1. Title: Soil Colors of The World.

Coordinator: Rodnei Rizzo

Supervisor: José A. M. Demattê

​

This project aims to develop a World Soil Color Map of agricultural lands. Although our primary goal is to perform a map, we also aim to calibrate soil color space models, which later will be made available and will allow users to retrieve the soil color notations based on satellite reflectance measurements. To accomplish such a challenge, we are inviting soil scientists from different parts of the globe. We already have access to representative soil samples from different soils, more specifically the visible, Near- and Shortwave Infrared (350-2500 nm) data. These data will be invaluable to increase the representativity of our dataset and consequently provide a robust algorithm. 

 

​

2. Title: Space–time topsoil attributes mapping for climate change mitigations at near-global scale.

Coordinator: Raul R. Poppiel

Supervisor: José A. M. Demattê

​

We will use Random Forest regression algorithm for mapping near-global topsoil organic carbon and other attributes in space and time at 250 m resolution between 1982 and 2020. The models will be calibrated using global and local soil datasets and environmental covariates. Relief and geology will be used as stable covariates, while vegetation and topsoil reflectance from Landsat satellites will be used as dynamics covariates. Global space–time topsoil maps will be associated with future climate projection layers (e.g. 2021-2040, 241-2060, 2061-2080 and 2081-2100 year periods) to assess the impact of land uses on climate change. The results may provide significant information to land managers and policy makers worldwide.