風場推估對ISC3空氣品質模式模擬影響分析

Abstract

國內開發案及空污分析經常採用ISC3模式評估污染源對週遭空氣品質之影響，唯若模擬時使用不適當的資料，易導致所得結果不正確及降低可信度，風速風向是空氣品質模擬的重要資料，然而由於氣象站站址有限，導致污染源附近往往無氣象站，而一般直接採用鄰近氣象測站資料的作法可能與真正風場有相當的誤差，因此有必要探討適當的風場推估方法。本研究收集Diagnostic 及Prognostic 兩類常用的風場推估方式，選擇四種推估方法：（1）Draxler法；（2）MINERVE 法－Cressman步驟；（3）MINERVE 法－MacLain步驟；（4）Objective analysis法（客觀分析），比較不同方法所推估結果之適用性。本研究並以數個案例區進行空氣污染物增量模擬以比較不同風場推估方法之優劣點，結果顯示在一些案例中，現行氣象站勢力範圍法所得結果與實際風場資料有相當差異，而四種風場推估方法無論是在風場或ISC3模擬均可得到較佳的推估結果，其中尤以MINERVE-Maclain推估方法最為接近污染源現地結果。

The ISC model is commonly applied for assessing air quality impacts of a potential pollution source on its adjacent areas. The model may produce incorrect or unreliable results while inappropriate model parameters are provided. The wind direction and speed are important parameters for air quality model simulation. However, a meteorological station is generally not available in close vicinity to a pollution source and data from the closest station are generally adopted. If the station is in a significant distance away from the source, the data collected on the station may not resemble the true wind field for the source. An appropriate method is therefore desired to give a good approximation based on data collected from adjacent monitoring stations. Among two common categories of diagnostic and prognostic methods, four wind field approximation methods are compared in this study: (1) the Draxler method; (2) The MINERVE method with the Cressman procedure; (3) The MINERVE method with the MacLain procedure; and (4) Objective analysis method. Several case studies were implemented to analyze the suitability of these four methods for use with the ISC model. Comparisons were made based on data generated from the four methods for wind field approximations and subsequent air quality simulation. According to the results obtained for case studies, significant differences are observed between the results obtained based on the wind fields adopted from the closest station and those based on the true wind fields. The four methods are capable in improving both the wind field approximation and ISC modeling, while the MINERVE-MacLain method performs the best.