Remote sensing environmental indicators for monitoring spatial and temporal dynamics of water and vegetation conditions: applications to the Brazilian biomes

Abstract

The SAFER (Simple Algorithm for Evapotranspiration Retrieving) algorithm and the radiation use efficiency (RUE) model were coupled to test large-scale remote sensing environmental indicators in the Brazilian biomes. The MODIS MOD13Q1 reflectance product and gridded weather data were used for the year 2016. The analyzed biomes were Amazon, Caatinga, Cerrado, Pantanal, Atlantic Forest, and Pampa. Significant differences on precipitation (P), actual evapotranspiration (ET), and biomass production (BIO) yielded differences on water balance (WB = P - ET) and water productivity (WP = ET/BIO). The highest WB and WP differences along the year were for the wettest Amazon, Atlantic Forest, and Pampa biomes, when compared with the driest Caatinga biome. Rainfall distribution along the year affected the magnitude of the evaporative fraction (ETf), i.e, the ratio of ET to reference evapotranspiration (ET0), however there was a gap between ETf and WB, what can be related to the time needed for recovering the good soil moisture conditions after the rainy seasons. For some biomes, BIO was more related to the levels of absorbed photosynthetically active radiation (Amazon, Atlantic Forest, and Pampa), while for others BIO followed more the soil moisture levels, represented by ETf (Caatinga, Cerrado, and Pantanal). The large-scale modelling presented suitability for monitoring environmental indicators, opening the room to detect anomalies for specific periods along the year by using historical images and weather data, with great potential to subsidize public policies regarding the management and conservation of the natural resources and possibility for replication of the methods in other countries.