Simulation of wheat growth using the 3D root architecture model SPACSYS: Validation and sensitivity analysis

SPACSYS, a model of C and N cycling in the soil–plant–atmosphere continuum that incorporates a detailed three-dimensional root growth sub-model, was tested for its ability to predict the growth and root length density (RLD) distribution of winter wheat. The root growth sub-model was parameterized using published or unpublished data for wheat and barley, and validated against an independent data-set from field experiments on wheat in the UK published in 1978. The model reproduced the experimental results well, accurately simulating total above-ground (r = 0.98) and root (r = 0.96) biomass. There was a slight tendency to over estimate root biomass before the start of stem extension and the rate of root loss after anthesis, suggesting that the partitioning coefficient for dry matter between shoot and root system overwinter and the rate of root mortality and the root decay constant after anthesis may require some adjustment. Simulations of rooting depth and RLD distribution over the season were also acceptable. Further simulation work investigated the sensitivity of root length and relative distribution in the soil profile to changes in specific root morphological traits. The total length of the root system was sensitive to changes in traits regulating the number of lateral branches a given root order can produce (the inter-branch distance, the maximum length of root and its apical non-branching distance) and the elongation rate of main axes. The distribution of the root system was most sensitive to the maximum elongation rate of main axes and their initial growth trajectory. The model will have applications in guiding the design of root system ideotypes for improved water and nutrient use efficiency and for investigating their effectiveness under a range of soil conditions and crop management regimes.

Research highlights
► SPACSYS is a crop simulation model with 3D root architecture.
► Validation against field data for wheat showed it accurately predicted above-ground and root biomass.
► Simulations of rooting depth and root length distribution over the season were also acceptable.
► Total root length was sensitive to changes in traits regulating the number of lateral roots.
► Root distribution was most sensitive to the maximum extension rate of main roots and their growth trajectory.