UK Consortium Project funded by the EPSRC S. Arnold (1&2), H. ApSimon (1), J. Barlow (2), S. Belcher (2), M. Bell (3), R. Britter (4), H. Cheng (5), R. Colvile (1), A. Dobre (2), S. Kaur (1), D. Martin (6), M. Neophytou (4), G. Nickless (6), C. Price (6), A. Robins (5), D. Shallcross (6), R. Smalley (3), J. Tate (3), A. Tomlin (3). (1) Imperial College London; (2) University of Reading; (3) University of Leeds; (4) University of Cambridge; (5) University of Surrey; (6) University of Bristol. MARYLEBONE RD FIELD CAMPAIGNS Two DAPPLE field campaigns, April-May 2003 and April-June 2004, were undertaken at the intersection of Marylebone Road and Gloucester Place, Westminster, Central London, UK. Marylebone Road is a 7 lane dual carriageway approximately 38 m wide running WSW-ENE and Gloucester Place is a 3 lane road approximately 20 m wide which is one way to the North. Building heights vary between m. Fig. 1. identifies the main DAPPLE measurement sites at the intersection. Data from these sites include: ● Wind field – up to 11 x 3-component ultrasonic anemometers were deployed at heights ranging from 1.5–180 m, within and above the street canyon intersection measuring the mean and turbulent (20 Hz) flow (Fig. 2). ● Pollution levels – 15 CO Learian street boxes were deployed in the intersection and surrounding streets between 4-10 m; vertical pollution gradients of CO, CO 2, O 3 and PM 2.5 were made at WCC between 0-15 m (Site 11); ● Personal exposure – measurements of CO, PM 2.5 and ultra-fines were made by people carrying instrumentation through the study domain at different locations on the pavement and via different routes and modes of travel (Fig. 7a); ● Traffic flow – manual traffic count validations of the SCOOT system were undertaken and will be developed into a micro-traffic flow simulation model; ● Tracer releases – inert tracers (SF6 and PMCH) were released in SSW winds from vehicles parked in York Street (Sites X1 & X2) and from WCC roof (Site X3). Time series (10 x 3 mins bags) of air samples at 16 locations throughout the study area were used to track the tracers (Fig. 6). INTRODUCTION The aim of the DAPPLE project is to enhance understanding of dispersion processes, over short distances, at a street canyon intersection. This information will be used to make improvements in predictive ability that will enable better planning and management of urban air quality, accidental and non-accidental releases, and the development of safer more sustainable cities. DAPPLE brings together a multidisciplinary UK consortium undertaking research in the field, wind tunnel, and via numerical simulations in order to provide a better understanding of the physical processes affecting the flow of air, traffic, people and their corresponding interactions with the dispersion of pollutants. This poster introduces the DAPPLE field measurements and a sample of the preliminary results. Fig. 1. Site plan of the DAPPLE field site. Westminster City Council (WCC), the base for the field campaign, is on the SW corner of the intersection. The sites were: 17 fixed instrument locations; 3 tracer release positions with up to 16 receptors; and mobile personal exposure and traffic flow measurements. Note: Below Fig. 1. has North in a conventional vertical orientation. The poster background image, taken from the BT tower (Site 180 m), is looking towards the DAPPLE site and so has an approximate westerly orientation. Fig. 2. Site 1, ultra sonic anemometer and street box on a lamppost in the central reservation of Marylebone Rd. WIND FIELD RESULTS – METEOROLOGY AND POLLUTION The flow at the intersection during oblique rooftop winds can be explained by the linear superposition of the parallel and perpendicular rooftop components; the parallel component giving the direction of channelling and the perpendicular component driving the in-street recirculation vortices (evidence for which can also be seen in the CO data from the street boxes, Fig. 3). The combination of the two wind component effects can also be used to explain in-street helical vortices (sometimes described as reflection phenomenon). Fig. 4a. shows a day when the rooftop winds (Site 10) change from a positive (SSW before 12:00) to a negative (W after 15:00) approach angle with respect to Marylebone Road (rotated to be 0 degrees). Despite the actual variation in wind direction being small the change in the approach sign results in a 180º reversal of the flow in Gloucester Place (Site 5). Sites 2 & 5, due to their constrained canyon locations, both exhibit channelled flow with single peak directional PDF’s (not shown). Site 1 (Fig. 2) is more open in westerly approach winds and produces a two peak PDF (Fig. 4b); the flow switching between Marylebone Road (0° - westerly) and Gloucester Place (-100° - southerly). TRACER RELEASE RESULTS – 15 MAY 2003 The concentration of tracer decreases with increasing fetch (R, which was between m) from the release site (Figs. 5a). There is an upper band to the maximum tracer concentration as a function of separation, C max = XR -2 (Fig. 5b). This relationship has also been documented in both the DAPPLE wind tunnel experiments and the numerical modelling results. The exact value of X in the equation does however vary. During the release the rooftop winds (U ref ) were approx. 3 ms -1 hence the travel speed to WCC (R=115 m) was 0.13U ref with a time of flight (t start - t 50%conc. ) of 4.5 mins. There was rapid vertical mixing of the tracer; the roof top and ground level sites at WCC recording similar concentrations (Fig. 5a). At all sites there is a coherent decrease in concentration mid-way through the experimental period. This can be related to a variation in wind direction from SSW to SW which may have switched the dominant wind direction at the intersections from a mainly southerly to a westerly flow. This would have temporarily diverted the passage of the tracer away from the Marylebone intersection and the majority of the receptors. Fig. 7a. Personal exposure monitoring across Marylebone pavement. The Health and Safety Laboratory visualisation equipment is in the pram (centre). Fig. 7b. An example still of the synchronised output from the video imagery and Ptrak particulate data. This information is used to relate exposure and person activity. CONCLUSIONS Despite the complexity of the urban intersection topology the main features of the mean wind field can be related to those described in idealised 2-D canyons. This information can be used to interpret in-street, time averaged, tracer and pollution concentrations. Further work is underway to relate these data to the finer scale, real-time, non-static, personal exposure measurements made during DAPPLE. For more details and publications please see Fig. 4a. Wind direction time series for the 9 th May Decomposition of the rooftop winds (which change from SSW to W during the day) is used to explain the in-street flows. Fig. 4b. Directional PDF for Site 1. Note, the 2 peaks would only be of even size for an approach flow of 45° to Marylebone Rd; lesser or greater angles produce a dominance in Marylebone Rd or Gloucester Pl, respectively. Fig. 5a. Non-dimensional tracer concentrations (C* = CU ref H 2 /Q, where: C is the mean concentration; U ref is ref wind speed at WCC Site 10; H is average building height (22 m); and Q is quantity of release (114 mg)) through time for 10 receptor sites. Fig. 5b. Non-dimensional concentrations as a function of separation R/H (where R is the straight line fetch/distance). Fig. 3. Street box data from the North (Site 3) and South (Site 4) pavements, and the central reservation (Site 2) of Marylebone Rd. Note, the cross-street variation in CO in the afternoon implying a street canyon vortex during oblique rooftop winds. Fig. 6. Blue tracer receptor box on Marylebone Road (near site 3) looking SW towards the WCC building. Marylebone Rd Fig. 7a. Fig. 7b. Fig. 5a. Fig. 5b. Fig. 4a. Fig. 4b. R= 75 m, Thornton Place R=115 m, WCC roof R=115 m, WCC ground R=120 m, Bickenhall St R=200 m, Marylebone Rd R=200 m, Bickenhall Mansions R=275 m, Baker Street