SENSITIVE ANALYSIS OF CALINE 4 HIGHWAY DISPERSION MODEL UNVER MIXED TRAFIC CONDITIONS

Abstract

Rapid urbanization and industrialization of cities have increased the vehicular traffic leading to increase in air pollution in urban areas. It has been estimated that in India road traffic contributes approximately 70% of air pollution in urban areas. To reduce the impacts of air pollution due to vehicular traffic, it is important to manage and improve the quality of air in such urban areas. Air pollution dispersion models are used to effectively and efficiently plan the management (environment management plan) of vehicular traffic pollution on particular area/ road corridor, along with monitoring of air pollutants. They not only aid in determining the present influenced area/ affected due to vehicular traffic pollution but also help in identifying the future scenarios under different traffic and meteorological conditions made by these models. Vehicular pollution modeling involves air pollution prediction estimates by simulating impact of emissions from vehicular activities in a given region under specified traffic and meteorological conditions. Throughout the world, including India the prediction of vehicular pollutant concentrations along highways and roads are carried out by using various Gaussian-based highway dispersion models. Based on the Gaussian dispersion model, several prediction models have been developed to predict vehicular pollution levels along the highways. The most popular amongst various highway dispersion models, are the CALINE model (latest being CALINE 4). CALINE 4 developed by Benson (1984) is extensively used throughout the world (including India) for various vehicular pollution estimate/ prediction along the highways. The CALINE 4 Model uses various inputs (viz., Traffic Volume, Emission Factor, Road geometry, Wind Speed, Wind Direction, Background Concentration) to predict the air pollution concentrations at pre-identified receptor locations along the highway. The present study focuses on sensitivity analysis of CALINE-4 model which is the fourth version simple line source Gaussian plume dispersion model. Ashram Chowk – CRRI highway Corridor of NH-2 was selected as the area of study. Inputs data (viz. traffic volume, traffic compositions, meteorological data etc.) required for CALINE 4 model was collected from field surveys data. Emission factors provided by CPCB (2000) and ARAI (2007) were used to estimate Weighted Emission Factor (WEF) to account for mixed traffic conditions. The CO concentration due to traffic along the xiii Ashram Chowk – CRRI highway corridor was predicted at the pre-identified receptor locations. The dispersion of CO concentrations was found to be present upto a distance of 150m from the edge of the mixing zone width (road width+3m on each side of the road). The predicted CO concentrations in all the cases (viz., 1-hour Standard Case Run Conditions, 1-hour Worse Case Run Conditions) were within the National Ambient Air Quality Standard, 2009 (NAAQS, 2009) (i.e. 2 mg/m3 for 8 hours and 4 mg/m3 for 1 hour for CO). The regression coefficient (r2) between predicted and observed 1-hour CO concentrations using CPCB emission factors for Standard Case Run Condition was 0.65 and for Worse Case Run Condition was 0.76. Similarly, the regression coefficient (r2) between predicted and observed 1-hour CO concentrations using ARAI emission factors for Standard Case Run Condition was 0.60 and for Worse Case Run Condition was 0.73. A sensitivity analysis of the CALINE 4 model had been performed to identify the most influential variables. CALINE 4 model was found to be relatively sensitive to wind angle (s) for small receptor distances. The highest CO concentrations were observed by a wind angle of ~10° as measured from the road centerline. Wind speed had a considerable effect, e.g., predicted CO concentrations were dropped by 75% - 80% as wind speed increased from 0.5 to 5 m/s. From unstable to stable conditions, average increase in CO concentration was 43%. The model consistently predicts lower CO concentrations for greater highway widths. This effect was most apparent for receptors near the roadway edge. Roadway height (from receptor location at ground level) had very less effect for small change in height but has considerable effect for more deeper or elevated roadway height. Sensitive Analysis of CALINE 4 had also revealed that among various input variable, source strength, wind speed, highway width and median width were most significant input variable and wind direction, roadway height, distance of receptor to roadway and atmospheric stability were the less significant input variables. Surface Roughness and Mixing height had negligible effect on predicted CO concentrations.