Public Abstract

DE-SC0021067: Functional-type modeling approach and data-driven parameterization of methane emissions in wetlands

Award Status: Active

Institution: The Ohio State University, Columbus, OH
UEI: DLWBSLWAJWR1
DUNS: 832127323

Most Recent Award Date: 05/09/2023
Number of Support Periods: 3
PM: Stover, Daniel

Current Budget Period: 08/15/2022 - 08/14/2024
Current Project Period: 08/15/2020 - 08/14/2024
PI: Missik, Justine

Supplement Budget Period: N/A

Public Abstract

Functional-type modeling approach and data-driven parameterization of methane emissions in wetlands Gil Bohrer, Professor, Ohio State University, (PI) Kelly Wrighton, Colorado State University (Co-Investigator) Jorge Vila, University of Louisiana at Lafayette (Co-Investigator) William Riley, Lawrence Berkeley National Lab (Co-Investigator) Eric Ward, USGS (Co-Investigator) Summary The role of natural wetland emissions in the recent sharp increase of global methane atmospheric concentration is hotly debated. Difficulties in modeling methane emissions are driven, to a large degree, by the spatially heterogeneous nature of wetland ecosystems and by the high temporal and spatial variability of methane fluxes. This variability in fluxes is a result of the underlying heterogeneity of the wetlands, combined with the complex and interconnected belowground and aboveground processes of methane production, consumption, and transport. Eddy covariance (EC) and chamber flux measurements, and remote sensing of wetlands and atmospheric methane concentrations are routinely used to parameterize and improve land-surface models that predict methane fluxes. However, model optimizations are typically done against observations of the net site-level flux, or coarse atmospheric methane concentration and do not independently address within-wetland heterogeneity or the various processes that add up to the net flux. Therefore, even models that make accurate CH4 emission predictions may be right for the wrong reasons. The goal of this project is to improve simulation of methane emissions from coastal wetlands in the E3SM Land Model (ELM). We will improve ELM along three spatial and conceptual axes: (i) Patch resolution – expand the subgrid surface-tile approach, currently widely utilized in representation of upland vegetation, to represent ecohydrological patch types, such as open-water, mudflats, emergent/floating vegetation, swap forest, etc. (ii) Vertical resolution – provide vertically detailed observations of belowground dissolved methane concentration gradients and utilize ELMs patch-level vertical soil column to resolve methane production, oxidation, and transport at high vertical resolution. (iii) Process resolution– increasing the number, type, and accuracy of independent process representations that combine within the soil column and interact with plants to affect net CH4 emissions. To support these model developments, we propose to expand an already extensive dataset of observations in 4 coastal wetlands along salinity and tidal influence gradient. Observations include flux measurements at multiple scales: whole-site EC fluxes (in 3 sites, already reporting to AmeriFlux), chamber flux measurements at each ecohydrological patch type, bubble accumulations, and methane transport through plants. These measurements will allow us to directly test and parameterize the model at multiple scales, including specific processes resolved at the patch level, and go beyond the typical whole-site parameterization. Observations also include vertical profiles of methane, dissolved organic carbon, oxygen, and temperature in the soil of each patch type, using in-situ dialysis samplers (peepers). Soil cores for chemistry, metagenomic and metatranscriptomics, and methane, and carbon isotope analysis, will provide estimates of the presence and relative importance of the rates of particular microbial processes of production and consumption of methane. We will conduct sets of parameterization, validation, and sensitivity analyses simulations using the PEcAn workflow. After the optimization of the model at our demonstration site, we will use the framework of the Global Carbon Project methane reanalysis simulations to test the sensitivity of global methane emission forecasts to resolving sub-site ecohydrological patches. Our project will significantly improve ELM coastal-wetland methane simulation capabilities by implementing the sub-site ecohydrological patch-type approach, improving process representation, and careful testing and calibration against multi-scale observations. We will demonstrate the process-level multi-scale parameterization approach and provide direct observations to determine and directly constrain parameters within the multiple methane biogeochemical and plant processes represented in ELM. We will demonstrate the need for global within-wetland patch characterization and delineation, and patch-level flux and vertically detailed concentration observations.

Functional-type modeling approach and data-driven parameterization of methane emissions in wetlands

Gil Bohrer, Professor, Ohio State University, (PI)
Kelly Wrighton, Colorado State University (Co-Investigator)
Jorge Vila, University of Louisiana at Lafayette (Co-Investigator)
William Riley, Lawrence Berkeley National Lab (Co-Investigator)
Eric Ward, USGS (Co-Investigator)

Summary
The role of natural wetland emissions in the recent sharp increase of global methane atmospheric concentration is hotly debated. Difficulties in modeling methane emissions are driven, to a large degree, by the spatially heterogeneous nature of wetland ecosystems and by the high temporal and spatial variability of methane fluxes. This variability in fluxes is a result of the underlying heterogeneity of the wetlands, combined with the complex and interconnected belowground and aboveground processes of methane production, consumption, and transport. Eddy covariance (EC) and chamber flux measurements, and remote sensing of wetlands and atmospheric methane concentrations are routinely used to parameterize and improve land-surface models that predict methane fluxes. However, model optimizations are typically done against observations of the net site-level flux, or coarse atmospheric methane concentration and do not independently address within-wetland heterogeneity or the various processes that add up to the net flux. Therefore, even models that make accurate CH4 emission predictions may be right for the wrong reasons.

The goal of this project is to improve simulation of methane emissions from coastal wetlands in the E3SM Land Model (ELM). We will improve ELM along three spatial and conceptual axes: (i) Patch resolution – expand the subgrid surface-tile approach, currently widely utilized in representation of upland vegetation, to represent ecohydrological patch types, such as open-water, mudflats, emergent/floating vegetation, swap forest, etc. (ii) Vertical resolution – provide vertically detailed observations of belowground dissolved methane concentration gradients and utilize ELMs patch-level vertical soil column to resolve methane production, oxidation, and transport at high vertical resolution. (iii) Process resolution– increasing the number, type, and accuracy of independent process representations that combine within
the soil column and interact with plants to affect net CH4 emissions.

To support these model developments, we propose to expand an already extensive dataset of observations in 4 coastal wetlands along salinity and tidal influence gradient. Observations include flux measurements at multiple scales: whole-site EC fluxes (in 3 sites, already reporting to AmeriFlux), chamber flux measurements at each ecohydrological patch type, bubble accumulations, and methane transport
through plants. These measurements will allow us to directly test and parameterize the model at multiple scales, including specific processes resolved at the patch level, and go beyond the typical whole-site parameterization. Observations also include vertical profiles of methane, dissolved organic carbon, oxygen, and temperature in the soil of each patch type, using in-situ dialysis samplers (peepers). Soil cores for chemistry, metagenomic and metatranscriptomics, and methane, and carbon isotope analysis, will provide estimates of the presence and relative importance of the rates of particular microbial processes of production and consumption of methane. We will conduct sets of parameterization, validation, and sensitivity analyses simulations using the PEcAn workflow. After the optimization of the model at our demonstration site, we will use the framework of the Global Carbon Project methane reanalysis simulations to test the sensitivity of global methane emission forecasts to resolving sub-site ecohydrological patches.

Our project will significantly improve ELM coastal-wetland methane simulation capabilities by implementing the sub-site ecohydrological patch-type approach, improving process representation, and careful testing and calibration against multi-scale observations. We will demonstrate the process-level multi-scale parameterization approach and provide direct observations to determine and directly constrain parameters within the multiple methane biogeochemical and plant processes represented in ELM. We will demonstrate the need for global within-wetland patch characterization and delineation, and patch-level flux and vertically detailed concentration observations.