Comparison of the performance of five indices for drought characterization in La Plata Basin. Perspectives towards a multi-scale monitoring system

RIVERA J. A., PENALBA O. C.

Conferencia; WCRP Open Science Conference; 2011

The quantitative knowledge of the properties of droughts in a region is an important aspect of theplanning and management of water resources. There are numerous natural drought indicatorsthat should be monitored routinely to determine the onset and end of drought and its spatialcharacteristics. This work compares the performance of five indices for drought monitoring in theLa Plata Basin (LPB) region, one of the largest basins in the world. The basin generates around 70%of the Gross National Product of Argentina, Brazil, Paraguay and Uruguay, and has a population ofover 100 million inhabitants. The LPB is also one of the major producers of hydroelectric power inthe world, generating around 95% of the energy used in Paraguay, 25% in Brazil, 40% in Argentina.Therefore, extreme low flows are critical for hydroelectric power and water resourcesmanagement. The economic wealth of LPB strongly depends also on agriculture, which set a majorchallenge: the increase in the prediction capacity of the impacts of drought for stakeholders.Five drought indices have been selected for this study. They include the Standardized PrecipitationIndex (SPI), the Deciles Index (DI), the Effective Drought Index (EDI), the Percentage of NormalPrecipitation (PRCT) and the Lack of Rain Index (LORI). These drought indices have been used toquantify deficits in water resources and as a drought monitoring tool in several regions of theworld. A common feature of the indices selected is that they all are calculated using precipitationdata only. Monthly rainfall data were obtained from the CLARIS LPB Data Base for 68 of itsstations, which were subjected to quality control procedures and have less than 10% of missingvalues. We used a common comparison period of 48 years (1961‐2008). The comparison wasperformed for the time scales of 3 and 12 months, in order to have a broad approach of thedifferent precipitation shortages. The assessment criteria to determine the most appropriate indexfor monitoring droughts include their interpretability, a detailed spatial and temporal analysis ofthe major historical droughts, and their flexibility in the time scale, among others. In conclusion,the SPI, which is the most widely used drought index, is likely the best choice for detecting theonset of drought and its spatial and temporal variation in LPB. However, the spatial and temporalresponses of the different indices point to the need of using several indices for drought monitoringin the study area. Besides the time scales evaluated, we suggest that the monitoring should beperformed over a broad time scales, which allows the effects of precipitation deficits on differentwater resource components to be assessed. These results could be useful for forecasting andmonitoring drought severity, to identify changes in drought frequency and magnitude, and fordeveloping a drought preparedness plan in the region. We also plan to use these results for thevalidation of model outputs and global and continental monitoring systems based in griddedprecipitation data, such as the VAMOS Atlas of Extremes for the Americas.