Proc eedings 8th International Symposium on Acoustic Remore Sensing and Associated Techniques of the Atmosphere and Ocean, Moscow, May 27-31, 1996, p. 5.1-5.8


The role of ground-based remote sensing techniques in providing meteorological input to modern dispersion models

Petra Seibert

Institute of Meteorology and Geophysics, University of Vienna
Hohe Warte 38, A-1190 Vienna, Austria

Tel.: +43-1-364453-2410, Fax: +43-1-364453-74, E-mail: xxx@xxx.at *)

ABSTRACT

With respect to traditional Gaussian dispersion models, the utilization of sodar data has mainly been limited to the mean horizontal wind and attempts to determine Pasquill-Gifford (PG) or similar stability categories. Modern dispersion models are based on the turbulent fluxes of momentum and heat and the mixing height. These parameters are usually determined from standard observations from surface weather stations and radiosondes, and used to compute the standard deviations of the turbulent wind components and their Lagrangian time scale at plume level. This provides ample opportunity for remote-sensing systems to improve the input to such models. Mean wind profiles are required in any model, and sodar data should be used to devise new parameterizations not based on PG categories. Mixing height, especially under convective conditions, is a parameter of primary importance for modern dispersion models, and will likely be determined from the backscatter profiles of boundary-layer radar profilers for operational purposes in the future. Temperature profile observations from RASS could be used in the determination of the stable boundary layer height and for plume rise calculations. The derivation of turbulence characteristics and turbulent fluxes at elevated levels would be of great value, but more field campaigns providing extended comparisons with in-situ measurements appear to be necessary before results can be used in dispersion modelling, especially in an operational setting.

RECENT DEVELOPMENTS IN DISPERSION MODELLING

About 10 years ago, it was realized that most dispersion models used for operational purposes like licensing procedures for industrial plants were substantially out of date with respect to their understanding of turbulent diffusion in the atmospheric boundary layer (ABL) (Weil, 1985). These models can be characterized as Gaussian models based on Pasquill-Gifford (PG) or similar dispersion categories with related sigma curves. The two most important criticisms were the failure to account for the qualitative difference between dispersion in the stable ABL (SBL) and the unstable, convective ABL (CBL), and use of stability categories instead of parameters directly related to the characteristics of mechanical and convective turbulence. New models which are based on the progress in boundary layer meteorology made in the past 20 years (see Nieuwstadt and van Dop, 1982, and Venkatram and Wyngaard, 1989) have now been introduced in a number of countries (see, e.g., Olesen et al., 1991; Hanna and Paine, 1989; Hanna and Chang, 1993; Carruthers et al., 1992). Within Europe, initiatives have been taken to promote so-called harmonisation in model development (Olesen and Mikkelsen, 1992; Kretzschmar et al., 1994; Kretzschmar, 1996).

In research-oriented dispersion models, the random-walk technique applied in Lagrangian particle dispersion models (LPDM) has become the standard for investigations of the impact of a limited number of single sources. In studies of regional air quality, especially if including chemical transformations, Eulerian models based on K-theory are still widely used. Lagrangian puff models are typically applied for receptor-oriented studies in the regional scale. Parameters needed for the dispersion calculations are often extracted from prognostic numerical models of the atmosphere; otherwise, they are calculated from parameterizations of the ABL which are very similar to those used in the advanced operational dispersion models.

Past efforts to use sodar data for dispersion modelling have focused on methods to derive PG categories from the sodar output. As the PG-based dispersion models are becoming obsolete, new approaches are required. In the new models, considerable efforts have been made to derive model input parameters from the set of meteorological observations which are usually available, leading to so-called meteorological preprocessors. The COST* Action 710 has been set up to promote and harmonize developments in this field (Beyrich et al., 1996; Pechinger et al., 1996). The potential role of remote-sensing systems in this context is discussed in the present paper.

DEVELOPMENTS IN REMOTE-SENSING SYSTEMS

A comprehensive, though somewhat outdated reference is Lenschow (1986). More recent developments are covered by the conference series on tropospheric profiling (see, e.g., MPG, 1994), and the review article of Clifford et al. (1994).

Sodar

Two main lines of development for sodars can be identified: the minisodar development and the improvement of signal processing. Minisodars, operating at frequencies from 3 to 6kHz, provide vertical resolution on the order of 5m with the first range gate as low as about 10m. Their results are less influenced by the volume-averaging effects. The signal processing improvements include the use of multiple frequency signals, more points in the FFT for Doppler shift identification, cross-correlation techniques in order to improve estimates of the standard deviation of the horizontal wind components, and others. A major limitation in the practical use still is the noise impact to the environment. The potential role of sodars in dispersion modelling has recently been discussed by Kallistratova (1994).

Radar

Pulsed clear-air Doppler radars, often simply termed "wind profilers", are becoming more and more an important tool in meteorology. In the USA, demonstration networks have already been installed while in Europe only a few singular systems are operational. So-called boundary-layer profilers or LAPs (lower atmosphere profilers), operating at frequencies between 900 and 1400MHz, are a promising tool for air pollution and other boundary layer applications. Vertical resolution and lower range gates have been improved, and the range between 100-200m and a few km can be covered with a vertical resolution between 50m and 200m. While in the beginning only the horizontal wind velocity was looked at (the radar profilers had been introduced with synoptic meteorology as the main application), now the vertical velocity and its variance as well as the backscatter profile are studied, too. A COST action (COST-76) is presently active to promote the use of VHF/UHF wind profilers in Europe; one of its results is a comprehensive bibliography including all types of ground-based remote sensing systems.

The range and resolution characteristics of sodars and radars suggest a co-location of the two systems in order to obtain data throughout the whole ABL. The partial overlap allows an estimate of measurement uncertainty. A number of publications have already explored this approach (Beyrich and G”rsdorf, 1995; Flowers at al., 1994). With the introduction of LAPs and their improved performance, sodars with lower range and better resolution are especially well suited for such purposes. However, in order to guarantee complete data coverage, the sodar should provide close to 100% data availability up to 200m, a goal not reached by typical minisodars at present.

RASS

Radio-acoustic sounding systems (RASS) are available in forms with a continuous sound source added to an electromagnetic profiler and with an electromagnetic source added to a Doppler sodar, and yield profiles of the virtual temperature up to about 1 km, depending on the system parameters. Their quality has been improved by taking into account the vertical motion of the air. Especially the form with a continuous acoustic source (usually operated for a couple of minutes each half-hour) may cause problems due to the noise.

Lidar

Lidars work with visible or near-infrared light and have been developed in a number of different forms. Pulsed and scanning Doppler lidars, with similar working principles as sodars or electromagnetic profilers, have been introduced to measure wind profiles. Beside the fact that they are the most expensive and delicate of all types of profilers, it seems that due to their characteristics they will probably be used mainly for probing higher atmospheric levels including the stratosphere, and as space-borne systems. An interesting feature is the differential-absorption lidar (DIAL) which allows to measure concentration profiles of atmospheric traces constituents. Lidars for the measurement of temperature profiles are under development.

GENERAL CONSIDERATIONS CONCERNING THE USE OF SPECIAL MEASUREMENT SYSTEMS IN ROUTINE DISPERSION MODELLING

There are different ways in which the special measurements provided by remote sensing devices can be used for routine dispersion modelling, which should be considered in the development of methods. They are discussed below.

Real-time applications

The ideal case is the permanent installation of a sodar or other special measurements system at a site of interest, so that the data can be fed in a real-time model. However, due to the costs this has mainly been limited to nuclear installations. As dispersion calculation for such plants often have to use prescribed regulatory models not representing the state of the art, and sites lack qualified meteorological staff, the practical value of the additional measurements may be limited.

Episode modelling

No problems are to be expected for the investigation of limited episodes in the frame of dedicated research projects. Here it is even possible to apply manual quality control or data evaluation.

Development, adaptation, and test of parameterization schemes

Sodar data have repeatedly baeen used in the development of parameterization schemes for the height of the stable boundary layer (Arya, 1981; Koracin and Berakowicz, 1988) and in order to test the power law exponents used in many dispersion models to describe the wind speed profile (Seaibert, 1990; Piringer, 1992a)a. If the accuracy of the sodar-derived data is good enough, this is a very useful application. However, often additional parameters such as the turbulent fluxes at the surface or 10m wind speed are needed in such schemes which are difficult to obtain with a sodar so that other supplementary instruments should be set up. A very interesting application which has not yet received much attention is the adaption of general parametrization schemes to a specific site or a specific day. For example, Beyrich (1995) has successfully used the sodar-derived development of CBL height in the morning hours to modify the parameters of a slab model of the CBL evolution. One could also imagine that measurements during a limited period of time could be used to derive a site-specific form of the vertical wind profile (direction and velocity) under different conditions (stability and flow direction) for use in subsequent long-term dispersion calculations.

Boundary layer parameters which can be derived from remote-sensing instruments

Mean wind

Reliable measurements of the profile of the mean wind by sodar are available from commercial instruments since many years, and the electromagnetic profilers extend the range for such measurements considerably. It is my impression, however, that considering the long time for which these data have been available, relatively little use has been made of them in practice. I am not aware of any regulatory dispersion model where the wealth of available sodar data has been used in the wind profile parameterization. It should be noted here that the widespread method of power law profiles with the exponent depending on the PG category lacks a physical basis and has been falsified using sodar data. However, the modern dispersion models are based on parameters such as friction velocity u*, characteristic convective velocity w*, and boundary-layer height which are appropriate also to describe the wind profile. In order to avoid uncertainties associated with the derivation of these parameters, the equipment of sodar measuring sites with sonic anemometers for a direct measurement of the relevant fluxes would be very desirable.

A special problem is the occurrence of low-level jets (LLJs) which cannot be described by standard approaches for the wind profile as used in dispersion models. With sodar measurements available at a site, one can either use them directly (though only during the measurement period) or to test more complicated formulations up to a 1-dimensional numerical boundary layer model.

Mixing height

Mixing height has been a relevant parameter for conventional dispersion models, and becomes even more important in advanced models where the intensity of turbulence under convective conditions is assumed to be proportional to the cubic root of the CBL height. Many papers have been written on the subject of mixing height determination from sodar data (for the most recent review, see Beyrich, 1996). Some sodar manufacturers also offer software which provides "mixing height" or "inversion height" as output. According to my experience with the PA2 sodar of Remtech S.A., this output appears to be useless because results are often inconsistent in time and look not very realistic.

In the SBL, the mixing height determination is plagued by two different problems. On one hand, turbulence in the outer SBL often is intermittent and patchy, so that it is very difficult to measure the height up to which turbulence extends, regardless which measurement system is used. The frequent occurrence of LLJs with associated wind shear and of gravity waves which can produce turbulence by wave breaking further complicates the situation. The other problem is that the backscatter intensity, on which mixing height estimation from sodar data in the SBL usually relies, depends not only on turbulent fluctuations of the refractive index but also on its mean gradient. The more stable the stratification the bigger is this contribution. This means that in weakly turbulent stable regions it may be impossible to decide whether the echo is due to turbulence or due to the stratification alone. It seems not unlikely that sometimes the mixing height determined from the sodar output is the top of the stable layer which has developed due to radiative cooling rather than the top of the mixing layer.

Until now it is common practice in boundary layer meteorology to parameterize the SBL height on the basis of the surface fluxes of momentum (and sometimes heat) as well as the Coriolis parameter f, e.g. as being proportional to u*/f. These formulae have also been adopted in modern dispersion models. However, there is growing scepticism towards the use of f as a central scaling parameter and the sole consideration of surface parameters. Alternative diagnostic formulations for the SBL height have been suggested (Mahrt, 1981; Vogelezang and Holtslag, 1996) which depend on the wind shear and temperature gradient across the SBL, e.g. in the form of a bulk Richardson number, or by replacing 1/f as the governing time scale by 1/N, where N is the Brunt-V„is„l„ frequency (Kitaigorodsky and Joffre, 1988). RASS is the only available instrument for operational measurement of the input data needed by these algorithms.

Under unstable conditions, the range of sodars is often insufficient to obtain echos from the stable layer capping the CBL. Unfortunately, in the afternoon of sunny days the mixing layer is deepest while the range of the sodars often is as low as a few hundred metres even if 1000m can be reached with the same instrument under favourable conditions. Attempts to derive the mixing height from similarity relationships and measurements of the vertical velocity standard deviation have to struggle with inherent uncertainties being hard to overcome (Seibert and Langer, 1996). The introduction of the LAPs, which usually reach up to the top of the CBL during the whole day, promises to fill this gap (Angevine et al., 1994a; Dye et al., 1995; Beyrich and G”rsdorf, 1995).

Aerosol lidars, whose backscatter profiles indicate the aerosol concentration profiles, have also been used in the determination of mixing height. The problem with this approach is that the height where a significant decrease of the aerosol concentration is observed does not need to coincide with the current mixing height. Instead, it may often represent the top of a reservoir layer, built up on a previous day (Marsik et al., 1995).

Turbulence and stability

Traditional dispersion models characterized turbulence by just one parameter, the dispersion category, e.g. according to PG. It could be derived for example from the 10m wind speed and the temperature gradient in the lower atmospheric layers. Modern dispersion models can either work with the surface fluxes of momentum and heat plus the depth of the ABL, and parameterize the vertical profile of the standard deviations and higher moments of the turbulent wind components and their Lagrangian time scale, or they use direct measurements of the standard deviations. Published investigations of the accuracy of sodar measurements of these quantities are inconclusive (Finkelstein et al., 1986; Chintawongvanich et al., 1989; Thomas and Vogt, 1993; Gaynor, 1994). Extracting turbulence information from electromagnetic profiler data still is in the beginning (Angevine et al., 1994b). It seems that more intercomparison campaigns are necessary, which should fulfil the following conditions:

-> no quality control beyond the operational software

-> duration of the campaign long enough to encounter all possible regimes of the ABL, e.g. highly convective, stable with weak winds, and strong wind situations

-> evaluation of the data separately for different ABL regimes

-> for selected intensive observation periods, in-situ measurements covering the whole range of the remote-sensing instruments, e.g. by aircraft or by a tethered balloon system able to measure turbulence (Derbyshire, 1995)

-> direct measurement of turbulent fluxes at the surface, and of all relevant parameters at elevated levels on a tower tall enough to extend beyond the region with mechanically dominated turbulence

-> LAP and/or nearby radiosoundings in order to determine the CBL height

Dispersion in the CBL critically depends on the skewness of the vertical velocity distribution. Though it is theoretically possible to derive this parameter from sodar measurements, it is well known that higher moments are increasingly difficult to determine in the field. Seibert and Langer (1996) found a significant bias in the vertical velocity measurements with a sodar under convective conditions which they attributed to selective sampling related to the lower backscatter cross-section in subsiding air as compared to rising thermals. Such effects critically affect the correct determination of second and third moments, and would need to be corrected before using sodar data as input to dispersion models. The same problems may also be relevant for radar profilers.

Under stable conditions, a part of the variance in the vertical velocity may be caused by gravity waves. Unless wave breaking occurs, waves will not contribute to turbulent diffusion of pollutants. However, operational sodar software does not differentiate between waves and real turbulence.

While modern dispersion models do not use the temperature gradient to characterize turbulence, it is highly desirable to know the temperature profile in order to utilize modern plume rise formulations. These formulations calculate not only the final plume rise or the plume rise as a function of travel time, they also try to determine the fraction dispersed into the ABL in the case of partial penetration of a buoyant plume into a capping inversion. If nothing better is available, temperature profiles from the closest radiosoundings are used, but it is obvious that continuous on-site profiles as obtained from a RASS are very useful in this respect.

The spatial representativity of surface measurements, especially of the turbulent fluxes, and the possibilities to obtain fluxes representative for a wider area are active research areas in boundary layer meteorology. These issues affect also dispersion modelling, as for elevated sources fluxes measured at a single surface site may not govern the state of the ABL at stack level. Remote sensing devices provide a unique potential to determine fluxes at elevated levels; in order to use them in dispersion models, thorough field tests (as characterized above) need to be carried out.

CONCLUSIONS

Modern dispersion models offer a wide range of applications for ground-based remote-sensing instruments. In addition to more effective incorporation of wind profile measurements in operational dispersion models, promising applications include the determination of the convective mixing height from the backscatter profiles obtained with electromagnetic boundary layer wind profilers, the determination of mixing heights under stable conditions with bulk Richardson numbers derived from RASS measurements, the measurement of statistical moments of turbulent wind components and turbulent fluxes at elevated levels (and thus representative for a larger area) by sodars or other profilers. However, in all these applications ready-to-implement, well-tested and verified solutions are not available at present. Carefully designed and conducted long-term field measurements are necessary to develop new methods, to test their suitability under the natural variability of atmospheric conditions, and to determine their accuracy. A closer co-operation between instrument developers, instrument users, and modellers will be a condition for undelayed application of new measurement techniques in operational models.

REFERENCES

Angevine, W.M., A.B. White, S.K. Avery (1994a) Boundary layer depth and entrainment zone characterization with a boundary-layer profiler. Bound.-Layer Meteor., 68, 375-385.

Angevine, W.M., R.J. Doviak, Z. Sorbjan (1994a) Remote sensing of vertical velocity variance and surface heat flux in a convective boundary layer. J. Appl. Meteor., 33, 977-983.

Arya, S.P. (1981) Parameterizing the height of the stable atmospheric boundary layer. J. Appl. Meteor., 20, 1192-1202.

Berkowicz, R., H.R. Olesen, U. Torp (1986) The Danish Gaussian air pollution model (OML) Description, test and sensitivity analysis in view of regulatory applications. Proc. 15th NATO/CCMS ITM on Air Poll. Modeling, St. Louis, Plenum Press, 453-481.

Beyrich, F. (1995) Mixing height estimation in the convective boundary layer using sodar data. Bound.-Layer Meteor., 74, 1-18.

Beyrich, F. (1996) Mixing height estimation from sodar data - a review. Proc. ISARS '96, Moscow.

Beyrich, F., U. G”rsdorf (1995) Composing the diurnal cycle of mixing height from simultaneous sodar and wind profiler measurements. Bound.-Layer Meteor., 76, 387-394.

Beyrich, F., S.E. Gryning, S. Joffre, A. Rasmussen, P. Seibert, P. Tercier (1996) On the determination of mixing height: a critical review. In: J. Kretzschmar, ed. (1996) Preprints of the 4th Workshop on Harmonisation Within Atmospheric Dispersion Modelling for Regulatory Purposes. VITO, Mol, Belgium.

Carruthers, D.J., R.J. Holyrod, J.C.R. Hunt, W.-S. Weng, A.G. Robins, D.D. Apsely, F.B. Smith, D.J. Thomson, B. Hudson (1992): UK atmospheric dispersion modelling system. In: H. v. Dop, G. Kallos, eds., Air Poll. Modeling and its Appl. IX, Plenum Press, New York, 15-28 (Proc. 19th CCMS/NATO ITM Ierapetra).

Chintawongvanich, P., R. Olsen, C.A. Biltoft (1986) Comparison of wind monitoring systems. Part ii: Doppler sodars. J. Atmos. Oceanic. Technol., 3, 594-604.

Clifford, S.F., J.C. Kaimal, R.J. Lataitis, R.G. Strauch (1994) Ground-based remote profiling in atmospheric studies: an overview. Proc. IEEE, 82, 313-355.

Dye, T.S., C.G. Lindsay, J.A. Anderson (1995) Estimates of mixing depth from boundary-layer radar profilers. Proc. 9th AMS Symp. Meteor. Instr. & Obs., Charlotteville, 156-160.

Finkelstein, P.L., J.C. Kaimal, J.E. Gaynor, M.E. Graves and T.J. Lockhart (1986) Comparison of wind monitoring systems, part II: Doppler sodars. J. Atmos. Ocean. Technol., 3, 594-604.

Flowers, W., L. Parker-Sedillo, G. Hoidale, J. Hines, F. Eaton, S. McLauglin (1994) Composite tropospheric wind profiles at White Sands Missile Range. W. Neff, ed., Proc. ISARS '94, Boulder 3-7/10/1994, p. 3-47 - 3-53.

Gaynor, J.E. (1994) Accuracy of sodar wind variance measurements. Int. J. Remote Sensing, 15, 313-324.

Hanna, S.R., J.C. Chang (1993): Hybrid plume dispersion model (HPDM) improvements and testing at three field sites. Atmos. Environ., 27A, 1491-1508.

Hanna, S.R., R.J. Paine (1989): Hybrid plume dispersion model (HPDM) development and evaluation. J. Appl. Meteor., 28, 206-224.

Kallistratova, M. (1994) Sodar data as short-range dispersion models input. In: W.D. Neff, ed., Proc. 7th ISARS 1994, Boulder, p. 6-1 - 6-10.

Kitaigorodskii, S.A., S.M. Joffre (1988) In search of a simple scaling for the height of the stratified atmospheric boundary layer. Tellus, 40A, 419-433.

Koracin, D., R. Berkowicz (1988) Nocturnal boundary-layer height: observations by acoustic sounder and predictions in terms of surface-layer parameters. Bound.-Layer Meteor., 43, 65-83.

Kretzschmar, J., ed. (1996) Preprints of the 4th Workshop on Harmonisation Within Atmospheric Dispersion Modelling for Regulatory Purposes. VITO, Mol, Belgium.

Kretzschmar, J., G. Maes, G. Cosemans, eds. (1994) Preprints of the 3rd Workshop on Harmonisation Within Atmospheric Dispersion Modelling for Regulatory Purposes. VITO, Mol, Belgium.

Lenschow, D.H., ed. (1986) Probing the Atmospheric Boundary Layer. Boston: American Meteorological Society, 269 pp.

Mahrt, L. (1981) Modelling the depth of the stable boundary layer. Bound.-Layer Meteor., 21, 3-19.

Marsik, F.J., K.W. Fischer, T.D. McDonald, P.J. Samson (1995) Comparison of methods for estimating mixing height used during the 1992 Atlanta field intensive. J. Appl. Meteor., 34, 1802-1814.

MPG (1994) Third International Symposium on Tropospheric Profiling: Needs and Technologies. Hamburg, 30/8-2/9/1994. Sponsored by MPG, NCAR, NOAA, DMG.

Nieuwstadt, F.T.M., H. van Dop (eds.) (1982,1984): Atmospheric Turbulence and Air Pollution Mode,lling. Reidel, 358 pp.

Olesen, H.R., P. Lofstrom, R. Berkowicz, A.B. Jensen (1991): An improved dispersion model for regulatory use - the OML model. In: H. v. Dop, G. Kallos, eds., Air Poll. Modeling and its Appl. IX, Plenum Press, New York, 29-38. (Proc. 19th CCMS/NATO ITM Ierapetra).

Olesen, H.R., T. Mikkelsen, eds. (1992) Proceedings of the workshop "Objectives of next-generation of practical atmospheric dispersion models". Danish Centre for Atmospheric Research, Riso, Denmark.

Pechinger, U., E. Dittmann, P. Johansson, G. Omstedt, A. Karppinen, L. Musson-Genon, P. Tercier (1996) COST-710 working group 1: status report and preliminary results. In: J. Kretzschmar (1996) Preprints of the 5th workshop on harmonisation within atmospheric dispersion modelling for regulatory purposes. VITO, Mol, Belgium.

Piringer, M. (1992) Sodar-derived mean vertical wind patterns of the lower PBL at several sites in Austria. WMO Instruments and Observing Methods Rep. No. 49, 374-379 (Proc. WMO Tech. Conf. on Instruments and Methods of Observation, Vienna, 11-15/5/1992).

Seibert, P. (1990): Use of sodar data in Gaussian dispersion models. Proc. EURASAP Int. Meeting on Application of Sodar and Lidar Techniques in Air Pollution Monitoring, Krakow, paper VI, 12 pp.

Seibert, P., M. Langer (1996) Deriving characteristic parameters of the convective boundary layer from sodar measurements of the vertical velocity variance. Bound.-Layer Meteor. 81, 11-22.

Thomas, P., S. Vogt (1993): Variances of the vertical and horizontal wind measured by tower instruments and sodar. Appl. Physics, B 57, 19-26.

Venkatram, A., J.C. Wyngaard (eds.) (1988): Lectures on air pollution modeling. AMS, 390 pp.

Vogelezang, D.H.P., A.A.M. Holtslag (1996) Evaluation and model impacts of alternative boundary layer formulations. To appear in Bound.-Layer Meteor.

Weil, J.C. (1985) Updating Diffusion Models. J. Climatol. Appl. Meteor., 14, 1111-1130.

*) This telephone, fax and e-mail information is obsolete. Please check our homepage for up-to-date contact information.


[ Project Catalogue ] [Ground-based remote sensing systems and environmental meteorology] [Top of page ]

Last update: 18 May 1998 by P. Seibert | URL of this page: http://www.boku.ac.at/imp/envmet/moscow.html