Parametrization of Land Surface Temperature Fields with Optical and Microwave Remote Sensing in Brazil's Atlantic Forest

Brazil is home to one of the largest and most biodiverse ecosystems in the world, primarily encompassed in its forests. The main region of interest in this project is the Atlantic Forest, and although this forest is only a fraction of the size of the Amazon Rainforest, it is nearly as biologically diverse and is located on the East to Southeast region of Brazil, bordering the Atlantic Ocean. As luscious and nutrient enriched this region is, the area of the Atlantic Forest has been diminishing over the past couple of decades, mainly due to human influences and effects of climate change. These changes are causing habitual and migratory changes in animals in various parts of the forest. One way to monitor climate change is by gathering and comparing changes in temperature, specifically by studying the land surface temperature (LST) datasets. The focus of this project is to compare LST based on optical and microwave remote sensing, particularly, using the LST retrieved from the Moderate-resolution Imaging Spectroradiometer (MODIS) and Advanced Microwave Scanning Radiometer EOS (AMSR-E) instruments. The motivation behind this project is to understand which instrument is able to model the actual land surface temperature, given their respective limitations. Once a formidable understanding is attained then one can use remote sensing based LST to model for various biodiversity related variables, such as biodiversity endemism and species distribution. Since AMSR-E is a microwave remote sensing instrument, products derived from it's measurements would not be altered due to cloud coverage, which is not the case for MODIS. This would indicate the temperature field of AMSR-E should be more accurate and temporally dense. The findings from the project showed that overall AMSR-E computed higher temperatures in the Atlantic Forest than MODIS and the in-situ. The spatial resolution of MODIS is far more finer than AMSR-E and therefore, MODIS can tell more about temperature dynamics in a smaller scale, however, there are gaps in MODIS LST recordings due to cloud coverage. The downscaled 1km AMSR-E can reveal more about spatial variability of temperature than its original 25km product. This 1km AMSR-E product should similar temperature fluctuations as that of the in-situ data.