Correlation of topography, soil properties and spatial variability of biomass - application of the self-organizing map method

Correlation of topography, soil properties and spatial variability of biomass - application of the self-organizing map method. Interactions of various geogenic factors and processes have a lasting effect on topography and vice versa on the development of biomass. This is obviously detectable in an agricultural plot, considering the spatial variability of biomass, here specified in terms of yield. The aim of the present study is to identify relationships among topography, chemical and physical soil properties and yield. We realized the analysis using the method of self-organizing maps (SOM), an unsupervised neural network method. The output of the SOM analysis provided a strong correlation between soil moisture and the amount of yield. Additionally, a relationship among soil depth and yield could be noted. The availability of plant nutrients, however, can positively affect the amount of yield, especially for zones with shallow soils. An inverse correlation was detected for soil depth and inorganic carbon content of the upper soil. Topographic properties expressed by the Topographical Index show a similar behaviour towards target variable yield, while exposition is of secondary importance. These results form a coherent image, which allows applying the SOM approach for further investigations.