Validação de elementos climáticos estimados por sensoriamento remoto e reanálises na Região Centro-Oeste do Brasil

Abstract

The Central-West region of Brazil plays a crucial role in the national economy, being recognized as one of the main agricultural production hubs in the country. With an extensive territorial area, complexity of biomes and geomorphological variety, the region contributes significantly to Brazil's climate variability. However, obtaining climatological data in this area is challenging due to the heterogeneity in the distribution of conventional meteorological stations and the complexity of local biomes. Given this scenario, the use of remote sensing and reanalysis products has been explored as an alternative to overcome the lack of conventional climate data. These methods provide accurate and continuous estimates of the spatial and temporal distribution of climate variables, filling the gaps left by the lack of traditional weather stations. Reanalyses, in particular, provide gridded data, enabling detailed analysis of climate patterns at different scales. The present study aimed to validate data estimated by different remote sensing products and reanalysis for essential climate variables, such as wind speed, precipitation, air temperature, relative humidity and vapor pressure deficit in the Midwest region. Using data from 120 automatic weather stations distributed across the region, the datasets evaluated included CPC, CHIRPS, ERA-5, GLDAS, GPCP, GPM, JRA-55, MERRA-2, NCEP/DOE and TRMM.The research was divided into three chapters dedicated to specific variables: wind speed (Chapter I), precipitation (Chapter II), air temperature, relative humidity and vapor pressure deficit (Chapter III). Each chapter covered different time scales and biomes, providing a comprehensive analysis of climatic conditions in the region. In Chapter I, dedicated to Wind Speed, it was observed that ERA5-Land presented the best performance on daily, monthly and annual scales for the Amazon, Cerrado and Atlantic Forest biome regions. However, for the Pantanal, GLDAS proved to be more effective, highlighting the need to consider the specific characteristics of each biome when interpreting the results. In Chapter II, referring to Precipitation, the CPC xlii achieved the best performance on the daily scale, while the model ensemble obtained the best performance on the monthly and annual scales. Overestimations and underestimations varied across different biomes, highlighting the importance of evaluating model performance in specific regional contexts. In Chapter III, addressing air temperature, relative humidity and vapor pressure deficit, distinct performance patterns were identified. The CPC excelled in maximum air temperature, while the ensemble was superior on monthly and annual scales. ERA5-Land showed consistency as the best performance in relative humidity and vapor pressure deficit across all scales and biomes. The study highlights the need for validation and adjustments appropriate to the region, considering the complexity of biomes and specific climate variability. The results provide valuable information for strategic sectors, such as renewable energy, civil infrastructure, agriculture and climate forecasting, contributing to a more robust understanding of climate conditions in the Central-West region of Brazil. The study highlights the continued need for in-depth research and specific adjustments for each biome, aiming to improve the reliability of climate estimates in the region.