Abstract
The aim of the study was to analyze the variability in the TVDI (Temperature-Vegetation Dryness Index) obtained from orbital sensors with distinct resolution in an agricultural region in southern Brazil. Three images of the Landsat 8 satellite OLI/TIRS (Operational Land Imager/Thermal Infrared Sensor) and 12 of Terra’s MODIS (Moderate Resolution Imaging Spectroradiometer), were used. Data collected in the field served as a basis for OLI/TIRS image classification and mapping of rice, soybean, grasslands, gallery forest and bare soil areas. The TVDI obtained by two parameterizations at different periods was evaluated, using the dispersions between Surface Temperature (TS) and NDVI (Normalized Difference Vegetation Index). The TVDI obtained for both sensors presented a similar pattern allowing differentiation of the targets. The average of all dates and classes, TVDI obtained from MODIS was 0.128 units higher than the obtained with the OLI/TIRS. When used OLI/TIRS there is a better detail in the representation of the moisture, but with less repetition throughout the crop cycle. Using TVDI-MODIS, it is possible to monitor moisture conditions on a regional scale, with less spatial detail, but in a continuous way over time. The TVDI estimated by the OLI/TIRS and MODIS can be used together, providing complementary information.
Keywords:
Surface temperature; NDVI; agriculture; water deficit
