The Importance of Stakeholder Outreach and Risk Communications in a Major Oil Spill: A Case Study of the Deepwater Horizon MC-252 Oil Spill

Tremendous public concern over the potentially devastating effects of the largest oil spill in U.S. history was clearly present in the early weeks of the DEEPWATER HORIZON MC-252 Oil Spill and continued on well after the source was secured. The story had wall to wall coverage on cable news channels and was the top news story for the major networks' news programs for months. The intense focus and interest in everything surrounding the response-operations, stakeholders, bystanders-was to be expected. The explosion, search and rescue, loss of the DEEPWATER HORIZON and ensuring discharge of oil elicited an emotional reaction to this compelling and dramatic story. In this type of high concern, low trust situation, it is essential for officials to provide information to the public in an effective manner. The Incident Commanders for Louisiana between May 28 and August 20 recognized the need to incorporate the tenets of Risk Communication in order to effectively communicate the actions taken by the response organization to reduce the impact of the oil spill in the safest, most efficient, effective and environmentally benign ways possible to the public, political officials and other stakeholders. This was challenging for several reasons:

1. Magnitude of the Incident--at its peak, the response organization included over 47,000 responders, thousands of vessels, hundreds of aircraft and spanned all five Gulf States. East Coast states from Florida to New Jersey were concerned about possible oiling due to the “Loop Current” and made preparations for responding to this “inevitable” oiling. This anxiety persisted for months. There were literally millions of stories to be told by responders, fishermen, out of work hotel workers, tourists, political leaders, academicians, etc. Some of these people had direct knowledge of response activities, but many did not—however, this did not preclude them from speculating publicly about response methods or how things were going.

2. Complexity of the Incident--There were many activities going on at once in response to this incident. Some were very visible, such as In-Situ burning, skimming and picking up tarballs on the beach. Other activities were not, such as the daily process for determining whether to use dispersants, subsea oil sampling and the entire incident planning process. When an incident is visible, such as a tornado, it is easy to see what happened and what is being done to respond to the damage (e.g. search and rescue, removing debris). In this case, the actions taken were not as easy to see. As a consequence, some people had the perception that either (a) nothing was being done, or (b) something was being done, but haphazardly or without prior thought.

3. Duration of the Incident-- Until the well was capped on July 15,^, 2010, there was a new major oil spill every day. The unending sight of oil gushing into the Gulf of Mexico was a constant reminder of the catastrophe, keeping the story fresh in everyone's mind.

4. Other Factors—the summer of 2010 was a “perfect storm” of other issues that were keeping the attention of the public, and these factors played a part in the way the response was perceived. These factors included the recession, the upcoming mid-term elections and the anti-federal government and Big Business sentiment. Additionally, some politicians and media took the opportunity to paint the response in the most negative light possible, and kept this narrative going for months, for political and ratings gains, because the public was receptive to this view.

As mentioned above, this incident presented several challenges that made it virtually impossible to accurately paint a picture of everything that was going on to fight the oil underwater, on the surface and on shore, including steps taken to protect the responders, wildlife and the surrounding communities:

1. There was an insatiable appetite for information from many stakeholders including political leaders at all levels, news media, responders and the general public. The capacity to satisfy the need was insufficient, as there were not enough personnel assigned to the Joint Information Center and Liaison Officer staffs for the crush of information requests. In addition, early processes/policies for engaging these stakeholders were not sufficient for an event of this magnitude and significance.

2. There was more than one source for information. While there were processes in place to communicate a consistent message, with 47,000 responders working for various companies, agencies and levels of government, it was impossible for every single member to have the entire picture, message or opinion. This gave the perception of disjointed activities not working towards a common goal.

3. Getting heard: TV is a visual, dramatic medium. When presented a choice of broadcasting a dispassionate official giving a detailed explanation of a complicated issue or an angry person pointing at an oiled shoreline, many times the official's story never aired, or it was reduced to an out of context sound bite.

4. It's not just TV: There has been an explosion of on-line magazines, blogs and news websites. It is impossible to monitor all of them. Therefore, if something inaccurate is posted, it may never be corrected due to lack of visibility by the Unified Command.

There have been many studies that demonstrate that a person's perceived risk has no correlation to the actual risk a hazard poses (Slovik, 2000). For instance, many people fear flying, but have no problem getting into a car and speeding down the highway, despite the fact that flying is a safer mode of transportation. Some of the factors that affect this perception include whether an activity or situation is voluntary or involuntary, natural or manmade and familiar or unfamiliar. Another factor that affects perception is trust—can you trust the source that is giving you the message? If an agency, or company, or individual has a reputation (true or not) of dishonesty, shoddy work or accidents, then their messages may not be given any credence. Officials need to be sensitive to these factors when communicating messages.

According to David Ropeik, Risk Communication is the “Actions, words, and other interactions that incorporate and respect the perceptions of the information recipients, intended to help people make more informed decisions about threats to their health and safety” (Ropeik, 2008). The authors broaden this definition to include any communication situation where complex, unfamiliar and “scary” information needs to be conveyed, regardless of whether the information is related to the health and safety of the stakeholders, for instance “will the government “cut and run” once Bottom Kill is achieved?”, which is something that we heard frequently from stakeholders once the well was capped. Some people, especially scientists, believe that if someone is worried about a health concern, for instance “will living next to this waste dump kill me?”, then all you need to do is present the air and water sampling data to them and they will calm down. In fact, if someone is worried, then no amount of data will calm them down until you first address their feelings (Ropeik, 2008). Officials need to show stakeholders that they care are willing to discuss and understand their concerns and finally, must take concrete steps to build trust with those stakeholders before they will accept the message. This is not a comfortable position for officials to find themselves. The officials must be willing to take criticism, sometimes delivered emotionally, without losing their cool. They must admit to mistakes and explain what is being done to rectify the situation. Once the trust is established, then the stakeholders will be willing to listen to clear, understandable data.

There were several issues that came up that elicited a high level of concern from the public. These issues included:

1. Oil flow rate out of the well. While the Incident Commanders planned and responded for a worst case scenario and therefore were not worried about the exact flow rate, the public wanted to know the flow rate to put a number value on the penalty that BP should pay.¹ When the Incident Commanders said “the flow rate is not important. We're responding to the worst case scenario”, it's possible that what the public heard was “We don't think assessing a penalty to BP for what they did is important”.²

2. National Contingency Plan vs. Stafford Act. Some stakeholders were wondering why the states didn't have the majority say in the response the way they do under a Stafford Act disaster declaration. The Incident Commanders were operating under the National Contingency Plan which requires the Federal On-Scene Coordinator to direct response actions. This caused some conflict, because the Governor and local officials defaulted to what they knew from prior disaster (hurricane) response, which was Stafford Act and they did not like that the federal government didn't do everything they requested.

3. Unified Command consisting of CG, state and BP. Since the enactment of the Oil Pollution Act of 1990, the CG, designated state environmental agencies and potential Responsible Parties have exercised and responded using a Unified Command consisting of the Federal and State On-Scene Coordinators, as well as the Responsible Party. This is to ensure that the party that caused the spill stays engaged in the response as well as pay for it. It is at once a hierarchical organization with the Federal On-Scene Coordinator ultimately responsible for ensuring the cleanup occurs in accordance with the National Contingency Plan, as well as a collaborative process which ensures all members of the Unified Command are heard. While this concept is well understood by responders, it is quite confusing to the general public and many officials not familiar with oil spill response, who wonder why the “guilty party” should be part of the Unified Command.

4. Effectiveness of booms and skimmers. Due to oil type and environmental factors, booms and skimmers, while effective, can never stop or pick up 100% of the oil they encounter. Because this was not well understood, stakeholders believed that any oil hitting the shoreline was due to responder incompetence/failure.

5. EXXON VALDEZ incident vs. DEEPWATER HORIZON incident. People made comparisons between the two events, perceiving that one was just like the other—thick, heavy persistent oil coating huge swaths of Alaska with tens of thousands of oiled wildlife. People could not understand that the oil was lighter, the water warmer and the shoreline was not uniformly heavily oiled—basically the two spills had very different factors and the damages were not identical.

6. Incident Command System. The use of the Incident Command System, which featured a Unified Command, gave many people the perception that the planning process precluded the swift removal of the oil. They believed the process slowed down operations, when in reality it had the opposite effect of increasing operational efficiency.

These issues provided conflict, uncertainty, fear—in a word, drama, which is why they remained both newsworthy and on the minds of important stakeholders. Because these issues engendered such an emotional response, the stakeholders' trust level for the responders in general, and of the Unified Command in particular, was negatively impacted.

The authors realized that we needed to rebuild the trust with the various stakeholders. Without that improved trust, they would not believe the information we needed to disseminate. We took several steps:

1. Parish President Liaison Officers (PPLOs). To improve responsiveness and information flow to/from the Incident Command, Unified Area Command and the National Incident Commander, senior Coast Guard officers were appointed as PPLOs and Governor's Liaison Officer. These direct lines of communication were able to quickly address the concerns of these officials, and helped to rebuild the trust.

2. Open Houses. Early on in the response, Incident officials attended typical “Town Hall” style meetings. The officials would sit on the stage, and audience members would, one at a time with the rest of the audience joining in, yell at the personnel on the stage. At the end of the event, no one's concerns were addressed, and everyone left angry. Recognizing the futility of such events, the authors directed the Unified Command to instead put on Open Houses, which were called “Science Expos”, in which tables on various topics were set up in a gymnasium or other open space, and the audience members could walk up to each table's display and have a one on one dialog with subject matter experts. For example, the table on dispersants featured a graphic that showed the chemical constituents in COREXIT 9500 (Table 1), as well as some common household products which contained these constituents, including Klondike Bars, Goo Gone and laxatives. This simple “show and tell” went a long way to demystify dispersants and explain to people that it was not some exotic poison being sprayed haphazardly. It was striking to notice the difference in people's body language, conversations and stress level between the start of one of these Open House events and an hour later, after their concerns were acknowledged respectfully and their questions were answered fully in clear, understandable statements.

3. Media Accessibility. While this was a challenge due to the sheer size of the response, number of media requests and other complications, the authors pushed for as much access as possible. This included frequent embeds, interviews, news conferences, as well as pushing public affairs officers out to field locations to increase media availabilities.

   Other Stakeholder Accessibility.

   • The authors improved the lines of communication between the Governor's Office of Homeland Security and Emergency Preparedness (GOHSEP) and the Louisiana National Guard. Some of these lines were technology based—when the authors discovered that parish GOHSEP representatives were putting new oil sightings on a web based program called WebEOC and had assumed we had access (and were wondering why we weren't responding to these reports), we took immediate steps to get access to the program in our Command Post and into the parish Branch offices. Some of the lines were process oriented—for instance, every day the Unified Command took a count of the boom in the state. The Louisiana National Guard did the same thing. The numbers never matched, and this became an issue that had visibility all the way to the White House. After some investigation it was discovered that how the boom was being categorized and counted, as well as the time the boom was being counted was not consistent between both boom counting groups. The process was changed to have both groups take their measurements together, using a consistent definition of the boom categories in the count. This eliminated the discrepancy and improved the trust between the Unified Command and the National Guard.

   • Community Leaders. Tours were provided to civic leader groups from the areas around the Houma Incident Command Post. This gave these leaders direct access to the Unified Command and they could see for themselves all of the planning and other work that was going on to fight the oil. We frequently heard “we had no idea all of this work was going on. We will be sure to tell our neighbors and associates” from the participants during these events.

   • School Children. Visits were made to local schools to explain oil spill response to the children. While the local media applauded this outreach effort, some bloggers accused the presenters of feeding “oil company propaganda” to the children. After that, some of the presenters were told to stop, which ended these events. This goes back to trust. Apparently, the local media outlets trusted the presenters; the bloggers did not.

*Figure and household product display were developed by Susan R. Shelnutt, RPh, PhD, Senior Toxicologist, CTEH®

1. Take Advantage of the “Honeymoon Period”. Early on in any large event, there is a period where stakeholders will ask “who, what, where and when” questions of responders. The “why, how and who's fault” questions will come later. This honeymoon period can last hours to days (Austin, 2007). It is important to establish trust in this period by being forthcoming, proactive and clear with information. If there is uncertainty, acknowledge that—better to say “I don't know” than be caught in a bluff later.

2. Anticipate the “hard issues”. If possible, craft easy to understand messages before an incident for complex issues that you can anticipate, e.g. dispersants, ICS. If the situation is novel, try to develop those messages as soon as possible. People tend to remember the first thing they hear on a subject. You want the message to be yours.

3. Identify your various stakeholders, and determine how you are going to keep them updated. Make sure this list is comprehensive so you don't forget any key constituencies.

4. Order enough Public Information Officers/Liaison Officers for your organization to be able to properly support the various stakeholder groups. This investment will pay dividends.

5. Ensure key people in your organization receive Risk Communications training. Some of the tenets might seem counterintuitive, so people need to learn them when things are calm and friendly. Exercise these skills with periodic drills, “murder boards” and mock news conferences.

6. Ensure message monitoring is done to understand what is being said about the response. This doesn't just include watching TV coverage but also reading blogs, on line magazines and other internet sources as well as print reporting. The purpose is two-fold:

   • To check the accuracy of what's being said—correct any inaccuracies that are uncovered.

   • To check the view on the “street”—to understand the rumors, misunderstandings and conspiracy theories that inevitably come from major incidents. This also serves as a gauge as to the trust level—if there are a lot of rumors/conspiracy theories, then chances are the trust level is low.

7. “Operationalize” communications. Stakeholder support cannot be an afterthought. There were several times where local officials felt blindsided by decisions because they were not notified ahead of time. Trust was lost not by willfully keeping stakeholders in the dark but by failing to make notifications a priority. By considering the information needs of stakeholders as operations are being planned, there is more chance that stakeholder needs will be considered and proper messaging will be done.

Despite all of the challenges of this response, the authors believe that the answer to improved stakeholder perceptions is more engagement, not less. Stakeholder engagement must occur at all levels of the response organization e.g. the National Incident Command, Unified Area Command as well as Incident Command Posts, and it must be exercised by all elements of the Unified Commands. Federal and local officials, as well as Responsible Parties at all levels need to be willing to engage stakeholder groups, including the media, often and must give them as much access as operations (and operational security) allow and apply the tenets of Risk Communications to ensure the key messages get through. Many people believe that it is bad to talk with the media. It is much worse not to. If we don't speak to them, someone else will. Whose message do we want to get through?

It doesn't matter how good a response is--how much oil is picked up, how many birds are rehabilitated and how fishermen were able to resume their livelihoods--if the various stakeholder groups perceive that they were kept in the dark, dispersants poisoned their seafood and fishermen were sickened by hydrocarbon fumes. We must strive to communicate with these groups effectively, honestly and earn their trust. It may not be easy, but it is essential to ensure the success of the response.