Study of lateral and vertical distribution of soil calcium carbonate using geostatistics and spline functions

In most studies on the distribution of soil carbon, three-dimensional distributions of soil inorganic carbon were neglected. Calcium carbonate is the most common carbonate in arid and semi-arid soils. Information on its spatial three-dimensional distribution is very important to determine factors controlling its distribution, to predict soil behavior and to improve soil management practice. This study aimed to map three-dimensional distributions of soil calcium carbonate (SCC) in an area of 3600 ha located in Silakhor plain (Lorestan province). An equal-area spline depth function (ESDF) was fitted to the measured SCC data of 103 pedons and the amounts of SCC at the five standard depths of the global soil map project were estimated.Then, ordinary kriging was employed to map the lateral distribution of SCC at all specified depths. Geostatistical analysis showed that spherical model was the best model representing spatial structure of calcium carbonate in all depths. All experimental variograms had a nugget to sill ratio less than 25 % which indicated strong spatial dependence for SCC. Anisotropy analysis indicated that ranges of variograms for all specified depths in the northwest-southeast direction were more than perpendicular direction. It indicated that SCC had more spatial dependence along Silakhor plain due to small variations in land use, slope and parent materials along the plain. Spline functions showed good performance in predicting vertical distribution of SCC (R2=0.88, RMSE=0.99). Both lateral continuous maps and spline functions indicated an increasing trend in SCC with increasing depth. In the eastern and southern parts, due to poor drainage and low leaching, SCC was high and its increasing trend with depths was not significant. Generally, results indicated that the combination of spline functions and geostatistical method is a promising approach to map three-dimensional distribution of SCC and to deal with some of the problem arising from legacy soil maps.