Thoughts on Chlorine Dispersion

Earlier this week I reported on a study that TSA will be conducting on the dispersion of chlorine gas from a catastrophic release from a chlorine railcar. TSA is concerned that there is an apparent mismatch between the chlorine dispersion model that is being used to evaluate the risk of a terrorist attack on such railcars and the actual dispersion seen in the two relatively recent chlorine accidents involving such railcars. This week I’ve been thinking about some things that might affect the dispersion that the studies need to take into account.

Weather Affects 

During my time in the military I spent a great deal of time working in 4.2” Mortar Platoons, most of it working in the Fire Direction Center (FDC); I had a special aptitude for doing the calculations necessary for sending mortar rounds to their desired point of impact.

One of the things that I spent some time learning and practicing was the calculations necessary to establish and maintain smoke screens with white phosphorous munitions. A major consideration that had to be taken into account in the attempt to maintain a thick, opaque smoke cloud was the weather in the target area. Not only did we have to take into account the wind speed and direction, but we also had consider the atmospheric temperature gradient; how the temperature varied as a factor of the distance above the ground.

One of the most favorable gradients was where there was a layer of warm air above a layer of cooler air; an inversion situation. This kept the smoke cloud close to the ground and provided for the slowest dispersion of the cloud. In doing their analysis of the accidents in South Carolina and Texas, TSA investigators need to take a hard look at how the local weather affected the dispersion of the chlorine gas cloud. They should also look at the possibility of the low temperature of the gas cloud creating a micro-inversion layer that helped to hold the cold chlorine gas close to the ground and impeding its dispersion.

Chlorine Reactivity

Another thing that needs to be considered in this study is the reactivity of the chlorine molecule. Chlorine is very reactive and chemically combines with a wide variety of both organic and inorganic chemicals; this is what makes chlorine so toxic. The calculations that go into the standard chlorine dispersion model shown in the Chlorine Institute’s Pamphlet 74 assume that the entire amount of chlorine released continues to expand throughout the dispersion pattern.

In reality, as the cloud disperses, the amount of chlorine available to be in the cloud decreases as the chlorine is consumed in a wide variety of chemical reactions. The amount that can be consumed is going to vary based on a wide variety of factors. They include the types of material available for reactions, the amount of water (both as liquid and vapor) available in the area (water acts as a solvent necessary for many of the reactions), and even the temperature (higher temperatures would be expected to increase reactivity) at the site.

Practical Effect 

I am certainly not claiming that the above described effects explain the anomalies apparently seen in the South Carolina and Texas incidents; they are just some of the things that the investigators need to examine. If the studies show no underlying problem with the current model and can explain the anomalies as being based upon situational variables, what will be the practical effect of the study?

First we need to understand that models like those presented in Pamphlet 74 are not intended to be absolute predictions of what will happen with a particular gas cloud. There are too many simplifying assumptions being made in the model design to make that a practical outcome. If the TSA studies show that the apparent anomalies can be explained by readily identifiable mitigating factors, that would not reduce the usefulness of the present model.

Emergency response planners can still use the model for planning for a generic catastrophic rail car release. This can guide initial evacuation decisions while the additional information needed to analyze the mitigating factors can be gathered at an accident scene. For railcar incidents at fixed facilities the situation becomes more complex. For initial planning prioritization decisions, the present model can still serve as a method for determining the population at risk of potential adverse effects from a catastrophic release.

Facilities with the largest at-risk populations should receive a measure of resource prioritization when allocating those resources for developing emergency response plans. Once work is begun on developing an ERP for a particular facility, then clearly defined mitigating factors can be investigated and quantified at that site. Physical layout of the facility and its surroundings could then be plugged into a more advanced model to map boundaries for initial evacuation planning and establishment of emergency notification systems.

Site modifications could be investigated that could have additional release mitigation effects. Additional modeling could be done for weather mitigation effects. Models could be run with a variety of weather conditions examined based on typical weather for the site. The results of these runs could be included as appendices to the ERP. This would allow an incident commander to make a more effective initial deployment of emergency responders. The information produced in these multiple weather runs would also allow everyone to understand what weather changes to be on the look-out for during an actual incident. Responses to those weather changes could be incorporated in the ERP.

Flawed Models If the TSA studies indicate that there are fundamental flaws in the current models, it will be absolutely critical that those models be appropriately modified. Then updated predictions for hazard areas will need to be widely distributed to potentially affected communities, agencies, facilities and transporters. Such information should change emergency planning prioritization.

TSA will need to resist the impulse to place any sort of restriction on the distribution of the new information. This would certainly be a situation where a clear case could be made for wanting potential terrorists to have access to the information. Any information that indicated that the populous at risk would be smaller than the current models predict would make these railcars less productive targets of potential terrorist attack. This would serve to reduce the likelihood of such attacks.

Increased Participation 

TSA is to be commended for pursuing this valuable set of studies. As I mentioned in my earlier blog, I think that the ISCD folks at DHS have a clear interest in the outcome of this study and should be invited to participate. TSA should also consider including participation of people at the EPA responsible for regulating water treatment facilities, as they also have a clear interest in the outcome of the study. DOT representatives from FRA and PHMSA would also be expected to be able to provide valuable information and have an interest in the outcome of this study.

Industry participation should also be considered. Organizations like the Chlorine Institute and the American Association of Railroad both have a well understood interest in this issue. They should be included in the study planning and should be able to provide valuable resources to aid the investigations. Finally, TSA should bite the bullet and consider inviting representatives of advocacy groups that are actively campaigning against the continued transportation of chlorine gas by rail to participate in the study. Not only do these groups have clear interest in the results, but if the models are modified to greatly reduce the expected hazard area for a catastrophic chlorine release, their participation will greatly reduce the public outcry against the results that must be expected.