STUDY OF DIFFERENCES BETWEEN VERTICAL ROOT MAPS OBSERVED IN A MAIZE CROP AND SIMULATED MAPS OBTAINED USING A MODEL FOR THE 3-DIMENSIONAL ARCHITECTURE OF THE ROOT-SYSTEM

Differences between observed and simulated vertical root maps were stu died in an attempt to evaluate the predictive ability of a simulation model of root system architecture under held conditions on mature plan ts, and to identify avenues for improvement. Some methodological probl ems associated with root mapping in the held are considered with a sen sitivity analysis. Comparisons were made on a maize crop (early maturi ng hybrid Fl cultivar 'Dea') 15 days after silking. Four vertical root maps, perpendicular to the row and midway between two successive plan ts, were observed. Simulated root maps for different locations along t he row showed essentially the same pattern, attesting of an approximat ely two-dimensional distribution of the roots in such a crop. Simulati on of the intersection of roots with thin layers (thickness from 0 to 20 mm) instead of a perfect plane allowed us to assess effects due to the roughness of actual trench walls, and possible artefacts in the ob servation of root intersections. The simulated root profiles were very sensitive to this thickness, especially in the 0-5 mm range, in both average values, and overall shape. Actual data were close to the 3 mm thick simulations. This value seems plausible under our field conditio ns. Differences between simulated and actual root maps were shown to b e mostly accounted for by the variations in soil bulk density. Thus, t his environmental parameter appears as the most important one to inclu de into the model for improving its predictions.