Back to Thinking Like a Scientist: Summary

Full Interview: Dr. Alan Mearns

A NOAA research team, including Alan Mearns, examines their sampling quadrat on a cobbled beach in Prince William Sound.

A NOAA research team, including Alan Mearns (in yellow pants), examines its sampling quadrat on a cobbled beach in Prince William Sound, Alaska. (Photo credit: OR&R, NOAA)

Have you ever wondered how marine life fares following an oil spill? How many die? How many survive to die later? How long does it take for an ecosystem to recover? What about the site of the Exxon Valdez oil spill—has the marine life along the coast recovered after 15 years? As a NOAA scientist, Dr. Alan Mearns has been asking and answering these types of questions since the spill occurred in March of 1989.

Dr. Mearns is a marine ecologist at NOAA's Hazardous Materials Response Division in the Office of Response and Restoration, a part of the National Ocean Service. His office is in Seattle, Washington. His job is to provide scientific information during the cleanup of oil and hazardous substance spills in coastal and marine waters. We spoke to Dr. Mearns about his work with oiled ecosystems and particularly his work in Prince William Sound, and his namesake—Mearns Rock. Here is what he had to say: 

Interviewer: When did you first start asking scientific questions about the aftermath of the Exxon Valdez oil spill?

Alan Mearns: Actually, I was involved in the spill from the very first day. I was in the NOAA HazMat (meaning Hazardous Materials) "war room" in Seattle on the morning of March 25, eight hours after the Exxon Valdez struck Bligh Reef. The HazMat Director was briefing the NOAA Administrator, telling him "This is the big one," and seeking support from NOAA. By that time, several NOAA HazMat staff members were already on flights to Alaska.

On that first day, my thoughts centered around how this very large oil spill was going to affect the marine life and wildlife of this productive and remote area. I also wondered how the arrival and activities of thousands of cleanup workers, hundreds of boats, news crews, and dignitaries were going to impact the area. I feared that these human activities might impact Prince William Sound as much as the oil spill itself!

Just a month after the spill occurred, I was asked to participate in a NOAA summer research cruise to survey shorelines and marine life in the spill area. The NOAA Research Vessel Fairweather was retrofitted to conduct biological studies for this purpose, and it began surveying the oiled coast of Prince William Sound in May. I joined the cruise in July. My task was to collect and process samples of sediments from several dozen sites along the Kenai Peninsula, Cook Inlet and Kodiak Island (See a Map of the Exxon Valdez oil spill). This was my first exposure to a major oil spill. During that cruise, I saw many sites, oiled and unoiled, and lots of wildlife. I have visited many of the oiled sites nearly every year since.

Interviewer: In your research with oiled marine life at the site of the Exxon Valdez oil spill, what was the first question you wanted to answer?

Alan Mearns: The first question was: Would the cleanup methods speed up the recovery of the shoreline marine life, or would they actually delay the recovery? Little did I know, then, that my colleagues and I would spend the next 10 years trying to answer this question.

In the spring of 1989, the government approved the use of high-pressure, hot-water washing to remove oil from the seaweed-covered, rocky, boulder, cobble and gravelly shorelines. Floating barges were built that were fitted with large pumps and water heaters to deliver this high-pressure hot water to the shoreline cleanup crews. The crews were using large fire hoses to blast oil out of the rocks and cobble, causing it to float on top of the water (called “refloating”) and washing it back out into the water, where it could be picked up by skimmers fitted on boats.


In the end, we returned to that first, very basic question I had asked on the first day:  To what extent would this type of cleanup enhance or delay recovery of shoreline marine life? We quickly broke down this simple question into numerous smaller, more answerable questions, and our long-term monitoring project was born. This is a typical scenario in a scientific investigation—scientists often take a very general question or an overwhelmingly large question and break it down into a series of smaller questions that they can answer through experimentation or observation.

Interviewer: Explain to us how you went about deciding how you would answer that first question.

Alan Mearns: We used the same approach that most scientists use when they first tackle a question. First, we read all the relevant published literature about the effects of both oil spills and shoreline cleanup activities on marine life. We found out, however, that not much was published on these topics. There had been some studies about the cleanup of oil spills, but we did not think the information was good enough to predict whether the hot-water cleanup method would be effective (in removing oil) or if it would really speed up “recovery.” To find out for sure, we decided that we needed a field-monitoring study at the site of the Exxon Valdez oil spill at both oiled and hot-water cleaned sites.

Fortunately, another research team had already started sampling and surveying the sites we needed. These included sites that were:

A lan Mearns and colleague establish a transect at Mearns Rock, a large boulder located in Snug Harbor on Knight Island in Prince William Sound, Alaska.

Alan Mearns and colleague establish a transect at Mearns Rock, a large boulder located in Snug Harbor on Knight Island in Prince William Sound, Alaska. (Photo credit: OR&R, NOAA)

Oiled (but not cleaned)

Oiled and cleaned (with high-pressure, hot-water washing)

Unoiled and not cleaned (called a “control”)

It was a grand experiment, with almost all the attributes required of good science, including exposed and control sites (for both oiled and cleaned), replication (more than one site in each category) and repeated sampling over a long period of time, provided we could argue for long-term funding.

We also had some data from parts of Prince William Sound that another research team happened to be studying before the Exxon Valdez oil spill took place—which was very lucky for us. We wanted to study all of these sites for several years using this “sampling design,” as scientists call it. One of the first questions my research team posed was: 

Would the sampling design be sufficient to allow us to clearly conclude when the marine life of oiled versus oiled and cleaned shoreline sites had actually recovered?

In other words, we needed assurance that the control sites were indeed comparable, so that over several years, we could fairly compare recovery at the oiled sites with the natural variation that we expected to occur in the absence of oiling and cleanup.

Additional sampling design questions we asked were:

How many study sites or sampling areas needed to be monitored in each category (oiled; oiled and cleaned; and control sites)?

How frequently must we monitor these sites?

How many replicate samples must we take at each location?

How often must we sample, to make sure we can claim that "recovery" has occurred?


First, we made sure that the control sites were indeed unoiled by checking detailed Coast Guard 1989 cleanup records. We needed to know if the unoiled control sites were similar to the oiled sites in terms of wave exposure and other forces of nature. We looked at detailed maps to see where each shoreline site was located, what kind of substrate it was (rocky, boulder, cobble, mud flat) and how it was exposed to wind and waves.  

Next, we evaluated how many sites would be needed in each category. We knew one each would not do because marine life and shoreline habitats, whether oiled or not, are extremely patchy. You could easily come to the wrong conclusion by chance alone. We sought three, four or five sites in each of the three categories.

Next, we considered how many replicate samples we needed at each site, at each time period, and at each of several tidal elevations (upper, mid and low intertidal). Again, one sample each would not do. Basic statistical analysis told us that we should have at least five replicate samples of each of the categories (oiled, oiled and cleaned, control). Fortunately, in most places, 5 to 10 samples had been taken at each site and each elevation in 1989.

Finally, we made our first "reconnaissance" trip to Prince William Sound in the spring of 1990, confirming that most sites were adequate, adding or adjusting sites and sampling locations, and just learning the "territory.” With two boats and a crew of nearly a dozen scientists and technicians, we went out every summer from 1990 to 2000, surveying these sites during the "spring tide" period (a time of very low, low-tides). After 2000, I led a small research team of three to continue photographing a dozen sites, including Mearns Rock.

Interviewer: What are the some of the field methods you use to study oiled marine life, especially along the coast? What technologies have you used? How have they assisted you in gaining a better understanding of oiled marine life?

Alan Mearns: Any shoreline field investigation requires the following:

a tide table to plan how much low-tide time we had at each site;

a statistically based monitoring plan (as described above);

the logistics for training staff and getting to and operating in the field locations (everything from boats and food to rain gear);

tools (such as GPS) for precisely locating stations and samples;

a team of people who are experts in identifying hundreds of species of marine plants and animals (called taxonomists);

people who are not taxonomists, but who can record the data;

taxonomic reference books;

photography equipment;

dozens of field notebooks and data forms;

biological survey tools (quadrats, core tubes, shovels, labels, jars, and preservatives); and

thick steel stakes (i.e., rebar) or marine putty to use as permanent markers at each site.


To survey the marine life on the shorelines, we used methods that gave us quantitative data like the kinds of species present, and their numbers or percent cover in each sampling area. On our first visit to each site, we randomly selected five or 10 points along a 100-foot line parallel to the water line (called a transect or transect line), using a “random numbers table.” At each point on a rocky shore transect, we laid down a one-quarter-square-meter quadrat (0.25 m on a side: you can see quadrats and a transect line in some of the photos on this page) and then identified and counted every plant and animal inside the boundary of the quadrat. Lastly, we used steel rebar or marine putty to permanently mark each quadrat location, so that we could relocate these exact points in subsequent years of our study and resample the same locations. This is known as a “fixed random” experimental design. The sample locations (quadrats) were randomly selected only the first time. In subsequent samples, the same locations are sampled again—so the locations were “fixed” in the remainder of the study.

This beach, near Knight Island, was heavily oiled and then cleaned in 1989.

This beach, near Knight Island, was heavily oiled and then cleaned in 1989. (Photo credit: OR&R, NOAA)

In areas of "soft" sediment (sand, gravel, or mud) we took a core sample and then, back at the lab, we carefully sorted through the sample and counted and identified all the clams, amphipods, polychaetes, and other marine organisms. Samples were also taken of the sediment to determine grain size distribution, organic content, and oil concentration (more specifically, polycyclic aromatic hydrocarbons or PAHs). Mussels were also collected from each sample location to test their tissues for PAH levels. We could also use the PAH information to “fingerprint” the oil to determine its origin. We did this at three elevations: upper, middle and low tide elevations.

Our fieldwork had to be done accurately but quickly, because we had a finite amount of time before the tide turned. If we stayed too long as the tide began to come in, we were flooded off the site! Surveying a dozen sites often took a week or more of especially low spring tides. We repeated this process for many sites, resulting in hundreds of quadrat samples, and thousands of numbers for each survey. One or two weeks of such fieldwork resulted in nearly a year of laboratory work, statistical analysis and report writing!

After three or four years, we thought we had our answer: The cleaned sites actually took a year longer to recover than the oiled, but uncleaned, sites. To make sure, we continued the study, eventually working with a second group of researchers. When we reviewed the new data and all the old data using newer statistical methods, we discovered that there were no real differences in recovery times! Thus, we had to modify our conclusions and recommendations: heavy cleaning did kill off marine life that otherwise survived the oiling, but the recovery time was about the same. Heavy cleanup made little difference in the end; it certainly did not remove all the oil, nor did it speed up recovery time of the shoreline marine life.

After 10 years we were ready to end the program. However, I noticed from our data, and from 10 years of shoreline photos, that something else was going on with the intertidal marine life in Prince William Sound. It looked like mussels, seaweeds, clams, barnacles and other organisms were going through a series of cycles. In some years, the shores were covered with mussels. Yet, in other years, mussels were nearly absent and the places were thick with seaweed. And, in other years, hardly anything at all grew on or inhabited some sites. This variation had no impact on our scientific study, however, because we were always comparing oiled and cleaned sites to unoiled sites.

The photos from one site, which featured a large boulder, clearly showed this long-term variation. My colleague and fellow biologist Gary Shigenaka gave this boulder a name: Mearns Rock! We used the Mearns Rock photos to capture the imagination of the boating public. Since 2000, I and several colleagues have continued to return and rephotograph our oiled sites, including Mearns Rock, in the company of local citizens.

Interviewer: How will you use the data you collected at Mearns Rock and other locations in Prince William Sound?

Alan Mearns: As you may have noticed, the Mearns Rock photo series is only "the tip of the iceberg"; a small piece in a much larger puzzle. Mearns Rock was one of our study sites that was oiled but not cleaned. One purpose of the Mearns Rock photos is to see if we can use these landscape-scale photos to replace some of the very intensive and expensive quantitative biology methods (described above) at future oil spills. We have the photos and the data we collected for the past 10 years, but have not yet begun to compare them. 

Another purpose of conducting our annual trips to Prince William Sound was to train local citizens in hopes that they would continue long-term monitoring and photography once we were gone. Each year, we invited members of the Whittier Coast Guard Auxiliary to join us as we traveled around Knight Island. Often, two or more vessels operated by citizens and their families accompanied us. We took them ashore, showed them the photo sites, and dug pits to look for signs of oil.

Showing the photos of Mearns Rock is only the first step in the scientific process. To begin to answer the "so what?" questions, we need to look at similar photos over the same period from other locations, especially locations that were not oiled. About half the people that look at these photos ask, "Do you have pictures of an unoiled site?" If you want to begin understanding why marine life has been changing, that is exactly the right question to ask! We hope to make photos from all the sites available in the coming year.

Interviewer: What is the present (2005) situation in regards to rockweed (Fucus) and mussel populations and other marine life in Prince William Sound, especially at Mearns Rock?

Alan Mearns: This past summer (June 2004) three of us, accompanied by local citizens, visited about 10 of our oiled, oiled and cleaned, and control sites, including Mearns Rock, which is on Knight Island. From 2001 to 2003, the percent cover of marine life at most sites was very low. This summer it looked like the rockweed was starting a new period of exuberant growth. Mussels were not particularly abundant. Overall, the abundance of conspicuous intertidal marine life has varied greatly at all sites, regardless of whether they were oiled, oiled and cleaned, or unoiled.

Interviewer: Overall, what changes have occurred in Prince William Sound and Mearns Rock since 1989?

Alan Mearns: During 1989, the oil killed off about half of the marine life on the oiled shores. Washing the shore with high-pressure hot water killed off most of what survived oiling. This washing removed a lot of oil, but not all of it. By the second summer (1990), seaweeds, barnacles, snails, limpets and other organisms were coming back, but a lot of oil remained in gravel under the shoreline surface. By the third and fourth summers (1991 and 1992) there was, with notable exceptions,  a prolific growth of seaweeds, including rockweed, and intertidal animals at all of our shoreline sites. After that, the abundance of algae and animals at oiled sites (both cleaned and not cleaned) varied considerably from year to year, just as similar variations occurred at the unoiled sites.


Interviewer:  Has Prince William Sound “recovered” from the 1989 Exxon Valdez oil spill? What is the evidence to support your answer?

Alan Mearns: As a scientist, I find it difficult to answer the general question, “Has Prince William Sound recovered?” Imagine if a storm completely trashed your garden and then it started to "recover." When would you say that it had completely "recovered"? After a year, you would have some plants and birds come back, looking like they did before, but other plants would be much slower to regrow, and some would have died, perhaps to be replaced by weeds for a season or two. You would have a change in the amount of shade and sunlight coming into various portions of the garden, changing the rates at which various plants grow and flower. And, you would have year-to-year variations in rainfall and soil moisture, which would also change the course of your garden’s "recovery."

So let me try to answer the question about specific groups of organisms affected by the oil spill rather than all of Prince William Sound. In terms of the abundance and diversity of shoreline marine life in oiled areas, meaning the seaweeds, barnacles, mussels and other invertebrates, they have recovered. In fact, that kind of recovery took place within three to four years after the spill, but we didn't know we could say that with specified scientific certainty until we "oversampled"  for a period of 10 years. A different answer might be this: all of that shoreline marine life was going to return, no matter what we did. But oil still exists in the shoreline sediments. Of the 10 sites we visited this year (2004), we dug small trenches in soft sediment and found oil sheen at six of the sites. One site has what looks like an asphalt pavement in the upper intertidal zone. The oil had mixed with gravel, then dried out.

Other researchers from government and industry have studied other groups of wildlife affected by the oil spill in Prince William Sound. The answers are mixed. Bald eagles recovered to their former abundance many years ago. Sea otters, however, may still be suffering at several locations, because they continue to feed on contaminated clams and mussels.

Some years ago, an article by writer Marguerite Holloway appeared in Scientific American. She used the French term, mis en mis—“window in a window”—to describe how people look at recovery. It depends on how close or how distantly you look at Prince William Sound and all the data compiled since March 25, 1989.

Interviewer:What were some of the challenges you faced in conducting your research at Mearns Rock?

Alan Mearns: One was simply getting there! Mearns Rock and the other 12 study sites are located on Knight Island, over 50 miles southeast of the nearest port in Whittier, Alaska. We have been very fortunate to get to the Mearns Rock site each year despite weather delays and engine troubles. May and June are often pleasant months in Prince William Sound, but storms can pop up anytime, especially in July. An engine failure nearly terminated our 2002 survey, but we completed it with the help of a citizen who volunteered his time and vessel.

Another is the timing of the tides. We need a good "minus" tide during daylight hours to photograph our sites. One year, the low tide was at dawn on a very dark and cloudy morning. Our film speed was not fast enough to get quality pictures and I had to rely on the low-light capability of a video camera to get passable photos. We always had to have a contingency plan, had to be ready to go when the tides and weather were right, and to hold back when they weren’t.

Also, there was the challenge of funding and support. We have been very fortunate that NOAA and other agencies have funded this work for over 10 years. But it has not been easy. Funding cycles go up and down, and in some years we had to argue and argue that the program should continue for at least one more year. Part of the "art" of being a scientist is being a good proposal writer and giving good speeches and presentations. Fortunately, the trips to Mearns Rock and other sites were relatively low in cost, and it was easier to justify a brief annual photo-survey than it was to justify a major annual biological survey.

Interviewer: Do you or other NOAA scientists continue to study the marine organisms on Mearns Rock? What about other locations in Prince William Sound?

Alan Mearns: Yes, our NOAA HazMat team continues to monitor sites in Prince William Sound. In addition, the Exxon Valdez Trustee Council and the Oil Spill Recovery Institute continue to make measurements in the sound.

In the spring of 2001, we had a surprising opportunity to try another approach to answering our question. A large earthquake in Prince William Sound caused a landslide on northern Knight Island, only a few miles from one of our study sites. The landslide buried about 400 meters of old shoreline, creating a new one composed of dozens of house-sized boulders. We used this new shoreline to monitor how marine life colonizes bare rock from the very beginning, and in the absence of mortality from an oil spill. Scientists call this a “natural experiment”—in other words, an experiment created by Mother Nature’s manipulation of a situation, rather than a planned manipulation by scientists.

We predicted that recovery of marine life at this new site would be identical to what we saw in 1989-1993 at the oiled and pressure-washed sites. Indeed, our three years of photos clearly show the same sorts of recolonization patterns that we saw elsewhere in the early 1990s, including an explosion of mussels in 2003, three years after the landslide.