Crop Yield Response to Fertilizer Relative to Soil Properties in Sub-Saharan Africa

Crop productivity in sub-Saharan Africa commonly encounters numerous unmitigated constraints that affect crop response to fertilizer use and relationships of soil test properties with nutrient responses. A dataset from 1265 blocks of 267 replicated trials conducted across 11 countries was analyzed to determine relationships of soil test properties with crop yield responses to applied nutrients. The dataset included cassava (Manihot esculenta L. Crantz) and five upland cereal and five grain legume crops with responses to N, P, K and a diagnostic treatment of Mg-S-Zn-B. Nitrogen was excluded for the pulses. Soil organic C (OC) was <10 g kg(-1) for 55% of the blocks. Mehlich-3 P was <12 mg kg(-1) for 42% of the blocks. Soil clay and OC contents accounted for 2 to 99% of the variation in soil properties. Yield with no fertilizer applied (Yield0) was most frequently related to OC and silt content. Combinations of soil test properties accounted for 0 to 65% of the variation in Yield0, for 0 to 45% of the variation in response to N, less for other nutrients, and least for P. Soil tests for availability of a particular nutrient were not related to response to that nutrient. Yield0 was positively related to yield responses to applied N. Fertilizer use decisions in sub-Saharan Africa need to be primarily based on robust crop-nutrient response functions relevant for a recommendation domain and avoidance of fertilizer use for non-responsive situations. Soil test results are not likely to be very useful for fine-tuning fertilizer use for fertilizer responsive situations.