Water deficit and water surplus maps for Brazil, based on FAO Penman-Monteith potential evapotranspiration

Abstract

The climatological water balance (CWB) proposed by Thornthwaite and Mather (1957) is a useful tool for agricultural planning. This method requires the soil water holding capacity (SWHC), rainfall (R) and potential evapotranspiration (PET) data as input. Among the methods used to estimate PET, the one proposed by Thornthwaite (1948) is the simplest and the most used in Brazil, however it presents limitations of use, which is caused by its empirical relationships. When Thornthwaite PET method is used into the CWB, the errors associated to PET are transferred to the output variables, mainly water deficit (WD) and water surplus (WS). As all maps of WD and WS for Brazil are based on Thornthwaite PET, the objective of this study was to produce new maps of these variables considering Penman-Monteith PET. For this purpose, monthly normal climate data base (1961-1990) from Brazilian Meteorological Service (INMET), with 219 locations in all country, was used. PET data were estimated by Thornthwaite (TH) and FAO Penman-Monteith (PM) methods. PET, from both methods, and R data were used to estimate the CWB for a SWHC of 100 mm, having as results actual ET (AET), WD and WS. Results obtained with PET from the two methods were compared by regression analysis. The results showed that TH method underestimated annual PM PET by 13% in 84% of the places. Such underestimation also led to AET and WD underestimations of 7% (in 69% of places) and 40% (in 83% of places), respectively. For WS, the use of TH PET data in the CWB resulted in overestimations of about 80% in 78% of places. The differences observed in the CWB variables resulted in changes in the maps of WD and WS for Brazil. These new maps, based on PM PET, provide more accurate information, mainly for agricultural and hydrological planning and irrigation and drainage projects purposes.