MIT Center for Global Change Science
| Report Series | Contact |
April 1996
Submitted to the Dept. of Earth, Atmospheric and Planetary Sciences on April 4, 1996 in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Meteorology
To obtain a copy of this Report in paper format, use the request form
or send an e-mail to
cgcs@mit.edu, specify Report 42, and provide a complete return postal address.
The surface fluxes of many radiatively and chemically important species are uncertain to a factor of 2-3, e.g., carbon dioxide or methane [IPCC, 1994]. In order to gain more information about the sources and sinks, inverse methods have been used in combination with chemical transport models and observations of the concentrations of trace gases at the surface [e.g., Hartley and Prinn, 1993]. These inverse methods rely critically upon the ability of the chemical transport model to accurately simulate the transport in the atmosphere. The goal of the thesis was the development of a chemical transport model (CTM) based on analyzed observed winds capable of simulating the transport of trace gases accurately enough to use an inverse method to determine the surface sources of important gases. The model was developed to be used with forecast center winds, either European Center for Medium-Range Weather Forecasting (ECMWF) or National Meteorological Center (NMC) reanalyses. These winds represent a combination of meteorological observations as well as numerical weather prediction model outputs, and represent the best information we have about the state of the global atmosphere at a given time. Unfortunately, the forecast centers do not archive mixing coefficients for important sub-grid scale mixing processes, so that the CTM must be capable of deriving both boundary layer mixing and moist convective transports. In the model developed here, which we name the MATCH model (Model of Atmospheric Transport and CHemistry), there is the ability to derive moist convective transport using one of three schemes: Hack [1992], Tiedtke [1989] or the Pan [NMC, personal communication, 1995] schemes. The non-local boundary layer scheme from Holtslag and Boville [1992] is used. The model is capable of model simulations at either T42 (approximately 2.8 by 2.8 degrees), with 19 layers for the ECMWF data or at T63 (approximately 1.8 by 1.8 degrees) with 28 vertical layers for the NMC reanalysis data.
The ability of the model to simulate interhemispheric transport, the seasonal cycle in transports and synoptic scale events is explored by comparing simulations of 222Rn and CCl3F with observations at surface stations (and upper level data where available). The model in general compares quite well with observations using either the ECMWF or NMC data. However, the ECMWF-based model results indicate too strong of an interhemispheric gradient in the CCl3F, which is not seen in the NMC-based model results.
The model is used in conjunction with an inverse method (recursive weighted least squares) to determine the sources of CCl3F using in the model either the ECMWF or NMC meteorological data. Here, the emissions of CCl3F are reasonably well known for the period considered (1990-91) and global CCl3F measurements are available from two networks, so that this serves as a test of the ability of the model to be used in such an inverse methodology. The results showed mixed success: the inverse method and the model, using NMC winds could deduce the correct global total emissions, and hemispheric total emissions, but was unable to distinguish between emissions from similar latitudes.
