Analysis of Joint Urban 2003 (JU2003) and Madison Square Garden 2005 (MSG05) meteorological and tracer data

As a first step towards a comprehensive analysis of the Oklahoma City Joint Urban 2003 (JU2003) data and the Manhattan Madison Square Garden 2005 (MSG05) urban field data, the authors have been analyzing the wind and tracer data and developing and testing basic scientific relations. The ultimate goals are to increase understanding of urban flow and dispersion, evaluate dispersion models with the data, and provide guidance to emergency responders.

The hourly mean and turbulent wind observations from over 150 anemometer sites in the JU2003 domain have been calculated and are separated into six groups defined by terms such as "near street level in built-up downtown area" or "rural area outside of city". The mean speeds and directions and the turbulent velocities for each of the ten JU2003 Intensive Operating Periods (IOPs) and for the entire period have been calculated and compared for each group. It is found that the turbulent velocities are much more robust and amenable to similarity analyses than the mean winds, which often have very small magnitudes at street level.

The SF6 tracer concentration observations from JU2003 have been analyzed following the methods applied earlier to the Urban 2000 data from Salt Lake City. Values of uCmax/Q were calculated for each release and arc distance, where Cmax is the 30-min average arc maximum concentration, Q is the continuous source emission rate, and u is the spatial averaged wind speed in the downtown area. The basic characteristics of the plot of averaged uCmax/Q agree reasonably well with similar plots for the Urban 2000 data. For example, the slope with distance is the same. Even though the magnitudes of the JU2003 uCmax/Q values are about a factor of two below those for Urban 2000, the JU2003 magnitudes are similar to those found in the recent London DAPPLE tracer experiment.

The science goals for MSG05, which took place in Manhattan on 10 and 14 March 2005, are to increase understanding of flow and dispersion in deep urban canyons and of rapid vertical transport and dispersion in recirculating eddies adjacent to very tall buildings in a large urban area. The average building heights in Manhattan are about four times what they are in Oklahoma City, and Manhattan is about four or five times broader in size. The two IOPs took place with six different PFT tracer gas releases near street level at five locations around MSG, with gas samplers at street level on two concentric circles at approximate distances of 200 and 400 m, and at rooftop on two tall buildings. Supporting meteorological instruments included seven sonic anemometers at street level, several sonic anemometers on building roofs, a minisodar in Manhattan and a wind profiler upwind in Hoboken, NJ. The concentration observations suggest that the plume initially spreads upwind and laterally for one or two blocks, due to the downdrafts at the windward edge of the tall buildings. Comparisons of observed street level wind flows and concentrations with CFD simulations are reasonable. The following guidance is suggested for emergency managers in Manhattan (or any other very large city): • The wind directions at street level can be any direction and are not necessarily the same as at rooftop level. • Plumes released at street level can initially be transported any direction (upwind and laterally as well as downwind) for typically one or two building heights (a block or two in Manhattan). • Strong vertical motions and mixing occur around very tall buildings, characterized by downdrafts and street-level divergence, and updrafts and street-level convergence, on the windward and leeward sides, respectively. • After a block or two of travel, the plume becomes relatively well mixed vertically over the depth of the buildings and is transported at all levels in the direction of the mean flow at the rooftops, but with a very broad lateral spread (covering a 45 to 90 degree sector).