Análisis espacial y temporal de la precipitación en imágenes satelitales Chirps en el contexto de la Orinoquía colombiana.

Abstract

This study focuses on the space-time analysis of precipitation data obtained from satellite images within the CHIPRS program in the Orinoquia river basin, with the validation of these data using information from pluviometry stations provided by IDEAM. The justification for this research is grounded in the limited availability of fluviometric data in the study area, which poses challenges and opportunities in the context of regional agricultural production planning. For the methodology, historical data extraction was carried out from Chirps satellite images spanning from 1981 to 2022. A prior selection of 178 pluviometry stations meeting the 10% historical data requirement was also conducted. Subsequently, data validation was performed through the calculation of correlation coefficients, revealing that 56.18% of the data showed strong relationships. The coefficient of determination indicates that 49.44% of the points explain the model's variability with values ranging from 0.74 to 0.90. The applied error metrics indicate acceptable biases in 72 points, although significant underestimation is observed in 14 study points. The Root Mean Square Error (RMSE) reflects moderate errors for 52.81% of the points of interest, while the Mean Absolute Deviation (MAD) shows low dispersion in 84.83% of the study points. Additionally, no significant differences were found that would invalidate satellite data compared to physical station measurements. However, significant differences are observed in the majority of study points during the months of July and October.