Toxic plume dispersion from thermal decomposition of ammonium nitrate

If heat exposure does not lead to detonation, ammonium nitrate or ammonium nitrate-based fertiliser may still thermally decompose. In this case, the main hazard is not blast overpressure, but the formation of a toxic plume containing decomposition products.

The exact plume composition depends on the fertiliser type, temperature, contamination, ventilation conditions, and decomposition behaviour. Nitrogen oxides are typically among the most important toxic products. For this example, nitrogen dioxide (NO₂) is used as a representative toxic component for the NOx plume.

This scenario demonstrates how a toxic plume from fire-induced decomposition can be modelled using a user-defined source term and the Plume Rise from Fires model.

Scenario

A fire occurs in the storage area with ammonium nitrate based fertiliser and heats part of the inventory. The heat exposure causes thermal decomposition of the fertiliser.

As the fertiliser decomposes, toxic decomposition products are released into a hot buoyant plume. The plume exits the storage through openings or damaged sections of the building and disperses downwind.

The objective of the simulation is to estimate the downwind concentration of NO₂ and evaluate the potential toxic impact at ground level.

Modelling approach

To set up your project file for the simulation of this scenario, you can follow the steps:

• Add background. Define the context and environment for the simulation.

• Add receivers (optional).

• Add equipment. Set the location of the NO₂ release on the map [3953986, 4015561].  

• Select models.

Model selection

For this scenario, the Plume Rise from Fires model is selected because the toxic gases are assumed to be released as part of a hot, buoyant plume generated by fire-induced decomposition. The heat produced during decomposition causes the plume to rise before it disperses downwind.

The model accounts for the heat production of the source and calculates the resulting plume rise and dispersion behaviour. It also considers atmospheric conditions, including the effect of an inversion layer, which can influence how the plume spreads and where ground-level concentrations may occur.

The model does not calculate the decomposition chemistry automatically. Therefore, the release rate, representative toxic component, release height, heat production, and other source-term parameters must be defined by the user.

The source term for this scenario is based on the Escombreras Valley fertiliser decomposition accident. In that event, NPK 15-15-15 fertiliser underwent thermal decomposition in a silo and generated a toxic cloud containing nitrogen oxides. The published analysis estimated an NO₂ generation rate of 2.16 kg/s and a fertiliser decomposition rate of 4.38 kg/s. For this example, NO₂ is therefore selected as the representative toxic component, and a mass flow rate of 2.16 kg/s is used as the toxic release rate. The associated heat production is represented by a convective heat release of approximately 7 MW (Baraza, X., Pey, A., & Giménez, J., 2020).

Inputs

Following values are used for demonstration purposes. In a project-specific study, the source term should be based on available information about the fertiliser composition, decomposition rate, heat production, ventilation conditions, release location, and expected decomposition products.

The key model input parameters include:

Results interpretation

The main output of this scenario is the distance to selected NO₂ concentration thresholds at the defined receiver height of 1.5 m. These results indicate how far the toxic plume may reach at breathing height.

The results can be visualised as concentration contours, which show the area affected by the toxic plume. Contours can be reviewed both at the receiver height and at the height where the maximum plume extent occurs. This helps assess ground-level exposure as well as the overall spread of the plume.

In addition, the Maximum concentration vs Distance graph can be reviewed to show how the NO₂ concentration changes downwind from the release location.

Because this is a hot plume release, the plume may initially rise above ground level before dispersing downwind or it may stay close to ground. The comparison between day and night conditions shows the influence of atmospheric stability on plume behaviour. In this example, the daytime case results in longer hazard distances for the selected AEGL concentration thresholds and lethality levels. Although unstable daytime conditions promote stronger atmospheric mixing and may dilute the plume more rapidly, they can also bring the plume back toward ground level more effectively. In the night-time case, the hot plume rises quickly and penetrates the inversion layer, reducing ground-level impact near the source and resulting in shorter hazard distances for this specific scenario.

Keep in mind that the results depend strongly on the selected source term, especially:

• NO₂-equivalent release rate

• Heat production

• Release height

• Atmospheric stability

• Wind speed

• Roughness length

Therefore, sensitivity cases are recommended.

Conclusion

This scenario demonstrates how toxic plume dispersion from thermal decomposition of ammonium nitrate-based fertiliser can be modelled in EFFECTS using the Plume Rise from Fires model. Unlike the detonation scenario, the main consequence is the downwind dispersion of toxic decomposition products, represented here by an NO₂ release based on the Escombreras accident source-term estimates. The results show that plume behaviour depends strongly on the interaction between heat production, atmospheric stability, wind speed, release height, and inversion effects. The comparison between day and night conditions shows that the most conservative meteorological case is not always obvious for hot buoyant releases. Therefore, project-specific assessments should clearly document the source-term assumptions and evaluate relevant sensitivity cases.

Download the project file

Explore the project file simulating the NO₂ dispersion from ammonium nitrate decomposition. Adjust map contours, select different graphs or multiple graphs at once, and evaluate how different hole sizes influence the received heat radiation dose. Inspect the receiver’s reports to assess the damage effect.

Click this link to download the project file

To view the project file, please open it using the EFFECTS software. If you do not have the software, you can download and use the free viewing demo version of EFFECTS via the link below.

Download EFFECTS free viewing demo

Resources

Baraza, X., Pey, A., & Giménez, J. (2020). The self-sustaining decomposition of ammonium nitrate fertiliser: Case study, Escombreras valley, Spain. Journal of hazardous materials, 387, 121674. https://doi.org/10.1016/j.jhazmat.2019.121674