Diffusion from Low-Level Urban Sources: Reexamination Using Recently Available Experimental Data

James L. McElroy U.S. Environmental Protection Agency, Las Vegas, Nevada

Search for other papers by James L. McElroy in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Experimental meteorological tracer data recently collected or declassified concerning dispersion from low-level sources in urban areas are examined in terms of the findings of the St. Louis Dispersion Study. The latter still provides a standard for use in urban air quality models and a basis of comparison for subsequent studies. Collectively, the results indicated that the quantitative findings of the St. Louis project are still valid. However, a tendency for the experimental data to be organized in terms of the local land use, especially nearer the tracer release points, provides evidence that it may be possible to catalog the resulting dispersion parameters in terms of land use and hence decrease the scatter of data points used to develop products such as stability related dispersion curves.

Corresponding author address: Dr. James L. McElroy, 82 Pheasant Run Drive, Sequim, WA 98382.

jasm@tenforward.com

Abstract

Experimental meteorological tracer data recently collected or declassified concerning dispersion from low-level sources in urban areas are examined in terms of the findings of the St. Louis Dispersion Study. The latter still provides a standard for use in urban air quality models and a basis of comparison for subsequent studies. Collectively, the results indicated that the quantitative findings of the St. Louis project are still valid. However, a tendency for the experimental data to be organized in terms of the local land use, especially nearer the tracer release points, provides evidence that it may be possible to catalog the resulting dispersion parameters in terms of land use and hence decrease the scatter of data points used to develop products such as stability related dispersion curves.

Corresponding author address: Dr. James L. McElroy, 82 Pheasant Run Drive, Sequim, WA 98382.

jasm@tenforward.com

Introduction

For a number of decades prior to and including the 1960s, experimental work, chiefly with meteorological tracers, had been performed to describe atmospheric diffusion and its relationship to meteorological parameters. Many of the experiments were concerned with the potential consequences of nuclear reactor accidents or the release of chemical or biological agents into the atmosphere. With few exceptions, such work had been carried out over relatively flat and uncomplicated terrain so that results would not be strongly dependent upon details of site location. Much of the work dealing with continuous, near-ground-level releases from point sources was summarized and parameterized by the British Meteorological Office in terms of stability categories and readily available meteorological data for use in Gaussian plume equations (Pasquill 1961). The resulting scheme was modified slightly and rearranged into a format widely known as the Pasquill–Gifford curves (Gifford 1961). The data were subsequently used to calculate the numerous concentration–distance curves that appeared in the Turner workbook (Turner 1970), in which more quantified specifications of some of the factors constituting the stability scheme were developed. Procedures to relate the stability classes to widely used surface-layer stability parameters and to directly incorporate surface roughness for a small range of values were subsequently developed (Golder 1972). Attempts have been made to improve the fundamental character of the dispersion estimates (e.g., Smith 1973), and other schemes have subsequently been devised (e.g., see Irwin 1983). However, the Pasquill–Gifford–Turner stability scheme/curves (PGT) continue to be used in modified form for a myriad of site-specific applications and represent the “universal” standard used as a basis for comparison. The direct use of such results over urban settings can be questioned since the urban surface presents a more severe and complex barrier to airflow than does semiarid terrain or open grassland and affects the thermal structure of air flowing over it differently.

Urban atmosphericenvironment

Adequate theory still does not exist to fully quantify these urban effects. Most of the relevant work concerns identification and depiction of major features of the urban environment that affect the meteorological structure and hence the turbulent intensities. Primary features delineated include the large, often irregular roughness elements within the urban canopy or fabric; the so-called heat island due to decreased vegetation or increased artificial fabric, a decrease of evapotranspiration commensurate with the decrease in vegetation, storage, and ensuing release of heat from buildings/paved fabric surfaces, and heat from anthropogenic sources; and the influx of airborne gaseous and particulate pollutants, as well as water vapor, into the atmosphere from anthropogenic activities.

Urban effects of importance relative to turbulent diffusion include the less stable thermal stratification within the so-called urban boundary layer, the increase in turbulence intensities, and the overall increase in thickness of the atmospheric mixing layer. The largest increases in the turbulence intensities occur in the lowest layers and decrease with height. The combined urban effects are generally larger with stable upwind rural conditions, which primarily occur at night. Results demonstrate that the urban effects can be determined largely through knowledge of stability in the upwind rural area and parameterizations involving the urban features.

St. Louis Dispersion Study

The U.S. Environmental Protection Agency and its predecessors had been given responsibility for the development of air quality models and the resulting application of these models in strategies for the alleviation of air pollution problems. However, when programs relevant to such activities began in earnest in the early 1960s, the bulk of the detailed urban studies were yet to be conducted. Thus, experimental studies were initiated to directly delineate the turbulent structure and attendant diffusive properties of the urban atmosphere. The largest and most comprehensive of such studies was a series of meteorological tracer experiments conducted in the metropolitan St. Louis, Missouri, area (Pooler 1966). The results of this series of experiments in terms of dispersion parameters were summarized by McElroy and Pooler (1968a,b) and McElroy (1969), with additional analyses provided by Briggs (1973) and Pasquill (1976). Like PGT for rural settings, the concentration–distance relations for St. Louis continue to be used in modified form for many site-specific urban applications and provide a standard reference base for analyses and discussion of subsequent studies.

The tracer-derived estimates of diffusion from low-level sources in St. Louis were described in terms of stability classes and meteorological indicators of turbulence intensity. In relation to studies over open country and the stability framework pioneered by PGT, it was concluded that for low-level point sources, the urban area affects crosswind dispersion primarily by increasing the initial size of the plume and affects vertical dispersion by significantly enhancing the diffusion, together with increasing the initial plume dimension; the enhancement in diffusion was considered to result from increased mixing induced by airflow over rough, irregular surfaces and fostered by instability caused by the urban heat island effects.

Other tracer studies in relation to the St. Louis study

A number of atmospheric tracer experiments for characterizing urban atmospheric diffusion have been conducted subsequent to the St. Louis Dispersion Study. Some have been carried out in other cities in the United States [e.g., Johnstown, Pennsylvania (Smith 1968), andFort Wayne, Indiana (Hilst and Bowne 1966)]. Others have taken place in Europe [e.g., Copenhagen, Denmark (Gryning and Lyck 1981), and Milan, Italy (Santomauro et al. 1970)], and in various Japanese cities [e.g., Onahama, Himeji, Omuta, and Nishiyoda (Sato 1977)]. In addition, data from classified urban tracer studies conducted primarily during the 1950s in several North American cities (e.g., Minneapolis, Minnesota, St. Louis, and Winnipeg, Canada) have been declassified and subsequently released (Stanford University 1953). Many of these latter experiments, in particular, were conducted in specific land-use classifications or areas (see Fig. 1b).

Results from these other studies are shown in two diagrams in Fig. 1 with respect to the estimated Gaussian standard deviation of concentration distribution (σz) as a function of downwind distance. The PGT curves (A–F) and the St. Louis best fit curves (B′–E′/F′) are displayed in the figure to provide baseline information and a source of reference to facilitate comparisons. The database was too large to easily present in a single diagram without substantial compression. Reference PGT classes have not been provided for the “declassified” datasets, for which tests were mostly conducted at night and generally within specific land-use areas (Fig. 1b); discrimination is rather by land use and overall low-level stability. Fewer data are available for which crosswind distribution parameters (such as Gaussian σy) had been determined or could be inferred. The available data (not shown here) usually have less scatter overall and about stability related diffusion curves than those for the vertical dimension.

A considerable portion of data scatter in the vertical dimension shown in Fig. 1 is felt, as indicated earlier, to be due to differences in dispersion owing to differing mechanical and thermal properties of specific urban land use. Scatter of data points, of course, can be related to other factors, such as natural variability within the same overall stability. In addition, the vertical distribution parameters were generally estimated from mass continuity using near-ground-level data (for which depletion may have occurred owing to deposition and “plume lofting”) and a Gaussian distribution assumption. It is generally very costly (e.g, to deploy and instrument towers and tethered balloons) and difficult (e.g., the presence of buildings, traffic, and power/phone lines) to make measurements of tracer material in the vertical dimension within the urban canopy. The prime exception in this database is for the Fort Wayne, Indiana, study for which the vertical diffusion parameters are based on vertical profile data. It should be noted that the Fort Wayne, St. Louis, and Johnstown, stable data (E–F) agree very well. Also, the neutral stability data (D) for Fort Wayne, St. Louis, Copenhagen, and Milan (not shown) are in general agreement (Fig. 1a). The data for the Japanese experiments in Fig. 1a appear to be more closely related to the PGT rural curves than to the other urban results; it is likely that the use of the land over which the experiments were conducted was more representative of rural Japan than of urban environs.

Overall, the scatter of data points in Fig. 1b within the first few hundred meters appears to be jointly discriminated by land use and by stability. Farther downwind, the nighttime data are bracketed by the St. Louis C′ and E′/F′ best-fit lines; data collected under“stable” conditions generally lie between the D′ and E′/F′ lines, while those collected under “unstable” conditions generally lie between C′ and slightly below D′. The unstable daytime data (Winnipeg) appear to agree well with St. Louis B′. It should be noted that data from tracer experiments conducted during the same time frame in rural environs near one of the cities (Winnipeg) during extreme daytime instability and extreme nighttime stable conditions (not shown), respectively, agree very well with PGT A/B and F curves.

Summary and discussion

Collectively, results from these tracer experiments appear to complement, as well as generally be quite similar to, those provided by the St. Louis study. Even though the scatter of data points is large, a characteristic of this type of information, the largest discrepancies appear to occur nearer to the source and hence may be related to details of the roughness and thermal properties of the specific urban canopy/fabric. Changes in site-specific land use owing to activities such as urban growth and urban renewal may, in many instances, preclude direct retrospective examination and quantification of such effects. However, studies could be conducted to quantify some effects, especially in terms of plume spread due to local land use and plume spread due to diffusion. One source of information readily available for most cities is literally house-by-house tabular and statistical land-use information, developed, for instance, by urban planning agencies and insurance companies. Other sources of semiquantitative and quantitative data that might be explored to assist in the development and refinement of relevant land-cover and land-use information are high-resolution satellite- and aircraft-based remote sensing data. Such data have been collected using devices such as aerial cameras, multispectral scanners, and synthetic aperture radar by civilian and national means sectors (i.e., security and defense complexes); declassification of pertinent material from the latter sector has begun and is expected to increase in scope in the near future.

Thus, the basic results of the St. Louis Dispersion Study still seem to be valid for low-level sources in urban areas, but modifications to its formulations, especially nearer the tracer release points, may be possible in terms of the local land use.

Disclaimer. The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development (ORD), prepared this article. It does not necessarily reflect the views of the EPA or ORD.

REFERENCES

  • Briggs, G. A., 1973: Diffusion estimates for small emissions. NOAA, ERL Rep. ATDL-79, 59 pp. [Available from U.S. Environmental Protection Agency, NERL, ASMD, MD-80, Research Triangle Park, NC 27711.].

  • Gifford, F. A., 1961: Use of routine meteorological observations for estimating atmospheric dispersion. Nucl. Saf.,2, 47–51.

  • Golder, D., 1972: Relations among stability parameters in the surface layer. Bound.-Layer Meteor.,16, 47–58.

  • Gryning, S. E., and E. Lyck, 1981: Results from elevated-source urban area dispersion experiments compared to model calculations. Proc. 12th Int. Technological Meeting on Air Pollution Modeling and Its Applications, Menlo Park, CA, NATO/CCMS, 550–567.

  • Hilst, G. R., and N. E. Bowne, 1966: A study of the diffusion of aerosols released from aerial line sources upwind of an urban complex. Final Rep. DA-42-007-AMC-38(R), 814 pp. [Available from TRC, Inc., Hartford, CT.

  • Irwin, J. S., 1983: Estimating plume dispersion—A comparison of several sigma schemes. J. Climate Appl. Meteor.,22, 92–114.

  • McElroy, J. L., 1969: A comparative study of urban and rural dispersion. J. Appl. Meteor.,8, 19–31.

  • ——, and F. Pooler Jr., 1968a: The St. Louis Dispersion Study: Volume I—Instrumentation, procedures and data tabulations. APTD Document 12, 352 pp. [Available from U. S. Environmental Protection Agency, NERL, P.O. Box 93478, Las Vegas, NV 89193-3478.].

  • ——, and ——, 1968b: The St. Louis Dispersion Study: Volume II—Analysis. PHS Environmental Health Series AP-53, 51 pp. [Available from U.S. Environmental Protection Agency, NERL, P.O. Box 93478, Las Vegas, NV 89193-3478.].

  • Pasquill, F., 1961: The estimation of the dispersion of windborne material. Meteor. Mag.,90, 33–49.

  • ——, 1976: Atmospheric dispersion parameters for Gaussian plume modeling: Part II, Possible requirements for a change in the Turner workbook values. EPA-600/4-76-30b, 44 pp. [NTIS PB-191 482.].

  • Pooler, F., Jr., 1966: A tracer study of dispersion over a city. J. Air Pollut. Control Assoc.,16, 677–681.

  • Santomauro, L., V. Maestro, and C. Barberis, 1979: Experimental Determination of Diffusion Coefficients Above an Urban Area. Observation Meteorologico di Bera-Milano, 12 pp.

  • Sato, J., 1977: The urban dispersion parameters of Gaussian plume model in neutral stability conditions. Pap. Meteor. Geophys.,2B, 97–103.

  • Smith, D. B., 1968: Tracer study in an urban valley. J. Air Pollut. Control Assoc.,18, 466–471.

  • Smith, F. B., 1973: A scheme for estimating the vertical dispersion of a plume from a source near the ground. Turbulence and Diffusion Note 40.

  • Stanford University and R. M. Parsons Co., 1953: Behavior of aerosol clouds within cities. Joint Quarterly Rep. 2-6 to U.S. Army Chemical Corps, 752 pp.

  • Turner, D. B., 1970: Workbook of atmospheric dispersion estimates. Office of Air Programs Publ. AP-26, 84 pp. [NTIS PB-191 482.].

Fig. 1.
Fig. 1.

Estimated Gaussian σz as a function of downwind distance for urban tracer experiments with superimposed PGT curves (A–F) and St. Louis best-fit curves (B′–E′/F′). Data points are related in (a) to PGT stability classes and (b) to overall stability [unstable (UNS) and stable (ST)] and land use [residential (RES), industrial (IND), downtown (DT), and citywide (CW)].

Citation: Journal of Applied Meteorology 36, 8; 10.1175/1520-0450(1997)036<1027:DFLLUS>2.0.CO;2

Save
  • Briggs, G. A., 1973: Diffusion estimates for small emissions. NOAA, ERL Rep. ATDL-79, 59 pp. [Available from U.S. Environmental Protection Agency, NERL, ASMD, MD-80, Research Triangle Park, NC 27711.].

  • Gifford, F. A., 1961: Use of routine meteorological observations for estimating atmospheric dispersion. Nucl. Saf.,2, 47–51.

  • Golder, D., 1972: Relations among stability parameters in the surface layer. Bound.-Layer Meteor.,16, 47–58.

  • Gryning, S. E., and E. Lyck, 1981: Results from elevated-source urban area dispersion experiments compared to model calculations. Proc. 12th Int. Technological Meeting on Air Pollution Modeling and Its Applications, Menlo Park, CA, NATO/CCMS, 550–567.

  • Hilst, G. R., and N. E. Bowne, 1966: A study of the diffusion of aerosols released from aerial line sources upwind of an urban complex. Final Rep. DA-42-007-AMC-38(R), 814 pp. [Available from TRC, Inc., Hartford, CT.

  • Irwin, J. S., 1983: Estimating plume dispersion—A comparison of several sigma schemes. J. Climate Appl. Meteor.,22, 92–114.

  • McElroy, J. L., 1969: A comparative study of urban and rural dispersion. J. Appl. Meteor.,8, 19–31.

  • ——, and F. Pooler Jr., 1968a: The St. Louis Dispersion Study: Volume I—Instrumentation, procedures and data tabulations. APTD Document 12, 352 pp. [Available from U. S. Environmental Protection Agency, NERL, P.O. Box 93478, Las Vegas, NV 89193-3478.].

  • ——, and ——, 1968b: The St. Louis Dispersion Study: Volume II—Analysis. PHS Environmental Health Series AP-53, 51 pp. [Available from U.S. Environmental Protection Agency, NERL, P.O. Box 93478, Las Vegas, NV 89193-3478.].

  • Pasquill, F., 1961: The estimation of the dispersion of windborne material. Meteor. Mag.,90, 33–49.

  • ——, 1976: Atmospheric dispersion parameters for Gaussian plume modeling: Part II, Possible requirements for a change in the Turner workbook values. EPA-600/4-76-30b, 44 pp. [NTIS PB-191 482.].

  • Pooler, F., Jr., 1966: A tracer study of dispersion over a city. J. Air Pollut. Control Assoc.,16, 677–681.

  • Santomauro, L., V. Maestro, and C. Barberis, 1979: Experimental Determination of Diffusion Coefficients Above an Urban Area. Observation Meteorologico di Bera-Milano, 12 pp.

  • Sato, J., 1977: The urban dispersion parameters of Gaussian plume model in neutral stability conditions. Pap. Meteor. Geophys.,2B, 97–103.

  • Smith, D. B., 1968: Tracer study in an urban valley. J. Air Pollut. Control Assoc.,18, 466–471.

  • Smith, F. B., 1973: A scheme for estimating the vertical dispersion of a plume from a source near the ground. Turbulence and Diffusion Note 40.

  • Stanford University and R. M. Parsons Co., 1953: Behavior of aerosol clouds within cities. Joint Quarterly Rep. 2-6 to U.S. Army Chemical Corps, 752 pp.

  • Turner, D. B., 1970: Workbook of atmospheric dispersion estimates. Office of Air Programs Publ. AP-26, 84 pp. [NTIS PB-191 482.].

  • Fig. 1.

    Estimated Gaussian σz as a function of downwind distance for urban tracer experiments with superimposed PGT curves (A–F) and St. Louis best-fit curves (B′–E′/F′). Data points are related in (a) to PGT stability classes and (b) to overall stability [unstable (UNS) and stable (ST)] and land use [residential (RES), industrial (IND), downtown (DT), and citywide (CW)].

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 262 51 1
PDF Downloads 74 24 0