Geostatistical-based mapping of topsoil texture in Fluvisols and Vertisols around Lake of Manyas

Abstract

It is critical to understand the spatial distribution of soil particle fractions to create sustainable soil management methods. The spatial distribution of particle fractions has long been studied using geostatistics. Soil particle fractions can be assessed and mapped using a variety of approaches, but selecting the best appropriate method for anywhere has always been a controversial topic in all soil mapping applications. In this study, there is an evaluation of the estimation performance of ordinary kriging (OK) and IDW (Inverse distance weighting) methods for digital mapping of soil particle fractions. It was determined that the clay content of the soil samples was between 24% and 76%, the sand content was between 2% and 69%, and the silt content was between 3% and 44%. The performance of the models was evaluated by the results of the root mean square error (RMSE) and the sum of the fractions. The best results were found using the OK method for Silt (RMSE: 5.36%), while IDW produced more high predictions for Clay (RMSE: 13.80%) and Sand (RMSE: 19.90%). In the control of composition structure, IDW is the method that most closely predicted the relative sum of the three fractions defined as 100%. Creating texture classes in a GIS environment and comparing the efficiency of the produced soil fraction maps is advised.