In the period 2021-2024, an international model inter-comparison exercise was undertaken to assess the performance of atmospheric dispersion models for simulating releases of pressure-liquefied ammonia. The exercise used data from two sets of ammonia field trials dating from the 1980s and 1990s: the Desert Tortoise and FLADIS trials. Twenty-one independent modelling groups from North America and Europe participated in the exercise and provided twenty-seven sets of results from a range of different dispersion models, including empirically-based nomograms, integral, Gaussian puff, Lagrangian particle and Computational Fluid Dynamics (CFD) models.
The exercise provided useful information on the performance of models that may be used for preparing risk assessments and emergency response plans. The work was also useful in informing plans for a future programme of field-scale ammonia release experiments in the US, known as the Jack Rabbit III trials. These planned trials will follow on from the successful Jack Rabbit I and II field trials that were conducted in the period from 2010 to 2016 at the US Army Dugway Proving Ground in Utah. Further details of the Jack Rabbit research programme can be found on the Utah Valley University website here: https://www.uvu.edu/es/jack-rabbit/
The coordinators of the JRIII modelling exercise (Joe Chang from RAND Corporation and Simon Gant from the UK Health and Safety Executive) provided four documents to participants. These described the model inputs needed to simulate the Desert Tortoise and FLADIS trials, and the requested outputs:
1. Description of the JRIII model inter-comparison exercise, Version 2.4, 30 March 2022 (download)
2. FLADIS geometry and wind directions, Version 1.0, 7 December 2021 (download)
3. Equivalent vapor-only source conditions for the Desert Tortoise trials, Version 1.2, 12 January 2022 (download)
4. Equivalent vapor-only source conditions: cloud aspect ratio for the Desert Tortoise trials, Version 1.1, 30 March 2022 (download)
The documents were produced over time in sequential order, starting with the first item in this list (the description of the modelling exercise), which is the central document describing the model inputs and requested outputs. After this was produced, the groups using CFD models requested further information on the obstacles present in the FLADIS trials and the second document was produced. Subsequently, some of the modelling groups using models that could only simulate vapour releases (i.e., not two-phase, liquid and vapour releases) requested vapour source conditions for Desert Tortoise trial 4, and the third document was produced. The Microsoft Excel spreadsheet that was used to calculate these source conditions can be downloaded from here. Finally, it was found that some model predictions were sensitive to the shape of the vapour source in the Desert Tortoise trials. This led to further analysis of the experiments and the development of revised input conditions for the Desert Tortoise trials, which were documented in the fourth document.
Each of the modelling groups participating in the exercise
was asked to provide (as a minimum) arc-max concentrations for Desert Tortoise
trials 1, 2 and 4, and FLADIS trials 9, 16 and 24. They were also invited to
provide a short summary of the models that they had used and a description of
how it had been configured. In Table 1 (below), there are links provided to
zipped files with this information that they provided – click on the name of
the organisation to download the relevant file. A complete package of all of
the submissions from the modellers can be downloaded as a zipped file by
clicking on this link.
Table 1 Summary of modelling contributions for the JRIII model inter-comparison exercise
|
# |
Organisation |
Model |
Model Type |
Desert Tortoise |
FLADIS |
|||||||
|
|
|
|
A |
B |
C |
D |
1 |
2 |
4 |
9 |
16 |
24 |
|
1 |
Air
Products, USA |
|
|
|
|
|
|
|
|
|
|
|
|
2 |
BAM,
Germany |
|
|
|
|
|
|
|
|
|
|
|
|
3 |
|
|
|
|
|
|
|
|
|
|
||
|
4 |
CEREA
(EDF/Ecole des Ponts), France |
|
|
|
|
|
|
|
|
|
|
|
|
5 |
|
|
|
|
|
|
|
|
|
|
||
|
6 |
DGA,
France |
|
|
|
|
|
|
|
|
|
|
|
|
7 |
|
|
|
|
|
|
|
|
|
|
||
|
8 |
DNV, UK |
|
|
|
|
|
|
|
|
|
|
|
|
9 |
DSTL, UK |
|
|
|
|
|
|
|
|
|
|
|
|
10 |
DTRA, ABQ,
USA |
|
|
|
|
|
|
|
|
|
|
|
|
11 |
EM
Solutions, Inc., USA |
|
|
|
|
|
|
|
|
|
|
|
|
12 |
|
|
|
|
|
|
|
|
|
|
||
|
13 |
Equinor,
Norway |
|
|
|
|
|
|
|
|
|
|
|
|
14 |
FOI,
Sweden |
|
|
|
|
|
|
|
|
|
|
|
|
15 |
Gexcon,
Netherlands |
|
|
|
|
|
|
|
|
|
|
|
|
16 |
Gexcon, Norway |
|
|
|
|
|
|
|
|
|
|
|
|
17 |
GT Science
& Software |
|
|
|
|
|
|
|
|
|
|
|
|
18 |
Hanna
Consultants, USA |
|
|
|
|
|
|
|
|
|
|
|
|
19 |
|
|
|
|
|
|
|
|
|
|
||
|
20 |
HSE, UK |
|
|
|
|
|
|
|
|
|
|
|
|
21 |
|
|
|
|
|
|
|
|
|
|
||
|
22 |
INERIS,
France |
|
|
|
|
|
|
|
|
|
|
|
|
23 |
JRC, Italy |
|
|
|
|
|
|
|
|
|
|
|
|
24 |
NSWC, USA |
|
|
|
|
|
|
|
|
|
|
|
|
25 |
Shell, UK |
|
|
|
|
|
|
|
|
|
|
|
|
26 |
SINTEF,
Norway |
|
|
|
|
|
|
|
|
|
|
|
|
27 |
Syngenta,
UK |
|
|
|
|
|
|
|
|
|
|
|
Model Type: A = Empirically-based nomograms/Gaussian plume model; B = Integral model; C = Gaussian puff/Lagrangian model; D = CFD. Shading in the right six columns indicates model was run for that trial.
The model predictions were compared to the experimental data in numerous plots. Many of these were shared in various conference presentations and a principal journal paper. However, there were too many of them to include in these articles. The full set of plots can be downloaded from this link as a zipped file or they can be selected individually by clicking on the links below:
· Desert Tortoise trial 1
o Arc-max concentrations: all models, CFD models, Gaussian puff and Lagrangian models, integral models, Gaussian plume models and empirical nomograms, PHAST results
o Scatter plots of predicted versus measured arc-max concentrations (all available data)
o Plume widths (all available data)
· Desert Tortoise trial 2
o Arc-max concentrations: all models, CFD models, Gaussian puff and Lagrangian models, integral models, Gaussian plume models and empirical nomograms, PHAST results
o Scatter plots of predicted versus measured arc-max concentrations (all available data)
o Plume widths (all available data)
· Desert Tortoise trial 4
o Arc-max concentrations: all models, CFD models, Gaussian puff and Lagrangian models, integral models, Gaussian plume models and empirical nomograms, PHAST results
o Scatter plots of predicted versus measured arc-max concentrations (all available data)
o Plume widths (all available data)
· FLADIS trial 9
o Arc-max concentrations: all models, CFD models, Gaussian puff and Lagrangian models, integral models, Gaussian plume models and empirical nomograms, PHAST results
o Scatter plots of predicted versus measured arc-max concentrations (all available data)
o Plume widths (all available data)
· FLADIS trial 16
o Arc-max concentrations: all models, CFD models, Gaussian puff and Lagrangian models, integral models, Gaussian plume models and empirical nomograms, PHAST results
o Scatter plots of predicted versus measured arc-max concentrations (all available data)
o Plume widths (all available data)
· FLADIS trial 24
o Arc-max concentrations: all models, CFD models, Gaussian puff and Lagrangian models, integral models, Gaussian plume models and empirical nomograms, PHAST results
o Scatter plots of predicted versus measured arc-max concentrations (all available data)
o Plume widths (all available data)
· Scatter plots of predicted versus measured arc-max concentrations
· Geometric mean versus geometric variance
· Factor of two agreement between predicted and measured arc-max concentrations
o Summary table for all trials
· Boxplot of arc-max concentrations
o Boxplot
· BOOT results
A table of the arc-max concentrations, including all of the models predictions and the experimental data, can be downloaded from these links as either a Microsoft Excel file or a Comma Separated Value (CSV) file.
Over the period from 2021 to 2024, there were numerous conference presentations on the JRIII model inter-comparison exercise. For ease of access, links have been provided below to download these presentations, where permission has been granted. In some cases, where presentations cannot be shared directly, a link to the conference website has been provided.
1. Chang J.C., Gant S.E., Meris R., McMasters S., Hanna S., Spicer T. and Mazzola T. (2021) Jack Rabbit III Modelers Working Group: Initial model inter-comparison exercise for 2021-2022, 25th Annual George Mason University (GMU) Conference on Atmospheric Transport and Dispersion (AT&D) Modeling, 2-4 November 2021 (download slides)
2. Chang J., Gant S., McMasters S., Jablonski R., Mearns H., Fox S., Meris R., Bradley S., Miner S., King M., Hanna S., Mazzola T., Spicer T., Hetherington R., McGillivray A., Kelsey A., Tucker H., Tickle G., Björnham O., Carissimo B., Fabbri L., Wood M., Habib K., Harper M., Hart F., Vik T., Helgeland A., Howard J., Bowman V., Silk D., Mauri L., Mackie S., Mack A., Lacome J.-M., Puttick S., Ibrahim A., Miller D., Dharmavaram S., Shen A., Cunningham A., Beverley D., O’Neal M., Verdier L., Burkhart S. and Dixon C. (2022) Summary of results from the Jack Rabbit III international model inter-comparison exercise on Desert Tortoise and FLADIS, 26th Annual George Mason University (GMU) Conference on Atmospheric Transport and Dispersion (AT&D) Modeling, Fairfax, Virginia USA, 26-28 July 2022, http://camp.cos.gmu.edu/ (download slides)
3. Gant S.E., Hetherington R., Chang J., McMasters S., Jablonski R., Mearns H., Fox S., Meris R., Bradley S., Miner S., King M., Hanna S., Mazzola T., Spicer T., McGillivray A., Kelsey A., Tucker H. and Tickle G. "Design of the Jack Rabbit III international model inter-comparison exercise on Desert Tortoise and FLADIS", 26th Annual George Mason University (GMU) Conference on Atmospheric Transport and Dispersion (AT&D) Modeling, Fairfax, Virginia USA, 26-28 July 2022, http://camp.cos.gmu.edu/ (download slides)
4. Gant S., Chang J., McMasters S., Jablonski R., Mearns H., Fox S., Meris R., Bradley S., Miner S., King M., Hanna S., Mazzola T., Spicer T., Hetherington R., McGillivray A., Kelsey A., Tucker H., Tickle G., Björnham O., Carissimo B., Fabbri L., Wood M., Habib K., Harper M., Hart F., Vik T., Helgeland A., Howard J., Bowman V., Silk D., Mauri L., Mackie S., Mack A., Lacome J.-M., Puttick S., Ibrahim A., Miller D., Dharmavaram S., Shen A., Cunningham A., Beverley D., O’Neal M., Verdier L., Burkhart S. and Dixon C. (2022) Summary of results from the Jack Rabbit III international model inter-comparison exercise on Desert Tortoise and FLADIS, 21st International Conference on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes (Harmo 21), Aveiro, Portugal, 27-30 September 2022, https://www.harmo.org/conference.php?id=21 (download slides or paper)
5. Gant S.E. Chang J., Hetherington R., Mauri L. and Mackie S. (2022d) Jack Rabbit III project and ammonia dispersion modelling, 2022 FLACS-CFD User Group Meeting, 1 Nov 2022, https://www.gexcon.com/event/jack-rabbit-iii-project-on-ammonia-dispersion-modelling/ (download slides)
6. Hetherington R., Tickle G., Bowman V., Silk D., Gant S., Kelsey A., McGillivray A. and Tucker H. (2022) DRIFT modelling of the Desert Tortoise and FLADIS ammonia trials for the Jack Rabbit III model inter-comparison exercise, 21st International Conference on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes (Harmo 21), Aveiro, Portugal, 27-30 September 2022, https://www.harmo.org/conference.php?id=21 (download paper)
7. Hetherington R., McGillivray A., Gant S. and Tickle G. (2024) Analyzing ammonia dispersion under varying atmospheric conditions using DRIFT, 2024 Annual Safety in Ammonia Plants and Related Facilities Symposium, American Institute of Chemical Engineers (AIChE), 9-12 September 2024, San Diego, California, USA, https://www.aiche.org/conferences/annual-safety-ammonia-plants-and-related-facilities-symposium/2024 (download slides or paper)
8. King M., Miner S. and Simpson S. (2022) HPAC modeling of Desert Tortoise and FLADIS in support of Jack Rabbit III, 26th Annual George Mason University (GMU) Conference on Atmospheric Transport and Dispersion (AT&D) Modeling, Fairfax, Virginia USA, 26-28 July 2022, http://camp.cos.gmu.edu/
9. McGillivray A., Harper M., Hart F., Puttick S., Ibrahim A., Verdier L., Gant S. and Hetherington R. (2022) PHAST modelling of the Desert Tortoise and FLADIS ammonia trials for the Jack Rabbit III model inter-comparison exercise, 21st International Conference on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes (Harmo 21), Aveiro, Portugal, 27-30 September 2022, https://www.harmo.org/conference.php?id=21 (download slides or paper)
10. Simpson S., King M. and Miner S. (2023) Modeling of Desert Tortoise and FLADIS using reanalysis weather in support of Jack Rabbit III, 27th Annual George Mason University (GMU) Conference on Atmospheric Transport and Dispersion (AT&D) Modeling, Fairfax, Virginia USA, 20-22 June 2023, http://camp.cos.gmu.edu/
11. Tickle G., Hetherington R., Gant S.E., McGillivray A. and Tucker H. (2023) Effect of humidity on the dispersion behaviour of pressure-liquefied ammonia jet releases, 27th Annual George Mason University (GMU) Conference on Atmospheric Transport and Dispersion (AT&D) Modeling, Fairfax, Virginia USA, 20-22 June 2023, http://camp.cos.gmu.edu/ (download slides)
The final summary of the JRIII modelling exercise is presented in an Open Access paper in the Atmospheric Environment X journal:
· Gant S.E., Chang J., Hetherington R., Hanna S., Tickle G., Spicer T.O., McMasters S., Fox S., Meris R., Bradley S., Miner S., King M., Simpson S., Mazzola T., McGillivray A., Tucker H., Björnham O., Carissimo B., Fabbri L., Wood M., Habib K., Harper M., Hart F., Vik T., Helgeland A., Howard J., Mauri L., Mackie S., Mack A., Lacome J.-M., Puttick S., Ibrahim A., Miller D., Dharmavaram S., Shen A., Cunningham A., Beverly D., O¿Neal D.M., Verdier L., Burkhart S., Dixon C., Nilsen S., Bradley R., Skarsvåg H.L., Fyhn E.H. and Aasen A., "Pressure-liquefied ammonia jet dispersion: Multi-model intercomparison using Desert Tortoise and FLADIS data", Atmospheric Environment X, 28 (2025) 100389. (Open Access pdf available from: http://dx.doi.org/10.1016/j.aeaoa.2025.100389. For supporting information, see http://www.admlc.com/JRIII or the mirror site http://www.gant.org.uk/JRIII)
One of the motivations for the JRIII modelling exercise was to address scientific knowledge gaps in toxic industrial chemical model systems. These gaps were described in an earlier journal paper, which can be freely downloaded from the Process Safety Progress journal here:
· Hanna S., Mazzola T., Chang J., Spicer T., Gant S. and Batt, R. (2021) Gaps in toxic industrial chemical model systems: Improvements and changes over past 10 years, Process Safety Progress, e12289, 16 pp., http://dx.doi.org/10.1002/prs.12289
The JRIII modelling exercise used published data on the Desert Tortoise and FLADIS trials from a number of different sources, which are listed below:
· Carissimo B., Jagger S.F., Daish N.C., Halford A., Selmoer-Olsen S., Riikonen K., Perroux J.M., Wurtz J., Bartzis J.G., Duijm N.J., Ham K., Schatzmann M. and Hall R. (2001) The SMEDIS database and validation exercise, International Journal of Environment and Pollution. 16 (1-6), p614-629, https://doi.org/10.1504/IJEP.2001.000654
· CERC (2000) SMEDIS Model Evaluation Protocol, Report SMEDIS/96/8/D, Version 2.0, Cambridge Environmental Research Consultants Ltd (CERC), 7 December 2000, https://admlc.com/smedis-dataset/
· Goldwire H.C. Jr, McRae, T.G., Johnson, C.W., Hipple, D.L., Koopman, R.P., McClure, J.W., Morris, L.K. and Cederwall, R.T. (1983) Desert Tortoise series data report 1983 Pressurized ammonia spills. Lawrence Livermore National Laboratory report UCID-20562. https://www.osti.gov/biblio/6393901
· Nielsen M., Bengtsson R., Jones C., Nyrén K., Ott S. and Ride D. (1994) Design of the FLADIS field experiments with dispersion of liquefied ammonia, Report Risø-R-755(EN), Risø National Laboratory, Roskilde, Denmark, June 1994
· Nielsen M. and Ott, S. (1996a) FLADIS field experiments: final report, Risø-R-898(EN), Risø National Laboratory, Roskilde, Denmark, July 1996
· Nielsen M. and Ott S. (1996b) A collection of data from dense gas experiments, Risø-R-845(EN), Risø National Laboratory, Roskilde, Denmark, March 1996
· Nielsen M., Ott S., Jørgensen H.E., Bengtsson R., Nyrén K., Winter S., Ride D. and Jones C. (1997) Field experiments with dispersion of pressure liquefied ammonia, Journal of Hazardous Materials, 56, p59-105, https://doi.org/10.1016/S0304-3894(97)00029-0
· Nielsen M. (1998) Dense gas dispersion in the atmosphere, Risø-R-1030(EN), Risø National Laboratory, Roskilde, Denmark, September 1998
Video footage from the Desert Tortoise trials is also available here.
For further information on the JRIII model inter-comparison
exercise, please contact: Joe Chang
or Simon Gant
Return to: http://www.gant.org.uk/research
