Researchers at the ATK Institute for Soil Sciences have developed a new national land evaluation model based on artificial intelligence (Random Forest). Using the traditional land valuation approach, the average yield was the dependent variable, and they also used remote sensing-based indicators and digital soil and climate maps. The importance of variables such as different soil parameters (soil type, pH, texture, organic matter, nitrogen, phosphorus and potassium content), average monthly precipitation, average monthly temperature and geographical coordinates were also considered in the model. The models were calculated for the three most important crops (wheat, maize, sunflower). In a final step, the resulting values were weighted by topography. The resulting maps contain values between 0 and 100 with a resolution of 100 metres. The proposed methodology can be used for integrated monitoring of biomass productivity in cadastral systems, land use planning and agricultural development programmes, among other possible applications. The research has been published in the open access journal Remote Sensing (Q1, IF: 5.349):
