Presentation Information
[MIS02-11]Evaluation of gridded wetland methane flux datasets with atmospheric CH4 simulations over Western Siberia
*Aleksandr Nogovitcyn1, Motoki Sasakawa1, Shamil Maksyutov1, Toshinobu Machida1,2, Takuto Taguchi1, Qianning Qin1, Irina Terentieva3, Mikhail Glagolev4,5, Aleksandr Sabrekov5 (1.National Institute for Environmental Studies, Japan, 2.Seto Inland Sea Carbon-neutral Research Center, Hiroshima University, Japan, 3.University of Calgary, Canada, 4.Institute of Forest Science, Russian Academy of Sciences, Russia, 5.Yugra State University, Russia)
Keywords:
Western Siberia,wetland methane flux,atmospheric methane observations,atmospheric transport model
Wetlands in Western Siberia are among the largest and most diverse in the world, making them important natural sources of methane (CH4). Accurately measuring CH4 emissions from these wetlands is crucial for reducing uncertainties in the global methane budget and improving climate predictions. Although numerous gridded wetland CH4 flux datasets have been developed using different approaches (e.g. using different wetland maps), their estimates often differ in Western Siberia, where pronounced seasonal dynamics in wetlands increase uncertainty. This study evaluates the performance of the wetland CH4 flux products by simulating atmospheric CH4 concentrations at observation sites in Western Siberia.
The 0.5° gridded wetland CH4 datasets used in this study are derived from: a) process-based models, b) machine-learning models that upscale eddy covariance data and chamber measurements, c) empirical models that upscale chamber measurements based on wetland maps for Western Siberia (Terentieva et al., 2016) and Russia (Terentieva et al., 2024). In addition, we used the Global Methane Budget (GMB) wetland emissions dataset, which is the ensemble mean of wetland process-based model datasets (Saunois et al., 2020). Observation-based empirical models originally provide annual wetland emissions for each grid cell without representing the monthly seasonal cycle of emissions, unlike other datasets. To ensure consistency among datasets, a mean seasonal cycle derived from the remaining wetland CH4 datasets was therefore applied to these gridded observation-based estimates.
NIES-TM–FLEXPART-variational inverse model (NTVAR; Maksyutov et al., 2021) was used to forward-simulate surface atmospheric CH4 concentrations observed at six JR-STATION sites (Japan-Russia Siberian Tall Tower Inland Observation Network; Sasakawa et al., 2010) in Western Siberia for the period 2009-2020. In order to compare the performance of different wetland flux datasets specifically over Western Siberia and to remove the influence of differing wetland fluxes from other regions, West Siberian wetland fluxes from each dataset were nested into a single global wetland CH4 flux dataset. This global dataset is based on the GMB wetland dataset, which was optimized together with antropogenic emissions (EDGAR) through inversion using atmospheric CH4 observations from ObsPack. The resulting nested products were used as wetland emission inputs for the transport model, while all other emission sources were kept same.
Our results show that observation-based wetland CH4 datasets for Western Siberia (Terentieva et al., 2016; Terentieva et al., 2024) exhibit some of the highest correlations between hourly observed and simulated CH4 concentrations during the summer period at study sites, compared to other datasets. This performance likely reflects their use of flux observations and detailed wetland maps specific to Western Siberia. The total annual emissions from Western Siberia estimated by those datasets (4.5 and 6.0 Tg, respectively) are less than those estimated by GCP (7.2 Tg). This evaluation highlights the importance of validating wetland CH4 flux datasets in Western Siberia, providing guidance for improving global methane budget assessments and climate projections.
References
Maksyutov et al. (2021) Atmospheric Chemistry and Physics, 21 (2) 1245-1266.
Sasakawa et al. (2010) Tellus B, 62, 403-416.
Saunois et al (2020) Earth System Science Data, 12, 1561-1623
Terentieva et al. (2016) Biogeosciences, 13, 4615-4626.
Terentieva et al. (2024) [dataset] Zenodo repository. https://zenodo.org/records/13997236.
The 0.5° gridded wetland CH4 datasets used in this study are derived from: a) process-based models, b) machine-learning models that upscale eddy covariance data and chamber measurements, c) empirical models that upscale chamber measurements based on wetland maps for Western Siberia (Terentieva et al., 2016) and Russia (Terentieva et al., 2024). In addition, we used the Global Methane Budget (GMB) wetland emissions dataset, which is the ensemble mean of wetland process-based model datasets (Saunois et al., 2020). Observation-based empirical models originally provide annual wetland emissions for each grid cell without representing the monthly seasonal cycle of emissions, unlike other datasets. To ensure consistency among datasets, a mean seasonal cycle derived from the remaining wetland CH4 datasets was therefore applied to these gridded observation-based estimates.
NIES-TM–FLEXPART-variational inverse model (NTVAR; Maksyutov et al., 2021) was used to forward-simulate surface atmospheric CH4 concentrations observed at six JR-STATION sites (Japan-Russia Siberian Tall Tower Inland Observation Network; Sasakawa et al., 2010) in Western Siberia for the period 2009-2020. In order to compare the performance of different wetland flux datasets specifically over Western Siberia and to remove the influence of differing wetland fluxes from other regions, West Siberian wetland fluxes from each dataset were nested into a single global wetland CH4 flux dataset. This global dataset is based on the GMB wetland dataset, which was optimized together with antropogenic emissions (EDGAR) through inversion using atmospheric CH4 observations from ObsPack. The resulting nested products were used as wetland emission inputs for the transport model, while all other emission sources were kept same.
Our results show that observation-based wetland CH4 datasets for Western Siberia (Terentieva et al., 2016; Terentieva et al., 2024) exhibit some of the highest correlations between hourly observed and simulated CH4 concentrations during the summer period at study sites, compared to other datasets. This performance likely reflects their use of flux observations and detailed wetland maps specific to Western Siberia. The total annual emissions from Western Siberia estimated by those datasets (4.5 and 6.0 Tg, respectively) are less than those estimated by GCP (7.2 Tg). This evaluation highlights the importance of validating wetland CH4 flux datasets in Western Siberia, providing guidance for improving global methane budget assessments and climate projections.
References
Maksyutov et al. (2021) Atmospheric Chemistry and Physics, 21 (2) 1245-1266.
Sasakawa et al. (2010) Tellus B, 62, 403-416.
Saunois et al (2020) Earth System Science Data, 12, 1561-1623
Terentieva et al. (2016) Biogeosciences, 13, 4615-4626.
Terentieva et al. (2024) [dataset] Zenodo repository. https://zenodo.org/records/13997236.