Dispersion Model Validation

Many different air pollution dispersion models have been developed worldwide. The validation of these models is also very important. It is usually carried out by comparing the models' predictions with the data from field and wind tunnel trials.

Some field trials

Maplin Sands

The Maplin Sands trials were conducted by Shell Research limited in 1980 and consisted of 34 spills of liquefied gases onto the sea. The aim of these
trials was the study of combustion and dispersion of flammable gases. Further details of the releases can be found in Puttock et al (1982) [1] and Colenbrander et al (1984a,b,c).[2] [3] [4]

The release site was an area of tidal sands in the Thames estuary and is shown schematically in Figure 1. The gas handling plant was on shore at the edge of flat farmland. The spill point was 350 m offshore and liquid was delivered along a pipeline from the gas handling plant. A 300m diameter dike was constructed around the spill point to retain water so that the trials could be conducted at low tide. The water level behind the dike varied by 0.75 m. The spill point was a 150 mm diameter pipe directed vertically downward terminating above the water surface.

Figure 1 Maplin Sands spill area

Thorney island

The Heavy Gas Dispersion Trials at Thorney Island were set up by the British Health and Safety Executive for the study of the dispersion of fixed volume releases of heavy gas (a mixture of Freon 12 and Nitrogen). The program was subsequently extended to include continuous release trials 45,46 and 47. However, trial 46 provided limited data due to wind shifts during the trial. Further details of the set-up and release can be found in McQuaid and Roebuck (1985) [5] and McQaid (1987) [6] .

The test site was a former Royal Air Force station at Thorney Island. The test area was largely clear for a length of 2 km and a width of 500 m and flat to within 1 in 100. Gas was ducted below ground from a 2000 m3 container to the release position. The release position consisted of a vertical duct emerging at ground level underneath a 2 m diameter cap situated 0.5 m above the ground. This arrangement was designed to release the gas with near-zero vertical momentum. The spill area is shown in schematically in Figure 2.

Figure 2 Thorney island spill area

Burro and Coyote

The Burro series of experiments (Koopman et al 1982)[7] were performed at the Naval Weapons Center (NWC), China lake, California in the summer of 1980. There were eight LNG spills of between 24 m3 and 39 m3 onto water. The Coyote series (Gold wire et al 1983)[8] followed a similar format and were intended for the study of vapor burn and Rapid Phase Transition (RPT) explosions that had been observed in the Burro trials. The Coyote trials consisted of five vapor dispersion and burn experiments and five experiments for investigation of RPT occurrences. Spill volumes were between 3.3 and 28 m3.

A complete description of the test site and spill area reports of Koopman et al (1982) and Goldwire et al (1983) and is shown schematically in Figure 3.
LNG was stored in a cryogenic tank capable of holding up to 40 m3 of LNG. A 25 cm diameter spill line ran from this tank to a 5.7 m3 spill tank with valve. A further 25 cm diameter line ran from the spill valve to a water test basin where it terminated 1.0 m above the surface of the water. A splash plate was fitted below this pipe outlet to limit penetration of the LNG horizontally across the water surface. The LNG was forced out under pressure by gaseous nitrogen. The water test basin had an average diameter of 58 m surrounding ground level. The terrain downwind of the spill pond sloped upward at about 7 degrees for 80 m before leveling out to about a 1 degree slope. Contour maps are given in Koopman et al (1982a,b) and Goldwire et al (1983).

Figrue 3 Burro and Coyote spill area

Falcon trials

The Falcon trials (see Brown et al 1990)[9] were a series of five large-scale LNG spill tests carried out by the Lawrence Livermore National Laboratory (LLNL). The trials were carried out at Frenchman Flat, an extremely flat area with little vegetation. The trials had the purpose of evaluating the effectiveness of vapor fences as a mitigation technique for accidental releases as well as providing a data set for model validation purpose. The spills (between 20 and 63 m3 of LNG) were onto a specially designed water pond equipped with a circulating system to maximize evaporation.
LNG was supplied to the spill area (see Figures 4 and 5) from two cryogenic 100 m3 storage tanks. The LNG was driven along spill pipes by means of nitrogen gas. The main 10'' diameter spill pipe terminated immediately above the center of the pond and then divided into 6'' diameter pipes as a multi-exit "spider" to provide a uniform distribution of LNG over the spill pond. The spider consisted of four arms of 11.6 m length oriented at 90 degrees to each other. Each arm was fitted with a restrictive orifice at the downstream end of the horizontal portion to prevent flashing in the pipe. The orifice diameter was 4.5'', except in the Falcon-4 trial when a 1.5'' diameter orifice was used. The outlet of each arm was directed vertically downwards. A 36'' diameter horizontal splash plate was located beneath each outlet, level with the pond water surface so as to direct the LNG horizontally. The height of the pipe outlet above the splash plate was not provided in Brown et al (1990), but appears to be greater than 12'' and much less than 36'' (estimated from schematic drawings).

The spill pond was 40 m by 60 m and filled to a depth of approximately 0.76 m. An 8.7 m high vapor containment fence 44 m by 88 m surrounded the spill pond, with the spill pond located at the downwind end of this fence. The fence was a proprietary fiberglass cloth impregnated with a mixture of Silicon, Teflon and graphite. Immediately upwind of the spill pond was a 13.3 m high and 17.1 m wide billboard. The billboard was made of the same fiberglass material as the fence.

Figure 4 Falcon spill area

Figure 5 Falcon vapor fence details

Further Reading

For those unfamiliar with air pollution dispersion modeling, the following books will provide a basic understanding:
  • D.B. Turner (1994), Workbook of Atmospheric Dispersion Estimates, 2nd Edition, CRC Press, ISBN 1-566-023-X.
  • M.R. Beychok (2005), Fundamentals of Stack Gas Dispersion, 4th Edition, privately published, ISBN 0-9644588.
  • K.B. Schnelle and P.R. Dey (2000), Atmospheric Dispersion Modeling Compliance Guide, 1st Edition, McGraw-Hill Professional, ISBN 0-07-058059-6.


  1. ^ Puttock, J S, Blackmore, D R and Colenbrander, G W, 1982, Field experiments on dense gas dispersion, J Haz Mats, Vol 6, pp 13 -41
  2. ^ Colenbrander G W, Evans A and Puttock J S, 1984a, Spill tests of LNG and refrigerated liquid propane on the sea, Maplin Sands 1980: Dispersion Data Digest; Trial 27, Shell Research Ltd, Thornton Research Centre, Report TNER.84.028, May 1984.
  3. ^ Colenbrander G W, Evans A and Puttock J S, 1984b, Spill tests of LNG and refrigerated liquid propane on the sea, Maplin Sands 1980: Dispersion Data Digest; Trial 34, Shell Research Ltd, Thornton Research Centre, Report TNER.84.030, May 1984.
  4. ^ Colenbrander G W, Evans A and Puttock J S, 1984c, Spill tests of LNG and refrigerated liquid propane on the sea, Maplin Sands 1980: Dispersion Data Digest; Trial 35, Shell Research Ltd, Thornton Research Centre, Report TNER.84.031, May 1984.
  5. ^ McQuaid J and Roebuck B, 1985, Large scale field trials on dense vapour dispersion, Commission of the European communities indirect action programme “Safety of thermal water reactors”, 1979-83, Final report on contracts 029SRUK and 036SRUK with the Health and Safety Executive, EUR10029en, Commission of the European Community, Brussels.
  6. ^ McQuaid J (Editor), 1987, Heavy gas dispersion trials at Thorney Island - 2, Proceedings of a symposium held at the University of Sheffield, Great Britain, September 1986, J Haz Mats, Vol 16, pp
  7. ^ Koopman R P, Cederwall R T, Ermak D L, Goldwire H C Jr, Hogan W J, McClure J W, McCrae T G, Morgan D L, Rodean H C and Shinn J H, 1982a, Analysis of Burro series 40 m3 LNG spill experiments J Haz Mats, Vol 6, pp 43 – 83. x
  8. ^ Goldwire H C Jr, Rodean H C, Cederwall R T, Kansa E J, Koopman R P, McClure J W, McCrae T G, Morris L K, Kamppinen L, Kiefer R D, (Urtiew P A and Lind C D), 1983, Coyote series data report: LLNL/NWC 1981 LNG spill tests dispersion, vapor burn and rapid-phase transitions, Vols 1 & 2, UCID-19953, Lawrence Livermore National Laboratory ( https://e-reports-ext.llnl.gov/pdf/195030.pdf
    and https://e-reports-ext.llnl.gov/pdf/195363.pdf )
  9. ^ Brown T C, Cederwall R T, Chan S T, Ermak D L, Koopman R P, Lamson K C, McClure J W and Morris, L K, 1990, Falcon series data report: 1987 LNG vapor barrier verification field, Gas Research Institute, Report No. GRI-89/0138, June 1990 (http://www.osti.gov/bridge/servlets/purl/6633087-9HP88a/)

Publishing Note

This article was written by Daoming Liu, a member of this wiki, and uploaded by him. It was then edited to improve its usage of American English by Milton Beychok, the organizer of this wiki.