Measuring and modelling yield and water budget components of wheat crops in coarse-textured soils with chemical constraints

Subsoil chemical constraints, including varying combinations of salinity, alkalinity, and high concentrations of boron and sodium, are widespread in southern Australia. We combined field and simulation experiments to: (a) quantify the effect of subsoil chemical constraints on the lower limit of plant available water in sandy regolithic, hypocalcic, calcarosols, (b) assess the ability of current modelling approaches to account for the effects of chemically inhospitable soils on crop yield and water budget, and (c) investigate the interactions between subsoil constraints and rainfall. Lower limits were derived from measurements of soil water content using frequency domain reflectometry in an experiment combining three rotations, two locations, and three seasons (Experiment 1), and from gravimetric measurements in fanner-managed crops during three to four seasons (Experiment 2). A two-step method was used accounting for the effect of clay content on lower limits, and analysing the residuals against likely constraints. Significant, positive associations between residuals and chemical constraints (Na, B, carbonates, pH) indicated lower limits greater than expected from texture, and less plant available water with increasing intensity of constraints. These empirical functions can be used in modelling, and in estimating target yields for management purposes. Yield and soil water content simulated with the APSIM model were compared with measured wheat yield (range: 0.6-3.7 t ha(-1)), and measured soil water content in wheat and canola crops (range: 0.0 10-0.205 cm(3) cm(-3)). These comparisons indicated that in soils with clay content less than or equal to31%, electrical conductivity (EC) less than or equal to 1.31 dS m(-1), pH(water) less than or equal to10.3, concentration of CaCO3 less than or equal to292 g kg(-1), B less than or equal to29 mg kg(-1), and Na less than or equal to2450 mg kg(-1), crop responses to subsoil constraints can be largely accounted for by higher lower limits and reduced crop transpiration. Higher B concentration would require accounting for reduced transpiration efficiency. Interactions between subsoil constraints and seasonal rainfall (April-October) were analysed in a simulation experiment including the factorial combination of 39 locations, 44 seasons and two soil profiles (with or without chemical constraints). Three types of site-dependent responses were found. Yield reduction caused by subsoil constraints was independent of seasonal rainfall in 18 locations (neutral response), was more severe in wetter seasons in eight locations (negative response), and decreased with increasing rainfall in 13 locations (positive response). Negative responses were more frequent in sites where conditions contributed to severe water deficits, i.e. lower seasonal rainfall, less available soil water at sowing, and greater evaporative demand. Estimates based on the assumption of a flat landscape with good soil infiltration leading to little runoff, showed that increase in deep drainage accounted for most of the reduction in crop transpiration associated with subsoil constraints. For a given amount of seasonal rainfall, the likelihood of a drainage event and its average size both increased with increasing frequency of large rainfall events, as characterised using power laws. (C) 2003 Elsevier B.V. All rights reserved.