Biocycle - Regional Inverse Modeling

Part of the challenge of creating a self-consistent, process-based description of CO2 exchange between the land and the atmosphere lies in testing our scaling assumptions and methods. This can be done locally by comparing simulations and observations at flux towers, for example. At much larger scales, we can compare extrapolated fluxes obtained from models and satellite imagery to continental fluxes inferred from global inverse models. Additionally, methods for extrapolating process models in space can be evaluated in detail by performing much higher resolution inversions of atmospheric data collected during regional campaigns.

We have been developing a set of techniques to combine the use of numerical models with regional CO2 measurements. The goal of regional inversion modeling is similar to global scale inverse modeling: to use available data from concentration and vertical flux measurements to estimate surface fluxes of CO2 However, in regional scale modeling, much smaller scales of source areas, from 10 to 1,000 km, are considered. Additional challenges are created for regional modelers because they must deal with limited domains. The inflow CO2 fluxes across lateral boundaries are usually several orders of magnitude higher than the surface fluxes of CO2 from these regional domains. These inflow fluxes can be treated as unknown parameters to be estimated in the inversion calculations, or they can be estimated from available measurements, or from the results provided by larger scale models. Another difficulty is related to a strong diurnal cycle of CO2 fluxes, which cannot be neglected at these scales.

The regional inversion framework is built around CSU RAMS and the Lagrangian Particle Dispersion (LPD) model. The LPD model is used to trace particles backward in time to derive influence functions for each concentration sample. The influence function provides information on potential contributions both from surface sources and inflow fluxes that make their way through the modeling domain boundaries into the CO2 concentration sample. Then the Bayesian inversion technique is applied in an attempt to estimate unknown surface emissions.

The modeling framework is especially useful for the design of field experiments and observational networks. It allows us to explore the feasibility of flux estimations, given a specific data set, and taking into account the uncertainty within the data. For this type of application, model generated concentration pseudo-data are used in inversion calculations to evaluate and compare different sampling strategies and data sets.

Several sampling strategies, including aircraft sampling and concentration time series from one or more towers, have been investigated under simple meteorological conditions (multi-day evolution of the planetary boundary layer over homogenous terrain). Current research focuses on a similar investigation, but uses meteorology from a multi-month RAMS simulation over real terrain in northern Wisconsin. Download documents, presentations, and data:
• Selected papers by Marek Uliasz on LPD Modeling
• Denning et al manuscript on local-scale SiB-RAMS simulations
• Nicholls et al manuscript on regional CO2 simulations in SiB-RAMS
• Uliasz et al draft manuscript on mesoscale inverse modeling
• Presentation on regional inverse modeling
• Animations of regional influence functions

Acknowledgements:

This research is part of the Chequamagon Ecosystem-Atmosphere Study (ChEAS) coordinated at the Pennsylvania State University. Atmospheric trace-gas measurements are made by NOAA/CMDL at the WLEF-TV tower and another tall tower (KETK-TV in Texas) under the MAGNETT program.

This research was supported by the South-Central Regional Center (SCRC) of the National Institute for Global Environmental Change (NIGEC), by the Terrestrial Carbon Program of the U.S. Department of Energy, and by the National Atmospheric and Oceanic Adminitrstion, Office of Global Programs. Any opinions, findings and conclusions or recommendations expressed herein are those of the authors and do not necessarily reflect the view of the DOE.