Caveats, Disclaimer & Dunno: Report To FERC re LNG Says Data Lacking In What Might Happen in Accident or Sabotage

(May 15, 2004) -- Accompanied by numerous caveats -- including an explicit caution that it "does not and was not intended to provide a measure of risk to the public" plus a lengthy disclaimer of future responsibility -- a report by consulting firm under contract to the Federal Energy Regulatory Commission (FERC) says there's a lack of data showing exactly what might happen in an accident or act of sabotage involving tanker vessel carrying Liquefied Natural Gas (LNG).

In its Executive Summary, the 128 page document ("Consequence Assessment Methods for Incidents Involving Releases from Liquefied Natural Gas Carriers") by Houston-based ABSG Consulting Inc. for the Federal Energy Regulatory states in pertinent part:

This work considers the flammable vapor and thermal radiation hazards created by unconfined LNG spills on water resulting from an LNG cargo release. This includes review of literature on experimental LNG spills and on consequence assessment methodologies that are applicable to modeling of incidents involving LNG spills on water...

The descriptions of the recommended analysis methods are left to the body of this report. In general these methods can provide only rough estimates of the magnitude of effects for incidents involving large LNG releases on water. This is typically the case with consequence assessments, and it is important to keep in mind that the recommended methods cannot provide precise estimates of effects because of variability in actual incident circumstances as well as uncertainty inherent in the methods used.

In the particular case of the methods of interest here (i.e., methods for large releases from LNG carriers), some important issues include:

• No release models are available that take into account the true structure of an LNG carrier, in particular the multiple barriers that the combination of cargo tanks and the double hulls in current LNG carriers provide
• No pool spread models are available that account for wave action or currents
• Relatively few experimental data are available for validation of models involving LNG spills on water, and there are no data available for spills as large as the spills considered in this study

In making the modeling recommendations in this report, our goal was to select methods that provide the most accurate estimates possible; however, we recognize the limitations of the models we suggest. As a result of these limitations, the project team made some selections that we believe lead to conservative estimates (i.e., tend to overestimate the consequences of LNG releases). Clearly there is an opportunity to develop pool spread models that consider more realistic analysis of the spill behavior on the water surface. In the long term, additional research will need to be performed to develop more refined models, and additional large-scale spill tests would be useful for providing better data for validation of models. However, definition of the specific research tasks and objectives needed to accomplish that research was not in the scope of this project...

It is also important to note that this study addresses the potential consequences of large scale LNG cargo releases without regard to the sequence of events leading to such an incident or their probabilities of occurrence. As such, this report does not and was not intended to provide a measure of risk to the public. A thorough risk assessment would consider both the probabilities and consequences of hazardous events. And finally it should not be assumed that the levels of hazards presented in this study are the assured outcome of an LNG vessel release, given the conservatisms in the models and the level of damage required to yield such large-scale releases...

While studying the results of this report, readers should keep the following key points in mind:

Models have limitations. The recommended models/approaches represent a reasonable set of tools to aid decision making. However, it should be recognized that models have an inherent level of uncertainty, as described in this report. In some cases, the scientific community’s understanding of the fundamental physical phenomena is better than its ability to simulate such phenomena. Also, models cannot make decisions for us. They inform decision makers, who must integrate information on many different factors in their decision-making process.

Consequence assessment is only one piece of the risk picture. Understanding risk requires an understanding of (1) what can go wrong, (2) what the consequences might be, and (3) how likely the losses are to occur. This report focuses only on consequence modeling of potential release scenarios, not how likely such scenarios are to occur. Decision making related to scenario risks should be considered in the context of both potential consequence and expected likelihood (or frequency). With regard to potential attack scenarios, the expected likelihood is a function of both threat (the likelihood that someone would try to carry out a specific type of attack) and vulnerability (the likelihood that a specific type of attack would be successful and produce the expected consequences of concern). As noted in the report, LNG vessel and associated facility operations are highly regulated and closely monitored/controlled by authorities, so many layers of protection exist against losses. The dependability of these layers of protection was not addressed in this project, but are important considerations in understanding the total risk picture.

Risk perception and risk acceptance are complex issues. How individuals and groups of individuals perceive or accept risks depends on many factors, which are often subjective with no clear right and wrong answers. Even when very precise/certain risk information is available, different people often react to the information in different ways. This project made no attempt to place value judgments on what risks people should or should not accept.