|| Checking for direct PDF access through Ovid
The Gregson et al. one-parameter model (GHM) is based on the log-log form of the soil water retention curve, below the air-entry value of ψ, ln ψ = a + b ln θ, where a and b are the intercept and slope, respectively. A strong linear relationship observed between a and b was expressed as a = p + qb. Given this relationship, the GHM was derived as ln ψ = p + b (ln θ + q). Given p and q values for a soil or group of soils, only one value of the ψ(θ) relationship needs to be known to calculate the only unknown parameter in the model - b, and hence, the entire ψ(θ) function. Typically, θ at the −33 kPa matric potential (θ-33 kPa) is used as the known ψ(θ) value. Here we provide a regression relationship between b and the available water content (AWC) to estimate b, since in many cases the AWC is available in the USDA soil survey reports, whereas θ-33 kPa is not. Using the b thus estimated in GHM gives only slightly larger errors in calculating the water content at different potentials than when using θ-33 kPa. Further we show that the intercept (a‘) and slope (b’) of a log-linear model, ln ψ = a‘ + b’ θ, are also linearly related and an alternate form of the one-parameter model (LLM) can be derived, ln ψ, = p‘ + b’ (θ + q‘), which uses AWC directly. The errors with this model are comparable to GHM. Unfortunately, LLM requires individual soil p’ and q‘ values and, because of more scatter in the intercept - slope relationship, pooled p’ and q‘ values for a group of soils are not as effective in LLM as they are in GHM.