Modeling impact of peak traffic and street geometry from transportation in Street Canyons - Case Analysis - Singapore

PREETY MUKHERJEE*#, LIM CHENG CHOON*@, YEE WEN LOH#, KAM WUI WUI#, THET MON

AYE#, NG WAI LING#

*# Lecturer, Ngee Ann Polytechnic, Building and Environment Divison, School of Engineering

@ Senior Engineer, National Environment Agency, Planning & Development Department # Students, Ngee Ann Polytechnic, Building and Environment Divison, School of Engineering

SINGAPORE

Abstract: Street Canyon Module of a regional scale dispersion model Indic Airviro was used to simulate series of peak hourly carbon monoxide (CO) across several meteorological conditions through a year using a dynamic emission database at two canyons at the central business district [CBD], Singapore.Worst case peak hourly CO with different street geometry for Pasquill stable meteorological condition at different wind speeds keeping the wind direction constant were also simulated. The average hourly total traffic flow and speed was obtained from Land Transport Authority, Singapore (LTA). The Singapore fleet average emission factors for each vehicle category [US EPA MOBILE 5 A] from an earlier study were used. On-site vehicle counts specific within the canyons were conducted using manual counters to use a representative, realistic dynamic hourly CO emission base representative at the 2 sites. Key-Words: Street Canyon, emission factors, vehicle, CO, emission database, modeling, worst case, stability class, wind speed 1 Introduction Traffic in Singapore is responsible for over 50 % of the total pollution [2, 3, 21, 43, 59]. Motor vehicle related pollution in Singapore has primarily related to diesel particulates, with CO levels mentioned as a key issue of concern [27]. CO was selected as a source fingerprint since it is inert and found to be the best tracer for the vehicular part of pollution in Singapore, with traffic being responsible for 97 % of CO emissions. Singapore with a level land area of 42 km by 23 km in the main island is a predominantly urban location. Street Canyon sections with vehicles in the city center, have limited and helical air circulation resulting in a unique distribution of pollutants across different sections within the canyon. The street geometry induces a wind flow pattern which undergoes recirculation within the canyon resulting in higher concentration of pollutants as compared with relatively open areas. The direction of the prevailing wind in relation to the street direction plays an important role in the recirculation of the wind in the canyon, affecting concentrations. Studies around the world have concentrated on empirical basis and on site [67, 42, 94, 74, 70, 31, 6, 77,

33, 58, 57, 45, 12, 35, 62, 49, 38, 4, 22, 56, 87, 34, 17, 39, 40, 13, 79, 82, 83, 20, 78, 44, 19, 5, 50, 61, 76, 18, 36, 86, 23, 32, 60, 37, 28, 11, 24, 8, 63, 64, 73, 75, 80, 81, 85, 88, 93] measurement with correlation to arrive at a pattern of concentration distribution within a street canyon section or mathematical and numerical modeling. This study tries to find the impact of worst case traffic flow with a very stable meteorological stability class at different wind speeds, and same wind direction at two parallel street canyon sections with a different height to width ratio at ground level and across the vertical profile within the two canyons. Peak hourly CO concentration over a series of different meteorological conditions is simulated over time at both the canyons as a comparison. 2 Description of the Model Regional scale dispersion model-Indic Airviro developed by the Swedish Meteorological and Hydrological Institute is used. This model is used by the National Environment Agency, Singapore [NEA],

2

through the Telemetric Air Quality Monitoring and Management System (TAQMMS) for monitoring the ambient pollutant levels at the existing monitoring station sites. For this study, all of the 3 functional blocks of the model, namely the Emission Database (EDB) and the Dispersion module and the Indico Administration for storing online ambient air quality and meteorological parameters from the monitoring stations were used. Modeling requires detailed collection of emission factors, traffic flow variations as a function of speed and time, and a coordinate system for describing the roads which is stored and processed under the EDB database. The dispersion module performs the dispersion of the pollutants using a prescribed model with either measured/assumed meteorological condition from the Indico Administration module. Street Canyon module was used for dispersion of CO due to mobile emissions. This module based on the "Stanford Model" originated from studies conducted in San Jose, USA, where measurements indicated a vertical eddy circulation whenever the wind blew perpendicular to the street directions [70, 88, 25, 63, 64, 26]. Canyon is defined as the section of the road between buildings on two opposite sides of the road. The model parameters describing the concentrations in lee wind and upwind side are given by:

( ) ( )

+++

=0

21225.0 LzxU

QKCL (1)

( )( ) HUW

zHQKCW

+=

5.0..(2)

where, CL and CW refers to concentration on the lee wind and upwind side; K is an empirical constant, set to 10; Q is the total traffic emission in g/m s; U is the roof level wind speed, in m/s; Lo is a length scale of the individual cars, set to 2m; W is the effective width of the canyon; H the typical building height in m; x, is the horizontal distance from the street emission segment; z is the receptor height in m Model calculations as described above are valid for all wind directions, except for those within 22.5 degrees of the principal canyon axis direction, at which the concentration is computed as the average of equation 1 and 2. The wind profile at the rooftop level at the given Canyon is usually extrapolated from the Changi Airport

meteorological station, located approximately 20 kms from the chosen canyon sections. For the worst case study we force the street canyon module to enforce assumed wind speed profiles under very stable stability class to generate CO concentration profile within the canyon, whereas for the other series of simulations through a year, the actual hourly meteorological conditions at the canyon sites are used for modeling. 3 Traffic Volume, Speed & Emission factors for the Selected Canyon Sites The canyon site in the CBD area of Singapore is chosen, in conjunction with NEA and LTA, based on the street geometry and availability of overall traffic volume and speed. Singapore fleet average emission factors based on US EPA Mobile 5 A, developed earlier [1, 53,54, 55] were used to compute the CO emission inventory. CO is chosen as a tracer pollutant, due to maximum CO emissions from petrol engine vehicles, a characteristic of typical vehicle type within chosen canyons. CO is a relatively stable pollutant for the reaction rates within street canyons [50, 53, 54, 55, 24, 23, 9, 19]. CO emissions are a function of the speed, and accounts for average speed on a given road between 20 and 105 kmph [53, 54, 55]. The site of the canyon on Cecil Street is between PIL Building and Keck Seng Tower and for Robinson Road between SIF Building and Robinson Pt [Fig. 1]. The trend of traffic flow pattern at the canyons is shown in Table.1. Cecil street shows higher traffic volume. Detailed dynamic emission inventory at the site used an hourly manual count of the vehicle composition using manual counters conducted for 6 weekdays through Dec 2003- March 2004. The detailed vehicle composition distribution is in Table.2. For worst case with stable stability class, Cecil street traffic flow was used to compute ambient CO at both the canyons.

Fig: 1 Cecil Street & Robinson Road Canyon (Adapted: Maps Powered by Streetdirectory.com)

3

Table 1. Traffic Volume at Cecil & Robinson Road Time of Day Cecil Street Robinson Rd Time of Day Cecil Street Robinson Rd

Ave Traffic Flow Ave Traffic Flow Ave Traffic Flow Ave Traffic Flow12am - 1am 254 375 12pm - 1pm 1695 1473 1am - 2am 153 224 1pm - 2pm 1441 12532am - 3am 118 169 2pm - 3pm 1668 13633am - 4am 84 133 3pm - 4pm 1540 12204am - 5am 52 92 4pm - 5pm 1343 11155am - 6am 69 81 5pm - 6pm 1361 13116am - 7am 161 193 6pm - 7pm 1530 15707am - 8am 773 764 7pm - 8pm 1383 17378am - 9am 1696 1325 8pm - 9pm 866 1116

9am - 10am 1488 1175 9pm - 10pm 721 94810am - 11am 1545 1243 10pm - 11pm 601 80811am - 12pm 1682 1396 11pm - 12am 374 615

Table 2. % Vehicle Composition [Manual counting]

Time of Day Cars Taxi Bus Cars Taxi Bus7am - 8am 74 20 6 74 20 68am - 9am 73 22 6 74 20 99am - 10am 51 44 5 56 40 410am - 11am 57 39 4 58 37 511am - 12pm 63 35 3 66 30 412pm - 1pm 64 33 3 70 27 41pm - 2pm 64 33 4 64 33 32pm - 3pm 70 27 4 66 31 33pm - 4pm 67 30 3 68 28 44pm - 5pm 70 27 4 69 24 85pm - 6pm 71 24 5 82 12 66pm - 7pm 85 12 4 82 15 37pm - 8pm 71 26 3 72 24 48pm - 9pm 68 28 4 61 34 59pm - 10pm 69 28 3 61 36 3

Cecil Street % Vehicle type Robinson Rd % Vehicle type

4 Street Canyon Geometry The canyon geometry at the two sites is as shown in Table 3. The two canyons give a different height to width ratio [42, 74, 76, 35, 20, 5] but the ratio satisfies the criteria set by AIRVIRO model (ratio between 2 to 3) [71]. The two roads are parallel to each other (Fig.1). Table. 3 Street Geometry at the two canyon sites

Parameters Cecil Street Robinson RoadRoad Width, m 13.6 17 Street Width (Road+Pavement), m

24 29.8

Avg. Building Height, m

58 47.9

Height to Width ratio 2.42 1.61

5 Ambient CO Modeling using Street Canyon Module The computed hourly average dynamic emission inventory for CO, along with the street canyon geometr