Link to NMFS Homepage Link to NOAA Homepage Keyboard Navigation Alaska Fisheries Science Center Homebanner

The Exxon Valdez Oil Spill: How Much Oil Remains?

  picture of a pit excavation in Prince William Sound Roughly 10,000 pits were excavated as part of the shoreline survey of Prince William Sound during summer 2001.

The 1989 Exxon Valdez oil spill (EVOS) in Prince William Sound, Alaska, released a minimum of 11 million gallons of Alaskan crude oil into one of the largest and most productive estuaries in North America.  During the summer of that year, the Alaska Department of Environmental Conservation (ADEC) estimated that 149 km of shoreline in Prince William Sound were heavily oiled and 459 km were at least lightly  oiled.  A year later a survey showed oiling had decreased 73 percent. Two years later in 1991 an interagency survey estimated only 1.4 km of shoreline were estimated to be heavily oiled. By 1992 the estimate of  heavily oiled shoreline was only 0.2 km.   After 3 years of unprecedented efforts to clean the polluted beaches and subsequent surveys showing declining contamination, it was expected that natural processes would disperse any remaining oil.

However, in 1993 the EVOS Trustee Council funded an additional survey that estimated 7 km of shoreline were still contaminated with subsurface oil. Smaller-scale studies dealing with continued clean-up efforts and restoration of oiled mussel beds conducted between 1995 and 1999 showed that oil was surprisingly persistent and often in a relatively unweathered state, containing high concentrations of toxic and biologically available polycyclic aromatic hydrocarbons (PAH).  Long-term monitoring in the oiled areas has also shown that fauna from higher trophic levels such as sea otters and sea ducks still have not recovered. It now appears the remaining oil deposits may have become a chronic source of low-level oil pollution within the spill-affected area.

Because a significant survey of Prince William Sound had not been conducted since 1993 and the cumulative extent of the remaining oil was unknown, concerns were generated by the public and scientific communities about the oil’s possible continuing effects on humans and fauna potentially exposed to the oil directly or indirectly.  Public perception of the amount of oil that remains on beaches within Prince William Sound has varied widely.  Without an accurate assessment of the extent of the remaining oil, subsistence food-gatherers, consumers of commercial fish products from the area, and tourists have used mostly anecdotal evidence as the basis for economic decisions regarding resource utilization in the affected area.  Scientists and resource managers also have lacked an accurate assessment of the amount of remaining oil in Prince William Sound necessary for determining further appropriate scientific studies and management actions.

Consequently, the Auke Bay Laboratory (ABL) with funding from the EVOS Trustee Council, took on the task of assessing the remaining oil along the shorelines of Prince William Sound during the summer of 2001. The primary objective of the project was to measure the amount of oil remaining in the intertidal zone of Prince William Sound.  Secondary objectives include determining the rate of decline of oil on these beaches, estimating the persistence of the remaining oil, and correlating the remaining oil with geomorphological features.


Table 1.  Target sampling effort for 2001
Prince William Sound shoreline survey.

Strata Km in Strata Km Sampled
Heavy 1990-1993 24 5
46 2
1989 only
43 0.7


Previous attempts to estimate the oil remaining on beaches affected by the Exxon Valdez oil spill have  relied mainly on Shoreline Contamination Assessment Teams (SCAT), field teams that perform comprehensive foot surveys of impacted beaches. The SCAT  survey crews estimated oiled areas based mostly on visual clues at the surface.  Although SCAT were useful for directing cleanup efforts immediately after the spill, it was determined that the SCAT methods would not be  useful for producing a quantitative estimate of subsurface oil contamination 12 years after the spill.

Sampling efforts for the 2001 survey  focused on  beaches assessed as  heavily and moderately oiled during the 1989-93 ADEC/EVOS Trustee Council surveys (Table 1 above).  Limited additional effort was allocated to areas that had not been surveyed since 1989 and where persistence was uncertain. The survey design examined about 20 percent of the areas heaviest hit by the spill, with an intensity that permits extrapolation to other hard hit areas for all of Prince William Sound. The survey design guarantees a credible minimum estimate of remaining oil in the area and will provide a confidence interval for the most likely amount remaining throughout the affected region.  This information is needed to predict oil persistence into the future and to determine associated risks to vulnerable biota.

The 2001 survey covered roughly 8,000 m of shoreline. Ninety-six sites were randomly selected from the total number of oiled beaches assessed during the ADEC/EVOS Trustee Council surveys.  A 72-ft vessel was chartered for 90 days (three 20-day legs and one 30-day leg) between May and September of 2001. In addition to ABL staff, three contract laborers per cruise were provided by the native villages of Tititlek and Chenega. The Alaska State Department of Natural Resources provided an archeologist on site to protect any cultural artifacts uncovered in the digging.

  photo of equipment used to dig pits Pits were dug manually with shovels and pry bars, mostly on cobbled beaches.

The 2001 survey adopted a stratified random/adaptive sampling (SRAS) design. Two random pits were excavated to a depth of 0.5 m (1.6 feet) in every stratified block (0.5-m verticle drop in tide height) within a grid system established at each site. If subsurface oil was discovered in any of the randomly stratified origin pits, then additional adaptive pits were excavated above, below, to the right, and to the left of the origin pit until the extent of the oil patch was determined.  Sampling methods utilized only manual labor; holes were dug with shovels and pry  bars.  Most beaches were cobble or armored with boulders, making digging physically challenging. The sampling effort translated into 6,775  random origin pits and several thousand more adaptive pits within 7.7 km of the spill-affected area (Table 2).

Table 2.  Summary of sampling effort.

No. of Sites
Surveyed (m)
No. of Random
Pits Excavated
No. of Sites
With Oil
No. of Sites
Without Oil
91 7,796 6,775 53 38

Two additional measurements were made from every oiled, random origin pit to aid in volumetric estimates. First, the oil was categorized into a standardized visual classification scheme developed by the previous surveys (light, moderate, or heavy oil residues). Then, the oiled zone within the pit was measured to the nearest centimeter. Also, a subset of  pits representing each visual oiling classification was weighed, homogenized, and collected for chemical analysis. A regression relationship based on the quantity of oil extracted and weighted from these samples will be generated to establish an estimate of the mean amount of oil in each oiling classification.

  picture of oil on a rock Oil was categorized into a standardized visual classification scheme.


Buried or subsurface oil is of greater concern than surface oil. Subsurface oil can remain dormant for many years before being dispersed and is more liquid, still toxic, and may become biologically available. A disturbance event such as burrowing animals or a severe storm reworks the beach and can reintroduce unweathered oil into the water. Results of the summer shoreline survey showed that the oil remaining on the surface of beaches in Prince William Sound  is weathered and mostly hardened into an asphalt-like layer. The toxic components of this type of surface oil are not as readily available to biota, although some softer forms do cause sheens in tide pools.

Table 3.  Number of random oiled pits, not including adaptive pits, stratified by 0.5-m tide heights.  Intertidal sampling zone ranges from +4.6-m to +1.8-m tide height.
0.5-Meter Vertical Drops (Tide heights) Total Number of Pits
Surface Subsurface
0.5 39 5
1.0 58 28
1.5 60 69
2.0 58 91
2.5 40 123
3.0 29 117
Total 284 433


Our  survey results indicate a total area of approximately 20 acres of shoreline in Prince William Sound are still contaminated with oil. Oil was found at 58 percent of the 91 sites assessed and is estimated to have the linear equivalent of 5.8 km of contaminated shoreline (Figure 1).  The overall 20 acre estimate of oil-contaminated beaches was more than twice the estimate from the EVOS Trustee Council survey conducted in 1993.  (The 1993 surveys covered more beaches, but dug far fewer holes).  Most of the oil found in 2001 was classified as lightly  oiled, but was still easily observed once it was uncovered (sheening, strong odor, and sticky) and did not require the aid of a mechanical sniffer or chemical analysis for positive identification.

In addition to the estimated area of remaining oiled beach, several other important points were evident.
1) Surface oil was determined to be not a good indicator of subsurface oil.
2) Twenty subsurface pits were classified as heavily oiled.  Oil saturated all of the interstitial spaces and was extremely repugnant. These “worst case” pits exhibited an oil mixture that resembled oil encountered in 1989 a few weeks after the spill - highly odiferous, lightly weathered, and very fluid.
3) Subsurface oil was also found at a lower tide height than expected (between 0 and 6 feet), in contrast to the surface oil, which was found mostly at the highest levels of the beach (Table 3).  This is significant, because the pits with the most oil were found low in the intertidal zone, closest to the zone of biological production, and indicate that our estimates are conservative at best.

  Photo of a pit with heavy oil residue A pit containing a heavy oil residue.


The possibility of continuing low level chronic effects of the Exxon Valdez oil spill seem very real now, although measurable population effects would be very difficult to detect in wild populations.  If there are continuing effects, it would be most likely restricted to populations residing or feeding in the isolated oil pockets.  Sea otters and harlequin ducks fall into this category. Researchers have been monitoring these populations’ poor recovery in heavily oiled areas since the 1989 spill.  The 2001 shoreline survey has provided new insights for possible sources of continued oil contamination.  This has stimulated future studies that will focus on the bioavailability of the oil and its impacts on species such as sea otters, harlequin ducks, and their intertidal prey.

We anticipate the significance of these results will be controversial and stimulate discussion.  Is the oil significant and to whom?  Are pink salmon or herring injured because of continued intertidal contamination?  Are near shore predators, like otters or sea ducks, at risk because they prey in this zone?  Are the area’s subsistence users avoiding appropriate beaches, or are they avoiding all beaches?  What can or should be done about the remaining oil?  If more cleaning is requested or required, will it do more good than harm?

The last beach assessment was completed in September 2001. Supporting chemical analyses will be completed in fall 2002, and a final report with statistical analyses and conclusions will be completed by April 2002.

By Jeff Short, Stanley Rice, and Mandy Lindeberg.


quarterly Jul-Sept 2001 sidebar Contents


Auke Bay Lab

National Marine
Mammal Lab

RACE Division

REFM Division


Quarterly Index

Quarterly Home