THE APPLICATION OF MONIN-OBUKHOV LENGTH IN THE ATMOSPHERIC STABILITY CLASS DETERMINATION FOR AIR DISPERSION MODELS (INDIANA)

The selection of stability classification has a critical impact on the calculated ambient pollutant concentrations. In this thesis, two methods of atmospheric stability class determination, Pasquill/Turner classification scheme and Monin-Obukhov Length Method, are used for Environmental Protection Agency Models PTMTP and RAM (both rural and urban versions). Three air dispersion models with two methods for determination of stability give six models to be tested. Comparisons between model predictions and class actual field observation are based on pertinent statistical parameters, correlation coefficients, cumulative frequency distribution plots, and statistical tests (F-test and t-test) for 1-hour, 3-hour and 24-hour averaging times. The data bank including sulfur dioxide concentrations, meteorological parameters, plant emission parameters, and plant physical parameters come from 1975 Purdue SO(,2) project for five selected coal burning power stations throughout the State of Indiana. Except Fairbanks, lower arithmetic means were provided by all models for five locations for three different averaging times. The correlation coefficients shown by all models range from 0.06 to 0.44 for Edwardsport; from 0.07 to 0.25 for Evansville; from 0.001 to 0.2941 for Indianapolis; from 0.001 to 0.070 for Michigan; and from 0.35 to 0.80 for Fairbanks for three different averaging times. All models display much increase in correlation coefficients at upper percentile. For 1-hour averaging time, all models except RRP for Indianapolis gave correlation coefficients ranging from 0.82 to 0.99 from the statistics at 90 percentile. Both RUL for Indianapolis and PTL at Fairbanks showed the highest correlation coefficient, 0.99. Considerable underprediction at low level concentration are exhibited by all models for three different averaging times for five power station sites due to the more spread of dispersion parameters, (sigma)(,y) and (sigma)(,z), for locations away from plume center line based on Gaussian plume assumption. For each pair of model, the cumulative frequency distribution curve of the one using Monin-Obukhov length method more approached to the actual measurement curve to give higher predictions at low level concentrations than the one using Pasquill/Turner Method, especially for urban cities, Indianapolis and Michigan City. In other words, better prediction at low level concentration were observed by Monin-Obukhov length method, especially for urban city, since shear stress and heat flux considered by Monin-Obukhov Length are most important to urban city