Tree root architecture and competition for water in an agroforestry system in semi arid Kenya
This study involved the combination of the competitive indices and fractal methods of root system study with soil moisture and sap-flow monitoring to determine whether "shortcut" methods of root system assessment could be used to predict competition with crops and complementarity of below ground resource sharing in agroforestry systems in semiarid Kenya. Grevillea robusta, Gliricidia sepium, Melia volkensii and Senna spectabilis aged between two and four years were grown in simultaneous linear agroforestry plots with maize and beans as the crop species. The fractal index was compared with other known indicators of competition, i.e. volumetric soil moisture content and crop yield. For all the trees studied, there was no significant relationship (P>0.05) between the ratio of the root diameter squared before branching and the total root diameter squared after branching (the parameter a), the quotient of the maximum diameter squared after branching and total diameter squared after branching (the allocation parameter Q) and the link length (length of the root axis between successive branching points). This indicated that the fractal rules held valid for these species. In all these species, α = I for roots greater than 1mm. However, for smaller roots, a was normally less than I. These findings indicated that the fractal methods were valid for coarse roots but seriously underestimated the length of fine roots.