Merging eddy covariance and remote sensing models to facilitate high resolution spatiotemporal monitoring of agricultural greenhouse gas budgets

Authors: WIESNER, SUSANNE - University Of Wisconsin; Duff, Alison; Niemann, Kristine; METZGAR, STEFAN - National Ecological Observatory Network (NEON); DESAI, ANKUR - University Of Wisconsin; STOY, PAUL - University Of Wisconsin

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 10/4/2021
Publication Date: N/A
Citation: N/A

Interpretive Summary: Reducing agricultural greenhouse gas (GHG) emissions is critical for climate mitigation, and ruminant livestock systems have relatively high GHG emissions within the agricultural sector. Management of vegetation cover has great potential to offset emissions from animal agriculture, although this strategy is often overlooked. Natural climate solutions (NCS) comprise strategies of land stewardship to increase carbon storage or offset farm emissions; this approach is promising but outcomes are subject to high uncertainty and influenced by local factors including climate, soil, and farm management. Using satellite land imagery data to quantify vegetation productivity and estimate carbon storage is relatively inexpensive, but has coarse spatial and temporal resolution. Eddy covariance instrumentation, which captures nearly continuous measurements of farm meteorological and emissions data, provides fine spatial and temporal resolution, but is cost prohibitive outside of research institutions. We used the US Dairy Forage Research Center's Prairie du Sac research dairy as a case study to understand how these two approaches (satellite remote sensing and eddy covariance measurements) can be combined to refine satellite remote sensing models and improve estimations of farm GHG budgets and assessment of natural climate solutions. Our results suggest that the fusion of these two approaches can improve predictions of vegetation productivity and potential carbon storage, and improve remote sensing models that can be applied effectively in agricultural landscapes.

Technical Abstract: Reducing agricultural greenhouse gas (GHG) emissions is critical, especially in ruminant livestock systems with high GHG burden. Vegetation can play a large role in mitigating emissions in these systems, but is often neglected. Mitigation strategies, like natural climate solutions (NCS, such as cover crops), and changes in soil carbon stocks are promising but subject to high uncertainty and often constrained by climate, soil, and management. Satellite remote sensing (RS) models quantify such mitigation, but often at a cost of spatiotemporal resolution. Eddy covariance (EC) measurements in turn have higher spatiotemporal resolution but offer lower financial accessibility. Here we use the U.S. Dairy Forage Research Center (USDFRC) dairy farm as a case study to understand how EC can help constrain Landsat RS models to improve farm GHG budget quantifications and NCS monitoring. We use the environmental response function (ERF) approach to constrain RS gross ecosystem exchange (GEE) and ecosystem respiration (Reco) models by updating input parameters such as maximum quantum yield of photosynthesis using EC data. Daily EC net primary productivity (NPP) matched Landsat NPP when daytime data were compared (slope and R2 0.77 and 0.88, respectively), as Landsat measures during daytime periods. Daily EC GEE and Reco were also in good agreement with RS models (R2 0.78-0.94, 0.86-0.95 & 0.74-0.89). Annual biomass predictions from EC-RS fusion estimates improved for all crop types when compared to field harvest estimates (R2 = 0.6 and slope improved from 1.09 to 0.98). The USDFRC farm vegetation was able to mitigate all emissions from within the farm boundaries (~9900 tons CO2eq), which included manure handling, fertilizer applications, electricity, and enteric fermentation, (farm NPP was ~10000 tons CO2), where natural vegetation types (i.e., forest, pastures and shrublands) contributed ~70% of NPP. Our results suggest that EC-RS fusion products can help improve NPP monitoring on spatiotemporal scales larger than EC and RS methods alone. EC-RS fusion products can help monitor farm GHG budgets by updating plant physiological parameters, which are often subject to high variability in agricultural systems due in part to seed varieties,genetic improvements, and other management decisions.