1
Transportation of Heavy Oils and the Risk of Spills
An assessment of the risk of spills involves evaluating the frequency and consequences of accidents. A formal assessment of consequences should be based on a wide range of factors, including loss of life, financial loss, and short-term and long-term environmental impacts. In this chapter, the quantity of oil spilled is considered. Between 1991 and 1996, domestic tanker operations were responsible for nearly 75 percent of the ton-miles of petroleum movements. The major component is the coastal movement of Alaskan North Slope oil to U.S. ports on the West Coast.
Definition of Terms
Group V oils are defined as persistent oils with a specific gravity of greater than 1.0 (Federal Register, 1996). Heavy oil is the term used by the response community to describe dense, viscous oils with the following general characteristics: low volatility (flash point higher than 65°C), very little loss by evaporation, and a viscous to semi-solid consistency (NOAA and API, 1995). Examples of heavy oils include Venezuela crude, San Joaquin Valley crude, Bunker crude, and No. 6 fuel oil. The term heavy oil, in this chapter, also refers to residual oils (No. 5 and No. 6 fuel oil, Bunker C, and slurry oil), asphalt, coal tar, coke, carbon black, and pitch.
The term nonfloating oil is used to describe all oils that do not float on water, including oils that are denser than the receiving waters and either sink immediately or mix into the water column and move with the water as suspended oil; as well as the portion of oil that is initially buoyant but sinks after interacting with
sand. The committee chose not to use the term sinking oil, which implies that the oil sinks directly to the bottom, because it would not include all of the types of oil and spill conditions of concern in this report. Emulsified fuels (anthropogenic fuels manufactured by mixing water with liquid oils or solid hydrocarbon products), for example, often contain a surfactant to stabilize the emulsion and can be dispersed in the water column.
Overview of Quantitative Evaluation
The historical frequency of oil spills in general and heavy-oil spills in particular can be estimated from spill statistics. The committee used the U.S. Coast Guard (USCG) database on oil spills, refined with collaborative data from the Mineral Management Service (MMS), to estimate the probability and mean size of oil spills. The U.S. Army Corps of Engineers (USACE) database on the waterborne transportation of petroleum products and other cargoes in U.S. waters was used to assess the volume of oil transported. By combining the statistics on spills with the data on cargo tonnage, the committee was able to estimate historical spill rates on a barrel-per-ton-mile. Because future spill rates may be influenced by fluctuations in traffic and trading patterns, as well as changes in vessel design and operation, these estimates should be reevaluated to predict future rates. The committee has combined the best available data with its own collective judgment in these estimates. It should be noted that in only 20 percent of spills of heavy oil does a significant portion of the spilled oil sink or become suspended in the water column.
Traffic and Trading Patterns
The USACE (1998a, 1998b) compiles detailed statistics on U.S. waterborne commerce, both foreign (imports and exports) and domestic (trade between U.S. ports). Domestic movements are further subdivided into coastal trade (involving carriage over the ocean) and internal trade (solely on inland waterways).
Figure 1-1 summarizes the data for all movements of crude oils and petroleum products during 1996 (the most recent data available). The USACE data are separated into 19 commodity codes, but for the sake of simplicity, the committee combined some categories (e.g., gasoline and kerosene) into seven categories (crude oil; residual fuel oil; coke, tar, pitch, asphalt; gasoline, kerosene; distillate fuel oil; naphtha, solvents; and lubrication, grease, wax). The substances in the residual fuel oil and coke, tar, pitch, and asphalt categories are heavy oils (i.e., they are either heavier than water or have the potential of sinking or becoming suspended in the water column upon weathering).
Crude oil accounted for 56 percent of the total tonnage of the petroleum commodities shipped in 1996; international trade accounted for 76 percent. The largest component of the domestic trade in crude oil was the coastal movement of
Alaskan North Slope oil to U.S. ports on the West Coast. Internal trade (solely on inland waterways) accounted for less than 5 percent of the total. Nearly all of the international tonnage and 70 percent of the domestic tonnage was shipped in tankers. The very heavy crude oils produced in the United States (e.g., California crudes, such as San Joaquin and Santa Maria) were transported primarily through overland pipelines. Some very heavy crude oils were also imported (e.g., from Venezuela and Mexico), but these are believed to comprise only a small fraction of the total volume of imported crude oil.
Residual fuel oils are represented by a single code in the USACE database, which includes Nos. 5 and 6 fuel oils and slurry oils. Residual fuel oils accounted for 12 percent of the total tonnage of petroleum products shipped in 1996; coke, tar, pitch, and asphalt accounted for another 5 percent of the total. The combined total for heavy oils was, therefore, 17 percent of the total movement of all oil and petroleum products. Approximately 90 percent of the domestic waterborne trade of these heavy oils was transported by barge (whereas more than 90 percent of the international trade was transported by tanker). Overall, therefore, about 44 percent of heavy oils was transported by barge and 56 percent by tanker.
Group V oils are transported along the Gulf Coast from Corpus Christi to
New Orleans, from the Gulf Coast upriver to the St. Louis area, and along the Ohio River to ports further inland. Some Group V oils are also produced in St. Paul, Minnesota, and transported down the upper Mississippi River and up the Ohio River. Heavy residual oils are transported to power generating facilities through the inland waterways and along the East Coast and Gulf Coast and are exported from California to the Far East. Asphalt is moved in tankers and tank barges along the coasts (primarily along the Gulf and East coasts) as both imports and domestic cargoes, and in barges along the inland waterways. Some very heavy crude oils (e.g., Venezuela Boscan crude) are imported to East and Gulf coast refineries. Carbon black feedstock moved among refineries on the Gulf Coast and was exported from California. Bunkering fuels for ships (typically No. 6 fuel oil) are moved intra-harbor on barges. Most large commercial ships (including container ships, dry bulk carriers, tankers, cruise ships, as well as some tugboats) use these heavy oils as fuel, although these oils are not included in the statistics on the waterborne commerce of petroleum.
In Figure 1-2, the movement of crude oils and petroleum products for calendar year 1996 are shown in metric ton-miles. The domestic ton-miles are calculated by multiplying the metric tons of cargo being transported by the number of miles actually moved on the water. The average length of a domestic voyage was about 900 miles. For imports and exports, a constant of 100 miles per voyage was
assumed to account for the exposure of the vessel when transiting coastal waters and navigating U.S. waterways and channels. Movements of residual fuels comprised only 8 percent of the total.
The U.S. waterborne commerce of petroleum gradually decreased from 1991 to 1996 (Figure 1-3), primarily as a result of cutbacks in the coastal tanker trade of crude oil. During this period, the movement of residual fuel oils declined by 45 percent, due partly to improvements in the refining process, which produces less residual oil per barrel of crude oil refined. The movement of coke, tar, pitch, and asphalt, however, increased by 47 percent. Preliminary USACE figures for 1997 indicate that the domestic trade for coke, tar, pitch, and asphalt was up nearly 70 percent compared to 1996.
Movements of petroleum by tanker and tank barge are summarized for calendar years 1991 through 1996 in Figure 1-4 and Table 1-1. Figure 1-4 shows that domestic barge traffic remained relatively constant during the period. Tanker import and export traffic increased by about 5 percent per year, reflecting increases in imports of crude oil; the tanker domestic traffic declined by about 7 percent per year.
The U.S. Department of Energy (DOE, 1998) estimates that the percentage of petroleum consumption met by imports will increase from 49 percent in 1997 to 65 percent in 2020. This increase is partially a reflection of anticipated reductions in domestic production as oil reserves are depleted and a projected 1.1 percent per year increase in domestic energy consumption. The higher demand will probably be met through increased imports of long-haul crude oil (NRC, 1998). Future trends in the movements of heavy residual oils and asphalt are more difficult to quantify. The committee heard several presentations on the interest of some utility companies in using emulsified fuels (e.g., OrimulsionTM) to generate power. Emulsified fuels do not float on water and are included in the
definition of nonfloating oils. Environmental groups responding to a proposal to burn a Venezuelan emulsified fuel for power generation in Manatee County, Florida, expressed concerns about cleaning up emulsified fuel spills once the oil had dispersed into the water column (Rains, 1998). Another concern was air quality because these fuels tend to be high in sulfur and other contaminants. At this point, it is difficult to project the consumption of emulsified fuels in the United States.
History of Spills
The historical data on oil spills in U.S. navigable waters was derived from both the USCG and MMS databases. The USCG database includes reported oil spills of all sizes in U.S. navigable waters. Although these data are
TABLE 1-1 Movements of Petroleum by Tanker and Tank Barge during Calendar Years 1991 through 1996
|
U.S. Waterborne Traffic in Metric Ton-Miles (× 1 billion) |
|||
|
1991–1996 Mean |
1996 Totals |
||
|
Barge |
Tanker |
Barge |
Tanker |
Crude Oil |
4.9 |
266.8 |
4.9 |
215.4 |
Residual Fuel Oil |
12.1 |
23.5 |
12.7 |
16.8 |
Coke, Tar, Pitch, Asphalt |
7.4 |
2.5 |
8.7 |
3.6 |
Other Petroleum Products |
37.8 |
60.5 |
27.8 |
76.1 |
Totals |
62.2 |
361.3 |
64.1 |
311.9 |
comprehensive, they have not been uniformly maintained over the years. The MMS database has been consistently maintained but only covers spills of more than 1,000 barrels from tankers and tank barges. By comparing the USCG data with the MMS data, the committee has modified the USCG data, as necessary.
Since 1991, there has been a dramatic reduction in the volume of oil spilled from vessels in U.S. waters (Figure 1-5). Losses from tankers since 1990 are less than one-tenth the volume of pre-1990 losses, and losses from barges are less than one-third the volume of pre-1990 losses. From 1973 to 1990, there were 18 spills of more than 25,000 barrels each. Since 1991, there has not been a single spill of this magnitude. This statistic may be fortuitous, however, and a very large spill is likely to occur in the future. Large future spills are likely to involve crude oil rather than heavy oil, however, because most heavy oils and asphalt are carried on barges and smaller tankers.
In light of the huge decrease in the number of oil spills since 1990, the committee based its projections on the 1991 to 1996 data. Because of inconsistencies in the data for small spills, the committee limited its analysis to spills of more than 20 barrels, which account for more than 98 percent of the spills in this period.
Spills in the USCG database are divided into two general categories based on their origin: vessels and facilities. Facilities include pipelines, ground
transportation onshore facilities (e.g., shoreside structures, such as terminals, refineries, and storage tanks), and offshore facilities (e.g., drilling rigs and production platforms). Vessels are subdivided into tankers, tank barges, and other vessels (ships not engaged in the transport of petroleum). The number and volumes of spills are summarized in Table 1-2. Although tankers were the primary source of marine oil spills prior to 1990, facilities have been responsible for a majority of the incidents and most of the total spill volume since then. Pipelines have been the source of more than 50 percent of the spill volume from facilities, and tank barges for more than 75 percent of the spill volume from vessels.
The USCG database provides a description of the substance spilled in each event. Table 1-3 summarizes data for all spills of more than 20 barrels of nonfloating oils (i.e., products with the potential to sink or become suspended in the water column when weathered or mixed with sediment). These products include asphalt, coal tar, carbon black, bunker C, and Nos. 5 and 6 fuel oils. Spills of nonfloating oils constitute about 23 percent of the total volume of oil spilled. From 1991 to 1996, the average number of spills of nonfloating oils was 16 per year, with an average volume of 785 barrels per spill. Tank barges were responsible for 28 percent of incidents and 80 percent of the total spill volume.
Releases of 20 barrels or more from facilities were generally spills of floating oils (either crude oil or gasoline). The largest reported spills of heavy oils from a facility was a spill of 929 barrels of No. 6 fuel oil in Pearl Harbor, Hawaii. In contrast, there were six spills from tank barges of more than 4,000 barrels each, all of them of heavy oils (either No. 6 fuel oil or slurry oil). The average volume of heavy-oil spills from barges was 2,254 barrels, and the largest spill during this period was about 18,000 barrels. Spills were widely distributed geographically (Figure 1-6), with the highest frequency from vessels in the Gulf of Mexico. Some of the oils categorized as heavy oils in the USCG and MMS databases are less dense than seawater and will remain afloat under certain environmental conditions. To determine the frequency of nonfloating-oil spills, the committee examined heavy-oil spills of more than 20 barrels (a total of 93 spills)
TABLE 1-2 Oil Spills of 20 Barrels or More in U.S. Waters by Origin (1991 to 1996)
|
No. of Incidents |
Total Spill Volume (barrels) |
Average Spill Volume (barrels) |
||
Tankers |
47 |
(8%) |
26,508 |
(8%) |
564 |
Tank barges |
100 |
(17%) |
100,785 |
(32%) |
1,008 |
Other vessels |
44 |
(7%) |
11,474 |
(4%) |
261 |
Facilities |
415 |
(68%) |
173,945 |
(56%) |
419 |
1991 to 1996 totals |
606 |
|
312,713 |
|
|
Average per year |
101 |
|
52,119 |
|
|
TABLE 1-3 Heavy-Oil Spills of 20 Barrels or More in U.S. Waters by Origin (1991 to 1996)
|
No. of Incidents |
Total Spill Volume (barrels) |
Average Spill Volume (barrels) |
||
Tankers |
17 |
(18%) |
6,442 |
(9%) |
379 |
Tank barges |
26 |
(28%) |
58,591 |
(80%) |
2,254 |
Other vessels |
22 |
(24%) |
3,877 |
(5%) |
176 |
Facilities |
28 |
(30%) |
4,083 |
(6%) |
146 |
1991 to 1996 totals |
93 |
|
729,913 |
|
|
Average per year |
16 |
|
12,166 |
|
|
to identify the spills in which a significant fraction of the oil did not float. These spills accounted for about 20 percent of the heavy-oil spills and about 50 percent of the volume of heavy oil spilled during this period. The relatively high volume of nonfloating-oil spills, as compared to the relatively low number of nonfloating-oil spills (20 percent), is attributable to a few large heavy-oil spills during the period. One spill in particular, the Morris J. Berman spill of nearly 18,000 barrels of heavy oil in 1994, strongly influenced the statistics.
The committee could not explain why the average volume of nonfloating-oil spills should differ from the average volume of heavy-oil spills and considers the high volume of nonfloating-oil spills to be an anomaly caused by the limited statistics. Assuming that nonfloating-oil spills comprise 20 percent of the heavy-
oil spills by number, the committee estimates that the average per year (20 percent of 16) will be three or four nonfloating-oil spills. Assuming that the average volume of nonfloating-oil spills is the same as for heavy-oil spills (i.e., 785 barrels per spill), the projected volume will be about 2,500 barrels per year.
Projections of Spills
To assess the risk of heavy-oil spills from vessels, the committee used ton-miles as a measure of exposure and the quantity of oil spilled as a measure of the consequences of accidents. Based on this approach, the spill rate is defined as the ratio of the historic volume of oil spilled to the historic movements in ton-miles and is expressed as barrels spilled per billion ton-miles. Tankers and tank barges were responsible for 89 percent of the heavy-oil spills from 1991 to 1996. The spill rates for all petroleum cargoes and for heavy-oil cargoes are presented in Tables 1-4 and 1-5, respectively. Barges had higher spill rates for all petroleum cargo than tankers during this period. The spill rates for heavy oil carried by barges were higher by a factor of two than the spill rates for all petroleum cargoes. The spill rates in Table 1-5 are for heavy oils, some of which remain afloat under certain environmental conditions. Only about 20 percent of the heavy oil spilled is expected to exhibit nonfloating behavior.
The volume of future spills will be affected by changes in the design and operation of tankers and barges. Decreases in both the number and volume of oil spills are expected as the fleet completes the transition to double-hull construction (NRC, 1998).
The spill statistics suggest that the barge industry has lagged behind the tanker industry in improving operations since the Oil Pollution Act of 1990 (OPA 90) was enacted. Major barge accidents from 1991 to 1996 had a variety of causes, including structural failure, capsizing, allisions, collisions, and groundings. The barge industry has instituted a number of voluntary programs to improve its environmental performance and safety record. These include the American Waterways Operators Responsible Carrier Program and partnerships with the USCG.
TABLE 1-4 Spill Rates for All Petroleum Cargoes in U.S. Waters (1991 to 1996)
|
Movement of Petroleum (billions of metric ton-miles per year) |
Oil Spill Volume (barrels per year) |
Spill Rate (barrels spilled per billion metric ton-miles) |
Tanker |
361.3 |
4,418 |
12 |
Barge |
62.2 |
16,798 |
270 |
TABLE 1-5 Spill Rates for Heavy Oil in U.S. Waters (1991 to 1996)
|
Movement of Heavy Oil (billions of metric ton-miles per year) |
Oil Spill Volume (barrels per year) |
Spill Rate (barrels spilled per billion metric ton-miles) |
Tanker |
26.1 |
1,074 |
41 |
Barge |
19.6 |
9,765 |
499 |
From 1991 to 1996, the percentage of tonnage carried in double-hull vessels was approximately 13 percent for tankers and 60 percent for barges. Theoretical comparisons with single-hull vessels (NRC, 1998) indicate that double-hulled tankers and tank barges will be involved in four to six times fewer spills. If all vessels trading from 1991 to 1996 had been double-hull, the number and volume of heavy oil spills could have been reduced by about 30 percent. In accordance with the provisions of OPA 90, the transition to double-hull vessels will be completed by January 1, 2015.
Total cargo movements in U.S. waters have increased at an average rate of 2 percent per year for the past 10 years. Further growth will tend to increase the number of spills from bunkers on freighters and other commercial vessels, and a move is under way to protectively locate bunker tanks on larger tankers and a few large container ships, which should lead to a reduction in the spillage of fuel oil.