http://lattes.cnpq.br/0256880651038969; MATSUNAGA, W. K.; MATSUNAGA, Wendy Kaori.
Abstract:
The agricultural zoning of climate risk (AZCR) is a fundamental tool for agricultural activities, as it identifies regions and times of lower climatic risk for planting and sowing crops. An AZCR representative of large areas requires a network of meteorological stations with dense spatial distribution, routine, and accurate observations. In Brazil, there is a good spatial distribution of rain gauges, but there is a scarcity in the spatial distribution of meteorological stations necessary to conduct an AZCR. One way to overcome the lack of data observed in situ is to use reanalysis data. However, it is necessary to show that it is possible to carry out a reliable AZCR. Therefore, the objective of this research is to elaborate an AZCR for corn in the state of Bahia using reanalysis data and quantifying its viability. Possible relationships between the Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) and AZCR were also analyzed. Finally, a AZCR was created for corn cultivation in Bahia, using climate projections data from the Coupled Model Intercomparison Project Phase 6 (CMIP6), precisely the CNRM-CM6-1-HR and EC-Earth3-Veg-LR models, both in the SSP2-4.5 scenario. The ERA5-Land reanalysis database was chosen. When performing the validation of the ERA5-Land reanalysis data, it was verified that rainfall was the only meteorological variable needed for an AZCR that did not present seasonal and interannual values and variability similar to those already observed. Therefore, it was decided to validate the rainfall data from the CPC/NOAA Precipitation Project. With data from the ERA5-Land and CPC/NOAA Precipitation Project reanalysis validated, the AZCR for corn was performed in the state of Bahia, leading to the conclusion that it is possible to develop a reliable and robust AZCR with reanalysis data. On the other hand, it was found that the vegetation indices follow the ZCAR in some areas, mainly those with original vegetation, and in others there was no coherence, especially in the most anthropic areas. The AZCR prepared with CMIP6 data using the CNRM-CM6-1-HR and EC-Earth3-Veg-LR models presented different results between themselves and also with the AZCR performed with reanalysis data.