A general tree-environment-crop interaction equation for predictive understanding of agroforestry systems

Agroforestry is a young science in need of theories and methods providing p redictive understanding. The present study is an attempt to contribute to t he development of a general agroforestry theory and accessory methods. A tr ee-environment-crop interaction equation is derived in which the overall to e effect on crop production is explained as a balance of (positive and nega tive) relative net tree effects on resource availability to the crop. Ln th is balance, each relative net tree effect receives a weight that is equal t o the degree of limitation of the resource in the specific environment. The equation thus makes explicit that the balance of available resources in th e environment is a major factor determining the biophysical outcome of an a groforestry technology. The equation enables the formulation of two rules t hat can be viewed as the agroforestry counterparts of classic crop producti on laws. The agroforestry counterpart of the law of limiting factors states : the more a resource becomes available in the tree-crop environment, the s maller its share becomes in the overall tree-environment-crop interaction. The agroforestry counterpart of the law of the optimum states: the more oth er limiting resources become available in the tree-crop environment, the gr eater the share of a resource becomes in the overall tree-environment-crop interaction. These rules help with extrapolation, without requiring full qu antification of the tree effect balance. They allow the development of a me thod for the analysis and synthesis of agroforestry experiments. The method is illustrated by analysing some published alley cropping experiments. Thi s analysis indicates that for alley cropping systems: (1) the hypothesis th at trees capture nutrients below the crop rooted zone resulting in a net be nefit for the crop is probably true for nitrogen and not true for phosphoru s, and that (2) competition for water by trees probably exceeds their water conservation effects. Suggestions for experimentation are given and the ci rcumstances under which yield benefits of agroforestry systems can be expec ted are discussed. (C) 2000 Elsevier Science B.V. All rights reserved.