Historical cropland expansion and abandonment in the continental US during 1850 to 2016

Aim: Land use and land cover changes (LCLUC) are among the most important driving forces that alter terrestrial ecosystem functions and their feedbacks to climate systems, but reliable, spatially explicit datasets over century-long periods are still lacking for fine-scale earth system modeling. We aimed to combine multiple data sources and reconstruct long-term land use history in the continental U.S., examining cropland expansion and abandonment since 1850. Location: Conterminous U.S. Time period: 1850 to 2016. Major taxa studied: Cropland. Methods: Cropland density maps, displaying the distribution and percentage of cultivated land each year (excluding summer idle/fallow, cropland pasture), were reconstructed by harmonizing multiple sources of inventory data and high-resolution satellite images. The cropland data are freely available to the public. Results: In total, national cropland expansion was 104 million hectares (Mha) from 1850 to 2016 and peaked at about 127 Mha in 1920. Forests and shrublands were the dominant land cover types that croplands were converted from during 1850 to 1880, which may be primarily attributed to agriculture development in the northeast U.S. Croplands began to expand into grasslands from 1870 onwards and the encroached area dramatically increased, mainly due to cultivation development in the Great Plain and midwestern areas. In comparison, the area of abandoned cropland in the U.S. was 65 Mha during the study period. We found cropland abandonment mostly occurred in the central and southeast U.S., while cropland expansion was centered upon the midwestern states, central California, and the Mississippi Alluvial Plain. Main conclusions: Nationally, cultivated lands shifted from the eastern to midwestern U.S. during the study period, contributing to the increasingly important role of the Midwest in the rise of food and biofuel production, enhanced greenhouse gas (GHG) emissions, and high nitrogen loads into the Gulf of Mexico. Our cropland database is essential for modeling assessments of LCLUC impacts, crop production estimation and socioeconomic analysis.