Modelagem de matéria orgânica do solo com imagens de radar do Sentinel-1

Abstract

The quantification of soil organic matter is essential for the sustainability of agricultural and forestry systems, contributing to the physical, chemical and biological attributes of the soil, with an impact on the stability and productivity of agro-ecosystems. field, which to be effective, needs to be very dense and with a relatively high financial cost. With the use of Remote Sensing and Geoprocessing techniques, it becomes possible to make estimates without sampling in the field. Radar systems make it possible to estimate soil organic matter, whose data are obtained through the Sentinel l system, belonging to the European Space Agency (ESA), with Synthetic Aperture Radar (SAR) sensors, whose products are distributed free of charge and operate in the C-band of the microwave spectral range, making this estimation possible. In this work, radar images were used to estimate soil organic matter, the study area was the watershed of the river Una, located in the southeastern region of the state of São Paulo. Data from the Sentinel-1 scene from January 2016 to December 2016 were used. As a result, the values of the backscatter data in the Vertical/Vertical (VV), Vertical/Horizontal (VH) polarizations, VVxVH and VV/VH product, was used in the Multiple Linear Regression model with stepwise strategy. The result shows the value of R² = 0.32 and significance of p <0.002, with the adjusted model, the t test and Pearson's correlation were performed, the adequacy of the regression model shows that the observed differences are not significant at the level of 5%, which guarantees model validation.