DEVELOPMENT OF AN AIR DISPERSION MODEL FOR COASTAL ZONES AND COMPLEX TERRAINS

DEVELOPMENT OF AN AIR DISPERSION

MODEL FOR COASTAL ZONES AND

COMPLEX TERRAINS

Dr. D.Y.C. Leung

Department of Mechanical Engineering

HKU306/95E

Introduction

Various methods have been devised for the prediction of atmospheric pollution, which lead to over 100 types of model for different applications. The Gaussian air dispersion model, or its various segmented plume and puff advection progeny, is the most popular and widely adopted model in the world. These models, however, often fail to predict accurately under complex wind and turbulence patterns found in coastal zones and complex terrains such as those presence in Hong Kong and some other similar cities in the world. This study concerns about the development of a powerful air dispersion model for Hong Kong to replace traditional Gaussian models.

Objectives

Due to the inadequacy of the Gaussian models to predict the complex situation in Hong Kong, there is a genuine need to develop a more powerful and accurate air dispersion model for Hong Kong. Thus the main objectives of this project are:

1.	To develop a powerful and accurate air dispersion model to suit the complex topography of Hong Kong, and
2.	To validate, calibrate and tune the model developed.

Model features

The model consists of two separate modules:

Flow module

a mesoscale model is adopted to allow the flexibility to study local as well as longer range cross border pollutant transport issues;
provides detailed flow and turbulence information for the use of the dispersion model.

Dispersion module

calculates pollutant concentrations using the flow and turbulence data generated from the flow module;
able to simulate dispersion under different atmospheric stabilities.

Flow module

the Canadian Mesoscale Compressible Community Model (MC2) is adopted;
3-D fully compressible;
non-hydrostatic;
semi-implicit semi-Lagrangian numerical scheme;
a staggered grid with uniform horizontal and non-uniform vertical resolution;
orography: a terrain-following Gal-Chen coordinate system;
application: from large scale extra-tropical storms (50-100 km mesh) to smaller mesoscale events (2 km mesh or lower);

Simulation tests

(a)	(b)

Fig. 1 Simulated wind field & other parameters at the 1st level of inner domain. (a) wind field and vertical velocity; (b) wind field and temperature.

table1.jpg (32031 bytes)

(A) from the Hong Kong radiosonde measurements by Hong Kong Observatory

(B) simulated by the model at the cell (21,13) of the inner domain

Table 1. Simulated and measured wind, temperature and humidity at selected isobaric surfaces at 8:00 pm (local time) on 11 Oct. 1994.

Fig. 2 Simulated and measured wind field (2 km resolution).

Dispersion module

a 3-D second order closure model
semi-implicit Eulerian Scheme;
spatial / temporal domains solved by finite element / finite difference methods;
variable resolution for both vertical and horizontal grid;
topography handled by isoparametric finite element method;
parallel code developed to enhance model efficiency and resolution;
applicable to meso-scale pollutant dispersion problems under different atmospheric stabilities with different terrain configurations;

Simulation tests

FIGURE_3.jpg (73315 bytes)

Fig. 3 Normalized pollutant concentration under convective atmospheric boundary layer.

FIGURE_4.jpg (50599 bytes)

Fig. 4 Normalized plume height for different emission heights under convective atmospheric boundary layer.

FIGURE_5.jpg (38821 bytes)

Fig. 5 Centreline pollutant concentrations for ground level emission under convective atmospheric boundary layer.

FIGURE_6.jpg (58027 bytes)

Fig. 6 Normalized pollutant concentration for flow over a cone.

FIGURE_7.jpg (45757 bytes)

Fig. 7 Normalized centreline pollutant concentration for flow over different terrain configurations.

Conclusions

1.	A sophisticated air dispersion model is developed, which consists of 2 modules for wind field and dispersion calculation.
2.	The MC2 meteorological model has been selected and adapted to the H.K. conditions, which is proved capable of simulating high resolution 3-D fields of wind and other meteorological parameters, using reasonable computer time and memory resources.
3.	A 3-D second order dispersion model is developed and validated extensively with experimental and field measurement results.

Funding Source: RGC $641,000

Co-investigator: Prof. A.T.Y. Chwang

Research staff: Dr. M. Niewadomski, Dr. C.H. Liu, Mr. W.K. Tse