Plant productivity is directly affected by the capacity of the root system to forage for soil resources.An enhanced understanding of root-soil interactions provides the potential to improve crop performance in specific soil environments.Interactions between roots and soil are, however, complex.The root-soil environment is heterogeneous and difficult to visualise and measure, root architecture and root growth responses are complex and dynamic, and processes from the ionic and rhizosphere scale right up to the whole crop and even catchment scale are involved.For these reasons, pot experiments are used in root studies to simplify the environment, target specific interactions and aid with visualisation and measurement.Significant challenges exist, however, in relating pot studies to the field, requiring upscaling from a spatially confined and artificially contrived environment to the reality of a more complex cropping environment.Simulation models provide an opportunity to upscale complex root-soil interactions from the pot to the field, but to do so they must represent the way that plant roots explore a restricted pot environment.In this study ROOTMAP, a 3D functional-structural model of root growth and resource capture, was modified to enable the simulation of barriers in soil, and the interaction of plant roots and soil water and nutrients with those barriers.This barrier-modelling utilises custom coding, with the support of Boost.Geometry (Generic Geometry Library) where appropriate.The barrier approach defines the 3D shape and location of any number of what are termed Volume Objects.Roots and soil can be: wholly contained within one Volume Object such as in the case of roots growing in a pot; a plant can have roots distributed between two Volume Objects such as in a split-pot experiment; and they can be wholly outside one or more Volume Objects for simulating the presence of rocks or other hard objects in soil.Volume Objects can be wholly impermeable, such as; pot walls that contain roots within them, or impermeable rocks or hardpan layers that roots grow around.Volume Objects can also have varying degrees of permeability for representing layers or areas in soil that have varying degrees of hardness and varying root penetrability.In this initial version of the code, barriers or objects can be represented as rectangular prisms, giving flat barrier layers or square or rectangular objects such as root/rhizo boxes, or as cylinders, representing curved pots or smooth curved objects in soil.The barrier modelling code calculates the deflection of a root tip when it intersects a boundary, representing the way that plant roots grow around and along object surfaces.It also calculates the effect of semi-permeable objects in soil on root growth into and around those objects.Water and nutrients are distributed through the soil environment by use of a variable 3D grid of sub-volumes or cells.The water and nutrient routines then search for the presence of a barrier or wall (Volume Object) intersecting each cell and the volume of the cell contained inside/outside the barrier is calculated.This combined with the permeability of the barrier determines the water and nutrient transfer within the cell.The result is a model which can simulate the root, water and nutrient dynamics in a bounded-environment.This provides an opportunity to represent root architectural development and root-soil interactions in pots and rhizo-boxes, and investigate how these studies relate to root growth and resource capture in un-bounded field soil.

Controlling rain-splashed crop diseases is an extremely difficult task. Their spread is a complicated process and large-scale field surveys to determine the extent of an incursion over a large area are often economically intractable. A failed attempt at control or eradication of a pathogen can be very costly. In 1996 there was a major incursion of lupin anthracnose in Western Australia, which crippled the albus lupin industry. At the time of the outbreak a wide-spread survey was undertaken to estimate the extent of the incursion. A containment protocol involving broad-scale crop destruction was put into place with the view of eradicating the disease. This eradication attempt subsequently failed due to wild lupins acting as a reservoir for the disease from road verges and non-arable land outside the cropping area. There was also evidence of long distance dispersal vectors such as native budworm. Had all the relevant information related to spread and spatial habitat suitability been collected and taken into account, the decision to destroy the crops may not have been made and significant economic losses to growers may have been avoided. Estimates of the current extent of an incursion based solely on incomplete empirical data are likely to be inaccurate, as are predictions of the future trajectory of an incursion that do not take into account all available information. Therefore any control or eradication attempt based on these estimates and predictions may be ineffective.Simulation modelling is an important method for making the best use of all available empirical data and integrating all available knowledge to predict the spread of rain-splashed crop diseases. With this prediction, an evaluation of the potential success of control or eradication measures may be estimated. This study describes a model that was built to simulate a situation analogous to that of the 1996 lupin anthracnose incursion in Western Australia, for the purpose of identifying general indicators of the eradicability of rain-splashed crop diseases.We extended the spatiotemporal model AnthracnoseTracer to simulate the spread of lupin anthracnose in a heterogeneous paddock environment analogous to the 1996 conditions. Three control methods aimed at eradication were investigated. A simple detection model was assumed, where the probability of detecting the disease is dependent on the level of passive surveillance and the detectability of the disease.As part of the preliminary analysis contained in this paper we investigated two scenarios to identify potential indicators of eradicability, based on the time taken to detect the disease. Our preliminary results indicate that rain-splashed pathogens are extremely difficult to eradicate and the chance of successful eradication appears strongly dependent on the level of surveillance of the susceptible areas and the detectability of the disease. The level of surveillance and detectability of the disease may both serve as general indicators of eradicability for rain-splashed crop diseases. We discuss further modelling analyses to be carried out to refine these indicators.

Root systems and their interactions with the below-ground environment are difficult to study. As a result, root research has typically lagged behind that carried out on above-ground parts of crop plants. Modern advances in computing technology are allowing 3D root architectural models to play a role in below-ground investigations. ROOTMAP is one such model, and has been used to investigate the influence that lupin root systems can have on water and nitrate distributions and nitrate leaching in field soils. Simulating a wide range of root architectures, ROOTMAP predicted that to reduce nitrate leaching to depth below lupin crops, a trade-off between surface and subsoil rooting density is required, with high rooting density in the topsoil alone insufficient to minimise leaching. Simulated plants that rapidly established a relatively high density of roots in the topsoil, reduced total nitrate leached with the break of season rains. Further increases in topsoil root density were, however, a poor investment of internal assimilates. A shift in resource allocation to subsoil root growth as the season progressed, gave plants a second chance to acquire nitrate previously leached to depth, and proved the most successful strategy for reducing total annual losses of nitrate. Given the ability to investigate rooting traits and quantify potential benefits of one rooting form over another, root modelling can be used to identify desirable root traits for which genotypes can then be screened. ROOTMAP is currently being used to investigate below-ground competition between wheat and annual ryegrass for both mobile (eg. water, nitrate) and immobile (eg. phosphorus) soil resources. This work will aid in the identification of key rooting traits that confer a competitive advantage to wheat.

Little is known about root architectural attributes that aid the capture of nitrate from coarse-textured soil profiles of high leaching potential. In this study, a range of root architectures from the herringbone to the dichotomous structure were simulated, and their capacity to take up nitrate leaching through a sandy profile was recorded. All root systems had equal total volume at each point in time, and so were considered cost equivalent. These simulations showed that the root architecture likely to maximize nitrate capture from sandy soils (under the Mediterranean rainfall pattern experienced in Western Australia) is one that quickly produces a high density of roots in the top-soil early in the season, thereby reducing total nitrate leached with opening season rains, but also has vigorous taproot growth, enabling access to deep-stored water and leached nitrate later in the season. This is the first published, spatially explicit attempt to assess the ability of different root architectures equivalent in cost, to capture nitrate from a spatially and temporally heterogeneous soil environment.

While it is widely accepted that both root system function and architectural form influence plant productivity, how they do this is still unclear. It has been suggested that large-scale foraging strategies may be (i) more efficient at capturing mobile ions, and (ii) restricted to resource-poor environments; in contrast, finer-scale foraging strategies that utilise root plasticity to exploit local soil volumes may be (i) more efficient at capturing immobile ions, and (ii) are typically found in relatively resource-rich habitats. Optimality of root function cannot, however, be so easily defined, with experimental data both supporting and refuting the above theory. Optimality is likely to vary with the type of nutrient and the level of heterogeneity in temporal and spatial supply. Also, since root systems have evolved as components of a whole plant-soil system, root form may be designed to optimise whole plant efficiency, particularly under competitive conditions, and not the efficiency of an isolated root system. Modem advances in root modelling techniques are helping to challenge previous hypotheses about the interaction between root architecture and nutrient capture. By incorporating some of the features of real, responsive root systems, root architectural modelling now suggests that foraging strategy may be largely determined by variability of soil nutrient supply in time or space rather than by the diffusive mobility of the limiting ion.

Uptake rates calculated from plants uniformly supplied with a nutrient will underestimate uptake under heterogeneous conditions. A split-root nutrient solution experiment was set up to compare the uptake rate of 2 lupin species (Lupinus angustifolius L., L. pilosus Murr.) under conditions of uniform and heterogeneous nitrate supply. Nitrate was supplied uniformly to the root system at 250 M (low), 750 M (high), or 1500 M (high), or in a split low/high or high/low combination between the upper and lower root system. While L. pilosus had a greater total nitrate uptake over the treatment period due to a higher total root length, L. angustifolius had 1.5–2.5 times greater nitrate uptake rate per unit of root length. L. angustifolius also had the capacity to increase the nitrate uptake rate in sections of the root system supplied locally with high nitrate, compared with a root system uniformly supplied with high nitrate. This increased uptake rate under heterogeneous supply enabled the plant to take up 74–94% of the total nitrate taken up when uniformly supplied with high nitrate, while only 58–72% would have been taken up without such a compensation mechanism. L. pilosus did not show this response. The difference between the response of these 2 species suggests that a range of nitrate uptake responses may exist across the lupin germplasm, and that it may be possible to select a lupin species with an enhanced ability to capture nitrate from the profile, thus decreasing nitrate losses from leaching.

Little is known about the ability of legume root systems to respond to the heterogeneous supply of nitrate. A split-root nutrient solution experiment was set up to compare the root growth response of 2 lupin species, Lupinus angustifolius L. (dominant tap root and primary lateral system) and L. pilosus Murr. (minor tap root and well-developed lateral root system), to differentially supplied nitrate. These 2 species represent the extremes of the root morphology types present across the lupin germplasm. Nutrient solution containing low (250 M) or high (750 M) nitrate was supplied either uniformly, or split (high and low) between the upper and lower root system. The average growth rate and total root length of L. pilosus was 1.7 times that of L. angustifolius. For both species, the increased proliferation of roots in a high nitrate zone was accompanied by a decrease in root growth in the low nitrate zone, giving approximately the same total growth as the uniform low nitrate treatment. This correlative growth rate response was 15% larger for the first-order branches of L. pilosus than L. angustifolius. While few second-order branches grew for L. angustifolius, the second-order laterals of L. pilosus showed a 2-fold correlative root growth and branching response to the split treatments, with no difference in growth between the uniform high and low nitrate treatments. The second-order laterals thus proliferated in response to the differential supply of nitrate and not the absolute concentration. While the growth rate and branching of the second-order laterals of L. pilosus exhibited a typical correlative response, first-order branching was inhibited in all split treatments, regardless of whether the roots were in the high or low nitrate zone. This response was not seen in L. angustifolius. The difference in the root growth response of the 2 root system types to differentially supplied nitrate suggests a potential in the lupin germplasm for developing a line capable of greater nitrate capture from the soil profile.