Using Oil Spill Dispersants on the Sea (1989)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

5 How Dispersants Are Used: Techniques, Logistics, Monitoring, and Application Strategies DESIGN OF DISPERSANT APPLICATION SYSTEMS Application systems are designed to meet, within practical limits, the following basic criteria that were discussed in Chanter 2: The dispersant must be sprayed on the oil. -sea - The dispersant must mix with the of} and move to the oil- water interface. · The dispersant must attain the proper concentration at the oil-water interface, ideally causing maximum reduction of interfacial tension. · Sufficient energy must be supplied from natural or artificial sources to disperse the slick into droplets. How these four goals are accomplished in practice is the subject of this chapter. Another factor that influences the application of dispersants is the surface area of the spill. Most oils spilled on water attain an average thickness of 0.1 mm or less within a few hours. Therefore, the surface area that must be sprayed to control the spill is approximately proportional to the volume of the of} spin. Spray Systems Systems must be able to apply dispersant uniformly to the slick; ideally the droplets will be small enough to descend to the of! surface 215

216 USING OIL SPILL DISPERSANTS ON THE SEA at a relatively low velocity without penetrating through the of} into the water below, yet large enough that they will not be carried away by the wind (SmedIey, 1981~. Thus, the primary objective in de- signing a spray apparatus is to achieve relatively uniform application without undue wind drift losses. Oil thickness or other factors do not influence the design of the apparatus (Exxon Chemical Company, 1985~. Application rate per unit area is usually varied by changing pump rate, boat or aircraft speed, or number of spray nozzles, or by repeat application (Before, 1985; Chau et al., 1986~. A new approach uses an air jet called a "dispersant spraying gun" which is capable of treating a 0.~- to 0.2-mm-thick of! slick from a distance of 30 m (Barbouteau et al., 1987~. Spray systems can be mounted on a boat, a fixed-wing aircraft, or a helicopter. They usually consist of spray nozzles mounted on a manifold (the spray boom), a reservoir for the dispersant, pumps, meters, valves, and other controls (API Task Force, 1986; Exxon Chemical Company, 1985; Lindblom, 1979~. Small-scale applications on offshore structures, along shorelines, and near piers and bulkheads can rely on hand spraying. Typical equipment ranges from a 5-gal (19-liter) portable tank with spray wand (ITOPF, 1982) to gasoline or electric pumps with reel-mounted hoses and sprayers. Spray Droplet Size Droplet size strongly influences dispersant effectiveness, but it is not easily controlled. Several studies have shown that if droplet size is larger than the of} film thickness, the dispersant tends to penetrate the film without mixing and is diluted by the water. If larger droplets predominate, effectiveness is diminished (Gill and Ross, 1980; Smediey, 1981~. Gill and Ross (1980) argue that the advantage of spraying a larger-sized droplet, which improves the probability of hitting the target, outweighs other considerations in successfully applying dis- persant via aerial spraying. Little quantitative data exist on how droplet size distribution affects the dispersant fraction that actuary reaches and interacts with of] on the water surface. Although exact droplet size can be measured in the laboratory by electromechanical methods, spray patterns from aircraft are normally studied over land, using Kromekote cards to preserve the droplet pattern (used in the APT 1979 southern California field trials; McAuliffe et al., 1981~.

HO W DISPERSA NTS A RE USED 217 Smedley (1981) found reduced effectiveness with larger droplet sizes, although not as much as would be expected based on film pen- etration. Penetration is dependent on ail viscosity: higher viscosity requires greater kinetic energy for the falling droplet to penetrate the slick. Smedley also pointed out that given the same droplet size distribution, dispersants tend to beless effective on thinner slicks, possibly because there is greater penetration. He found, however, that all droplets smaller than 400 I'm in diameter have about the same effectiveness. This is similar to the typical range of slick th~ck- nesses, 20 to 200 I'm (Chapter 2~. There is some evidence that hydrophobic dispersant droplets that penetrate a film are not lost; they resurface on the underside of the film and hence are particularly effective (McAuliffe, private communication). Airborne test data show that 100-pm droplets are not effectively deposited from aircraft because of windage loss, whereas 500-pm droplets are effectively deposited (Figures 5-l and 5-2; Smediey, 1981~. Therefore 350 to 500 I'm appears to be a reasonable size range. This range is produced naturally with more viscous formulations under typical air shear and is practical, since the droplets will hit the target and not drift away on the wind (Lindblom and Cashion, 1983~. In any event, only limited control over droplet size is possible. Fine droplets can easily be achieved in the air, but windage loss defines the lower usable droplet size. The upper size is controlled within limits by aircraft speed, pump capacity, and nozzle size, but large drops are broken up by air turbulence (Hornstein, 1973~. Distance between drops also plays an important role in deter- mining efficiency. To achieve good coverage (i.e., acceptably small drop-to-drop distances) requires a droplet diameter smaller than 700 I'm (Smediey, 1981~. For an of] film 0.02 to 0.2 mm thick, generally acceptable for dispersant application (Exxon Chemical Company, 1985), a dosage between 20 and 70 liters/ha (2 and 7 gal/acre) is recommended. Mechanical system design factors that strongly affect droplet size distribution are nozzle orifice diameter, number of nozzles, fluid pumping rate, and aircraft air speed (Lindblom and Cashion, 1983~. These factors affect the shear acting on the fluid as it passes through the nozzle and enters the air stream behind the aircraft. Mechanical shear at the nozzle outlet, under laminar flow, varies inversely with the cube of the orifice diameter: Shear rate (sec~ ~ ~ = 8V/d = 1.7 x 105Q/d3,

218 USING OIL SPILL DISPERSANTS ON THE SEA | ~ CENTER LINE OF FLIGHT PATH Gin ~ ~ ELEVATION OF AIRCRAFT T 46 m ABOVE GROUND 2.0 LL IL G 3 1.5 IL o - cn ~ 1.0 o a: 0.5 o ·\ CROSSWIND OF 8.33 m/s (18.6 mph MEASURED AT 10 m ABOVE GROUND) a. / ~91% . NO VORTEX EFFECT 500 MICRON DROPLETS / 'act / OF VOLUME ~ DEPOSIT AT '— 105 SECONDS \ ~_\ i 1 `97% \ I OF \ VOLUME\ I DEPOSIT ~ I AT 105 \ SECONDS FULL VORTEX EFFECT 500 MICRON DROPLETS 0 100 200 300 400 500 600 700 DOWNWIND DISTANCE FROM RELEASE POINT (METERS) FIGURE 5-1 Droplet (500 microns) deposition with a crosswind. Source: Smedley, 1981. where V = linear velocity of fluid at the orifice in mm/see, Q = flow rate in liters/min (or 3.785 x gal/min), and d = orifice diameter in mm. Orifice, flow rate, and the number of nozzles used are chosen to keep the mechanical shear below 10,000 sec~i (Exxon Chemical Company, 1985~. Fortunately, the range of droplet sizes naturally produced with more viscous formulations under typical air shear is in the range recommended above, 350 to 500 am (Lindblom and Cashion, 1983~. An even stronger effect results from air shear, which increases with aircraft speed or decreasing fluid exit velocity from each nozzle: Fluid exit velocity (m/sec) = 21 L/nd2, where L = total system flow in liters/min, n = number of nozzles,

HOW DISPERSANTS ARE USED CENTER LINE OF FLIGHT PATH ELEVATION OF AIRCRAFT, 46 m ABOVE GROUND 2.0 LL] A) 1 5 CD LL o - cn J 1.0 o C) 0.5 o 219 CROSSWIND OF 8.33 m/s (18.6 mph MEASURED AT 10 m ABOVE GROUND) NO VORTEX EFFECT 100 MICRON DROPLETS 0.2% DEPOSIT AT 105 SECONDS \ FULL VORTEX EFFECT 100 MICRON DROPLETS ./ A: 79% - ~ OF VOLUME DEPOSIT AT 105 SECONDS 0 100 200 300 400 500 600 700 DOWNWIND DISTANCE FROM RELEASE POINT (METERS) FIGURE 5-2 Droplet (100 microns) deposition with a crosswind. Source: Smedley, 1981. and d = inside diameter of nozzle (mm). Subtracting the fuid exit velocity from the air speed of the aircraft (also in m/see) yields the differential velocity. Differential velocity is the single most important factor affecting droplet-size distribution and therefore depositional efficiency. Extensive testing by API and the Exxon Chemical Com- pany has shown that depositional efficiency drops off markedly if clifferential velocity is more than about 64 m/see (Lindblom, 1987; Smediey, 1981~. Anything that causes a decrease in median droplet size, such as a decrease in dispersant viscosity, increase in differen- tial velocity, or increase in nozzle shear, will decrease depositional efficiency and accuracy. For a given pump rate, a decrease in the number of nozzles in- creases mechanical shear but decreases differential velocity. However

220 USING OIL SPILL DISPERSANTS ON THE SEA since the air shear determines ultimate (deposited) droplet size dis- tribution, it is only necessary that the nozzle orifice be large enough that it will not produce very small droplets. Extremely high pres- sure drop through the nozzle, which increases flow rate and thus shear rate at the nozzle, should be avoided. Spray-boom pressures on aircraft-mounted systems rarely exceed 2.7 atm (Lindblom and Cashion, 1983~. The preceding effects have been assumed to be independent of aircraft type (Meyers, private communication). Vortex fields, however, may vary from aircraft to aircraft and are known to have a substantial effect (Smedley, 1981~. Dispersant Type Four physical properties of a dispersant formulation affect drop- let size during aerial dispersant spraying (I.indblom and Cashion, 1983~: · Viscosity is by far the most important. Volatility is not likely to be important for any concentrate dispersant formulations, but can be significant for hydrocarbon-based formulations. Density has an essentially negligible effect, except when drop- lets penetrate the slick. · Surface tension has a small effect on droplet size, but is of minimal concern for aerial spraying of dispersant. As the viscosity of sprayed fluid decreases, the median diameter of the droplet distribution also decreases and can become very small, producing a mist or fog that drifts far from the target, which can lower depositional efficiency to less than 50 percent (Figures 5-1, 5-2, and 5-3~. On the basis of viscosity at 60°F, dispersants can be grouped into three classes (Lindblom and Cashion, 1983~: Five to twenty-five centistokes, typical of most hydrocarbon- based products, generally is not recommended for aerial application. · Thirty to sixty centistokes including hydrocarbon-based prod- ucts with high surfactant concentration as well as less-active formu- lations in other solvents, can be effectively sprayed from helicopters or small airplanes. · Higher than 60 cSt, including concentrates with high surfac- tant content, can be applied using all types and speed of aircraft.

HO W DISPERSA NTS A RE USED 350 300 _ 250 o .O - a, 200 a) E ~ 150 a, E o > 100 50 o 221 _ " _/ ;" -I'' In o o O m _ z ~ , 0 z C3 O I (1) ,/ CO C) O O C a: 30 z o o ~ O I (2) , . .K O ~ z UJ O I' 'a Weiss and Worsham (II) _~~ _ ,~Mugele ,~ C,, _.. Inaebo Thai Mayer a' 1 l Weiss and Worsham (I) Density - 1.0 g/cm3 Relative Velocity -120 knots Nozzle Diameter - 2 cm Surface Tension - 30 dye/cm I I . 0 0.2 0.4 0.6 0.8 1 1.2 Dynamic Viscosity (Poise) 1.4 1.6 1.8 2 FIGURE 5-3 Comparison of various droplet- - ze models. volume mean diameter versus dynamic viscosity. Weiss and Worsham presented two possible equations to correlate available data, as noted by (I) and (II). Source: Lindblom and Cashion, 1983. In most cases, glyco} ether- or water-based dispersants are used at sea, but hydrocarbon-based dispersants appear (from laboratory tests) to be more effective on mousse and highly viscous of! (Chapter 2~. Undiluted chemicals require a lower flow rate, and nozzles must be properly chosen and sized accordingly (Becker and Lindblom, 1983~. Dosage Control Dosages recommended by manufacturers are determined by ex- perience and limited in range by viscosity and other properties of the formulation as well as the application system itself. Dosages can also be based on laboratory determinations of optimum dispersant-oi} ra- tios; 1:20 is typical (Cormack, 1983c, and private communication). This calculation requires an assumption of slick thickness to con- vert observed area to of} volume. Typically a thickness of 0.1 mm

222 USING OIL SPILL DISPERSANTS ON THE SEA is chosen, equivalent to 940 liters/ha (approximately 100 gal/acre) of oil. A dispersant-oil ratio of 1:20 would then require application of 50 liters/ha (5 gal/acre) of dispersant, in agreement with the 20 to 70 liters/ha (2 to 7 gal/acre) guideline quoted above. Since oil slicks are far from uniform in thickness, a constant dose per unit area wiD produce a correspondingly nonuniform dispersant-oil ratio. Unfortunately, there is no practical way to vary application rate to achieve a constant dispersant-oil ratio. Another reason dosage is regulated by area sprayed is that the amount of oil being treated is usually not accurately known (Horn- stein, 1973~. At the time of spraying, the amount of oil the disper- sant contacts depends on slick thickness, which is only approximately known, and on the accuracy with which the slick is hit by the spray. Deliberately aiming for the thickest part of the spin, for example, near a wrecked tanker or a blowout, has been recommended as an efficient strategy (Mackay, private communication). Injecting disper- sant directly into a blowing well through a previously installed port below the wellhead has been proposed, and some field tests have been encouraging (Audunson et al., 1987; Kolnes, 1986~. Recent tests of depositional accuracy were made using Krome- kote cards to obtain total amount deposited per unit area and the percentage of total fluid pumped that reached the target (Lindblom, 1987~. Data from these tests indicate that 45 to 90 percent depo- sitional efficiency can be expected (most of the data were in the 65 to 85 percent range), with spray altitude on the order of 15 m and differential velocity (air-speed minus exit velocity of fluid from spray nozzle) of 64 m/see (210 ft/sec) or less. Boats Workboats, barges, fireboats, and tugboats have been adapted to spray dispersants, but the use of surface vessels is complicated by slow speed and vessel motion caused by sea conditions. Only relatively small spins (less than 1,000 bbl) can be treated by boats, even with a nearby chemical reloading point. For example, a boat operating at 5 kn, spraying a 12-m (40-ft) swath can only treat about 1.3 km2 (0.5 sq mi) in 12 hr. If slick thickness averages 0.l mm, the boat could treat about 830 bbl per day assuming that the boat has dispersant storage and fuel capacity to operate the entire day without having to interrupt spraying operations (APT Task Force, 1986; Belore, 1985; Chau et al., 1986; McAuliffe, 1986~. Use of multiple craft or vessels

HO W DISPERSA NTS A RE USED TABLE 5-1 Amount of Oil Spray Boats Could Disperse in 10-hr _ray Width System Meters Feet Amount Dispereeda at, Meters" Barrels One boat 5 16 56 350 One boat 10 23 111 8700 One boat 80 100 333 2,100 a Calculation based on boat traveling at 6 kn, an average click thickness of 0.1 mm, and 100 percent efficiency. 223 with greater swath widths increases treatment area proportionately (Table 5-1~. Many improvements have been made since the early Warren Spring Laboratory work (Chapter 1) in the design of dispersant spraying equipment for boats. Examples include spray booms that do not require overhead rigging, low-drift nozzles, electric-start pumps, and flow meters (Frank Ayles & Associates, 1983, 1984~. Most systems now use water-compatible concentrate dispersants diluted with seawater during application. A seawater pump allows the dispersant to be added to the water stream where it is mixed with the water and applied through the nozzles of a spray boom mounted as far forward as possible. This is done to avoid the bow wave, which may push of} out of reach of the spray at typical boat speeds of 2 to 10 kn (Exxon Chemical Company, 1985~. High-pressure water jets have been tested in the laboratory as a means of improving dispersant electiveness by applying mixing energy to the dispersant and water system. In the laboratory, at a dispersant-oi} ratio of 1:100, 80 to 100 percent of the slick was dispersed at the equivalent of low boat speeds (e.g., 2 kn), compared to 10 percent using the Warren Spring breaker-board system (Before, 1987~. If the dispersant is to be added from a positive displacement metering pump, only the minimum flow rate of water necessary to fill and maintain pressure in the spray system is used, thus keeping water dilution to a minimum. Booms should be rigged with multiple nozzles arranged to pro- duce a fan-shaped spray pattern, with sufficient pressure to maintain this pattern until it reaches the water level. The required pressure de- pends on the distance between the nozzles and the water. The spray pattern should be flat and perpendicular to the direction the boat is . ~

224 USING OIL SPILL DISPERSANTS ON THE SEA traveling, and the spray should strike the water in droplets like fine rain. The fan-shaped sprays from adjacent nozzles should overlap just above the oil-water surface and the inboard spray strike the hub just above the water line (APT Task Force, 1986; Exxon Chemical Company, 1985; I.indblom, 1979; Lindblom and Barker, 1978; Spray- ing Systems Co., 1984, 1985~. Cone-shaped spray patterns are less efficient because they concentrate the dispersant unevenly along the edges rather than in the middle (Spraying Systems Co., 1984, 1985~. In a properly designed system, the effect of wind on the spray pattern, spray-boom location relative to the bow wave and distance above water, vessel speed, nozzle design, pump rate, and pressure are considered (Exxon Chemical Company, l9SS). However, even with a properly designed system, there are disadvantages to using water as the carrier for the dispersant. Field tests with boats have indicated much lower effectiveness rates with diluted dispersants when compared to neat applications (McAuliffe et al., 1981~. In heavy seas, application becomes progressively less effective due to pitch and roll, which alters the spray pattern on the water (Cormack, private communication). Direction of spraying operations from the boat deck is limited by poor visibility. In an emergency, water-dilutable (glycol ether-based but not hydrocarbon-based) dispersant concentrate can be educted into a boat's firefighting system, but dosage control and distribution is difficult (Exxon Chemical Company, 1985~. Eduction of dispersant depends on the flow rate of the water stream. The fire system normally provides much more water than is required so its flow must be reduced by valve control or by providing a "bleed" in the stream. For either of these procedures, flow calibration is required. Aircraft Aircraft can be equipped with chemical tanks, pumps capable of spraying 20 to 100 liters/ha (2 to 10 gal/acre), digital readout flow meters, spraybooms (generally airfoil in shape attached to the aircraft), and nozzles. Unless the dispersant viscosity is higher than approximately 60 cSt, aircraft speed should not exceed 100 kn (~85 km/hr or 115 mph). At faster than about 100 kn, spray nozzle tips are not necessary because the wind shear at the nozzles is sufficient to break the stream into small droplets, and the droplet size tends to be less than 350 ,um if the dispersant viscosity is less than 60 cSt. The spray-boom altitude should not be over 9 m (30 ft) for helicopters

HO W DISPERSANTS ARE USED TABLE 5-2 Dispersant Spray Capabilities of Various Aircraft 225 . Aircraft Tank Volume Metered Barrels Amount Dispersed per Flighty Meters Barrels Various helicopters Agriculture spray planes Piper Pawnee, Cesena Agtruck, and Ayres Thrush 0.4-1.1 Turbo Thrush 1.5-2.6 DC-3, Fokker F-27, and Canadair CL-215 Four-engine aircraft (DC-4, DC-6) Hercules C-130, ADDS System 0.4-2.3 2.5-15 3.0-4.5 8-46 50-290 2.5-7.3 8-22 50-140 9.5-16 30-52 190-330 19-28 5.7-11 36-72 21 13 60-90 380-570 115-230 730-1,400 420 2,600 aAssumes dispersant-oil ratio is 1:20 and 100 percent efficiency of dispersion. SOURCE: Adapted from McAuliffe (1987b). and small aircraft, although the pilots of larger aircraft prefer to fly at 15 m (50 ft) or higher (Lindblom and Cashion, 1983~. Table 5-2 shows the dispersant spray capabilities of various aircraft. Small airplanes used in agricultural spraying generally have very small load capacity, often less than 380 liters (100 gal), although some newer planes stave I,500- to 2,300-liter (400- to 600-gal) tanks, as shown in Table 5-2. For dispersant spraying, agricultural spray aircraft need to be fitted with adequate pumps, meters, and aft-facing nozzles. Most pumps and nozzles used for agricultural spraying produce too fine a spray, although the airplanes may be quickly converted if the correct equipment is included in a contingency plan. Small aircraft are best used for rapid response to small spills and test applications while larger equipment is being readied and brought to the spill site. I.arge two- and four-engine airplanes (i.e., DC-3, DC-4, DC-6, Cl'-215, and C-130) are most useful for large spins because of their greater range, capacity, speed, and potential for great areal coverage (Lindblom and Barker, 1978~. The Airborne Dispersant Delivery System (ADDS, also known as ADDSPACK) unit developed for the Lockheed C-130 is the only spray system currently available that does not require permanent

226 USING OIL SPILL DISPERSANTS ON THE SEA installation of wing-mounted booms or other integral parts. It can be used with almost any available C-130 that has a rear cargo door that can be opened to horizontal in flight. Liquid capacity is 20,800 liters (5,500 gal), and the boom width is 12.5 m. Spray rates from 0.4 to 90 liters/ha (0.04 to 9.6 gal/acre) can be achieved; typical coverage is 30 ha/min (74 acres/min) at 130 to 150 kn. This system has been tested extensively over land and water and was used effectively during a recent major spill incident (Oi! Spill Intelligence Report, 1986a). Like all other systems, it must be well maintained and carefully operated (Fingas, 1986; Lindblom, 1987; Lindblom and Cashion, 1983~. Logistics are not trivial for aircraft deployment. Even though ADDS systems are on standby in Florida (agreement between Biegert Aviation and the Clean Caribbean Cooperative), Singapore (Oi} SpiD Intelligence Report, 1987a), and in England (Oi} SpiB Intelligence Report, 1986b), an aircraft must be found to carry the system at the time of a spill. In general, civilian aircraft must be relied upon. Helicopters Helicopters are limited by the volume of dispersant they can carry and their relatively short range over water. Integral (airframe- attached) spray systems require a specially equipped helicopter. Sus- pended bucket units are the most versatile (see Figure 5-4~. Most of them have a capacity of under 1,100 liters (300 gal) (Cormack, 1983c), although the French "Sokaf" bucket has a capacity of 3,000 liters (794 gal) (Bocard et al., 1987~. The most important advantages of helicopters are maneuver- ability and convenience- they do not need a nearby airport. They can also be cost-effective for small spills, especially where offshore platforms provide staging areas and landing sites, although local or- dinances that prohibit helicopters from carrying sling loads across highways may pose potential problems in some areas. These situa- tions necessitate transporting spray equipment by truck to a place where it can be picked up at the shore (Clean Seas, Inc., 1981; Onstad, private communication). Hydrofoils and Hovercraft Hydrofoils and hovercraft have been suggested as potential ves- sels for dispersant application because of their speed and intense

HO W DISPERSA NTS A RE USED 227 FIGURE ~4 Rotortech Spray Bucket and booms; capacity approximately 100 gal used during Beaufort Sea Trials by Environment Canada. Photograph by Colin Jones (committee). agitation of nearby water (Flaherty, private communication). Tests using hydrofoils suggest the following advantages (Vacca-Torelli et al., 1987~: ~ High speed permits the hydrofoil to be on the scene in less than half the time of conventional boats. Furthermore, at higher speeds, hydrofoils do not have a bow wave, which breaks up the oil slick ahead of standard vessels traveling faster than 10 kn. · The hydrofoil has greater stability for its speed than other waterborne craft. ~ Very long spray booms can be used since the hy(lrofoiT is less susceptible to wave motion. Because of the elevation of the craft, however, it may be difficult to position nozzles close enough to the water and the spray would be susceptible to wind-in(luced ([rift. At one time, dispersant spraying gear was carried aboard the Canadian Coast Guard's cargo-carrying hovercraft Voyageur as a possible response to oil well blowouts, but it was never used (Gill,

228 USING OIL SPILL DISPERSANrTS ON THE SEA 1978). The Warren Spring Laboratory has employed sidewall hov- ercraft on an experimental basis, but they are more vulnerable to rough seas than most workboats (Nichols, private communication). Another problem specific to hovercraft is that spray is blown away from the craft by the propulsion system's strong air currents (Jones, private communication). It also appears that turbulence caused by the hovercraft may reduce the dispersant contact time enough to lessen effectiveness. Boats Versus Aircraft Boats, aircraft, and helicopters Al have a role to play in the proper application of dispersants. For applying dispersant to large spins in the open ocean, large aircraft appear to be the only practical means. This is indicated by Table 5-2, which shows the area typically covered by an oil spill as a function of the size of the spin. For some small spills, for example, less than 38,000 liters (10,000 gal), the most practical and cost-effective method of dispersant application is a helicopter with a spray bucket. Boats can spray dispersants on small spills, but their effectiveness may be diminished by limited range and coverage area, weather and sea state (spray-boom length is limited by vessel roll), and the difficulty of determining where to spray. Spotter aircraft generally should be used to direct aircraft and boat spraying operations. The necessity (in most boat-mounted systems) to spray with diluted dispersant generally diminishes the overall effectiveness of the dispersant. In some situations, however, such as a spill caused- by a shipwreck where there is an ongoing salvage operation, boats wig not only be available but may be best suited to spray close to the wreck. This may also be true for of! weD platforms where a minor spin occurs. Calibration Both boat and aircraft application systems require calibration to provide proper dosage control, allow optimum dosage, vary dosage as conditions change, and provide accurate documentation. Dosage from boat-mounted systems is determined by swath width, boat speed, and water or chemical pump rates (Exxon Chemical Com- pany, 1985~. Published pump curves, tables of nozzle output versus pressure, eductor percentage settings, and even calculated swaths

HO W DISPERSA NTS A RE USED 229 provide only general guidance, not true calibration. Manufacturers of spray units generally do not provide the necessary calibration (Lindblom, private communication). Water can be used to calibrate units designed for diluted dispersants and for aerial application sys- tems (Exxon Chemical Company, 1985~. For calibration of eductors and chemical metering pumps, the dispersant to be used should be employed if possible. It is especially important that eductors be properly cleaned and maintained and frequently recalibrated (Onstad and Lindblom, 1987~. Calibration of commercial equipment at U.S. EPA's Oil and Hazardous Materials Simulated Environmental Test Tank (OHMSETT) facility showed that flow rates among nozzles varied significantly, some as a result of plugged or defective internal parts. The flow meter integral to the system was also inaccurate. Calibration procedures and development of operating charts are described by Onstad and Lindblom (1987~. The results of calibration tests should be made available to equip- ment operators and response coordinators in easy-to-use form, such as tables, charts, and curves. This format allows for rapid actions in dosage control that do not depend on numerical calculations or interpolations (Onstad and Lindblom, 1987~. System evaluation should be conducted under realistic operating conditions and data recorded during the tests so that a basis for calibrating and adjusting the system can be formulated. Field trials can help quantify the effects of pumping rates, swath width, and speed (Strum and Nash, 1987~. Recommended eductor settings and pump curves, or calculated swath widths, provide a starting point for calibration, but do not substitute for field tests. Aerial swath width can be estimated as the width of the spray boom at 9-m (30-ft) altitude, or I.2 to 1.5 times the distance between terminal nozzles for an aircraft operating into the wind at 15 m (50 ft) (Exxon Chemical Company, 1985; Lindblom, 1987; Lindblom and Barker, 1978~. Crosswind swaths can only be estimated roughly by guidance aircraft. MONITORING EFFECTIVENESS OF DISPERSANTS Directly monitoring the fraction of oil removed from the surface and dispersed in the water is the best method of determining ef- fectiveness if it can be done reliably. This concept of effectiveness, equivalent to obtaining a mass balance on the amount of of! dispersed in the water column, has given good estimates in a few elaborate field

230 USING OIL SPILL DISPERSANTS ON THE SEA tests; but it is complicated, requires set-up time, and is not practical in real spins. However, the concentrations of dispersed of! in the water are also relevant to assessing potential toxicity. Water sampling programs have been useful for scientific purposes and have been proposed as a regulatory requirement by local agencies; for example, in California (API Task Force, 1986; Boehm and Fiest, 1982; McAuliffe et al., 1980, 1981~. But water sa~npling programs are expensive and of limited value unless carefully planned and executed. Even when the protocols, mechanisms, and equipment for such a program are set up beforehand, as in research field tests, the logistics of getting the appropriate people and equipment to the scene and carrying out the program satisfactorily are difficult (Meyers and Onstad, 1986~. Distribution of oil in the water column during oil spill experi- ments off the coast of Norway in June 1985 was determined by use of instruments that detected light scattering by of} droplets in water and the ultraviolet fluorescence of dissolved of! components (Gen- ders, 1986~. Fluorescence in water is the sum of natural fluorescent materials as well as oil. Backgrounds in the North and Baltic seas varied from 0.5 to 10 ppb and occasionally up to 50 ppb. Attempts have been made to separate backgrounds by using different excita- tion and reception wavelengths. Background is usually a factor of at least 10 less than the signal in the case of significant of] pollution (Hundah] and JojersTev, 1986~. Remote Sensing Other real-time monitoring methods that might lead to prac- tical measures of effectiveness have been proposed. Ideally a spiD- detection and monitoring system should be able to ~ provide continuous, day and night, ad weather, wide area real-time surveillance; J7 ~ detect any of! spin that occurs in the marine environment both on and below the surface; confirm that the detected substance is in fact oil; map the areal extent of the spill; obtain the thickness distribution and quantify the amount of oil spired; identify the source and type of pollutant being discharged; ~ provide precise navigation information for spill source loca- tion and for positioning vessels;

HOW DISPERSANTS ARE USED 231 · provide objective data on dispersant effectiveness and on the presence and behavior of threatened organisms (especially birds and mammals) in the area; and . document Al collected data. Since no one sensor can do all these things, a multisensor system is required. A good sensor wig detect oil, identify it as oil, and determine the of} type (Garrett et al., 1986; Intera Environmental Consultants, 1984; Moniteq, 1985~. Multispectral devices and active laser fluorosensors are now used for more sophisticated detection systems. The laser fluorosensor is an active device that gives basic of} type and operates in any weather, day or night (Hoge and Swift, 1980). Aircraft provide adequate spatial coverage and can be dispatched to a spill with ah the necessary systems operative. They are often the platform of choice for monitoring (Fast, 1985~. Aerial or remote sensing has a distinct advantage over shipboard observations because of the rapid and extensive coverage possible. Oil can be detected from surface vessels by a variety of methods (e.g., visible color, fluorescence, and radar), but the limited area visible from a boat makes these techniques less than ideal. visual and Near-Visible Observation Reflection, absorption, and scattering from water with and with- out spiked oil, and the ability of oil to affect the interference and polarization of reflected light allow visual and near-visible obser- vation of oil. During experiments in the North Sea, for example, observers aboard aircraft noted that ~ day after treatment a chemi- cally dispersed slick had disappeared, but an untreated control slick was still clearly visible (Lichtenthaler and Dating, 1985~. Videotap- ing is one simple, inexpensive, and available technique that can be used from airplanes or boats to document the extent and configura- tion of a spill. It can also record how of} responds to dispersants, the effects of wind, waves, and currents, and the presence and behavior of birds, marine mammals, and other organisms. Using a polarizing filter to enhance the reflectance differences, even thin oil films can sometimes be observed (Burns and Herz, 1976~. The primary drawback of videotape, like visual or photographic observation, is dependence on natural lighting. Sunlight reflecting off

232 USING OIL SPILL DISPERSANTS ON THE SEA water can reveal the shape of very thin slicks, but lack of contrast be- tween of} and water on overcast days greatly reduces the effectiveness of visual observations. Although there are many examples of qualitative observations, obtaining ground (or sea) truth to calibrate a quantitative measure- ment system is a difficult challenge. One day after the 1986 Beaufort Sea tests, various observers could not agree about how much surface area was covered by the thick portions of the slick. The two observers who were in aircraft estimated 20 percent of the slick was thick oil, but the observer on the water estimated 80 percent. In addition, observers on the water saw numerous of] particles or flakes, or small tar balls, detail that could not be discerned in the remote-sensing imagery (Fingas and Jones, private communications). Under relatively calm conditions, experimental slicks tend to spread more extensively when treated with dispersant than do un- treated controls (Lichtenthaler and Daling 1985; Sergy, 1985~. Be- cause of spreading, herding, and the irregular shape and thickness of a large slick, it is difficult to tell visually how much of the of] has been removed from the surface, and, therefore, it is not always possible to determine effectiveness accurately by this method (Cormack et al., 1986/87; Fingas, 1985~. Infirared Sensing Because oil absorbs and retains more heat (from solar infrared ra- diation) than does water, it is possible with the use of infrared-sensing devices to detect of} on the water surface by means of temperature differences. Sensitivity is greatest when the of} is warmer (in sunlight, early afternoon); but when the of} cools to water temperature (in the evening or morning), sensitivity may drop to zero. High-resolution devices, including television-type cameras, are now available that detect the thicker parts of of} slicks. The suc- cessfu} use of infrared image intensifiers for night vision in combat suggests that night flights might track a slick with infrared monitors in darkness or low visibility, which would allow response forces to position themselves for the next day's activities. Such a procedure was tested at night and detected of] at the natural of} seeps of the Santa Barbara Channel (Onstad, private communication). However, the 1986 Beaufort Sea tests, using a different camera, cast doubt on the value of night flights in tracking slicks (Jones, private communi- cation).

HO W DISPERSANTS ARE USED 233 The most commonly used sensors are infrared plus ultraviolet. Remote-sensing airplanes are used in Canada, France, the Nether- lands, Norway, Sweden, the United Kingdom, and the United States (Cormack et al., 1987; Fast, 1987; McCoD et al., 1987; Lavache, private communication). The ideal system is portable from plane to plane, fitting into window openings, screens, or camera hatches (Fingas, private communication). Microwave Sensors Both microwave emissivity and reflectance vary with of} spin presence. Dual-channel-imaging microwave radiometers have mapped of] spills thicker than 100 am (HoDinger and Menelia, 1973~. Active microwave sensors have been used successfully, especially in Europe (Croswel1 et al., 1983; I,oostrom, 1986/87~. Radar Side-Iooking airborne radar (SLAR) is used as a means of de- tecting of! slicks at sea. SLAR can operate through a cloud cover and outline the slick, but cannot distinguish a thick slick from a thin sheen. It is a key element in remote-sensing packages used to detect of] spills in the Netherlands, Norway, Sweden, the United Kingdom, and the United States (Cormack et al., 1987; Fast, 1987; Schriel, 1987; White et al., 1979; McAuliffe, private communication). Detection of slicks by SLAR depends on damping of capillary waves by the floating oil. Reliable detection depends on a variety of factors such as the size of the slick and environmental conditions, but slicks have been detected at distances up to 20 km (Cormack et al., 1987~. In the context of monitoring dispersant effectiveness, it may be that the primary use of SLAR would be the important job of locating a slick rather than measuring the subsequent disappearance of the of! from the water surface when treated by dispersants. Summary of Monitoring Techniques The best method of surveying appears to be from an aircraft using SLAR for initial mapping, followed by infrared line scans to determine relative slick thickness and ultraviolet line scans to produce a picture that can be correlated with visual observation (Fast, 1985, 1987~. With such data, aircraft can direct spraying units to the

234 USING OIL SPILL DISPERSANTS ON THE SEA correct portion of the slicks (Cormack et al., 1986/87; Fingas, private communication). Remote-sensing information on very thin slicks has produced evi- dence for the enhanced spread of dispersant-treated oils (Cormack, et al., 1986/87; Lichtenthaler and Ding, 1985; Cormack, private com- munication). Remote sensing is partially successful in experimental situations, and appears to be at the stage of commercialization of the needed measurement devices. Equipment has been developed for the U.S. Coast Guard (Kim and Hickman, 1975), the Swedish Coast Guard (Fast, 1985), the Canadian Environmental Protection Service (Intera Environmental Consultants, 1984), and the British government (Nichols and Cormack, private communications). This equipment can measure from aircraft some of the parameters nec- essary to determine slick thickness and configuration, but it is very expensive compared to other response equipment and considerable development is required before remote sensing becomes routine. One or more backup systems have been suggested because there are many opportunities for sensitive instruments to malfunction (Fingas, pri- vate communication). Regulatory Requirement The U.S. National Contingency Plan does not require or regulate documentation of cleanup, and subjective visual evaluation is nor- maDy sufficient. The Region {X contingency plan (Pavia and Smith, 1984) requires documentation, as do some local policies, but does not specify detailed methodology. Canadian regulations imply that effectiveness of cleanup should be monitored, but no techniques are specified (Fingas, private communication). STRATEGY OF DISPERSANT APPLICATIONS Development of contingency plans for spill response requires a great deal of thought in advance of any incident. The necessary equipment, personnel, and chemical dispersant must be available to go to work immediately, whenever the need arises. Contingency plan- ning strategy should consider different spin sizes, weather conditions, and ah available control measures (see Chapter 6~. Since timely application is essential for success, it is necessary that ah personnel involved in decisions be educated in the scientific and technical information and be briefed on their role and the con- sequences of failure to act promptly or decisively. It would also be

HO W DISPERSA NTS A RE USED 235 desirable for them to have participated in field trials, spraying real dispersants on real oil. Dispersibility of Oil Dispersibility of the oil is the logical place to start in the planning process. Many important characteristics of an oil from a known source can be obtained from existing data. Laboratory or field testing of dispersant on a sample of the spired oil can give even more information, but it is difficult and time consuming to obtain an adequate oil sample from an actual stick. An exception would be if the source were known ahead of time, as with a production rig or a port where a well-defined type of of! is handled. Spraying a portion of the slick is probably the best method of determining effectiveness (Smith and Pavia, 1983; Stacey, 1983~. SpiD Size and Configuration Spill size and configuration must be known for effective disper- sant application. Generally, the goal of dispersant use is to protect sensitive marine areas or coastline by controlling the most threaten- ing portions of the slick. This can be done by treating portions of the slick, such as its leading edge or selected windrows of oil. The por-- tions of the slick most effectively treated with dispersant are about 0.02- to 0.2-mm thick (Exxon Chemical Company, 1985~. The state of California Contingency Plans require clear evidence of a leading edge (an unusual configuration for a large spill) and threat to marine mammals or shoreline before permission for disper- sant application can be granted. By the time a clear threat can be demonstrated, however, it may be too late to apply dispersant. The Exxon Chemical Company (1985) suggests that boat spray- ing operations generally proceed from the spill's outside edge and work gradually toward the center, so that the slick is not disrupted by the boat's wake. However, most of the oil volume is in a relatively small thick portion, while the outlying portions consist of thinner sheen layers. Circling from the outside in results in much of the dispersant being sprayed on the sheen, rather than on the bulk of the oil. Belore (1985) and Chau and Mackay (1985) have suggested re peated application of dispersant to thicker portions of the stick This has been verified in field tests (LichtenthaTer and Daling, 1985 McAuTiffe et al., 1981~.

236 USING OIL SPILL DISPERSANTS ON THE SEA Lindblom, Meyers, and Onstad (private communications) rec- ommend that small spills generally be treated through their thickest part and large ones treated across a leading edge threatening shore- lines or environmentally sensitive areas. They recommend that sheen be allowed to disperse naturally. Often the thick areas consist of a number of separate patches or windrows, and must be identified and located by aerial reconnaissance. Visual observations must be in- terpreted cautiously (as discussed earlier)—after treatment, of! can spread more thinly over a larger area and the slick may appear to be about the same size as before, even though much of the of} has been dispersed. Large spills cannot be treated in their entirety, but dispersants can be used tactically under favorable conditions to protect sensitive shoreline areas. Even without dispersant application, the concentra- tions of oil in a relatively self-conta~ned body of water can remain high for months after a spill. For example, 2 months after the Argo Merchant spin on Georges Bank in December 1976 the dissolved hydrocarbon levels were as high as 14 to 60 ~g/liter; and 1 to 25 g/liter 5 months afterward. The "normal" unpolluted level for that area appears to be 5 ~g/liter (Farrington and Bochum, 1987~. Would the use of dispersants have increased or decreased the amount of hy- drocarbons retained in the water column? That question cannot be answered without a great deal of additional knowledge about water circulation, exchange with the open ocean, and sedimentation and resuspension rates in that area. Aerial Spraying Strategy When flying directly into the wind, the effective swath width is roughly T.2 to 1.5 times the overall length of the spray boom. Crosswind application, which is common in agricultural spraying, may be useful when very large spills are treated by large aircraft with high-volume pumps (Smedley, 1981~. This technique gives a much greater swath width than upwind spraying but has a potential drawback for smaller spills unacceptable and off-target dispersant drift. For crosswind application, increased pump rate and nozzle adjustment are necessary to allow for a wider swath. Other Strategies It is logical to be prepared, especially for small operational spills, that is, on onshore exploration or production platforms, or loading

HOW DISPERSANTS ARE USED 237 platforms. Here of} and dispersant can be pretested. This is a condition for "maximum effectiveness." During the Chevron Main Pass Block 41 Platform Cspill, a water spray system consisting of a large number of fire monitors mounted on a barge (used for fire control), applied dispersant on the blowing weds and on the of} slick in the immediate vicinity of the platform (McAuli~e et al., 1975~. The same type of system was also used at other platform blowouts. The barge was kept upwind to prevent the spray from covering the barge and to produce the most effective downwind coverage of the platform and surrounding water. During the early stages of an offshore continuous discharge, the oil can be treated initially, at least during daylight hours, by aircraft and boat spraying. Oil that is released during the night and moves as a slick away from the platform can be sprayed the next day. This technique may be sufficient to control smaller of} releases, but it may not control large spins. An offshore platform blowout, either at the surface or subsea, provides a special opportunity for developing techniques to apply chemical dispersants (Audunson et al., 1987~. There is time to as- semble a spray system at the source of the oil release, and the platform or surrounding support vessels allow for 24-hr operation of spraying. Applying dispersant near the of} discharge point has the added advantage of spraying fresh of! and generally thicker oil, thereby more efficiently using the chemical dispersant and better controlling the slick. As discussed in other sections of this report, chemical dispersion becomes more difficult when the of! has weathered and as the slick thins and breaks into patches and particles. One proposal is to add dispersant or emulsion inhibitor to the cargo of a tanker before the of! is released to the water (Gordon and Milgram, 1986; Ross et al., 1985~. This approach does not seem practical it would require enormous amounts of dispersants, could affect the chemistry of of} refining and combustion, and the dispersant would have to be mixed with the of! before it was loaded in tankers. Command and Control An aircraft, given a cloud base of more than 60 m (200 ft) and an indication of where to look, will usually be able to find of! Boating on the water. A boat or ship with dispersant spray equipment may find

238 USING OIL SPILL DISPERSANTS ON THE SEA some oil, but has little hope of locating major patches in the area without spotter aircraft. Radio contact between aerial spotters and spraying boats is therefore essential (API Task Force, 1986; ITOPF, 1982; McAuliffe et al., 1981~. Spotter aircraft are also necessary when aerial spraying is con- ducted. The spraying aircraft flies adjacent tracks and turns the spray on and off in order to hit the slick without wasting chemical on open water. This requires coordination between spotter and spray units (ITOPF, 1982~. The potential use of remote sensing by spotter aircraft as discussed earlier can possibly assist in locating the thicker portions of the slick and provide an indication of effectiveness of the dispersant. Weather Wind and waves not only affect the physicochemical processes of (lispersion, they also affect spraying operations. High winds may blow dispersant spray off target. Heavy seas, however, can be advantageous since breaking waves more rapidly disperse the of} (Bonwmeester and Wallace, 1986a; Raj and Griffith, 1979~. Fog and low clouds cause the most difficulty because they obscure visibility and can stop aircraft operations. Application techniques in adverse weather conditions are included in some contingency plans. (EPA Regions OX and X plans allow consideration of dispersant application when mechanical means are impossible.)