The efficiency with which applied resources are utilized in sub-Saharan African cropping systems is especially critical as the resources are generally scarce. Research efforts to improve farm productivity increasingly focus on resource interactions and trade-offs operating at farm-scale. Farm-scale models that integrate summary models of the various subsystems (crops, livestock, household) are proposed to analyse the complexity of management systems. NUANCES-FIELD is a summary model of the crop/soil system that calculates seasonal crop production based on resource availability, capture and utilization efficiencies. A detailed mechanistic crop growth model, APSIM, was used to generate parameters and variables that can be introduced as descriptive functions in NUANCES-FIELD. To such end, we first parameterized and tested APSIM based on several field experiments carried out on different soil types in western Kenya farms where nitrogen and/or phosphorus were applied. The model was further configured to generate nitrogen and phosphorus response curves as a function of soil condition (carbon content, clay content, phosphorus-sorption characteristics) and the effects of alternative weed management scenarios in relation to labour availability. Nitrogen, phosphorus and rainfall capture efficiencies ranged between 0.22-0.85 kg kg(-1), 0.05-0.29 kg kg(-1) and 0.10-0.53 mm mm(-1), respectively, depending on soil nutrient and physical conditions. Variation in the integrated seasonal fraction of radiation intercepted (intFRINT) with plant density was adequately described by the function y = 0.058x + 0.11 within a range of 1.5-5.5 maize plants per m(2). Investigation of weed management using the APSIM model identified a weed-free period of at least five weeks from maize emergence for minimum yield loss from weed-crop competition. The simulation exercises confirmed that resource-use efficiencies sharply decrease on moving from relatively fertile fields 'close' to the homestead towards degraded 'remote' fields within the same farm, giving impetus to expedite the search for better targeted management strategies for spatially-heterogeneous farms