Deepwater Horizon avian mortality study: Is it possible to estimate total losses?

A study modeling the possible track of birds who died at sea after the Deepwater Horizon (DWH) blowout suggests that the death toll may run as high as 800,000. BP America disputes the results, saying the numbers suggested did not correspond to what the Natural Resource Damage Assessment (NRDA) findings show.

Final information from the NRDA will need to be compared with this study

A problem in interpreting data from this study is that the legal case surrounding the Deepwater Horizon incident prevents some NRDA information, for example the Data Inputs from Beached Bird Assessment, from being released to the public until the case is settled. This means that the researchers were working with a limited database and their findings must be considered preliminary.

As can be seen in the comments in the New York Times article, scientists not directly involved in either the study or the NRDA are divided in their opinion on the results. However, details on the methods used, particularly the type of statistical analysis and self-assessment, can help independent scientists to determine whether the estimates stand up to further scrutiny, even before the final NRDA data are made available.

Both sides have a stake in the answers to the bird mortality question

Another important point is that the scientists behind this study are working for conservation organisations involved in environmental impact compensation claims, thus they and the NGOs they are working for, have a financial interest in the information provided in this paper.

BP’s financial stake seems clear, although funds for damage assessment and restoration are held by a trust. The final cost will be determined based on the results of the pending litigation. For BP there is also the concern regarding public perception of their response to the environmental impacts of the spill.

The type of statistical evaluation used can significantly influence the results produced and each side will look at the data available with an eye to supporting their position.

Many variables in the NRDA process

A look at the variety of factors that influence the number of birds which make it to shore, whether alive or dead, after a spill gives insight into the problem of estimating mortality. The California Department of Fish and Game lists a number of mitigating factors which may come into play:

  • Removal of carcasses by scavengers
  • Removal or burial by the public
  • Loss at sea due to winds and currents
  • Departure from the area
  • Beach transit (birds may not remain on the beach where they first came ashore)
  • Unsearched areas
  • Search efficiency
  • Rewash (narrow beaches may be swept clear with each high tide)

With 40,000 square miles of territory to search, it is likely that birds were missed. Several of the factors listed above may have been particularly significant in this spill:

  1. Departure from the area: The spill began during migration and the Gulf coast is an important migratory stopover. Birds on migration typically only spend a day or two feeding and resting at stopover sites before moving on. Obtaining data on degree of oiling and survival of these birds may not be possible.
  2. Loss at sea due to winds and currents: The study authors contend that many birds were washed out into the Gulf rather than onshore resulting in lower counts than might have occurred had the winds and currents been more onshore.
  3. Unsearched areas may include the marshes and mangroves along the coast where access is limited.

Other factors affecting recovery rates include the speed with which beach surveys were initiated after the spill, how quickly carcasses and live birds were recovered, and the experience levels of those involved in search and collection. It is also important to note that BP is basing their survival rates on short-term post release monitoring, a basis that a number of scientists feel overestimates long-term survival of oiled birds.

Evaluating the evidence preseented in scientific studies

This is the hardest part for for the average person, weighing the value of the information available. Most of the results from the NRDA work are not in the public domain at this time. Until all that information is made available and scientists with no personal stake in the situation can review the data and make further assessments, the information must still be classed as preliminary. For those wishing to better understand how to interpret such information, the last resource below may help.

Even when all the data is available, it will be some time before the full picture of the DWH impact emerges. Studies just now coming out regarding the Exxon Valdez (1989) and Prestige (2002) oil spills make it clear that damage assessment is difficult and that many factors, including type of oil, where the oil came ashore, local weather, season and species affected, can influence the long-term impacts of a spill.

Resources:

Still Counting Gulf Spill’s Dead Birds (New York Times article)
BP Response to New York Times Article
BP White Paper on NRDA and Mortality Estimate Tracking Studies
Estimating Bird Mortality (California Department of Fish and Game public information page)
Haney, JC, HJ Geiger and JW Short. 2014. Bird Mortality from the Deepwater Horizon oil spill. I Exposure probability in the offshore Gulf of Mexico. Marine Ecology Progress Series. 513:225-237. Doi:10.2254/meps10991
Interpreting Research Studies -Guttmacher Institute.

(Note: although this paper is concerned with social science research, the information regarding how to evaluate a research paper is common to all sciences.)