ABSTRACT:

Agriculture plays a vital role in the US, generating billions of dollars in revenue and providing essential resources such as food, fiber, and fuel. At the same time, this sector consumes more water than all other sectors combined, and contributes significantly to greenhouse gas emissions. Accurately mapping cropland areas at a high spatial resolution, and understanding cultivation trends are crucial for promoting sustainable land management. However, it is not always clear where these crops are cultivated at a fine spatial resolution. Existing data sources either provide extensive temporal coverage with low spatial detail or vice versa. To address this, we employed a data-fusion approach, combining strengths of existing data sources. We produced annual irrigated and harvested areas for 30 major crops in the US from 1981 to 2019 at 2.5 arc minutes by combining county level US Department of Agriculture records with gridded land use datasets. We evaluated our dataset by comparing it with existing large-scale gridded cropland data, revealing alignment and divergence across data products depending on the year, area, and crop. Our dataset serves as a critical tool for analyzing and understanding long-term trends in cropland cultivation, thereby contributing to more informed and sustainable agricultural practices