Spectral soil analysis for fertilizer recommendations by coupling with QUEFTS for maize in East Africa: A sensitivity analysis

Laboratory analysis of soil properties is prohibitively expensive and difficult to scale across the soils in sub-Saharan Africa. This results a lack soil-specific fertilizer recommendations, where recommendation can only be provided at regional scale. study aims assess feasibility using spectral provide recommendations. Using range spectrometers [NeoSpectra Saucer (NIR), FieldSpec 4 (vis-NIR) with contact probe or mug light interface, FTIR Bruker Tensor 27 (MIR)], 346 archived samples (0–20 cm) known chemical collected from Ethiopia, Kenya Tanzania were scanned. Partial least square regression (PLSR) was used develop prediction models for selected including pH, organic carbon (SOC), total nitrogen, Olsen P, exchangeable K. These predicted properties, associated uncertainty, derive recommendations maize Quantitative Evaluation Fertility Tropical Soils (QUEFTS) model parameters Most (pH, SOC, K) well (Concordance Correlation Coefficient values between 0.88 0.96 Ratio Performance Interquartile 1.4 5.9) by all but there performance variations spectrometers. Use data development gave promising when compared obtained conventional analysis. For example, performing NeoSpectra over/under-estimated up 8 24 kg ha-1N respectively, though insignificant variation estimation P among We conclude that technology determine major satisfactory precision, sufficient specific decision making East Africa, possibly even portable equipment field.