Model-Data Fusion for improving quantification of methane fluxes from wetlands with LPJ-GUESS
Ecosystem models calculating wetland emissions incorporate process knowledge
(bottom-up modeling) and allow estimation of methane (CH4) fluxes at various scales ranging from local to regional and global. The main aim of this project is to apply model-data fusion (formally called ‘data assimilation’) techniques to improve quantification and understanding of the natural CH4 cycling in boreal wetlands.
The specific aims are to:
1. Develop a CH4 Data Assimilation System around the LPJ-GUESS dynamic global vegetation model based on a Markov Chain Monte Carlo approach.
2. Generate improved model process parameter estimates, with quantified posterior uncertainties by assimilating observed CH4 fluxes from wetlands.
3. Quantify how assimilation of observations increases confidence in the analysis of the CH4 emissions from wetlands.
4. Provide new scientific insights and process understanding, e.g. what controls the sensitivity of CH4 emissions from wetlands to external forcing.
5. Provide a CH4 budget for the boreal region for the dominant natural emission source (i.e. wetlands).
6. Provide input to the Global Carbon Project’s CH4 budget as well as deliver elaborated products (i.e. maps of CH4 fluxes from wetlands) to e.g. the Carbon
Portal of the Integrated Carbon Observation System (ICOS) derived from CH4 flux observations at, among others, ICOS sites.
Through the model-data fusion approach it is particularly relevant to the research area on data assimilation and multimodel integration by combining measurements with ecosystem models. The research will improve our understanding of wetlands processes and CH4 cycling as included in models and will improve upscaling processes from local to regional/global scales. It will thus refine the predictions of future CH4 wetland emission to inform policy making especially with regard to the Paris Agreement of the UNFCC and subsequent NDCs.