How to Estimate Soil Field Capacity Without Using a World Map?.

Field capacity (FC) as a soil hydrologic parameter is widely used in climatological, hydrological, agronomic, and ecological models at coarse spatial scales. FC is commonly equated to soil water contents of small soil cores at specific soil water potential. The values of this specific potential are amazingly different in different parts of the World, e. g. -5 kPa in UK, -6 kPa in Brazil, -10 kPa in Australia, -33 kPa or -10 kPa in the United States, variable in Russia, etc. The objective of this work was to revisit this common assumption. We used the USDA-ARS database on FC derived from experiments in 30 locations across US. The 360 datasets contained the FC as water content measured after two days drainage after ponding, soil core water retention, and contents of three fractions of sand, two fractions of silt, and clay, and bulk density. Most of textural classes were well represented. The water content at -33kPa, θ₃₃, gave a biased representation of FC which was larger than θ₃₃ by about 0.08 cm³cm^-3. A poor correlation was observed between the θ₃₃ and FC, R² was 0.42 for all samples and 0.20 for all but sandy soil samples. To relate FC directly to soil properties, we applied regression trees that allow one to select the most efficient predictors. Medium sand, fine sand, coarse silt, and total silt appeared to be the best predictors, and the determination coefficient of the linear relationship between predicted with regression trees and measured FC was R²=0.72. Including the depth in the list of predictors further improved the correlation. The direct FC estimation from soil basic properties can be not only more accurate but also more suitable for the scales at which this parameter is used, as the information for such direct estimation is available from soil surveys.