Multi-year extreme drought events are part of the history of the Earth system. Legacy data on the climate drivers, geomorphic features, and agroecosystem responses across a dynamically changing landscape throughout a region can provide important insights to a future where large-scale catastrophic events may occur more frequently. One of the most devastating multi-year droughts occurred in the central grasslands region of North America in the 1930s. This regional-scale climatic event combined with land management practices to result in broad-scale plant mortality and massive dust storms that impacted the entire continent. However, not all areas were affected similarly, even across relatively short distances with similar climate, soils, and land management. Spatial discontinuities in impacts occurred across a 100-km transition from high plant mortality and high rates of soil erosion in eastern Nebraska to areas in western Iowa with reductions in grass cover and biomass, but low rates of plant mortality and erosion. Because this time period preceded modern agriculture and the extensive plowing of native tallgrass prairie, the landscape was composed of both native prairie grassland and cropland. Responses were compared during the drought to responses for the same region in the 1920s (pre-drought) and the 1940s (post-drought). A large number of legacy datasets from the U.S. Department of Agriculture and other sources were spatially integrated to test two hypotheses: (1) local factors of climate and soil properties explained agroecosystem responses in the pre- and post-drought periods, (2) local factors were insufficient to explain agroecosystem responses during the multi-year drought. Analyses supported these hypotheses and found that landscape features, such as large alluvial plains that reduced connectivity by sand and deposition by wind, were more important than local factors to explain different responses along the gradient during the drought. Similar results were found for both native grassland and cropland. These findings have important implications for landscape configurations and connectivity in the face of large-scale, extreme events.