Oil Spill Fingerprinting and Source Identification by EUROCRUDE
25 Sep 2000
Van Enckevort, Harry
Oil and oil products are recovered, transported and transferred in large quantities through the world. The pollution of the marine, coastal and harbour environments by these products continues to be an issue around the globe. Areas of the Danish coast are continually affected by crude oil pollution. The pollution might arise by natural seepage from the sea floor, but typically it occurs as a result of an accident, such as a well, shipping or bunkering incident, or by deliberate discharge from the bilges of vessels.
Bilge discharges are subject to the MARPOL Regulations. A discharge that is in compliance with these regulations should neither lead to pollution of the shoreline nor to visible oil on the feathers of seabirds. Discharges that contain high concentrations of oily material are illegal and the perpetrator might be sought. Responsibility is sometimes declared without recourse to evidential proof, but evidence of culpability is generally required.
Legal proceedings may follow and the desire for successful court prosecutions has been a strong driver to improve the efficacy of means to identify the culprit.
I am not going to talk about cases where the responsible party is patently obvious because a ship has run aground, or you have eye witness accounts and photographic evidence of oil running down the side of the ship. But even then, the concept of an oil spill fingerprint and forensic evidence has its uses. For example, in the Exxon Valdez incident, oil spill fingerprinting showed that Prince William Sound was not the pristine environment it was thought to be. Oils from other sources were present, including those from historical accidents, commercial boating traffic and natural seepage of oil from the sea floor. This might have consequences for determining the extent of liability and charges for cleanup.
Oil is a complex mixture of chemicals. It reflects the origin of the materials from which it was made and the geological, geochemical and other processes it has undergone. Crude oils are formed from vegetative material, be this land-based plants or water-based plants such as algae. Plants are veritable chemical factories. They make a wide variety of different products - some are plant specific, some in common with other plants but in different proportions between different plants.
The composition of a crude oil is therefore dependent on the plant life at the time of sedimentation. Different flora in different areas gives a different mix of chemicals.
There are then the modifying geological and geochemical factors that give a crude oil from any one source its particular characteristics. Exposure to oxygen, pressure and temperature affect the chemical processes that occur in the formation of the crude oil and its maturation. There are filtering and chemical sorting processes that occur during the migration from the source rock to an underground reservoir, a path that may be kilometres long. And, there are the changes that occur by biodegradation, water washing, mixing, and relative mobility issues in the reservoir.
Although superficially similar, there would seem to be great potential to develop an oil fingerprint and uniquely identify an oil and provide forensic evidence in oil spill cases to prove or disprove a vessel as the source of a spill. This has been the subject of scientific study for many years, and a raft of chemical tests have been applied to crude and refined oils in an attempt to do just this.
You may have heard of scientific techniques such as UV fluorescence, Infra red spectroscopy, nickel/vanadium ratios, iron, copper and /or sulphur content. In practice, these tests are of low discriminating power – they are not good at telling you when oils were different or narrowing down the field of suspects
Hydrocarbon profiling is a more useful technique. Oils are about 75% hydrocarbons, there are many different individual hydrocarbons, and the hydrocarbon distribution can be used to profile and identify a product type, for example, petrol and diesel. It can also distinguish between some of the different fuel oils and some of the crudes.
For spilled products, the situation is more complicated. Spilled oils spread out to cover large surface areas, many of the hydrocarbons are very volatile, oil components can dissolve in the water, chemical changes can occur on exposure to sun and air, biochemical degradation can occur, and these is the possibility of contamination by other products. The spreading of the oil can accelerate the rate of all these processes. The longer the spilled product is in the environment the more difficult it becomes to identify or exclude vessels as possible suspects in a spill incident using hydrocarbon profiling.
The evidence available for legal proceedings using these techniques is often along the lines of “could have come from this ship”, “cannot exclude the possibility it has come from another ship”, “can’t tell whether it has or has not come from this ship”, and so on. Many may have seen the legal adjudicator err on the side of caution, the result being no prosecution, when there is no useful forensic scientific evidence.
Apart from the chemicals mentioned so far, petroleum oils also contain a diverse and complex group of chemicals known as biological markers, or “biomarkers” in short. These are natural chemicals formed by the plants and cannot be made artificially by man. There are some 100 biomarkers, and even though they are present at only very low concentrations, their complexity and the natural variations in their relative proportions potentially could provide a unique fingerprint.
With advances in scientific capability, there have been attempts to develop formal systems of fingerprinting, and until recently the most advanced of these was the NordTest method. This was developed in 1983 by 5 Nordic laboratories, and later modified. The focus was successful prosecution, recognising that the scientific evidence would not necessarily identify the offender, but was an important part of the legal process. The method considered forensic evidence requirements and came with detailed recommendations for sampling, handling samples, transport, chain of custody, storage and analysis procedures. Analysis was by standard scientific methods that had been available for some 10 years prior so court acceptance was not an issue.
The NordTest method was devised for water borne oils, and could analyse small volumes of spill. It was suitable for crude and refined oil products with boiling points higher than kerosene at 200C. This did not mean that you couldn’t do an analysis on the lighter products. Rather, the reliability and efficacy of the testing is much reduced because the chemical patterns are much simpler, have smaller differences between them and the problems of evaporation and water washing were much more significant.
The NordTest took a tiered approach using a hydrocarbon profile for initial testing and to separate those sources that were not responsible for the spill from those which might be. It then looked at 29 different biomarkers that were either not affected by weathering, or were affected predicably to try to reach a more definitive conclusion. Weathering, if an issue, could be taken into account with a mathematical calculation.
The NordTest advanced the use of forensic scientific evidence and the interpretation of results in oil spill cases. In conjunction with other evidence, it could identify the perpetrator of a spill - as usual, where there was any doubt it went in favour of the suspect. It became and important part of the legal process in some countries despite not being able to provide a unique fingerprint. Another of the limitations was that it remained necessary to have a sample of spilled oil and samples from a suspect ship.
Enter the EUROCRUDE system of fingerprinting. EUROCRUDE was an attempt to achieve a more effective identification and to work in those cases where there was no suspect ship. It was developed through the collaboration of 6 European countries. It recognised that there had been advances in scientific capability since the NordTest was developed, and there was greater accessibility to computers and software packages for statistical analysis and comparisons.
The goal was to develop a database of fingerprints for crude oil produced or transported in European waters, at least in the first instance, and to evaluate or develop statistical methods and computer software for comparing and matching fingerprints in a database.
The biomarkers were the focus of interest and the team went thorough an extensive exercise by testing many different crude oils to identify which of the biomarkers would enable them to produce a fingerprint and the highest evidential value. As part of this they developed standardised methods that would give reproducible fingerprints suitable of use in a database and which any laboratory with the appropriate scientific instrumentation could use. The revisited the effects of weathering and decided which biomarkers were most effective for weathered oils.
The result? They found that fifty six biomarkers provided a fingerprint sufficient to identify the country and oil field of origin. Sometimes it could even identify the oil well of origin. It also used biomarkers with greater resistance to weathering so time delays between spill and sampling would not be crucial. This meant that even when the vessel was long gone from the scene, shipping records could still be used to identify the culprit. It was not necessary to have samples from the ship as the records could identify which ships were carrying what crude, where they were bound, and therefore whether they were potentially in the area at the time of the spill.
Again there was no new science so court acceptability was not an issue – the methods have been in use for oil exploration work and scientific testing for some time. It also worked for refined products and was more effective than the NordTest as it used a larger number of biomarkers to form the fingerprint.
And, also importantly, it was discovered that the biomarkers in a vessel’s bilge contents and discharge are also a fingerprint of the ship. They reflect the history of the ship – the products it has carried and used over time. So by comparing “fingerprints” of different spilled oils with material from the bilges of ships, it is possible to determine from which ship a spilled oil came. For this work, samples from ships are required for analysis. And, like the NordTest method, hydrocarbon profiles can be used in the first instance to separate suspect ships into “eliminated” and “possible” groups.
Again, it must still be evident that despite the advances in the power of method, science doesn’t do it on its own. It is only a piece in the jigsaw of any puzzle or case to solve which ship is responsible or not responsible for a spill. The scientist must work in conjunction with others to build the case.
And so I am now going to pass Captain Robin Keer-Keer, Harbour Master at Environment Canterbury, New Zealand, to describe a case where the principles of EUROCRUDE were applied to a spill and a guilty plea was entered in court proceedings.
It was a blustery cold Saturday morning at 0700 on Saturday 3rd August 1997 in mid winter with a southerly wind blowing at 30 knots after the passage of an active cold front. The air temperature was 4C, with moderate rain and at times sleet. The on duty pilot rang to report an oil spill in the Lyttelton inner harbour in the area of number 2 wharf and the tug and ferry wharves.
The pilot tells me there is a lot of oil around and he has put the pilot launch crew on standby. His description of the spill is ‘thick black oil.’
I was still in bed listening to the radio news! This was not the kind of news I wanted to hear about.
Environment Canterbury has a small, trained team of staff to respond to the environmental cleanup and the investigation and evidence gathering from a possible polluter.
With my ‘Environment’ hat on I realised it was in our interest to get on site as soon as possible, especially with the air temp being as low as 4C. and the sea temp being about 8C – 9C. The effective window for a dispersant clean up was not going to be open for very long and the wind would make the water choppy.
On arrival at the spill site, the first small pointer in our favour became apparent in an otherwise cold and wintry day. The Southerly wind had driven the oil in front of it and confined it to a small surface area, under the wharf and against the shoreline.
This made collection of sea samples quite quick and simple. The rubber hoses, which were connected to a high-pressure water supply on the wharf, could reach it. They were able to mix dispersant by way of a venturi suction pipe from a 20-litre container, and spray it onto the sea surface.
Cleanup commenced as soon as sea samples were obtained. They were given a unique number, labelled and put in secure stowage in Schott glass laboratory bottles. The cleanup operation commenced at 0900 and continued until 1400 hrs. A launch was also used to get at some of the awkward places the oil had managed to find its way into.
The second phase of this operation was the evidence gathering and if possible locating a source of where the oil had originated from. From handling the oil when obtaining the sea samples, the smell and feel of the oil, I estimated the oil to be light fuel oil rather than gas or lube oil.
Looking around the harbour for what vessels were in the area of sufficient size to burn light fuel oil in their main engine, I was able to eliminate the tugs, small fishing vessels and a passenger commuter ferry, as they were all diesel powered. This left four foreign trawlers each about 85 metres in length overall within the possible area.
I took the pilot vessel to go around the harbour and examined the seaward side of each vessel for any signs of an oil overflow, and found nothing.
The sampling started on the vessel directly upwind. I got the agent on board and interviewed the Master and Chief Engineer, explained that there had been an oil spill in the harbour and asked if their vessel had been responsible for the spill. In all vessels visited the staff said they had no knowledge of the spill and the oil had not come from their vessel.
We requested the ship’s staff accompany us while samples were obtained from the pumps and bilges on each vessel - this amounted to three pumps per vessel plus samples from the engine room bilge. All four vessels were sampled and these samples together with samples from the sea were sent for oil fingerprint analysis to see if a match could be found.
Two vessels were quickly eliminated using hydrocarbon profiling as not being in any way connected. A further level of tests and analysis was carried out using the EUROCRUDE biomarker method for the remaining two vessels.
An oil fingerprint match was obtained to one of the vessels and the case went to court. The scientific evidence was the main evidence, everything else was circumstantial. The defendant pleaded guilty and was fined $6500 plus cleanup costs of $5980
Lyttelton Harbour Saturday morning 5th June 1999 at 10.30 a.m. The duty Pilot calls in that a black oil spill is floating on the sea in an area bounded by Z berth in Gladstone Pier.
The weather at the time was fine and clear, air temperature 8C. Good visibility, sunny.
Most of the light fuel oil, which was floating on the sea surface, was to be found under the wharf. There were five ships in the possible area, no recent shipping movements and a visual inspection of all vessels from the seaward and wharf side showed no signs of any oil flowing down the side of any vessel.
Three of the five vessels were laid up, rafted alongside each other and with a crew of six people between them. The fourth vessel was a Russian trawler moored astern of the three rafted together, these vessels were about eighty-five metres in length. The fifth vessel was a Chinese squid jigger due to sail from the port the same day.
Samples were obtained from each vessel from the pump discharges and from the engine room bilge. Control samples were obtained from the sea and all these samples were sent for oil fingerprint analysis. A match was obtained from one of the three vessels rafted together.
The case went to court, on 20th September 1999 a conviction was obtained the fine was $8000 plus cleanup costs of $5952.
In this case the vessel which pumped the oil into the sea had a clean sample of seawater from each pump, and it was the sample from the engine room bilge which provided the evidence required. In my opinion after pumping the oil into the sea the pump had been changed from bilge suction to sea suction to clean the pump of any signs of oil. Without the oil fingerprint analysis we would never have got evidence to the standard required for court. End of story.
In selecting a laboratory for oil spill finger print analysis, the points I considered in my selection were as follows:-
• A laboratory accredited to an International standard.
• A laboratory with suitable hardware application ( gas chromatic mass spectrometer )
• Suitable software available at the laboratory.
• Qualified staff with experience in forensic analysis.
• Clear and definite reports, which a Judge sitting in his court can easily understand.
• Laboratory staff with experience in cross examination in court.
The last point is most important – that the analysis staff is capable of stating their own opinions and will not necessarily agree with everything that is suggested to them by cross examining council and thereby creating an element of doubt. If any doubt exists it correctly belongs to the defence and the decision will benefit the defence.
Gentlemen you can make all the right moves and have a good case and loose it under cross-examination through inexperienced laboratory staff. Great care needs to be taken in selecting the correct team.
Thank you Robin.
The scenarios presented by Robin make two points that I would like to clarify:
The first is the careful consideration given to making a thorough assessment at the scene of which ships are possible suspects. EUROCRUDE offers the potential to only sample the prime suspect for bilge discharges. However, the adversarial nature of our legal systems may suggest, depending on your confidence, all suspect vessels which cannot be excluded on other factors be sampled. Taking samples while vessels are available is certainly prudent. Hydrocarbon profiling alone will be able to eliminate many and avoid extensive biomarker testing.
The second point is the potential that a fingerprint match offers for providing clear and definite reports that unambiguously identify the source of the spilled oil. For the first case Robin described, the conclusions derivable using EUROCRUDE for biomarker fingerprinting and giving due consideration of the circumstances, could be and were stated clearly as:
“The spilled oil:
• has not come from the FV Ognevka, the FV Planerist, or the FV Dorada
• has come from the FV Olenino.”
The refinement, further development and wider distribution of EUROCRUDE continues abroad. The crude oils on the database are being supplemented by crude oils sourced world wide to develop a global database. A network of international laboratories with the suitable scientific equipment and expertise is to be developed. There are plans to extend the system to oil products and to incorporate the effects of weathering. Also, software capable of backtracking and trajectory forecasting of slicks is being developed.