Machine Learning Models for Prediction of Soil Properties in the Riparian Forests

Spatial variability of soil properties is a critical factor for the planning, management, and exploitation resources. Thus, use different digital mapping models to provide accuracy plays crucial role in providing physicochemical maps. Soil spatial forest stands not well-known Iran. Meanwhile, riparian buffers are important several services such as high water quality, nutrient recycling, buffering agricultural production. Accordingly, this research, 103 samples were taken using Latin hypercubic method Maroon Behbahan lands vicinity evaluate nitrogen, potassium, organic carbon, C:N ratio, pH, calcium carbonate, sand, silt, clay, bulk density. Different machine learning models, including artificial neural networks, random forest, cubist regression tree, k-nearest neighbor used compare estimation properties. Moreover, three main sources information remote sensing images, elevation model, climate parameters ancillary data. Our results indicated that model has best estimating In contrast, tree phosphorous, clay. Further, networks showed silt contents. revealed geospatial terrain parameters, satellite images could be well data physiochemical forests lands. conclusion, specific needs each property highly accurate maps with less error.