| Summary: | 碩士 === 國立臺灣大學 === 大氣科學研究所 === 90 === This thesis dedicated to study the imfluence of meteorological simulation with regard to modeling atmospheric aerosols. Series of numerical experiments made by PSU/NCAR Fifth Generation Mesoscale Model (“MM5”) and Taiwan Air Quality Model (“TAQM”) have been carried out with three major scientific goals: (1) to study the impact of assimilating additional meteorological observational data; (2) to study the discrepancy of applying different cumulus parameterizations; (3) to study the discrepancy of applying different PBL and surface schemes. The results demonstrated that the uncertainty of modeling air quality fraught with meteorological simulation is significant.
Numerical simulations during the selected air pollution episode (Nov. 21~28, 1996) present that assimilating additional meteorological observational data can reinforce the data quality of the initial condition of the model, but does not necessarily improve the simulation. In addition, assimilating the additional soundings obtained from IOP-1 (Intense Observing Period-1, Dec. 10~15, 1998) leads to similar conclusion. FDDA (Four Dimensional Data Assimilation), however, could modify the false simulation caused by lateral boundary perturbation. The PV diagnostics has been applied to quantitatively evaluate its impact upon regional circulation around Taiwan. With respect to the sensitivity test of MM5 physical schemes, the results showed that the meteorological field is not sensitive to the choice of cumulus parameterization in this case; the only notable difference is a 12-hour time lag among simulations transporting aerosols from mailand China to Taiwan while cumulus parameterizations varied. Adopting PBL (Planetary Boundary Layer) and surface schemes created diverse local meteorological fields and PBL characteristics and thereby affected aerosol’s behavior. Additionally, different PBL and surface scheme can cause a five-time difference of yellow sand emission rate calculated by the Yellow Sand Assessment Module.
Since the uncertainty of modeling atmospheric aerosols due to meteorological simulation has been ackenowledged, this study suggests that (1) optimizing physical schemes; and (2) identifying the most sensitive area to obtain informative data for modeling are two necessary for improving air quality modeling. Notably, the effect of any nonlinear interactions between physical schemes in MM5 and chemical transformations in TAQM remains unknown that requires further study.
|
|---|
