Report on the Tasks Scheduled for the July-December 2002 period Task numbers according to the working plan.

PROJECT LA 26: PROGRESS REPORT ABSTRACT The mean water level of the Rio de la Plata responds to mean sea level, wind field and discharge of its tributaries. Models HADCM3, ECHAM4/OPYC3, CSIRO-mk2 and GFDL-R30 represent adequately the mean annual SLP fields and most of their annual cycle. Therefore, SRS A2 scenarios of these models were used to develop SLP and wind scenarios until the year 2100. A regional trend in SLP and wind was detected in the observed fields (1950/2000) and in GCM SLP outputs 1960/2000). This trend can be summarized as a southward shift of the western border of the South Atlantic High that persists during the XXI century in the four models. Digitized information of geology and geomorphology of the coastal areas of the Argentine was completed, while topography is being improved using a GP system. Social vulnerability indicators were developed for the administrative units affected by future floods under an extreme climate change scenario. The future RP water level is being studied with two hydrodynamic. One of them is completely validated and ready for the development of future scenario. It continue the inter-consultation process with qualified stakeholders and a workshop was organized to present the Project and discuss with them the pertinence of the Project objectives to social and governmental demands. Report on the Tasks Scheduled for the July-December 2002 period Task numbers according to the working plan. 10. Photo interpretation of dry, normal and flood events The objective of this task was to contribute to the construction of topographic maps in the Samborobóm bay since available altitude information had a resolution of only 2.5-m. According to the IPCC (2001) predicted sea level rise, this value is not enough to precisely define areas that could be affected by future marine ingressions. This task was changed to an analysis of Landsat images to determine lines of equal height according to the boundary of water. As explained in the Second Report, this task will complement measures to be taken with a high resolution GPS. Results indicate that along the coast in most of the bay, the altitude of the land after the first one or two kilometers from the coast is about 50 to 100 cm over the actual river level. 11. Data bases construction in a GIS environment. Point 3 of the Joint Document with LA 32 This is a continuous activity in the Project including outputs of the tasks. This activity continued in two ARC VIEW systems. Social, economic, and geological information is being merged with basic cartography and satellite information 12. Topographic measurements and fieldwork will be carried out to produce detailed altitude level maps of coastal areas subject to possible floods During the previous six-month period, topographic maps of the coastal area and its zone of influence between the southern limit of the Rio de la Plata and the Paraná delta were digitized from 1:50.000 IGM maps. Interpolation was made when necessary using the geomorphology maps. In this period the same work was done with the San Antonio Cap area because is under the influence of the Río de la Plata being a transition zone with the marine environment. Project LA 26 1

These maps resolution are being upgraded with field measurements taken with a GPS system. The field activity in the northern part of the Great Buenos Aires City was completed. The GPS was acquired with collateral funds in last March. 13. Election of future climate scenarios in cooperation with Project LA32 Point 6 of the joined document As expressed in the first report, the mean water level of the Rio de la Plata (RP) estuary responds to the wind field and to the discharge of its tributaries. Therefore, the variables required from climate scenarios are surface mean winds over the RP and over the outer adjacent ocean and precipitation over the RP basin. Surface winds are not a standard output of the GCM scenarios. However, this variable is strongly coupled with sea surface pressure (SLP) fields. Consequently, wind scenarios are being developed through SLP scenarios. In the Second Report was described that four experiments (HADCM3, CSIRO-mk2, ECHAM4, GFDL-R30) simulate adequately the mean regional SLP field. Therefore, we elected these models for future scenarios. As a first step, we are using the SRS A2 scenarios up to the year 2100. Surface wind scenarios As explained in the Second Report, it is required daily data. While the daily data from GCM experiments are obtained, an alternative approach was undertaken that is based on downscaling from daily SLP using NCEP/NCAR data. Thus, a regression model for regional surface daily wind was constructed, and its results are being evaluated. Future wind scenarios will be constructed using a perturbation technique that accounts for the mean SLP GCM scenarios. Other direct approach is reported in the next-semester tasks. A regional trend in winds and SLP was detected both in data and in GCM outputs. A paper is being prepared in this subject. A resume is presented in Annex 1 Precipitation scenarios The ETA model is being run with high resolution over South America for the 1990-1999 period. The intention is assessing if the underestimation of precipitation is caused by the low resolution of the GCM models. This activity will take the whole year 2003. January results indicate a good mean precipitation over most of the Plata basin, except in the SACZ area. Other alternative for precipitation scenarios is being reported under next semester tasks. In the meantime, it was recommended to the AIACC Projects dealing with agriculture (La 27 and LA 29) to use the HADCM3 scenarios with some sort of correction accounting for systematic errors. 19. Writing of a paper manuscript on the study of weather storms over the region embracing the Rio de la Plata using NCAR/NCEP reanalysis for the period 1950-2000 A manuscript on this subject entitled: Surface level atmospheric conditions associated to extreme surge tides in the Rio de la Plata estuary was prepared by Gustavo Escobar, Walter M. Vargas and Susana A. Bischoff was submitted to the International Journal of Climatology. A copy was sent with the Second Report 21. Writing of a paper manuscript on extreme streamflow events in the Paraná and Uruguay rivers The paper The greatest discharge events in the Paraná River and their climate forcing by Inés Camilloni and Vicente Barros was published by the Journal of Hydrology, 278 (2003) 94-106. It was submitted in the previous semester, and therefore its copy was sent with the Second Report The manuscript The largest floods in the Uruguay River and their climate forcing by Inés Camilloni and Ruben Caffera was submitted to the J. of Hydrometeorology in May 2003. Copy of the paper in Annex 2. Project LA 26 2

22. Consultation about demands and communication of Project partial results with stakeholders through mail and email (March 2003) This task was almost completed in 2002 but continued with a consultation to a greater group of stakeholders. In March 5, 2003, a workshop was held to present the Project results to stakeholders and to receive from them suggestions and demands. There were about forty participants from 30 governmental and non-governmental organizations. In the same workshop, it was also presented results from the University of Buenos Aires Project on Floods on the Paraná and Uruguay Rivers that provide the collateral funds to this Project. The report of the workshop is in Spanish in the web page of the Project http://www-atmo.at.fcen.uba.ar/~lcr/proyectos/aiacc/. The English version of this document is being prepared and will be attached in the next report. 23. Geomorphology description and study of the evolution of the coastal area and construction of geomorphology maps In the Second Report was reported the geological and geo-morphological maps of the coastal area of Samborombón Bay, which included the digitization of the different units and their lithological features represented in the same format as the topographic maps. In this semester, it was completed the respective two maps of the area Paraná Delta Punta Piedras. Therefore, all the Argentine coast of the Rio de la Plata was completed. The geo-morphological maps in scale 1:40.000 were made base upon satellite images, airplane photography and field activities. This map interpretation allows identifying the areas affected by regular astronomical tides and storm tides and wetlands. The geological map was based upon previous bibliography and field tasks as well as with the use of airplane photography and Landsat 5TM (Thematic Mapper) images This map allows through its interpretation to extract the necessary data to estimate coastal erosion. 25. Develop of a Social Vulnerability Index to flooding in a GIS environment. Point 3 of the Joint Document with LA 32 Number of people living in the area of possible affectation Last semester we arrive to a preliminary figure for the AMBA Sector (Census 1991). Now we complete the elaboration from the INDEC census with the number of inhabitants for the rest of the area, and a more detailed design for the 5 meters topographic line. The population that can be affected by a probable sea level rise for climatic change in the area of the Project is near 1.500.000 inhabitants. Social Vulnerability Index The social vulnerability index developed during 2002 was improved by the inclusion of indicators of production and consume. The indicators have been clustered into sub-indexes: demographic indicators, conditions of life indicators and work/ production/ consumption indicators The demographic sub index includes the following indicators: total population, population's density, index of potential dependence: the young, and index of potential dependence: the elderly. The administrative units that have a higher demographic vulnerability sub index are Buenos Aires City, General Sarmiento and Vicente López. The conditions of life sub index includes the following indicators: population s percentage in homes with unsatisfied basic needs (NBI), percentage of homes with boss woman, total rate of infantile mortality, and population s percentage without access to health services. The administrative units that have the higher conditions of life vulnerability sub index are San Fernando, Tigre, General Sarmiento and Castelli. The work, production, consumption sub index includes the following indicators: unemployment rate, Aggregated gross product, registered cars rate (inhabitants / car rate), and percentage of workers without provisional taxes. The administrative units that have the higher Project LA 26 3

conditions of vulnerability according to this sub index are Partido de la Costa, General Sarmiento and Esteban Echeverría. The last steep in this task is to develop a comprehensive social vulnerability index and to apply it to the administrative units. 27. Selection of sea level scenarios according to TAR IPCC This task was started in this semester. It will be considered not only the mean sea global rise scenarios, but also the predicted regional increase in the South Atlantic. The main effort was in developing the mean sea level with respect to the reference geode in the domain of the hydrological models being used in the Project and for different seasons. These levels are being calculated from the TOPEX set of data for the 1990` decade. It was found substantial differences in height both in space and along the year. Once this task will be completed, it results will be used as baseline to add the regional increments as predicted by coupled GCMs. The outcome of this step will be the input for the hydrological model. 28. Validation of the ETA model and the hydrodynamic models The use of the ETA model to develop surface winds was abandoned, at least momentarily, since calibration from NCEP/NCAR surface winds were satisfactory. The calibration process of model RPP-2D, through successive stages (purely astronomical tide, normal tide condition, and storm tide condition), was explained in the previous Report. Preliminary, and encouraging, results were shown. The calibration is now completed. The following improvements were made: i. Bathymetric information was rebuilt based on more accurate data sources. ii. iii. The astronomical tide boundary condition at the southern border was redefined by using results from a global tidal model. The adjustment of wind fields from NCEP/NCAR was refined through detailed experimentation. Model RPP-2D was first exploited to analyze sensitivity of system response to probable changes in the forcings: sea level rise, change in predominant seasonal wind pattern and tributary discharge increase. More information on this task is in Annex 3. 31. Development of strong waves scenarios This task was completed in the past semester. In Annex 4 is attached a copy of the manuscript Impact of a possible regional wind change on the wave climate of the Rio de la Plata by Walter Dragani and Silvia Romero that is being prepared for submission to the Int. J. of Climatology. 35. Engineering implications and prospective of the evolution of coastal area. This includes both the new coastal geography arising from sea level rise and the advancement of the Paraná Delta front This task was initiated collecting data from different institutions, both public and private that have large infrastructure on the coast. In addition, soil uses, goods and infrastructure localized in the study area are been survey, in order to asses the impact of present floods and the future impact caused by climatic change. Travel activities V. Barros made a short trip (one-day) to present the Project activities in a workshop held in Mar del Plata to discuss the future Argentine satellites. The Co-PI M. Caffera made one short trip to Buenos Aires to work with I. Camilloni and V. Barros on the extreme discharges of the Uruguay River. Project LA 26 4

Dr Kokot and Dr. Codignotto made a short trip to the Samborombón bay area to make fieldwork. Dr Barros and Dr Kokot made a short trip to Santa Fe to discuss a cooperative agreement with the Faculty of Water Resources of the University of Litoral. This agreement will favor the Project with the help of experienced professionals on the use of GPS Madeleine Renom from Uruguay made two trips to Buenos Aires to work on climate scenarios with I. Camilloni and to attend a course on the South American Climate at the University of Buenos Aires. She was also funded by the LA 32 Project. Description of Difficulties Encountered and Lessons Learned Task 12 Because of the lack of available high-resolution altitude maps, the topography was digitized with a resolution of 1.25 m. This is not enough for the Project requirements and therefore, field measurements are being taken with a GPS. Task 13 Other problem is the availability of daily data from GCM experiment scenarios. We are trying of get this data by ourselves, but help from AIACC would help. Description of Tasks to be performed in the Next Eight-Month Period Tasks numbers as in the working plan. It is indicated when the task was already initiated as reported in part B. 11. Data base construction in a GIS environment. Point 3 of the Joint Document LA 32 (Initiated). This will be a continuous activity during the Project duration, including outputs of the different tasks. 12. Topographic measurements and fieldwork will be carried out to produce detailed altitude level maps of coastal areas subject to possible floods (Initiated). Fieldwork with a GPS will be conducted to improve the actual topography resolution 13. Election of future climate scenarios in cooperation with Project L32 (Initiated) Wind: If available daily GCM outputs, it will be used to generate daily winds from daily SLP, to be used as input in the hydrodynamic model Precipitation: Sensitivity analysis indicates that only the greatest and exceptional discharges have some influence on the Rio de la Plata level. On top of that, Task 20 showed that these exceptional discharges are closely related to El Niño events. Therefore, a shortcut could be taken, to estimate the frequency of the exceptional discharges if the GCMs properly represent the ENSO phases. 16. Development of future scenarios for the Paraná delta growth (Initiated). The past evolution of the Delta during the last 300 years is being completed. Future evolution of the Delta will depend on many variables; one of the most important is the future level of waters Project LA 26 5

in the inner estuary of the Rio de la Plata. Therefore, scenarios will be constructed after tasks 32 and 33 will be well completed. 25. Develop a Social Vulnerability Index to flooding in a GIS environment. Point 3 of the Joint Document (Initiated) The remained activity is to construct a comprehensive social vulnerability index and to apply it to the administrative units. 27. Selection of sea level scenarios according to TAR IPCC (Initiated) 32. Development of mean and extreme scenarios of water level with hydrodynamic models 33. Development of extreme scenarios of water level with hydrodynamic models under different storm weather scenarios 34. Writing a manuscript on the future mean and extreme scenarios of water level 35. Engineering implications and prospective of the evolution of coastal area. This includes both the new coastal geography arising from sea level rise and the advancement of the Paraná Delta front (Initiated) 36. Identification of critical zones of social vulnerability to flooding 41. Joint workshop with Project LA 32 as explained in Point 11 of the joint document 43. Integrated analysis of the lower basin and mouth of the Santa Lucía using the Samborombóm experience as explained in point 8 of the joint document 45. Indicator analysis and social vulnerability diagnosis Anticipated Difficulties in the Next Eight-Month Period Task 13 The use of daily data from GCM scenarios can be delayed because of the difficulty to get these data and because of the large consume of computer time involved Connections with the National Communication The plans for the Second Communication are being made in the Foreign Affairs Ministry by Dr. Estrada Oyuela. Recently, it was also incorporated the Secretary of Environment. Vicente Barros and Walter Vargas participate of informative meetings at the Foreign Ministry were the first steps of the planning activity are reported to key experts. There is no yet a formal commitments, but we learnt in the informative meetings that it will be a selection of groups of national experts for each vulnerability activity based on their qualifications. Project LA 26 6

Annex 1 CLIMATE SCENARIOS Sea level pressure (SLP) fields trends were analyzed considering SRES-A2 GCM outputs and NCEP/NCAR reanalyzes. This analysis was performed considering only the four GCMs that best agree with the NCEP/NCAR reanalyzes for the region delimited by 20 S-47.5 S and 67.5 W-45 W. These models and the available periods are presented in Table 1. Model Period HADCM3 (1) 1951-2099 CSIRO-Mk2 (2) 1951-2099 ECHAM4/OPYC3 (3) 1991-2099 GFDL-R30 (4) 1961-2099 Table 1. GCMs with best agreement with NCEP reanalysis in Southeastern South America. Available periods are included. Rotated principal component analysis (PCA) was applied to the mean seasonal SLP fields of the reanalyzes and GCMs outputs of models (1), (2) and (4) for 1961-2000. Model (3) was excluded from this analysis because data start in 1990. The two first PCs of the reanalyzes data (Figure 1) explain more than 90% of the total variance and represent correctly the seasonal variability. These components also show a shift to the south of the axis of maximum South Atlantic SLP. The principal modes of variability and the southward displacement of the South Atlantic high observed Figure 1. Principal components of seasonal SLP NCEP reanalysis -25.0 PCA 1 - NCEP - 46.1 % -25.0 PCA 2 - NCEP - 45.1% -27.5-27.5-30.0-30.0-32.5-32.5-35.0-35.0-37.5-37.5-40.0-40.0-42.5-42.5-45.0-67.5-65.0-62.5-60.0-57.5-55.0-52.5-50.0-47.5-45.0-45.0-67.5-65.0-62.5-60.0-57.5-55.0-52.5-50.0-47.5-45.0 in the reanalysis data are also observed in the PCs of the GCMs outputs. This is confirmed in Table 2 where the spatial correlation between the first two PCs derived from the reanalysis SLP data and the GCMs outputs are presented. Model Spatial correlation PC 1 PC 2 HADCM3 0.90 0.85 CSIRO-Mk2 0.89 0.73 GFDL-R30 0.88 0.77 Table 2. Spatial correlation coefficients between the first two PCs obtained for the NCEP-NCAR SLP data and GCMs outputs. The PCA was also performed for the seasonal SLP outputs of the four models for the complete periods indicated in Table 1. The first two PCs for the four GCM SLP outputs are quite similar and show a future scenario with a positive trend in PC2 that represent a southward displacement of the South Atlantic high. As example, Figures 2 a and b show 1

-27.5-27.5-30.0-30.0-32.5-32.5-35.0-35.0-37.5-37.5-40.0-40.0-42.5-42.5-45.0-67.5-65.0-62.5-60.0-57.5-55.0-52.5-50.0-47.5-45.0-45.0-67.5-65.0-62.5-60.0-57.5-55.0-52.5-50.0-47.5-45.0 Figure 2a. Principal components of SLP outputs for GFDL-R30 (1961-2099). the first two PCs for the SLP outputs of GFDL-R30 and the decadal explained variance by both PCs. The displacement of the South Atlantic high implies an increase in the frequency of easterly winds that could be responsible for the rise of the Río de la Plata level in Buenos Aires. 0.7 0.6 GFDL-R30 (1961-2099)- Explained Variance M1 M2 Lineal (M2) Lineal (M1) 0.5 0.4 0.3 1961 1971 1981 1991 2001 2011 2021 2031 2041 2051 2061 2071 2081 2091 Figure 2b. Decadal explained variance by the two first principal components of SLP of GFDL-R30 outputs (1961-2099). M1 correspond to PC1 and M2 to PC2. Table 3 presents the historical and future scenarios of meridional SLP gradient over the Río de la Plata according to the NCEP-NCAR reanalysis and the GCMs indicated in Table 1. This variable is proportional to the wind intensity and results show that although there are discrepancies among the NCEP reanalysis and the different GCMs outputs, all of them agree with the fact that an increase in frequency of easterly winds will be also accompanied by an increase in their intensities. 2

Period 1950-59 1960-69 1970-79 1980-89 1990-99 2000-2050 2050-2099 NCEP 1.74 2.22 2.10 3.00 2.52 - - HADCM3 1.94 1.88 2.20 2.36 2.16 2.76 3.88 CSIRO-Mk2-0.74 0.42 0.74 0.72 1.02 1.34 ECHAM4 - - - - 4.18 4.78 5.26 GFDL-R30-0.62 0.82 1.10 1.00 1.28 1.76 Table 3. Meridional annual SLP gradients over the Río de la Plata according to NCEP-NCAR reanalysis and GCMs outputs. Annex 3 HYDRODYNAMIC MODEL RPP-2D Hydrodynamic model RPP-2D, already introduced in the previous Report, was taken as the basic model to make experimentation. As already explained RPP-2D, based on software HIDROBID II, is a 2.5 km x 2.5 km resolution 2D- Horizontal model, which includes the Río de la Plata and an extended area of its maritime front (Figure 1). Its performance was shown to be comparable to more sophisticated, but much higher time-computer demanding, 3D models, like HANSOM-CIMA. Figure 1 Calculation domain of model RPP-2D FINAL CALIBRATION OF MODEL RPP-2D The calibration process of model RPP-2D, through successive stages (purely astronomical tide, normal tide condition, and storm tide condition), was explained in the previous Report. Preliminary, and encouraging, results were shown. 3

The calibration is now completed. The following improvements were made: 1. Bathymetric information was rebuilt based on more accurate data sources. 2. The astronomical tide boundary condition at the southern border was redefined by using results from a global tidal model. 3. The adjustment of wind fields from NCEP/NCAR was refined through detailed experimentation. As an illustration, Figure 2 shows the comparison between water level Tide Table data and model results for a pure astronomical tide scenario (only the oscillatory component is presented); the agreement is considered quite satisfactory. The comparison for a storm event is presented in Figure 3, showing a good agreement taking into account the relatively poor wind information. In Figure 4 velocity records and predictions are compared, showing an acceptable degree of agreement for such a sensitive quantity. Level (m) 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00-0.20-0.40-0.60-0.80-1.00-1.20-1.40 01-Feb-97 04-Feb-97 07-Feb-97 10-Feb-97 13-Feb-97 16-Feb-97 19-Feb-97 22-Feb-97 25-Feb-97 28-Feb-97 Time (days) Modeled Predicted Figure 2 Comparison of water level from Tide Table and model in Buenos Aires, for pure astronomical tide 4

3.50 3.00 2.50 2.00 Level (m) 1.50 1.00 0.50 0.00-0.50-1.00-1.50 09-Nov-89 10-Nov-89 11-Nov-89 12-Nov-89 13-Nov-89 14-Nov-89 15-Nov-89 16-Nov-89 Time (days) Torre Oyarbide 1989 Modeled Torre Oyarvide 1989 Observed Figure 3 Comparison between recorded and predicted water level at Torre Oyarvide station for the November 1989 storm 1.00 0.75 0.50 0.25 Velocity (m/s) 0.00-0.25-0.50-0.75-1.00 27-Ago-96 28-Ago-96 29-Ago-96 30-Ago-96 31-Ago-96 01-Sep-96 Time (days) Observed Modeled Figure 4 Comparison between recorded and calculated flow velocities, for a normal tide scenario. Component South- North Period 27/ Aug/ 96 01/ Sep/ 96 (station 30-519) PRELIMINARY MODEL EXPLOITATION Model RPP-2D was first exploited to analyze sensitivity of system response to probable changes in the forcings: sea level rise, change in predominant seasonal wind pattern and tributary discharge increase. The results for Buenos Aires are summarized in Figure 5. It is observed that the system is practically blocked to sea level rise; the change in wind pattern could produce mean water 5

level increases up to about 10 cm; the discharge increase could lead to a mean level rise above 20 cm. 6

level (m) 0.12 0.10 0.08 0.06 0.04 0.02 0.00 level (m) 0.60 0.50 0.40 0.30 0.20 0.10 0.00 MG BA MO SC MP MG BA MO SC MP (a) (b) level (m) 0.80 0.60 0.40 0.20 0.00 level (m) 0.08 0.06 0.04 0.02 0.00 MG BA MO SC MP MG BA MO SC MP (c) (d) 50 cm + 30000 m3/s 75000 m3/s Wind (FUT-PRE) Figure 5 Mean water level increase due to (a) wind pattern change, (b) sea level rise, (c) and (d) tributary discharge increase. MG: Martín García; BA: Buenos Aires; MO: Montevideo; SC: San Clemente; MP: Mar del Plata. 7

The largest floods in the Uruguay river and their climate forcing Inés A. Camilloni 1 Departamento de Ciencias de la Atmósfera y los Océanos. Universidad de Buenos Aires Centro de Investigaciones del Mar y la Atmósfera (CIMA/CONICET-UBA) Buenos Aires, Argentina and Ruben M. Caffera Unidad Docente Meteorología. Facultad de Ciencias. Universidad de la República Montevideo, Uruguay Mayo 2003 1 Corresponding author address: Inés A. Camilloni Departamento de Ciencias de la Atmósfera y los Océanos Ciudad Universitaria. Pabellón 2. Piso 2. (1428) Buenos Aires. Argentina. Tel. (5411) 4576-3398 Fax (5411) 4576-3356/3364 ext. 12 e-mail: inesc@at.fcen.uba.ar

Abstract The Uruguay river is the second tributary in importance of the La Plata river with a basin that includes a vast region of Southeastern South America. The objective of this article is to examine the largest discharge events in the Uruguay river and to explore their atmospheric forcing. Extreme daily Uruguay river levels registered during the austral warm season (October to March) are related to intense rainfall in the upper basin particularly in the period -12 to -9 days before the date of maximum level at the Salto Grande gauging station while those occurred during the austral cold semester (April to September) are mostly due to large rainfall amounts registered over and upstream Salto Grande in two separate periods before the flooding dates (-12 to -9 and -4 to -1 days). The composites of the low-level wind fields for the warm season events show a northwestern flow over Bolivia and Paraguay that converges with a weak northern component over the Uruguay basin during the 12 days before the flooding date. The composites for the cold season events show an intense northwesterly flow over the same region that continues over southeastern South America and penetrates in the Atlantic ocean for the days -12 to -9 and -4 to -1 in accordance with the two heavy rainfall periods identified over the Uruguay basin. The circulation pattern during the -8 to -5 period shows northern flow and weak convergence over the basin. The enhanced precipitation due to moisture flux convergence in the South American low-level jet (SALLJ) region could be responsible for almost half of the largest floods of the Uruguay river at Salto Grande. For both the warm and cold seasons, the SALLJ occurrence largest frequencies are observed during the periods -12 to -9 and -4 to -1 before the flooding dates with a major decrease in the frequency during the intermediate period of the cold season events. In most cases, the extreme daily Uruguay river levels induced extraordinary discharges at the monthly scale. The largest monthly discharge anomalies at the Paso de los Libres gauging 2

station occurred mostly during warm ENSO phases explained by the large positive rainfall anomalies observed in the region during these events. 3

1. Introduction The Uruguay river is the second tributary in importance of the La Plata river with a basin of 365,000 km 2 that includes parts of Brazil, Uruguay and Argentina and a mean flow of about 4500 m 3 /s. It begins near the Atlantic Ocean at a height of approximately 1800 m above sea level and ends at the La Plata river (Figure 1). The major tributary of the Uruguay is the Negro river, which originates in Brazil and crosses Uruguay with a northeast to west direction. During recent years, many authors have explored different features of the hydrology of the Uruguay basin. García and Vargas (1996, 1998) investigated the spatial variability of average monthly and yearly flows at selected gauging stations in the river. They also searched for significant changes in the streamflow trends, and found a positive trend since 1970 and changes around the years 1917-1918 and 1943. Genta et al (1998) identified long-term trends in the monthly streamflow records of four major rivers in Southeastern South America including the Uruguay river that were consistent with a significant decrease in the amplitude of the seasonal cycle. Robertson and Mechoso (1998) analyzed the interannual and decadal cycles in river flows of Southern South America. They focused the analysis on annual mean streamflows and found interannual cycles (at around 3.5, 6 and 9 years) and a nonlinear trend. Other studies were focused on the relationship between river discharges and climatic forcing. For instance, Mechoso and Perez Iribarren (1992) analyzed the relationship between the Southern Oscillation (SO) and the Uruguay and Negro rivers. They found that streamflows in both rivers present a tendency to be below average in years with high SO index and a tendency to be slightly above average in El Niño-Southern Oscillation (ENSO) years. Robertson and Mechoso (1998) demonstrated that the 3.5 years cycle in the Uruguay and Negro flow series is associated to an ENSO-like pattern in the Pacific sea surface temperature anomalies with above-average streamflow coinciding with a warm phase in the eastern equatorial Pacific. Bischoff et al. (2000) studied the variability of the Uruguay river flows and the influence of the ENSO phenomenon. 4

They found significant differences in the river flow between warm and cold events in the equatorial Pacific during November, December, July and August. Recently, Krepper et al. (2003) analyzed the river flow and the precipitation in the Uruguay river basin and its subcatchments in order to address how the signals of the subcatchments relate to the overall basin signal. They found that the around 6-year and the 3.5-year ENSO signals are the two statistically significant peaks for the river flow and that both signals are present only in the precipitation of the upper two-thirds of the basin. However, despite the aforementioned contributions, there is a lack of a description of the largest floods of the Uruguay river and their causes. A better understanding of the climate forcing of these events can help to prevent future disasters and huge economic loss. The aim of this paper is to examine the extreme discharge events in the Uruguay river for the period 1950-2000 and to explore their atmospheric forcing. The paper is structured as follows. Section 2 summarizes the main regional climate features and the datasets are described in section 3. Section 4 discusses the main hydrological features of the Uruguay river and the differences with the regime of the Paraná river. Section 5 discusses the climatic forcing of the extreme discharge events in the Uruguay river and the relation between these extreme events and the SALLJ. Finally, section 6 examines the relation between the largest monthly discharge anomalies and ENSO and section 7 summarizes results and conclusions. 2. Regional climate background Northern Argentina, Paraguay, Uruguay and southern Brazil experience different types of transient disturbances. Cold frontal passages are common transient weather events over the continent and convective activity is triggered as they pass over northern Argentina, Paraguay, Uruguay and southern Brazil. Oliveira (1986) performed monthly statistics of frontal penetrations and their effect on the convective activity over subtropical South America. The 5

frontal passages are well spread over all seasons and although their convective activity is very low during the winter months, they are responsible for a large part of the rainfall. Southeastern South America also experiences the effects of sudden development of mesoscale convective complexes (MCCs) especially in the period November-April (Velasco and Fritsch, 1987). These complexes have a life cycle of less than a day (Satyamurty et al., 1998). An important feature of the South American climate is that as the Atlantic subtropical high migrates southward in early summer, convection develops along the South Atlantic Convergence Zone (SACZ). This band of intense convective activity emanates from the Amazon region extending from the tropical South America southeastward into the South Atlantic Ocean (Kodama, 1992; Figueroa et al., 1995; Nogués-Paegle and Mo, 1997; Liebmann et al., 1999; Barros et al., 2000). Nogués-Paegle and Mo (1997) found evidence of a seesaw pattern in the convection over the SACZ with each phase lasting no more than 10 days. This pattern indicates that the intensification (weakening) of the SACZ is associated with rainfall deficit (abundance) over the subtropical plains of South America. Doyle and Barros (2002) found that the midsummer interannual variability of the low-level tropospheric circulation and of the precipitation field in subtropical South America is associated to the sea surface temperature (SST) anomalies in the western subtropical South Atlantic Ocean. Composites corresponding to extreme SSTs in the area 20ºS-30ºS and 30ºW-50ºW show two different lowlevel circulation and precipitation patterns. The composite of cases with positive SST anomalies shows two regional maximum in the precipitation field with one of them centered at about 30ºS and 55ºW and the other one located along the continental extension of a southwardly displaced SACZ. Another of the key atmospheric features that contributes to the precipitation over Southern South America is a northwesterly low-level jet that is situated east of the Andes mountains and transports atmospheric moisture from the Amazon basin toward subtropical South America. The 6

South American low-level jet (SALLJ) is particularly pronounced over Bolivia, extending southeastward to the La Plata basin comprising Paraguay, southeast Brazil, Uruguay and northern Argentina. The low-level jet provides the moisture transport for the formation of clouds (Berri and Inzunza, 1993) and precipitation in MCCs that generally moves eastward from their source region in northern Argentina and Paraguay to affect southwestern Brazil and Uruguay with intense rainfall (Satyamurty et al., 1998). The observational evidence of the SALLJ is scarce because of the poor radiosonde coverage. Because of this, current understanding of the SALLJ is founded on diverse basis, including operational analyses (Nogués-Paegle, 1981; Sughara, 1994; Wang and Paegle, 1996), NCEP/NCAR reanalyses (Douglas et al., 1999; Marengo and Soares, 2002), high resolution regional models (Douglas et al., 1999) or considering special observation projects like the Pan American Climate Studies (PACS) Sounding Network (SONET) during a two-months period of 1998 (Douglas et al., 1999) or the Wet Season Atmospheric Mesoscale Campaign (WETAMC) of January-April 1999 conducted in the southwest corner of the Amazon basin (Marengo et al., 2002). The SALLJ occurs all year long but 75% of the cases during 1950-2000 were detected during the austral warm season November-February at the core region near Santa Cruz (17.76ºS, 63.15ºW) while at the exit region near Mariscal Estigarribia (22.02 S, 60.6ºW) the jet is more frequent during the austral winter and spring seasons (Marengo and Soares, 2002). According to Berbery and Barros (2002), the moisture flux over South America during the austral summer has a southward maximum between 900 and 950 hpa and a somewhat higher maximum at 850 hpa during winter. The winter SALLJ is probably the responsible for providing moisture and heat ahead of frontal zones and cyclonic systems that produce much of the winter precipitation (Vera et al., 2002). During the austral summer, the meridional transport of tropical air into midlatitudes is linked to the Chaco Low located approximately at 25ºS and 65ºW. During this season, this system 7

intensifies due to positive net radiation favoring the transport of tropical air masses to the south and the presence of strong convection activity fed by moisture from tropical regions (Seluchi and Marengo, 2000). The SALLJ displays a maximum near Santa Cruz and in some occasions a secondary maximum is found in southern latitudes called Chaco jet events. During the Chaco events the SALLJ penetrates downwind up to 30ºS and exhibits a strong moisture flux convergence and enhanced precipitation over Paraguay, northeastern Argentina, southeastern Brazil and Uruguay (Nicolini and Saulo, 2000, Nicolini et al, 2002). Likewise, Nicolini and Saulo (2000) showed that in those cases the composite fields of the 850 hpa wind and sea level pressure show an intensification of the Chaco thermal-orographic low extending farther south up to Uruguay. 3. Data Daily Uruguay river levels at the Salto Grande monitoring station (see Figure 1) were obtained from the National Direction of Hydrology of Uruguay for the period 1950-2000. In order to evaluate the possible effect of anthropogenic management of the river flows immediately upstream from Salto Grande, daily Uruguay river levels at the Argentine gauging station Paso de los Libres (Figure 1) were also considered for the available period (1997-98). Monthly streamflows at Paso de los Libres for 1909-2000 were analyzed to explore the relation between the largest monthly discharges of the Uruguay river and ENSO events. The Argentine Department of Water Resources provided these data. The data used to examine the atmospheric conditions during the major discharge events of the Uruguay river were the 0000, 0600, 1200 and 1800 UTC and daily averaged gridded reanalyses from the National Centers for Environmental Prediction-National Center for Atmospheric Research (NCEP-NCAR) (Kalnay et al., 1996) with a horizontal resolution of 2.5º latitude by 2.5º longitude. Daily averages of the National Oceanic and Atmospheric Administration 8

(NOAA) satellite outgoing longwave radiation (OLR) were used as a precipitation proxy to validate with independent data results obtained with the NCEP-NCAR precipitation reanalyses. 4. Major discharge events The Paraná and Uruguay rivers are the main tributaries of the La Plata river. Therefore the comparison between their regimes is important due to the different time scales associated to their extreme events. The greatest discharge events in the Paraná river present a longer time scale than in the Uruguay and consequently they have been studied on a monthly basis (Camilloni and Barros, 2003). The time scale of the greatest discharge events in the Uruguay river have approximately the same duration of the atmospheric perturbations (3-10 days) that induce the associated extreme rainfall events. This is due to the steep slope (~0.086 m/km for the Uruguay river around Salto Grande in comparison to ~0.036 m/km of the Paraná river in the proximity of Corrientes) and the narrow transverse section that favor the runoff (Tossini, 1959). The differences between the daily regimes of both rivers can be visualized in Figure 2 where the daily river level values for 1997-98 at Corrientes in the Paraná river (Figure 1) and Paso de los Libres in the Uruguay River are presented. Although both rivers exhibit the same low frequency modulation, the Paraná river presents lower interdiurnal variability. Figure 3 also provides an insight of the differences between both rivers as presents the spectral densities of the monthly discharge anomalies at Corrientes and Paso de los Libres for 1946-2000. The largest spectral densities of the Paraná river are found for periods of 4.5 months or longer and are almost null for periods shorter than 3 months. In the case of the Uruguay river, large spectral densities are also observed for the shortest periods. Figures 2 and 3 indicate that the greatest discharge events in the Uruguay river are mainly noticeable in a daily scale and eventually in a monthly scale as it will be discussed in section 6. 9

The critical Uruguay river level for flooding at the Salto Grande monitoring station established by the National Direction of Hydrology of Uruguay is 8 meters. Table 1 lists all the events during 1950-2000 with the eighteen highest river levels that in all cases were higher than 14 meters. The dates presented in this table were selected as the ones during the flood event with maximum height. Most of the extreme events presented in the table lasted for less than a week with the exception of the events occurred in 1983 and 1998 simultaneously with the strongest El Niño events of the century. The monitoring station Salto Grande is located downstream the Salto Grande dam shared by Uruguay and Argentina. In order to identify the possible effect of anthropogenic regulation of the discharges by the dam after 1980 when Salto Grande started to operate, Figure 4 presents a comparison between the daily river levels at the monitoring station Paso de los Libres upstream the Salto Grande dam and the Salto Grande values for a flooding period of 1997-98. Although it is possible to appreciate a slight modulation in the river flows at Salto Grande, the regulation activities at the dam are relatively small and do not hide the flood tide. Figure 4 also allows identifying that the lags between the beginning of the increase of the river level at Paso de los Libres and the date of maximum level at Salto Grande were of no more than 12 days. Therefore, to analyze the largest river levels of the Uruguay at Salto Grande, only the daily atmospheric fields of different variables up to 12 days before the flooding date will be studied. Table 1 shows that although the number of extreme events occurred during the warm semester (nine) is the same than during the cold season, the highest river level values are mostly observed during the cold semester. In fact, 8 out of the 10 highest river levels occurred between April and September. 10

5. Climate forcing of the major discharge events a. Rainfall, OLR and low-level atmospheric circulation In order to explore the climatic forcing of the flood events included in Table 1, they were divided in two groups of nine events each according to the month of occurrence, in warm season events (occurred during the austral spring and summer from October to March) and cold season events (occurred during the austral autumn and winter from April to September). Figures 5 to 7 present the composites for the warm and cold seasons events of the values of different meteorological variables for the averages of three four-day periods comprising the days -12 to -9, -8 to -5 and -4 to -1 prior the selected flooding dates presented in Table 1. Figure 5 presents the daily mean accumulated precipitation for the composites defined previously. Rainfall patterns show a quite different spatial distribution between the warm and cold season events. For the events occurred during the austral spring and summer, daily rainfall values are between 6 and 12 mm day -1 upstream Paso de los Libres in the period between -12 to -5 days before the maximum river level at Salto Grande with a maximum in the period -12 to -9. The four days immediately before the flooding date of the warm season events show lower daily rainfall values in the basin than the previous eight days although a high precipitation rate continues to the north outside the Uruguay basin. During the cold events, the largest precipitation values are observed in a southernmost position than during the warm events although they show a similar intensity during the -12 to -9 and -4 to -1 days. The rainfall amounts are quite small during the period -8 to -5 indicating that the extreme discharge events in Salto Grande could be caused by two wet periods with an intermediate one probably associated to a cold frontal passage. This is confirmed when inspecting individual cases (not shown). These results are consistent with the OLR composites presented in Figure 6. Considering that OLR values less than 240 Wm -2 are indicative of convective activity in the region (Kousky, 1988; Barros et al., 2000), the figure shows that the main convective centers over the Uruguay 11

basin are located in a southernmost position and are more intense during the cold events than during the warm ones. Figures 5 and 6 show a different temporal and spatial distribution of convection and rainfall over the Uruguay basin between the warm and cold events during the days previous to the flooding date. Consequently, floods registered during the warm semester (October to March) are due to intense rainfall in the upper basin particularly in the period -12 to -9 days before the flood date while those occurred during the cold one (April to September) are mostly due to the large rainfall amounts registered over and near upstream Salto Grande in two separate periods (-12 to -9 and -4 to -1). Figure 7 shows the four-day composites of the 925 and 850 hpa wind fields for the warm and cold events respectively. The different tropospheric levels considered for both periods were selected according to the results of Berbery and Barros (2002) showing that the maximum moisture flux over South America during the austral summer has a maximum between 900 and 950 hpa and a higher maximum at 850 hpa during winter. The composite for the warm season events show a northwestern flow over Bolivia and Paraguay that converges with a weak northern component over the Uruguay basin during the first two four-day periods. The composites for the cold events show an intense northwesterly flow over Bolivia and Paraguay that continues over northern Argentina and southern Brazil and penetrates in the Atlantic ocean for the periods -12 to -9 and -4 to -1. This circulation pattern is probably associated to warm and humid advection and it is in accordance with the two heavy rainfall periods over the Uruguay basin identified in Figure 5. The circulation pattern during the -8 to -5 period is somewhat different with a predominantly northern flow and weak convergence over the Uruguay basin. b. Relation between extreme discharge events and the SALLJ The SALLJ has been identified as a key factor in the water vapor transport and the triggering of organized convection associated to enhanced precipitation over Southeastern South America. 12

This section explores the occurrence of SALLJ events up to 12 days before the flooding dates indicated in Table 1. Although Bonner's (1968) criterion for low-level jet identification is based on observations, his methodology was adapted to NCEP reanalyses for studies to identify the SALLJ (Marengo et al, 2002). In this study, the conditions that must be satisfied to identify a day with SALLJ are: (a) meridional winds must be more predominant than zonal winds (v>u) immediately east of the Andes, (b) the meridional component must be southward (v<0), (c) wind speed must exceed 12m/s in some pressure level between surface and 850 hpa, and (d) vertical wind shear between 850 and 700 hpa (for the austral autumn and winter cases) or between 925 and 700 hpa (for the austral spring and summer cases) must be at least 6 m/s. These conditions must be observed at least at one of the four daily individual reanalysis data (0000, 0600, 1200 and 1800 UTC). The SALLJ was identified in 42.6% of the days among the 12 days before the flooding dates occurred during the austral autumn and winter. This percentage is slightly higher than the 40.7% obtained for the spring and summer cases. Consequently, almost half of the largest discharge events of the Uruguay river at Salto Grande could be related to enhanced precipitation due to moisture flux convergence in the SALLJ exit region. Table 2 summarizes the occurrence of SALLJ during the warm and cold seasons for three four-day periods before the extreme discharge dates at Salto Grande indicated in Table 1. Both seasons show the largest frequencies of occurrence of SALLJ during the periods -12 to -9 and -4 to -1. During the cold season events, these frequencies are more than twice the frequency of the intermediate period in accordance with the two heavy rainfall periods identified in Figure 5. 6. Relations between the largest monthly discharge events and ENSO The largest monthly discharge anomalies of the Uuruguay river for 1909-2000 at Paso de los Libres the gauging station were selected following the criterion that the discharge anomalies 13

were larger than at least three times the monthly standard deviation for the respective month. The eighteen events that satisfied this condition are listed in Table 3. The magnitude of these extraordinary discharges minimizes the possible impact of water management by upstream dams. Comparison between the maximum daily Uruguay river level at Salto Grande (Table 1) and the extreme monthly discharge anomalies at Paso de los Libres (Table 3) indicates that 10 out of the 18 extreme daily values resulted in extraordinary discharges at the monthly scale. Out of the other 8 extreme daily cases, six occurred during months that rank among the thirty-five largest monthly discharge anomalies of the period with values higher than twice the corresponding standard deviation for that month while the two remaining daily cases also show positive anomalies at the monthly scale. The cases listed in Table 3 constitute events that must necessarily have been caused by considerable precipitation anomalies. The Uruguay basin is part of a region that has a strong precipitation signal during ENSO events (Kousky et al., 1984; Ropelewski and Halpert, 1987, 1996; Kiladis and Díaz, 1989, Pisciottano et al., 1994) and therefore it can be expected a hydrological response to ENSO. To facilitate the discussion of this aspect, Table 3 includes a classification of the events according to the season and ENSO phase and the SST anomalies in the El Niño-3 region. The starting and end months of ENSO events that occurred after 1950 were defined according to Trenberth (1997) and the ENSO years before 1950 were taken from Kiladis and Diaz (1989). There is a clear relationship between ENSO phases and the major discharge anomalies in the Uruguay river. In Paso de los Libres, 13 out of 18 occurred during El Niño events. In addition, none of them occurred during La Niña phase. Table 3 shows that although the largest monthly discharge anomalies at Paso de los Libres occurred mostly during a warm ENSO phase, there is not a dominant season since there are 5 cases during the austral autumn (March to May) of the year following the onset of an El Niño event (autumn +), 4 14

during the austral spring 0 (September to November), 3 during the austral summer + (December to February) and only 1 during the austral winter+ (June to August). This result shows a different response to ENSO events between the Uruguay and the Paraná basins as the analysis of the largest monthly discharge anomalies of the Paraná river at Corrientes shows that the top discharges occurred during the austral autumn + (Camilloni and Barros, 2003). Figure 8 presents the seasonal rainfall anomalies in Southeastern South America for the El Niño phases with monthly extreme discharge events in the Uruguay basin considering all the El Niño events occurred after 1950. All seasons show positive rainfall anomalies over the Uruguay basin which are statistically significant at the 95% level for three of the four analyzed phases. The largest significant anomalies occurred during autumn + in the upper part of the basin at about 25ºS. During spring 0 and summer +, the largest anomalies occurred to the north of 30ºS upstream from Salto Grande. This figure suggests that the extreme monthly discharges at Paso de los Libres are a consequence of large rainfall anomalies to the north of that station with a relatively small contribution of local rainfall events. 7. Concluding remarks The Uruguay basin is part of a region in Southeastern South America that has a strong precipitation signal during ENSO events. The hydrological response to ENSO is evident as most of the largest monthly discharge events of the Uruguay river at Paso de los Libres gauging station occurred during the El Niño phase. Nevertheless, the analysis of the individual extreme discharge events on daily basis show that these floods last between 3 and 10 days and consequently it is necessary to study their climate features in the synoptic scale. The number of extreme events occurred at the Salto Grande gauging station at the daily scale during the warm semester (October to March) are the same than during the cold season (April to September). However, the highest river levels are mostly observed during the cold semester. 15

Floods registered during the warm season are due to intense rainfall in the upper basin particularly in the period -12 to -9 days before the flooding date, consequently the hydrological forecast of these events is possible many days in advance. The largest discharges occurred during the cold semester are mostly due to the large rainfall amounts registered over and upstream from Salto Grande in two separate periods (-12 to -9 and -4 to -1 days before the flooding date) and the hydrological prediction must necessarily depend on an appropriated weather forecast. Almost half of the largest discharge events of the Uruguay river at Salto Grande could be related to enhanced precipitation due to moisture flux convergence in the SALLJ region. The frequency of occurrence of SALLJ during the 12 days before the flooding date of the cold season extreme events is slightly larger than during the warm semester. This result shows the relevance of the cold season SALLJ in the contribution to moisture transport and atmospheric convection and also indicates a departure from the behavior of the North American LLJ that is only present and exhibits large moisture fluxes during the warm season (Berbery and Barros 2002). Acknowledgements This paper was funded by the University of Buenos Aires under its Program of Strategic Projects; by the AIACC, a Project of the GEF implemented by the UNEP and co-executed by START and the TWAS and by IAI CRN-55 (PROSUR). References Barros, V.R., M.H. Gonzalez, B. Liebmann and I.A.Camilloni. 2000. Influence of the South Atlantic Convergence Zone and South Atlantic Sea Surface Temperature on interannual summer rainfall variability in Southeastern South America. Theor. and Appl. Meteor. 67, 123-133. 16

Berbery, E.H. and V.R. Barros. 2002. The hydrological cycle of the La Plata basin in South America. J. Hydromet. 3, 630-645. Berri, G.J. and B.J. Inzunza. 1993. The effect of the low-level jet on the poleward water vapor transport in the central region of South America. Atmos.Envirnom.27A, 335-341. Bischoff, S.A., N.O. García, W.M. Vargas, P.D. Jones, and D. Conway. 2000. Climatic variability and Uruguay river flows. Water International 25, 446-456. Camilloni, I. and V.R. Barros. 2003. Extreme discharge events in the Paraná River and their climate forcing. Journal of Hydrology (in press). Doyle, M.E. and V.R. Barros. 2002. Midsummer low-level circulation and precipitation in subtropical South America and related sea surface temperature anomalies in the South Atlantic. J. Climate 15, 3394-3410. Figueroa, S., P. Satyamurty and P.L. Silva Dias. 1995. Simulation of the summer circulation over the South American region with Eta coordinate model. J. Atmos. Sci. 52, 573-584. García N.O. and W.M. Vargas. 1996. The spatial variability of runoff and precipitation in the Río de la Plata basin. Hydrological Science Journal 41, 279-299. García N.O. and W.M. Vargas. 1998. The temporal climatic variability in the Rio de la Plata basin displayed by the river discharges. Climatic Change 38, 359-379. Genta, J.L., G. Pérez Iribarren and C.R. Mechoso. 1998. A recent increasing trend in the streamflow of rivers in Southeastern South America. J. Climate 11, 2858-2862. Kalnay, E and coauthors. 1996. The NCEP/NCAR Reanalysis 40 years- project. Bull. Amer. Meteor. Soc. 77, 437-471. Kiladis, G.N. and H.F.Diaz. 1989. Global climatic anomalies associated with extremes in the Southern Oscillation. J.Climate 2, 1069-1090. 17

Kodama, Y.M. 1992. Large scale common features of subtropical precipitation zones (the Baiu frontal zone, the SPCZ and the SACZ). Part I: Characteristics of subtropical frontal zones. J.Meteor.Soc.Japan 70, 813-836. Krepper, C.M., N.O. García and P.D. Jones. 2003. Interannual variability in the Uruguay river basin. Int. J. of Climatology 23, 103-115. Liebmann, B., G. Kiladis, J. Marengo, T. Ambrizzi. and J. Glick. 1999. Submonthly convective variability over South America and South Atlantic Convergence Zone. J. Climate 12, 1877-1891. Mechoso, C.R. and G. Pérez Iribarren. 1992. Streamflow in Southeastern America and the Southern Oscillation. J. Climate 5, 1535-1539. Nicolini, M. and A.C. Saulo. 2000. ETA characterization of the 1997-98 warm season Chaco jet cases. Preprints of Sixth International Conference on Southern Hemisphere Meteorology & Oceanography. Santiago, Chile, 336-337. Nicolini, M.; Saulo, A.C.; Torres, J.C. and P.Salio. 2002. Enhanced precipitation over Southeastern South America related to strong low-level jet events during austral warm season. Meteorologica 27, 59-69. Nogués-Paegle, J. and K. Mo. 1997. Alternating wet and dry conditions over South America during summer. Mon. Wea. Rev. 125, 279-291. Oliveira, A.S. 1986. Interacões entre sistemas frontais na América do Sul e a conveccão da Amazônia (Interactions between the South American frontal systems and the Amazonian convection) INOE-4008-TDL/239, Instituto Nacional de Pesquisas Espaciais, 115 pp. Pisciottano, G.; A.Diaz; G.Cazes and C.R.Mechoso. 1994. El Niño-Southern Oscillation impact on rainfall in Uruguay. J. Climate 7, 1286-1302. Robertson, A.W. and C.R. Mechoso. 1998. Interannual and decadal cycles in river flows of Southeastern South America. J. Climate 11, 2570-2581. 18

Satyamurty, P., C.Nobre and P.L. Silva Dias. 1998. South America. Meteorology of the Southern Hemisphere. D. Karoly and D. Vincent (eds.). American Meteorological Society, 410 pp. Seluchi, M. and J. Marengo. 2000. Tropical-midlatitude exchange of air masses during summer and winter in South America: climatic aspects and examples of intense events. Int. J. of Climatology 20, 1167-1190. Tossini, L. 1959. Sistema hidrográfico y Cuenca del Río de la Plata. Contribución al estudio de su régimen hidrológico. (Hydrogrphic system and Río de la Plata basin. Contribution to the study of the hydrologic regime). Anales de la Sociedad Científica Argentina. Marzo-Abril 1959, III y IV, Tomo CLXVII, 41-64. Trenberth, K.E. 1997. The definition of El Niño. Bull. Amer. Meteor. Soc. 78, 2771-2777. Velasco, I. and J.M. Fritsch. 1987. Mesoscale convective complexes in the Americas. J.Geophys.Res. 92, 9591-9613. Vera, C.S., P.K. Vigliarolo and E.H. Berbery. 2002. Cold season synoptic-scale waves over Subtropical South America. Mon.Wea.Rev. 130, 684-699. 19

Captions Figure 1. Uruguay and Paraná basins and location of gauging stations (1: Salto Grande, 2: Paso de los Libres, 3: Corrientes). Figure 2. Daily river levels at Corrientes (Paraná river) and Paso de los Libres (Uruguay river) for 1997-98. Figure 3. Spectral densities of the monthly discharge anomalies at Corrientes (Paraná river) and Paso de los Libres (Uruguay river) for 1946-2000. Figure 4. Daily Uruguay river levels at Salto Grande and Paso de los Libres for a period between December 1997 and June 1998. Figure 5. Composites of daily mean accumulated precipitation for the averages of three fourday periods comprising the days -12 to -9, -8 to -5 and -4 to -1 prior the selected flooding dates for the warm (left)and cold (right) season events. Figure 6. As fig. 5, for OLR. Figure 7. As fig. 5, for 925 hpa wind (left) and 850 hpa wind (right). Figure 8. Rainfall anomalies in Southeastern South America during El Niño phases with extreme monthly discharges at Paso de los Libres (significant anomalies at the 95% level are shaded). 20

Table 1. Largest Uruguay river daily levels at Salto Grande during 1950-2000. Table 2. Relative frequency (%) of occurrence of SALLJ for three four-day periods before the flooding dates indicated in Table 1. Table 3. Extreme monthly discharge anomalies of the Uruguay river at Paso de los Libres. ENSO phase and El Niño-3 SST anomalies are also included. 21

Date Height (m) 16 April 1959 20.18 25 July 1983 17.85 16 April 1986 17.71 7 June 1992 17.71 19 June 1972 17.16 8 May 1983 16.91 10 January 1998 16.70 8 September 1972 16.43 26 April 1998 16.26 20 October 1997 16.01 17 November 1982 16.01 25 April 1987 15.91 13 November 1997 15.76 9 March 1998 15.59 18 October 1979 15.46 13 February 1998 15.33 9 November 1979 15.04 30 October 1954 14.63 Table 1. Largest Uruguay river daily levels at Salto Grande during 1950-2000. 22

period before the flooding date -12 to -9 days -8 to -5 days -4 to -1 days Warm season events 41.7 33.3 47.2 Cold season events 50.0 22.2 55.6 Table 2. Relative frequency (%) of occurrence of SALLJ for three four-day periods before the flooding dates indicated in Table 1. 23

Date of maximum monthly discharge anomaly Discharge anomaly (m 3 /s) SST Niño 3 anomaly (ºC) May 1941 (Autumn +) 17338 1.88 July 1983 (Winter +) 15838 1.04 November 1997 (Spring 0) 15592 3.59 May 1983 (Autumn +) 15196 1.91 October 1929 (Neutral) 15129 0.53 November 1982 (Spring 0) 14689 2.66 February 1998 (Summer +) 13615 2.57 October 1997 (Spring 0) 12946 3.29 June 1990 (Neutral) 12633 0.08 April 1998 (Autumn +) 11932 1.68 September 1928 (Neutral) 11591 0.07 September 1972 (Spring 0) 11301 1.52 April 1932 (Neutral) 9926 0.50 March 1998 (Autumn +) 9410 2.05 March 1983 (Autumn +) 9267 2.01 August 1983 (Neutral) 8894 0.92 January 1998 (Summer +) 8703 3.32 January 1912 (Summer +) 7692 1.39 Table 3. Extreme monthly discharge anomalies of the Uruguay river at Paso de los Libres. ENSO phase and El Niño-3 SST anomalies are also included. 24

3 1 2 ATLANTIC OCEAN Figure 1. Uruguay and Paraná basins and location of gauging stations (1: Salto Grande, 2: Paso de los Libres, 3: Corrientes). 25