Efeito da mudança climática no padrão dos erros gerados pelas estimativas remotas de precipitação do produto IMERG-GPM

Abstract

Rainfall monitoring is essential for hydrological, hydraulic, and environmental studies, such as flood forecasting, dam failure, and hydrodynamic simulations. However, fieldmeasured rainfall data are not always adequately available. Satellite Precipitation Products (PPS), such as IMERG-GPM, offer free and accessible estimates, partially filling this gap. Since these are estimates, they are subject to errors, which can be influenced by climate variables such as temperature and solar irradiance. Although there is evidence of these influences, quantitative analyses on the subject are still scarce. Thus, the objective of this study is to statistically analyze how solar irradiance and temperature have been influencing the error pattern of estimates through metrics that correlate the three variables. The Sertão Alagoano region was used as the study area due to its low rainfall monitoring. For this purpose, solar irradiation and temperature data were collected using an algorithm that interpolates data measured in the field or provides data in a gridded data format, the BR-DWGD. Subsequently, an algorithm was developed that uses bilinear interpolation to obtain satellite precipitation estimates for any point in the study area. The Spearman Correlation Test was also used to determine the relationship between the variables, the Mann-Kendall Test to assess trends, and the Principal Component Analysis of the data. The Spearman Test returns that solar radiation presented the greatest relationship in the average and regional data, with 𝑟𝑠 = 0.213, and 𝑟𝑠 = 0.532, respectively. The Mann-Kendall Test returned that the maximum temperatures for the region of interest show an increasing trend over the years, with a 𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 0.02, and a 𝑧𝑠 = 2.41. These results demonstrate a correlation between the climate variables and the rainfall error, and that at least one of the climatological variables has been influenced by climate change. Finally, the PCA indicates a correlation between the errors and solar radiation, evidencing what was presented in the previous test, in addition to mapping the spatial distribution of the data.