Simultaneous Optimisation of Global Fluxes of Methane and Carbon Dioxide using Satellite Data
Pandey, Sudhanshu1; Houweling, Sander1; Krol, Maarten C.2; Röckmann, Thomas2; Aben, Ilse1
1Netherlands Institute for Space Research (SRON), NETHERLANDS; 2Institute for Marine and Atmospheric research Utrecht (IMAU), NETHERLANDS

The inverse modeling technique is used to translate satellite (GOSAT, ENVISAT) and ground based (WMO-GAW, TCCON) observations into corresponding emissions at the Earth's surface. However, in the case of long-lived greenhouse gases, such as methane, it is a challenge for the satellites to meet the required level of accuracy. A successful method is the so-called proxy retrieval (Frankenberg et al., 2005; Butz et al., 2011), which makes use of the ratio of methane over carbon dioxide. Systematic errors in satellite data, e.g. due to atmospheric scattering, affect both carbon dioxide and methane similarly. Hence, taking the ratio largely eliminates this error. It is assumed that the contribution of carbon dioxide is well enough understood to use this method for studying methane. However, with the improved measurement quality obtained using the GOSAT instrument this assumption is starting to become an important limitation (Schepers et al., 2012).

We present a new inverse modeling method (called the 'ratio' method from here on) based on the 4DVAR technique (Meirink et al., 2008). The aim is to optimize the ratio of methane and carbon dioxide, i.e. without translation to CH4. The cost function optimization is done w.r.t. the sources and sinks of both CH4 and CO2, which circumvents prescribing model-derived CO2 fields. Since the measurements consist of the ratios of these compounds, the optimization problem is nonlinear and, hence, we cannot make use of the widely used conjugate gradient method (Hestenes and Stiefel, 1952). In our case, we use the M1QN3 method (Gillbert and lemarechal, 2006), which is based on quasi Newton method and is effective in optimizing nonlinear problems.

The method has been successfully tested in a simplified optimization setup, and is currently being applied to the global transport model. In our presentation we will further explain our method, and show first results of synthetic data inversions demonstrating its performance and potential for applications to real measurements

1. Butz, a., Guerlet, S., Hasekamp, O., Schepers, D., Galli, a., Aben, I., Frankenberg, C., et al. (2011). Toward accurate CO 2 and CH 4 observations from GOSAT. Geophysical Research Letters, 38(14), L14812. doi:10.1029/2011GL047888
2. Frankenberg, C., Meirink, J. F., Van Weele, M., Platt, U., & Wagner, T. (2005). Assessing methane emissions from global space-borne observations. Science (New York, N.Y.), 308(5724), 1010-4. doi:10.1126/science.1106644
3. Gilbert, J. C., & Lemarechal, C. (2009). The module M1QN3 version 3.3. Technical report, INRIA
4. Hestenes, M. R., & Stiefel, E. (1952). Methods of conjugate gradients for solving linear systems. Journal of Research of the National Bureau of Standards, 49(6), 409. doi:10.6028/jres.049.044
5. Meirink, J. F., Bergamaschi, P., & Krol, M. C. (2008). Four-dimensional variational data assimilation for inverse modelling of atmospheric methane emissions: method and comparison with synthesis inversion. Atmospheric Chemistry and Physics, 8(21), 6341-6353. Retrieved from
6. Schepers, D., Guerlet, S., Butz, a., Landgraf, J., Frankenberg, C., Hasekamp, O., Blavier, J.-F., et al. (2012). Methane retrievals from Greenhouse Gases Observing Satellite (GOSAT) shortwave infrared measurements: Performance comparison of proxy and physics retrieval algorithms. Journal of Geophysical Research, 117(D10), D10307. doi:10.1029/2012JD017549