Five foundational tools for managing metadata from the USDA Long-Term Agroecosystem Research (LTAR) network

Authors: Kaplan, Nicole; Armendariz, Gerardo; AZAD, SHEFALL - Archbold Biological Station; Carlson, Bryan; White, William; Abendroth, Lori; Coffin, Alisa; Gordon, Vanessa; Maul, Jude; Osterholz, William; Sears, John

Submitted to: Journal of Environmental Quality
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/25/2025
Publication Date: 4/29/2025
Citation: Kaplan, N.E., Armendariz, G.A., Azad, S., Carlson, B.R., White, W.A., Abendroth, L.J., Coffin, A.W., Gordon, V.S., Maul, J.E., Osterholz, W.R., Sears, J.L. 2025. Five foundational tools for managing metadata from the USDA Long-Term Agroecosystem Research (LTAR) network. Journal of Environmental Quality. 1-15. https://doi.org/10.1002/jeq2.70027.
DOI: https://doi.org/10.1002/jeq2.70027

Interpretive Summary: The United States Department of Agriculture (USDA) Long-Term Agroecosystem Research (LTAR) Network is comprised of nineteen research sites in different agroecosystems. Each site generates various kinds of data collected by people in the field as well as sensors installed in the environment. Data are then stored locally at each site in various formats. This makes it challenging to conduct cross-site analysis, because it is difficult to know what data are available as well as when, where and how data were collected. The LTAR is currently designing a system that can serve the network, but it does not have a centralized staff for documenting and harmonizing data from across the sites for data submission. So data managers and scientists worked together to create four tools for finding, harmonizing, mapping and sharing data along with information about the data (i.e. metadata). Tool development is focused on (1) inventorying datasets managed by each site, (2) creating controlled vocabularies for measurements to facilitate cross-site comparisons and analyses, (3) geolocating the data collection, site and experimental boundaries, (4) reporting quantitative research impact of published literature, and (5) using dashboards to visualize what data are available, where they were collected, and their collection period. These efforts are described by LTAR site data managers and scientists, which serve as a foundation for a network management system and facilitate data analysis at cross-site, regional, and national levels.

Technical Abstract: The United States Department of Agriculture (USDA) Long-Term Agroecosystem Research (LTAR) Network is comprised of nineteen research sites in different agroecosystems with over 940 terabytes of highly diverse, multidisciplinary data including longitudinal ground-based time-series, remotely sensed imagery, and environmental sensor-based measurements. Participants leverage long research histories, referred to as legacy data, at individual sites for network-level analyses. These legacy data allow for assessments and trend analysis of environmental quality and resource allocation at cross-site, regional, and national-levels. However, the analytical power for network research is severely constrained due to legacy data managed and contained in local siloed systems. Although centralized infrastructure is currently in development, the lack of a network information management system and centralized staff is a significant challenge. Data managers are hired and assigned to each LTAR site and this cohort has formed teams to develop foundational tools for finding, harmonizing, mapping, and sharing data and metadata. Tool development is focused on (1) inventorying datasets managed by each site, (2) creating controlled vocabularies for measurements to facilitate cross-site comparisons and analyses, (3) geolocating the data collection, site and experimental boundaries, (4) reporting quantitative research impact of published literature, and (5) using dashboards to visualize what data are available, where they were collected, and their collection period. These efforts by LTAR site data managers and scientists serve as a pivot point around which collective work at cross-site, regional, and national-levels can occur. Harmonized data and metadata provide a robust foundation for the development of a network information management and synergistic data science solutions.