calving processes at grounded ice cliff. Annals of Glaciology, 24: 116-121. Lamsal, D., T. Sawagaki and T. Watanabe (2011) Digital terrain modelling using Corona and ALOS PRISM data to investigate the distal part of Imja Glacier, Khumbu Himal, Nepal. Journal of Mountain Science, 8: 390-402. Mattson, L.E., J.S. Gardner and G.J. Young (1993) Ablation on debris covered glaciers: an example from the Rakhiot Glacier, Punjab, Himalaya. IAHS Publication, No. 218 (Symposium at Kathmandu 1992 Snow and Glacier Hydrology): 289-296. McDonald, R.A. (1995) CORONA: Success for space reconnaissance. A look into the Cold War, and a revolution for intelligence. Photogrammetric Engineering and Remote Sensing, 61: 689-720. Mool, P.K., S.R. Bajracharya and S.P. Joshi (2001) Inventory of Glaciers, Glacial Lakes and Glacial Lakes Outburst Floods. ICIMOD, Kathmandu, Nepal. Nakawo, M., H. Yabuki and A. Sakai (1999) Characteristics of Khumbu Glacier, Nepal Himalaya: recent changes in the debris-covered area. Annals of Glaciology, 28: 118-122. Narama, C., M. Duishonakunov, A. Kääb, M. Daiyrov and K. Abdrakhmatov (2010) The 24 July 2008 outburst flood at the western Zyndan glacier lake and recent regional changes in glacier lakes of the Teskey Ala-Too range, Tien Shan, Kyrgyzstan. Natural Hazards and Earth System Science, 10: 647-659. Nuimura, T., K. Fujita, K. Fukui, K. Asahi, R. Aryal and Y. Ageta (2011) Temporal changes in elevation of the debris-covered ablation area of Khumbu Glacier in the Nepal Himalaya since 1978. Arctic, Antarctic, and Alpine Research, 43: 246-255. Quincey, D.J., S.D. Richardson, A. Luckman, R.M. Lucas, J.M. Reynolds, M.J. Hambrey and N.F. Glasser (2007) Early recognition of glacial lake hazards in the Himalaya using remote sensing datasets. Global and Planetary Change, 56: 137-152. Racoviteanu, A.E., M.W. Williams and R.G. Barry (2008) Optical remote sensing of glacial characteristics: a review with focus on the Himalayas. Sensors, 8: 3355-3383. Racoviteanu, A.E., F. Paul, B. Raup, S. Khalsa and R. Armstrong (2009) Challenges and recommendations in glacier mapping from space: results of the 2008 GLIMS Workshop, Boulder, Colorado, USA. Annals of Glaciology, 50(53): 53-69. Reynolds, M.J. (2000) On the formation of supraglacial lakes on debris-covered glaciers. IAHS Publication, No. 264 (Symposium at Seattle 2000 - Debris-covered Glaciers): 153-161. Richardson, S.D. and J.M. Reynolds (2000a) An overview of glacial hazards in the Himalayas. Quaternary International, 65/66: 31-47. Richardson, S.D. and J.M. Reynolds (2000b) Degradation of ice-cored moraine dams: implications for hazard development. IAHS Publication, No. 264 (Symposium at Seattle 2000: Debriscovered Glaciers): 187-197. Röhl, K. (2008) Characteristics and evolution of supraglacial ponds on debris-covered Tasman Glacier, New Zealand. Journal of Glaciology, 54(188): 867-880. Sakai, A. and K. Fujita (2010) Formation conditions of supraglacial lakes on debris-covered glaciers in the Himalayas. Journal of Glaciology, 56 (195): 177-181. Sakai, A., M. Saito, K. Nishimura, T. Yamada, Y. Izuka, K. Harada, S. Kobayashi, K. Fujita and C.B. Gurung (2007) Topographical survey of end moraine dead ice area at Imja Glacial Lake in 2001 and 2002. Bulletin of Glaciological Research, 24: 29-36. Sakai, A., K. Nishimura, T. Kadota and N. Takeuchi (2009) Onset of calving at supraglacial lakes on debris covered glaciers of the Nepal Himalayas. Journal of Glaciology, 55: 909-917. Shrestha, A.B. and R. Aryal (2010) Climate change in Nepal and its impact on Himalayan glaciers. Regional Environmental Change, doi: 10.1007/s10113-010-0174-9. Slama, C. (1980) Manual of Photogrammetry, Fourth edition, The American Society for Photogrammetry and Remote Sensing. Falls Church. Surazakov, A.B. and V.B. Aizen (2006) Estimating volume change of mountains glaciers using SRTM and map-based topographic data. Institute of Electrical and Electronics Engineers (IEEE) Transactions on Geoscience and Remote Sensing, 44: 2991-2995. Tadono, T., M. Shimada, H. Murakami and J. Takaku (2009) Calibration of PRISM and AVNIR-2 onboard ALOS Daichi. Institute of Electrical and Electronics Engineers (IEEE) Transaction of Geoscience and Remote Sensing, 47: 4042-4050. Vuichard, D. and M. Zimmermann (1987) The 1985 catastrophic drainage of a moraine dammed lake, Khumbu Himal, Nepal: cause and consequences. Mountain Research and Development, 7: 91-110. Watanabe, T., J.D. Ives and J.E. Hammond (1994) Rapid growth of a glacial lake in Khumbu Himal, Himalaya: prospects for a catastrophic flood. Mountain Research and Development, 14: 329-340. Watanabe, T., D. Lamsal and J.D. Ives (2009) Evaluating the growth characteristics of a glacial lake and its degree of danger of outburst flooding: Imja Glacier, Khumbu Himal, Nepal. Norsk Geografisk Tidsskrift, 63: 255-267. Yabuki, H. (2003) Monitoring of Imja glacier lake in the east Nepal using the satellite image. Scientific Program and Abstract of

94 T. SAWAGAKI et al. International Union of Geodesy and Geophysics: 48. Yamada, T. (1998) Glacier Lake and Its Outburst Flood in the Nepal Himalaya. Data Center for Glacier Research, Japanese Society of Snow and Ice, Monograph 1. Yamada, T. and C.K. Sharma (1993) Glacier lake and its outburst flood in Nepal Himalayas. IAHS-AISH Proceedings and Reports Publications, 218: 319-330. Takanobu SAWAGAKI Takanobu SAWAGAKI is an Assistant Professor of the Faculty of environmental Earth Science, Hokkaido University. His major fields of study are Glacial Geology, Quaternary Science and Geographical Information Systems. He has broad experience in circumpolar and alpine regions as a member of Japanese Antarctic research expeditions, the Academic Alpine Club of Hokkaido, and his own research expeditions in cold regions around the world. Damodar LAMSAL Damodar LAMSAL obtained his Ph.D. at the Graduate School of Environmental Science, Hokkaido University, in September 2011. His major areas of interest include modeling of glacier surface morphology, glacial lake development, and evaluation of the potentiality of glacial lake outburst floods (GLOFs) and their hazards. He has expertise in geographic information systems (GIS) and remote sensing (RS), especially digital photogrammetry. Alton C BYERS Alton C. BYERS, Ph.D., is a mountain geographer, climber and photographer specializing in applied research, integrated conservation and development programs. He received his doctorate from the University of Colorado in 1987, focusing on landscape change, soil erosion, and vegetation dynamics in Sagarmatha (Mt. Everest) National Park, Khumbu, Nepal. He joined The Mountain Institute (TMI) in 1990 as Environmental Advisor, and has since worked as Co-manager of the Makalu-Barun National Park (Nepal Programs), Director of Appalachian Programs, Director of TMI s 400-acre Spruce Knob Mountain Center in West Virginia, and Director of Science and Research at TMI since 2004. Teiji WATANABE Teiji WATANABE is a Professor of the Faculty of Environmental Earth Science and the Graduate School of Environmental Science, Hokkaido University. Having a background in alpine geomorphology, he specializes in mountain geoecology, with interests in interaction of humangeo-ecosystems in high mountain areas and natural resource management in protected areas. He has conducted extensive fieldwork in the Himalayas, the Karakorums, the Alps, the Pamirs, and Japanese mountains. (Received 20 November 2011, Accepted 13 March 2012)