Data Fusion Strategy for Mapping Environment and Climate Variables of Brazil

The study aims at utilizes the machine learning methods in cartography with a case study of climate and environmental mapping of Brazil. Rapid advances in machine learning applied to Earth observations have resulted in the application of scripting and programming languages for cartographic visualization and modelling. This research applies the Generic Mapping Tools (GMT) scripting toolset for advanced environmental mapping of Brazil. The data includes TerraClimate dataset of 2020. The GMT is an advanced cartographic tools that operates mapping from the console using scripts. Selected codes of the used scripts are presented in the research for technical explanation of the workflow. The results show correlation among the parameters and demonstrate climate and environmental trends notable for different biomes of Brazil: Amazônia, Caatinga, Cerrado, Pampa and Pantanal. The study presents 10 new maps made using GMT. Based on the obtained results, the increase of precipitation is notable in the Amazônia, along with the highest temperatures in the northern Brazil (Negro river basin) which corresponds to the increase in soil moisture and runoff. The evapotranspiration is higher in the southern regions than those in the north. On the contrast, the Caatinga region shows the minimal values of evapotranspiration, soil moisture and runoff. The main advantage of scripting cartography, demonstrated in this research, consists in automated data processing which pushes climate studies towards a data-driven research. Automated mapping technically facilitates the workflow due to the fast and smooth handling of various formats and types of data. The results contribute to the environmental analysis of climate in Brazil that has applications in agricultural and food studies and shows technical use of GMT.