Can saturated paste be used to predict bulk density in annual cropping systems in California?

Abstract

AbstractSoil bulk density (BD) is important for measuring changes in soil chemical, physical, and biological properties; however, the measurement is tedious to collect and requires specialized equipment. Database measurements for soil surface BD do not always correspond to present field conditions as field management can alter BD in time. Saturation percentage (SP) is a routine lab measurement. The objectives of this study are to (1) understand if a relationship between BD and SP can be developed and (2) build a model that predicts BD based on a routine low‐cost lab analysis. We collected 83 soil samples from different experimental sites around California's Central Valley. At each site, BD, SP, soil organic matter (OM), and soil total organic carbon were measured. A set of models were generated and compared based on their Akaike information criterion (AIC) and adjusted R2. The best two models are presented in this paper, and their accuracy and precision in estimating BD were further compared by calculating the root mean square error (RMSE) and the R2 of the predicted versus values measured in the field. We determined that a strong relationship between BD and SP exists (R2 = 0.70) and that a cubic model that includes SP and OM resulted in the best model to predict BD in California soils. Inclusion of additional data may further strengthen this model or make it applicable for other grower regions.