Unraveling the effects of management and climate on carbon fluxes of U.S. croplands using the USDA Long-Term Agroecosystem (LTAR) network

Understanding the carbon fluxes and dynamics from a broad range of agricultural systems has the potential to improve our ability to increase carbon sequestration while maintaining crop yields. Short-term, single-location studies have limited applicability, but long-term data from a network of many locations can provide a broader understanding across gradients of climate and management choices. Here we examine eddy covariance measured carbon dioxide (CO2) fluxes from cropland sites across the United States Department of Agriculture's Long-Term Agroecosystem Research (LTAR) network. The dataset was collected between 2001 and 2020, spanning 13 sites for a total of 182 site-years. Average seasonal patterns of net ecosystem CO2 exchange (NEE), gross primary productivity (GPP), and ecosystem respiration (Reco) were determined, and subsequent regression analysis on these “flux climatologies” was used to identify relationships to mean annual temperature (MAT), mean annual precipitation (MAP), cropping systems, and management practices. At rainfed sites, carbon fluxes were better correlated with MAP (r2 ≤ 0.5) than MAT (r2 ≤ 0.22). Net carbon balance was different among cropping systems (p < 0.001), with the greatest net carbon uptake occurring in sugarcane (Saccharum spp. hybrids) and the least in soybean (Glycine max) fields. Crop type had a greater effect on carbon balance than irrigation management at a Nebraska site. Across cropping systems, grain crops often had higher GPP and were more likely to have net uptake when compared to legume crops. This multi-site analysis highlights the potential of the LTAR network to further carbon flux research using eddy covariance measurements.