How Do We Know What We Grow? Interrogating the Datafication of Agricultural Landscapes in the United States

Abstract

This article analyzes the data processes that render US agricultural landscapes knowable at scale as objects of anthropological inquiry. We focus our inquiry on the US Department of Agriculture's Cropland Data Layer (CDL), a widely used, moderate-resolution raster data set classifying national agricultural land use annually. The CDL's crop categories are based upon—but depart significantly from—those of another federal agricultural office, the Farm Service Agency (FSA). We visualize several transformations from the FSA's data categories to the CDL's to identify which crop varieties are preserved during this process and which are coarsened into higher-level categories. These patterns illustrate the logics underlying the CDL's data categorization schema. Constrained by the technical limits of remote sensing technology, these data most often obscure the presence of specialty, native, and food crops, rendering them unknowable at a national scale and entrenching long-standing productivist values into the country's agri-data infrastructure. Bounded by the same path dependencies shaping the very agricultural landscapes they codify, these data themselves become barriers to recognizing where agricultural transformations may already be under way.