Delimitações de terras secas no Brasil utilizando dados de reanálise.

Abstract

Most of Northeast Brazil is characterized by a semi-arid climate with high temperatures and low rainfall levels. These climatic conditions, combined with human actions such as deforestation, land use, and occupation in the region, have significant impacts, leading to environmental degradation, loss of biodiversity, and increased vulnerability to extreme climate indices. To study the climate and conduct analyses of land use and land cover behavior, data that encompass the physical and climatic characteristics of the region with continuous observations are necessary. From the perspective of meteorological stations, this is not yet fully observed in Northeast Brazil, but it can be supplemented using reanalysis data provided they are validated. In this context, this work aims to: validate ERA5-Land and CPC/NOAA reanalysis data using meteorological station data; assess the performance of Thornthwaite's hydrological/climatic indices and the United Nations Environment Programme aridity index (IaUNEP) in delineating Drylands; investigate the use of vegetation indices as additional variables in Drylands climatic classification. For the validation of reanalysis data, the methodology employed consisted of analyzing statistical indices: mean percentage error, mean absolute error, standard estimation error, root mean square error, and correlation coefficient, with reference data being observed at stations and reanalysis data to be tested. Thornthwaite's climatic indices were determined using Thornthwaite and Matter's (1955) water balance methodology, while for IaUNEP, the ratio between observed or reanalysis precipitation and reference evapotranspiration obtained through the Penman-Montheith-FAO method was used. For vegetation indices, the NDVI and EVI equations were employed. As a result, it was observed that ERA5- Land reanalysis data represent the observed data very well, except for precipitation. For this reason, CPC/NOAA precipitation data, which were successfully validated, were used instead. From the obtained results, it was observed that using Thornthwaite's climatic indices, 87.4% of the Northeast area is classified as Dry Lands, with 26.7% of Dry Lands being arid. Meanwhile, using the IaUNEP, 71.2% are Dry Lands, with 0.3% being arid. Due to this discrepancy, a new climate classification index, called the absolute aridity index (Iab), was proposed, which yielded satisfactory results. Overall, the spatial configurations of the climatic indices obtained were consistent with the physical and climatic characteristics of the region, and the spatial patterns of vegetation indices generally followed the regional climate classification. It is concluded that ERA5-Land reanalysis data, except for precipitation, and CPC/NOAA precipitation data are reliable and can be used in climatic studies, especially in conjunction with observed data. Vegetation indices can be used as additional variables in climate classification in the Northeast.