By 2023, deforestation in the Cerrado biome surpassed 50% of its original area, primarily due to the conversion of native vegetation to pasture and agricultural land. In addition to anthropogenic pressure, climate change has intensified hydrological stress by reducing precipitation and decreasing river flows, thereby threatening water security, quality, and availability in that biome. The Annual Water Yield (AWY) model from the InVEST platform provides a tool to assess ecosystem services by estimating the balance between precipitation and evapotranspiration (ET). In this study, we applied the AWY model to the Urucuia River Basin, analyzing water yield trends from 1991 to 2020. We evaluated climate variables, land use dynamics, and river discharge data and validated the model validation using observed stream flow data. Although the model exhibited low performance in simulating observed streamflow (NSE = −0.14), scenario analyses under reduced precipitation and increased evapotranspiration (ET) revealed consistent water yield responses to climatic variability, supporting the model’s heuristic value for assessing the relative impacts of land use and climate change. The effects of deforestation on estimated water yield were limited, as land use changes resulted in only moderate shifts in basin-wide ET. This was primarily due to the offsetting effects of land conversion: while the replacement of savannas with pasture reduced ET, the expansion of agricultural areas increased it, leading to a net balancing effect. Nevertheless, other ecosystem services—such as water quality, soil erosion, and hydrological regulation—may have been affected, threatening long-term regional sustainability. Trend analysis showed a significant decline in river discharge, likely driven by the expansion of irrigated agriculture, particularly center pivot systems, despite the absence of significant trends in precipitation or ET.
Despite the several existing initiatives to obtain evapotranspiration data for several scientific analysis, it remains a challenge when dealing with data scarcity, especially when it is coupled with many data gaps of long-time climatological data series. In this study, we developed two Python scripts to perform reference evapotranspiration calculations in data scarcity conditions, filling potential data gaps, and testing them in the Urucuia River watershed.The first script (ETO_calc.py) is able to calculate evapotranspiration both by the Penman-Monteith method, adapting it according to user data availability, and by the Hargreaves-Samani method. The second script (Filling_gaps.py) fills out potential data gaps based on input from neighboring stations and tests the adequacy of the filling. The first script was tested by comparing the data generated by it with the online calculator 'ETCalc', while the second was tested by filling in a station's data from five neighboring stations, comparing the original data with the filled data. The output of the first script was practically identical to the online calculator used for comparison, which confirms the adequacy of the estimation made by ETO_calc.py. As for the second script, the regional weighting method and both linear and multiple regression methods were the methods that best filled in the missing data for the main station. This is due to their better fit to the historical variation of the data compared to the other tested methods. The results allow us to conclude that both scripts proved to be suitable tools for estimating evapotranspiration data.