A remote sensing approach to understand controls on cropland productivity (Mexico)

Crop productivity, or yield, is a critical trait of agricultural systems. The response of yield to changes in management or environmental factors, however, is often understood only at the small scales at which experiments and field observations are feasible. Growing demand for food, rapid environmental changes, diminished rates of yield progress, and concerns about environmental impacts of agriculture will create a tremendous need for improved agricultural management and policies over the next century. These advances will in turn rely on an understanding of cropland responses at broad scales, such as for regions comprised of thousands of individual fields.; In this project, a new approach to understanding broad scale responses of crop productivity was developed that utilizes remote sensing data acquired by the Landsat satellite sensors. The study focused primarily on the intensive wheat systems of the Yaqui Valley in Sonora, Mexico, although the methodology was intended to be flexible and readily applied to other regions. The project consisted of three main components: (1) techniques to map crop yields using Landsat imagery were developed and tested; (2) these maps were combined with geographic datasets on soil, management, and climate variables to empirically assess the relative importance of various yield controls at the regional scale; (3) the relationship between yield patterns (both in space and time) and underlying controls on productivity were analyzed to facilitate interpretation of yield maps in situations where ancillary datasets are sparse. In the Yaqui Valley, yield variability was found to be predominantly driven by management variables that interact strongly with growing season weather conditions. These results suggest that efforts to improve soil quality or address a single management variable are likely to have limited impact on broad-scale productivity, while seasonal forecasts would have a significant positive effect on yields in this region. The approach developed in this project represents a relatively inexpensive and rapid tool with which the most significant constraints to crop productivity at the regional scale can be uncovered, and the relative impact of potential investment and policy strategies assessed.