A lecture by Prof. Dr. Milan Kilibarda was held at the Mathematical Institute of the Serbian Academy of Sciences and Arts (SANU), focusing on modern spatial interpolation methods, with a particular emphasis on Random Forest Spatial Interpolation (RFSI) – an innovative approach that combines artificial intelligence and spatial statistics.
The lecture included a comparative overview of classical interpolation techniques (e.g., kriging), as well as the advantages offered by regression kriging and machine learning. The RFSI method is especially valuable when the number of available predictors is limited, as it incorporates values from neighboring observations directly into the model.
This method has broad application in environmental modeling, including the interpolation of temperature, precipitation, and soil quality. It is also directly relevant to the objectives of the ForestCO2 project, particularly in the context of high-resolution mapping of biomass and soil organic carbon.
