IRRIGATED AGRICULTURE MAPPING IN A SEMI-ARID REGION IN BRAZIL BASED ON THE USE OF SENTINEL-2 DATA AND RANDOM FOREST ALGORITHM

Abstract. Irrigation is important for agricultural production and often decisive this, especially in arid semi-arid areas, where precipitation insufficient. In Brazil, irrigated agriculture responsible 46% of withdrawals from water bodies 67% the consumption total volume collected, representing highest consumptive use country. Remote sensing technologies have great potential developing methods monitoring areas. However, mapping areas still a challenge, due to complexity diversity irrigation crops, country with continental dimensions like Brazil. techniques Brazil been applied mainly center pivot Cerrado, paddy rice south But few or no applications, involving crops by other methods, semi-arid, carried out. The objective this work was investigate method classifying semiarid region based on Sentinel 2 imagery random forest algorithm. We proposed novel robust methodology showing preliminary results that it's possible identify class-f1-score 74% complementary 95% center-pivots.