Comparing observed and simulated water storage during drainage to select hydraulic parameters for volcanic soils

Abstract

In soil-water modelling, hydraulic parameters must be chosen carefully to simulate water fluxes accurately. This study aimed to parameterise the Soil Water Infiiltration and Movement (SWIM) model for sandy volcanic soils that are irrigated with wastewater from Rotorua, New Zealand. Three plots 1·5 by 1·5 m were dyked, flooded until soils approached saturation, and then covered to prevent evaporation. Water-content measurements were taken to 1 m depth during 20 days of drainage. For each profile, this drainage process was simulated using 9 combinations of hydraulic conductivity [K(θ)] and soil water desorption θ(ψ) parameters. The θ(ψ) parameters were fitted with different ranges of desorption data, using Campbell’s approach. K(θ) parameters were obtained from flow measurements using soil cores or disc permeameters, through use of Campbell’s method, and by applying a unit-gradient analysis to the drainage data. Model output was compared with measured water contents graphically and statistically. Several combinations of parameters provided good agreement between simulated and observed data, but no single combination gave the best prediction for all 3 profiles. Campbell’s approach to determine the slope of K(θ) function provided excellent agreement for 2 of the profiles. In the third profile, a nearly steady water content during drainage at 1 m depth was closely simulated using parameters from the unit gradient approach. Total measured drainage ranged between 8·6 and 15·9 cm amongst the 3 profiles. Results provide 3 parameterised profiles for modelling water movement in these soils, which receive >4000 mm of rainfall and irrigation annually.