Modelling trees' effect on maize in the Grevillea Robusta + maize system in Central Kenya

In the Central Highlands on Kenya, Grevillearobusta (A. Cunn.) trees are commonly planted on maize (Zeamays L.) fields. This type of production system covers about 750 000ha in the highland area around Mt Kenya. Knowing the influence oftrees on maize yield would help to regulate the density and size distribution ofthe tree cover. The spatial distribution of trees varies from rather uniform toextremely aggregated patterns, calling for a spatial modelling approach. Thisstudy employed three different spatial competition index types to model theeffect of trees on the maize yield at a given location. The maize yield wasexpressed as a function of competition index, which was calculated from thediameters and distances of trees. The data were collected from 14 maize fields,and included measurements on 857 trees and 2514 one-m² maize plots. The models indicated that high competition by treesdecreases maize yields considerably, but at the field level the decrease issmall with normal tree stockings (about 200 trees per hectare). Characteristicto the models was low degree of explained variance, which was partly due to highsampling error in maize yield measurement and low variation in competition(places with high competition by trees were seldom planted with maize). One ofthe maize yield models was combined with an earlier simulation program forG. robusta. The simulator allows one to predict thetemporal development of any G. robusta – maize fieldin Central Kenya. This revised version was published online in June 2006 with corrections to the Cover Date.