Studying the uncertainty in computationally expensive models has required the development of specialized methods, including alternative sampling techniques and response surface approaches. However, existing techniques for response surface development break down when the model being studied exhibits discontinuities or bifurcations. One uncertain variable that exhibits this behavior is the thermohaline circulation (THC) as modeled in three-dimensional general circulation models. This is a critical uncertainty for climate change policy studies. We investigate the development of a response surface for studying uncertainty in THC using the Deterministic Equivalent Modeling Method, a stochastic technique using expansions in orthogonal polynomials. We show that this approach is unable to reasonably approximate the model response. We demonstrate an alternative representation that accurately simulates the model's response, using a basis function with properties similar to the model's response over the uncertain parameter space. This indicates useful directions for future methodological improvements.

© 2006 Baywood Publishing Co., Inc.

Methane (CH4) and carbon dioxide (CO2) are the two most radiatively important greenhouse gases attributable to human activity. Large uncertainties in their source and sink magnitudes currently exist. We estimate global methane surface emissions between 1996 and 2001, using a top-down approach that combines observed and simulated atmospheric CH4 concentrations. As a secondary study, we describe our participation in a CO2 inverse-modeling intercomparison.

The available methane time-series data used in this work include observations from 13 high-frequency stations (in-situ) and 74 low-frequency sites (flask). We also construct an annually-repeating reference emissions field from pre-existing datasets of individual methane processes. For our forward simulations, we use the 3-D global chemical transport model MATCH driven by NCEP meteorology. A prescribed, annually-repeating OH field scaled to fit methyl chloroform observations is used as the methane sink. A total methane source of approximately 600 Tg/yr best reproduces the methane growth rate between 1993-2001. Using the reference emissions, MATCH can reproduce the observed methane variations at many sites. Interannual variations in transport, including those associated with ENSO and the NAO, are found to be important at certain locations.

We adapt the Kalman Filter to estimate methane flux magnitudes and uncertainties between 1996 and 2001. Seven seasonal processes (3 wetland, rice, and 3 biomass burning) are optimized at each month, while three aseasonal processes (animals/waste, coal, and gas) are optimized as constant emissions. These optimized emissions represent adjustments to the reference emissions. For the entire period, the inversion reduces coal and gas emissions, and increases rice and biomass burning emissions. The optimized seasonal emission has a strong peak in July, largely due to increased emissions from rice producing regions. The inversion also attributes the large 1998 increase in atmospheric CH4 to global wetland emissions, consistent with a bottom-up study based on a wetland process model. The current observational network can significantly constrain northern emitting regions, but is less effective at constraining tropical emitting regions due to limited observations. We further assessed the inversion sensitivity to different observing sites and model sampling strategies. Better estimates of global OH fluctuations are also necessary to fully describe the interannual behavior of methane observations. Carbon dioxide inversions were conducted as part of the Transcom 3 (Level 1) modeling intercomparison. We further explored the sensitivity of our CO2 inversion results to different parameters.

The study of the uncertainties in future climate projections requires large ensembles of simulations with different values of model characteristics that define its response to external forcing. These characteristics include climate sensitivity, strength of aerosol forcing and the rate of ocean heat uptake. The latter can be easily varied over a wide range in an anomaly diffusing ocean model (ADOM). The rate of heat uptake in a three-dimensional ocean general circulation model (OGCM) is, however, defined by a large number of factors and is far more difficult to vary. The range of the rate of the oceanic heat uptake produced by existing Atmosphere-Ocean General Circulation Models (AOGCMs) is narrower than the range suggested by available observations. As a result, simpler models, like an ADOM, are useful in probabilistic climate forecast type studies as they can take into account the full uncertainty in ocean heat uptake.
        To evaluate the performance of the ADOM on different time scales we compare results of simulations with two versions of the MIT Integrated Global System Model (IGSM): one with an ADOM and the second with a full three dimensional OCGM. Our results show that in spite of its inability to depict feedbacks associated with the changes in the ocean circulation and a very simple parameterization of the ocean carbon cycle, the version of the IGSM with ADOM is able to reproduce important aspects of the climate response simulated by the version with the OCGM through the 20th and 21st century and can be used to obtain probability distributions of changes in many of the important climate variables, such as surface air temperature and sea level, through the end of 21st century. On the other hand, the ADOM is not able to reproduce results for longer term climate change and specifically those concerned with details of the feedbacks on the heat and carbon storage. Such studies will require the use of the OGCM and uncertainties in those results will be limited.

This paper summarizes the main features of a linked CGE-MARKAL model system capable of simulating the macro-level economy and micro-level technology detail of the transport sector. Emphasis is given to issues of calibration of such a hybrid system, with references provided to already published papers based on this research for coverage of other details.

Copyright IAEE