Linear Regression Models to Estimate Soil Liquid Limit and Plasticity Index From Basic Soil Properties

In Soil Survey, there is a need to estimate liquid limit (LL) and plasticity index (PI) for areas where data are not available. The objectives were to determine if LL and PI prediction equations could be developed from readily available soil properties in Soil Survey, and to test two different data stratification approaches to improve predictability. Measured data in the National Soil Survey Characterization database and multiple linear regression were used for model development. Clay content (<2 µm) and cation exchange capacity were the primary variables used to predict both LL and PI. To predict LL, four equations were developed from 10 taxonomic soil order strata (aggregate of seven soil order strata, Andisols, Spodosols, and Vertisols) that explained between 68% and 81% of the variation in LL, with the Andisols order having the lowest predictability. To predict PI, 10 unique taxonomic soil order equations were developed (Aridisols, Alfisols, Entisols, Inceptisols, Mollisols, Oxisols, Ultisols, Andisols, Spodosols, and Vertisols) that explained between 15% and 77% of the variation in PI, with the Andisols order having the lowest predictability. A few prediction equations were developed from the taxonomic mineralogy strata, which produced models with similar predictability to that of the soil order equations. Validation of the best fitting models with an independent data set showed no significant difference from unit 1 slope and 0 intercept. Predicting LL and PI from readily available soil properties resulted in mostly moderate to strong prediction equations. The most useful equations are those with R2 > 0.60. These prediction equations can be useful in Soil Survey when there are no available data.