Unusual fine root distributions of two deciduous tree species in southern France: what consequences for modelling of tree root dynamics?
Résumé
The spatial distribution of fine roots of two deciduous tree species was investigated in contrasting growing conditions in southern France. Hybrid walnut trees (Juglans regia×nigra cv. NG23) and hybrid poplars (Populus euramericana cv. I214) were both cultivated with or without annual winter intercrops for 10 years on deep alluvial soils. Soil samples for measuring the fine root distribution of both trees and crops were obtained by soil coring down to 3-m depth at several distances and orientations from the tree trunk. The distribution of live fine roots from walnut and poplar trees was patchy and sometimes unexpected. In the tree-only stands, fine root profiles followed the expected pattern, as fine root density decreased with increasing depth and distance from the tree trunk. However, many fine root profiles under intercropped trees were uniform with depth, and some inverse profiles were observed. These distributions may result from a high degree of plasticity of tree root systems to sense and adapt to fluctuating and heterogeneous soil conditions. The distortion of the tree root system was more pronounced for the walnut trees that only partially explored the soil volume: in the tree-only stand, the walnut rooting pattern was very superficial, but in the intercropped stand walnut trees developed a deep and dense fine root network below the crop rooting zone. The larger poplars explored the whole available soil volume, but the intercrop significantly displaced the root density from the topsoil to layers below 1 m depth. Most tree root growth models assume a decreasing fine root density with depth and distance from the tree stem. These models would not predict correctly tree–tree and tree–understorey competition for water and nutrients in 3D heterogeneous soil conditions that prevail under low-density tree stands. To account for the integrated response of tree root systems to such transient gradients in soils, we need a dynamic model that would allow for both genotypic plasticity and transient environmental local soil conditions