Agricultural and Forest Meteorology 114 (2002) 31–43 Soil, climate, and management impacts on regional wheat productivity in Mexico from remote sensing David B. Lobell a,b,∗ , J. Ivan Ortiz-Monasterio c , C. Lee Addams b , Gregory P. Asner a,b a Department of Global Ecology, Carnegie Institution of Washington, Stanford, CA 94305, USA b Department of Geological and Environmental Science, Stanford University, Stanford, CA 94305, USA c International Maize and Wheat Improvement Center (CIMMYT), Wheat Program, Apdo. Postal 6-641, 06600 Mexico D.F., Mexico Received 12 July 2002; received in revised form 11 August 2002; accepted 14 August 2002 Abstract Understanding sources of variability in net primary productivity is critical for projecting ecosystem responses to global change, as well as for improving management in agricultural systems. However, the processes controlling productivity cannot be fully addressed with field- or global-scale observations. In this study, we performed a regional observational experiment using remote sensing to analyze sources of yield variability in an irrigated wheat system in Northwest Mexico. Four different soil types and 3 years with contrasting weather served as the two main experimental factors, while remotely sensed yields provided thousands of observations within each treatment. Analysis of variance revealed that 6.6 and 4.6% of the variability in yields could be explained by soil type and climate, respectively, with a negligible fraction explained by soil-type–climate interactions. The majority of the variability in yields (88.6%) was observed within treatments and was attributed mainly to variations in management. The impacts of management were observed to depend significantly on both soil type and climate, as revealed by distributions of yields within each treatment. The results indicate that changes in management will have the greatest impact on regional production, and will also play a large role in determining the impact of any changes in climate or soil. This work also demonstrates the use of consistent remote sensing estimates to perform regional studies unfeasible with field-based approaches. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Carbon cycle; Climate impacts; Remote sensing; Wheat; Yield; Yield loss 1. Introduction Spatial and temporal variations in the productivity of terrestrial vegetation can have profound impacts on humans and their environment, with effects ranging from atmospheric chemistry and climate to global food ∗ Corresponding author. Present address: Department of Global Ecology, Carnegie Institution of Washington, Stanford, CA 94305, USA. Tel.: +1-650-325-1521; fax: +1-650-325-6857. E-mail address: dlobell@globalecology.stanford.edu (D.B. Lobell). production (Pielke et al., 1998; Schimel et al., 2001). This variability arises from changes in numerous fac- tors, including soil physical and chemical properties, temperature, precipitation, solar radiation, and human management. In an effort to project the impact of fu- ture changes in these controlling factors, it is critical to understand the relationship between productivity and each factor, along with their interactions. Interactions between human activity and envi- ronmental controls are particularly relevant in man- aged ecosystems such as agriculture. For example, 0168-1923/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII:S0168-1923(02)00138-7