How important is the description of soil unsaturated hydraulic conductivity values for simulating soil saturation level, drainage and pasture yield?

Accurate simulation of soil water dynamics is a key factor when using agricultural models for guiding management decisions. However, the determination of soil hydraulic properties, especially unsaturated hydraulic conductivity, is challenging and measured data are scarce. We investigated the use of APSIM (Agricultural Production Simulation Model) with SWIM3 as the water module, based on Richards equation and a bimodal pore system, to determine likely ranges of the hydraulic conductivity at field capacity (K-10; assumed at a matric potential of -10 kPa) for soils representing different drainage characteristics. Hydraulic conductivity measurements of soils with contrasting soil drainage characteristics and values for K-10 were extracted from New Zealand's national soil database. The K-10 values were then varied in a sensitivity analysis from 0.02 to 5 mm d(-1) for well-drained soils, from 0.02 to 1 mm d(-1) for moderately well-drained soils, and from 0.008 to 0.25 mm d(-1) for poorly drained soils. The value of K-10 had a large effect on the time it took for the soil to drain from saturation to field capacity. In contrast, the saturated hydraulic conductivity value had little effect. Simulations were then run over 20 years using two climatic conditions, either a general climate station for all seven different soils, or site-specific climate stations. Two values for K-10 were used, either the APSIM default value, or the soil-specific measured K-10. The monthly average soil saturation level simulated with the latter has a better correspondence with the morphology of the seven soils. Finally, the effect of K-10 on drainage and pasture yield was investigated. Total annual drainage was only slightly affected by the choice of K-10, but pasture yield varied substantially.