Converting pHÂ 1:1 H2O and 1:2CaCl2 to 1:5 H2O to contribute to a harmonized global soil database

•We determined pH 1:5 W for 563 samples previously analysed for pH1:1 W and 1:2CaCl2.•We examined the relationships between pH 1:5 W versus pH1:1 W and 1:2CaCl2.•We examined the influence of soil orders/mineralogy/texture/depth on relationship.•Linear regression for predicting pH1:5 W from pH1:1 W and/or pH1:2CaCl2 was selected.•Soil orders/mineralogy/texture/depth did not substantially improve the relationship.

The GlobalSoilMap initiative calls for the generation of continuous maps for soil properties, including pH in a 1:5 suspension of soil in water (pH 1:5W) based on a standard method, ISO 10390. The United States Department of Agriculture-Natural Resources Conservation Service (USDA-NRCS) employs a 1:1 suspension of soil in water (pH 1:1W), and a 1:2 suspension of soil in CaCl2 (0.01 M) (pH 1:2CaCl2) for routine pH analysis (Soil Survey Staff, 2009). The objective of this study was to determine the most efficient way to convert these pH values to the GlobalSoilMap standard. For this analysis, 563 soil samples from the USDA-NRCS-National Soil Survey Center (NSSC) soil archive, which had been previously analysed for pH 1:1W and pH 1:2CaCl2, were selected for determination of pH 1:5W, pH 1:5CaCl2 and electrical conductivity (EC) in 1:2 suspension of soil in water (EC 1:2W). The samples represented 11 soil orders, 8 mineralogy classes, 5 family particle size classes, 4 genetic master horizons, and 7 depth intervals. For each category, 25-30 samples were selected to represent a comprehensive pH range. Regression analysis showed strong and significant relationships (R2 > 0.92) between pH methods across all categories. The simple linear regression equation, pH 1:5W = − 0.51 + 1.06 pH 1:1W, had an RMSE = 0.44 pH units. Smoothing spline, did not significantly improve pH 1:5W predictions, nor did the incorporation of EC. Genetic horizons and soil depth intervals did not have a significant effect on pH 1:5W. The linear regression models for predicting pH 1:5W using pH 1:1W or pH 1:2CaCl2 as predictors emerged as the best candidates for a standard pedotransfer function. Using pedotransfer functions such as these will allow for the simple conversion of existing measured and estimated pH 1:1W or pH 1:2CaCl2 values from NRCS databases to the GlobalSoilMap standard of pH 1:5W.