Beyond Agroecological Delineation: A Functional Understanding of Productivity Drivers Across Us Row Crop Agroecosystems

Abstract

CONTEXTAgricultural systems are typical examples of how diverse biophysical (climate, soil, topography) and socioeconomic components interact and collectively drive different production outcomes. However, previous approaches in defining agricultural ecosystems often overlooked the interactive role of the complex variability of these components, as well as their contribution to productivity across agroecological landscapes.OBJECTIVEHence, the overall objective of this study was to exploit the spatial heterogeneity of these biophysical and socioeconomic components to generate distinct typologies by harmonizing variability across factors and understand their critical role in determining productivity outcomes in agroecosystems.METHODSWe delineated such typologies for the US agricultural landscape and applied them to four major row crop systems (i.e., corn, soybean, winter wheat, and cotton) using high-dimensional statistical modeling. We first use hierarchical clustering to create climate, soil, and topographic typologies across the U.S. landscape and spatially integrate them to define agroecosystem typology (AET). State typology was then added to represent socioeconomic aspects. County-level long-term (1991-2020) crop productivity data was disaggregated into typologies using remotely sensed net primary productivity. A dummy variable regression model was applied with crop productivity as the dependent variable and typologies as independent dummies. The model was estimated via linear dummy least squares. Separate regression models were developed for each crop, tested for robustness, and used to quantify the productivity contributions of each typology across space.RESULTS AND CONCLUSIONSWe found that AET drives the most heterogeneity across row crop systems thereby controls greater proportion of yield variability, with 40–53% contributions across the studied crops. Furthermore, the influence of agroecological heterogeneity on crop productivity does not scale proportionally with the spatial extent of cultivated area. The contribution of these typologies to crop productivity results in a diverse range of interactions that reflect crop-specific productivity challenges and opportunities within U.S. agroecosystems.SIGNIFICANCE           An agroecosystem is a coupled domain encompassing the interplay of climate, soil, and topography. This dynamic and interactive entity must be holistically accounted when assessing its responses to environmental or socioecological changes. Moreover, the diverse productivity contributions of various typologies underscore the critical need for spatially explicit adaptive strategies to address crop production challenges effectively. This approach can be extended to identify regions where specific conservation management practices are effective or ineffective by incorporating crop management layers into the analysis as such data become available in the future.

Keywords: Agroecosystems, Productivity heterogeneity, High-dimensional statistical modeling, Typologies