Defining features of diverse and productive agricultural systems: An archetype analysis of U.S. agricultural counties

Prior research suggests that greater spatial diversity in crops and land use is associated with higher crop yields and improved ecosystem function. However, what leads to the emergence of agricultural systems that meet both productivity and ecological health goals remains an open question. Understanding the factors that differentiate these places from other agricultural systems is key to understanding the mechanisms, pathways, consequences, and constraints to employing diversification as a tool for increasing agricultural sustainability. In this study, we employ archetype analysis to examine the factors uniquely associated with the conjoint existence of high crop diversity and high crop productivity. We identify five agricultural system classes that represent a range of diversity and productivity combinations using k-means cluster analysis then use random forests analysis to identify factors that strongly explain the differences between the classes—describing different agricultural production regimes. Our exploratory analysis of the difference in agricultural system factors across classes suggests (1) crop diversity and its preconditions are associated with the highest yields, (2) biophysical conditions bound diversity-productivity realities, (3) productivity comes at a petrochemical cost, and that (4) crop rotations are a key diversification strategy. Overall, our results suggest that despite clear biophysical constraints on transitions to high diversity—high productivity systems the role of actionable factors on crop production regimes is stronger, providing reason to be hopeful about transitions to agricultural production regimes fit for new climate realities.