Evaluating maize yield response to fertilizer and soil in Mexico using ground and satellite approaches

Abstract

Crop management recommendations that are tailored to local conditions offer promise for improved farmer livelihoods, input use efficiency, and crop output. An important step towards achieving localized recommendations is to establish repeatable, low-cost approaches to diagnosing nutrient constraints. Satellite remote sensing is particularly attractive in this context for its ability to monitor crop conditions and provide yield estimates at much lower expenses than extensive field measurements. Here we evaluate nutrient constraints for maize yields in central Mexico by utilizing an extensive field dataset on soil properties, fertilizer practices, and yields that were collected as part of the MasAgro program. We also compare insights from satellite-based yield estimates to those derived from crop-cut yields. Overall we find strong positive associations between maize yields and fertilizer inputs for N, P, and K in farms at lower elevations, characterized by larger, commercial farms. In higher elevations where farms are typically smaller and use less responsive varieties, we still find significant but slightly less positive associations with N and P. In both regions, we observe significant negative associations between yield and soil pH, consistent with an unmet need for micronutrient fertilizers in the region. Satellite-based yields were able to reproduce the positive associations and large yield response to fertilizer N and P, but were not sensitive enough to detect significant effects of secondary factors, namely fertilizer K and soil pH. We conclude by suggesting specific interventions that could be experimentally tested in the region.