Geostatistical modeling—a tool for predictive soil mapping

Geostatistics has been extensively used to determine the spatial variability of soil properties based on its capacity to quantify and minimize sampling errors. Researchers have been examining the spatiotemporal variability of soil variables since the 1970s. By predicting spatial correlation with minimum variance and using regional variables as a foundation, geostatistics has been proven to be among the most effective tools for assessing the spatial pattern and variation of soil characteristics. Kriging is a fundamental geostatistical methodology for generating the best linear unbiased estimate for spatially dependent variables. Geostatistical modeling methods such as kriging, semivariance functions, and variograms are used in the spatial interpolation of the collected data to generate a digital soil map. To quantify the measurement at an unsampled location, digital soil mapping (DSM) consists of three main components: information in the form of experimental and field objective measurements, a framework in the form of spatial and nonspatial soil inference systems, and output in terms of spatial soil information systems, which includes the outcome in the form prediction raster as well as also prediction uncertainty. DSM may be used to produce specific soil interpretations, improve or update current soil surveys, and measure soil prediction uncertainty. DSM effectively generates measured spatiotemporal variability and minimizes the requirement to combine different soil characteristics according to a set of mapping scales. Because of their ability to articulate the spatial continuity of a discrete soil property as well as to quantify associated uncertainties in these estimates and also to map a large area and a large number of samples, DSM and geostatistics have replaced the conventional statistics in predictive mapping and monitoring to evaluate soil heterogeneity and variation.