Development of a gridded yield data archive for farm management and research at the USDA Beltsville Agricultural Research Center

Authors: Dulaney, Wayne; Anderson, Martha; Gao, Feng; Stern, Alan; Meyers, George; Daughtry, Craig; White, William - Alex; Akumaga, Uvirkaa; Showalter, Jennifer; Moglen, Glenn

Submitted to: Agrosystems, Geosciences & Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/2/2024
Publication Date: 6/1/2024
Citation: Dulaney, W.P., Anderson, M.C., Gao, F.N., Stern, A.J., Meyers, G.E., Daughtry, C.S., White, W.A., Akumaga, U., Showalter, J.J., Moglen, G.E. 2024. Development of a gridded yield data archive for farm management and research at the USDA Beltsville Agricultural Research Center. Agrosystems, Geosciences & Environment. 7(1). Article e20474. https://doi.org/10.1002/agg2.20474.
DOI: https://doi.org/10.1002/agg2.20474

Interpretive Summary: High quality observations of crop yield are invaluable for agricultural research: for evaluating crop models, for studying crop response to climate and soil variability, and for providing guidance for future management. While many yield datasets exist at the county and even field level, comprehensive multi-year and multi-field datasets at sub-field scales appropriate for evaluating remote sensing models and precision management strategies are very rare. This paper describes the protocol used to generate a 7-year (2015-2021) dataset of annual yield maps developed at 5-m resolution over 43 production fields at the USDA-ARS Beltsville Agricultural Research Center in Beltsville, MD. The maps were developed from yield monitor data collected during harvest and were rigorously quality controlled and interpolated to 5-m grids using block Kriging methods. The gridded yield data archive will be housed at the National Agricultural Library as a dynamic dataset that will continue to expand in breadth and depth. It will be a part of and provide support to internal USDA yield monitoring investigations as well as complementary external efforts.

Technical Abstract: High quality, gridded maps of crop yield, which discern the spatial variability in productivity across individual farm fields, are extremely valuable in numerous agricultural and remote sensing applications. The availability of these data was greatly facilitated by the development and adoption of grain yield monitors starting in the early 1990s. However, the utility of raw yield monitor data is limited if additional processing has not been carried out to ensure data quality. In addition, a readily accessible data repository that allows for examining long term yield variability in response to climate and management is often not available. We present a protocol for producing a multi year (7 years currently processed), gridded yield data archive (5-m resolution) derived from yield monitor data acquired from over 40 production fields at the Beltsville Agricultural Research Center (BARC) in Beltsville, Maryland. Construction of the archive involved the projection of the yield monitor data into a geographic coordinate system suitable for mapping, the identification and removal of data outliers, and the creation of gridded, spatially interpolated yield maps using block Kriging. All map products were produced in common, non proprietary file formats for easy access. Preliminary assessments of the processed yield monitor data have underscored the necessity of data filtering and have shown the influence of previous land use practices and biophysical properties, such as topography and soil moisture availability, on yield response. We discuss the use of spatially interpolated, archival yield maps in the implementation of precision farming techniques such as the development of site specific variable rate prescriptions as well as remote sensing applications including the selection of field sites for long term agricultural experiments and the assessment of crop modeling approaches. The gridded yield data archive will be housed at the National Agricultural Library as a dynamic dataset that will continue to expand in breadth and depth. It will be a part of and provide support to internal USDA yield monitoring investigations as well as complementary external efforts.