Drivers of sorghum response to fertilizer microdosing on smallholder farms across Burkina Faso

Fertilizer microdosing (FM) is being disseminated to resource-limited smallholder farmers within the framework of sustainable intensification in order to boost crop productivity. However, scaling up of this technology revealed that yield response variability may be a barrier to its adoption. A total of 351 on-farm trials across three agro-climatic zones (Sissili: 800-1000 mm, Oubritenga: 600-800 mm and Zandoma: ≤ 600 mm mean annual rainfall) over two years in Burkina Faso were therefore conducted in order to identify the factors driving sorghum (Sorghum bicolor (L.) Moench) yield response to FM (2 g/hill of N-P-K 14-23-14 and 1g/hill of urea). The trials were stratified to consider soil types (indigenous classification) as well as the distance between homesteads and fields as a criterion of fertility. Additional variables included previous crop management practices and weed pressure. On average, FM application increased yields by 385, 571 and 388 kg/ha in Sissili, Oubritenga and Zandoma, respectively. The proportion of fields with a value-cost ratio (VCR) ≥ 2 was 34% in Sissili, 56% in Oubritenga, and 30% in Zandoma. In the three zones, yield response to FM tended to increase with increasing soil depth and plot duration of cultivation, as well as with the proper timing of weeding and fertilizer application. Within Sissili lower responses were observed on stony-gravelly and sandy soils compared to black soils, and also on plots previously amended with high amounts of fertilizer compared to soils with moderate or no previous fertilizer application. For plots with no previous fertilizer application, better responses were observed on plots with legumes as antecedent crop. In Oubritenga, responses were greater on clayey soils compared to other soil types. In Zandoma, yield response was greatest in lower floodplain soils compared to other soils, and increased with increasing delay in sowing. Overall, crop response to FM was affected by (i) soil type and topographic position, (ii) plot history through the duration of cultivation, the antecedent crop and the previous fertilization rate, and (iii) plot management. The relative importance of these factors was site-specific and can be easily communicated to farmers.