Design, inference and application of several runoff-rainfall statistical models at six sub-basins within La Plata Basin

GULIZIA, CARLA; HANNART, ALEXIS; CAMILLONI, INÉS

La Haya

Conferencia; 7th International Scientific Conference on the Global Water and Energy Cycle; 2014

GEWEX

The link between climate change and freshwater resources is a major concern of the scientific community and the societyas a whole. However, climate change is only one of several factors thatthreaten water resources. Land use changes, agrochemicals management and dams installation,among others, may also alter the hydrological system. The present study focuses in South America, particularly in La Plata basin, since this region is extremely important for being one of the main sources of freshwater in the world and because it has a great socio‐economic importance.A better understanding of the hydrological variability of the basin in the context of climate change and variability is essential to generate information that maybe relevant to hydrologists and water resources managers. Consequently, the overallobjective of this study is to advance in the knowledge and statistical characterizationof hydrological variability of the great rivers of La Plata Basin. Particularly,it was aimed to evaluate to what extent the precipitation variability helps to explain the runoff variability corresponding to the closing points of 6 defined sub‐basins within La Plata Basin: Paraguay, Pantanal, Lower and Upper Parana, Uruguay and Iguaçu. For thispurpose, the design, inference and application of several runoff-rainfall statistical models (RRM) was performed based on the waterbalance equation under linear regression settings for monthly to annual temporalscales where the soil water storage can be considered negligible. Runoff simulations at six river sub-basins within La Plata basin were assessed exhaustively. Moreover the delay of rivers discharges of La Plata Basin was analyzed through unit hydrographs for each sub‐basin. Evapotranspiration was estimated through various parameterizations proposed in the literature, including one that considers the effect of vegetation in order to assess the runoff response to land use changes. This last issue wasevaluated using eighteen variables derived from the Global Land Modeling (GLM) Dataset,which provides global gridded estimates of the underlying land conversions (land-usetransitions), wood, harvesting, and resulting secondary lands annually. Overall,both of the proposed statistical models capture remarkably well the variability of flows in all the studied sub‐basins, though major deficiencies were found in Paraguay and Pantanal sub-basins. Moreover, simulations present some discrepancies in representing the associated extremes. Uruguay and Iguaçu rivers have shown both the variability and magnitude fairly well represented. Besides, the RRM, which considers the effect of land use changes produced in the sub-basins, seems to explain partially better the positive trends observed in runoff, except for the Uruguay River. Furthermore, results from this latter RRM show a reduction in the differences found between simulated and observed runoff extremes for most of the river sub‐basins.