E M I S S I O N S F R O M IN S I T U B U R N I N G OF C R U D E OIL

IN T H E A R C T I C

T H O M A S D A Y , D O N A L D M A C K A Y * , S T U A R T N A D E A U and R O B E R T T H U R I E R

Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario, Canada M5S 1A4

(Received 26 June 1978; revised 2 October 1978)

Abstract. The results of an exploratory study of the effects of in situ crude oil spill burning on air quality in the Beaufort Sea region of the Arctic are presented. A scenario is postulated defining the amounts of oil released, the size and number of burnable oil pools, and the duration of the burning period. Estimates are made of the likely emissions of soot, CO, SO 2 and metals based on literature and some experimental work. Assumptions are made about plume rise and dispersion which permit downwind concentrations of emissions to be calculated and compared with air quality objectives. Although the calculated concentrations may contain significant error because of the many assumptions, it is believed that the data demonstrate that concentrations of SO 2 and CO will be acceptably low, concentrations of soot and metals will often be undesirably high within 10 km of the fires, but will be acceptably low at greater distances. It is concluded that burning may be a method of substantially reducing the adverse environmental impact of oil spills in the Arctic.

1. Introduction

Concurrent with increased petroleum exploration off-shore in the Canadian Arctic have been studies of the environmental impacts of oil and gas blow-outs and discharges, and reviews of the feasibility of cleaning up oil spills in this hostile environment (Logan et al., 1975, Pimlott et al., 1976, Ross et al., 1977, and Norcor, 1975). The present area of exploratory interest is the continental shelf above which water depths reach 25 m and first year ice cover exists for approximately 280 days of the year, reaching a thickness of 2 m. Drilling takes place during the short open water season from mid-July to late September. Further North, the permanent and thicker polar pack circulates forming the Beaufort Gyre and a transition or shear zone of irregular actively deforming ice exists. From an environmental viewpoint, the obvious primary effort is to prevent oil release by carefully controlled drilling operations, and indeed, extreme precautions are taken. There remains a low probability of a blowout which could have a profound environmental effect. Since it may be difficult to deploy conventional oil spill counter measures equipment in the Beaufort Sea, one of the most attractive options is in situ burning of oil, either as it reaches the water surface directly, or when it migrates to the surface through melting ice in the spring. Burning would result in the emission of substantial quantities of combustion products into the atmosphere with adverse effects on down-wind air

* To whom correspondence should be addressed.

Water, Air, and Soil Pollution 11 (1979) 139-152. 0049-6979/79/0112--0139 $02.90 Copyright © 1979 by D. Reidel Publishing Co., Dordrecht, Holland, and Boston, U.S.A.

140 T. DAY E T A L .

quality. In elucidating this issue the approach taken is to establish a scenario for burning, estimate emissions, calculate down-wind concentrations of combustion products using conventional plume dispersion equations, and examine the issue of

superposition of plumes in time and space from a number of burning pools. In support of this calculation, some laboratory and field experiments were undertaken.

2. Oil Spill Scenario

In assessing the environmental impact of oil burning, it is necessary to make some assumptions about the volume, nature and location of the oil to be burned. The approach adopted is to assume the 'standard' Beaufort Sea oil and gas blowout scenario. Ice movement in this area was studied by Norcor (1977) and the probable behavior of a winter blowout postulated. This postulate forms the basis of the present scenario.

It is assumed that a blowout starts late in the drilling season (August and September), too late to permit the drilling of a relief well. Oil discharges from the well at a rate of 2500 barrels day-1 for the first 30 days, falling to 1000 barrels day - during the remaining months of the winter period. The total volume spilled during

this 8 mo period will be 287,000 barrels (45,700 m3). The natural gas associated with the oil will be vented through the ice and will play no role in the burning process.

In the event of an oil well blowout, the oil will 'paint ' the undersurface of the

moving ice. The thickness of the oil layer will depend on the lateral spreading of the oil plume, the oil release rate and the ice velocity. During the months October to May inclusive, there is a probability that a strip of ice about 240 km long and about 100 m wide will be painted with oil. The area affected will be 24 km 2, thus an

average oil film thickness will be 1.9 mm. It is believed (Norcor, 1977) that only 35°70

of the oil will achieve a thickness greater than 5 mm, which is judged to be the thickness necessary to permit ignition and combustion.

For this study, it is assumed that the oil reaches burnable thickness on the first year ice about June 1st, and continues to flow for 20 days. Little is known about oil migration under multiyear ice, but recent work by Milne (1975) suggests that the oil

may reach the surface in September. It is now possible to assemble an approximate mass balance for the oil in this

scenario, making the following assumptions: 1. The oiled area is a tortuous strip and lies entirely in a rectangle 80 km by 30

km. The net ice movement is considerably less than the total distance travelled, especially during the static period of February and March (Norcor, 1977).

2. Seventy per cent of the oil lies under first year ice and will surface in June. 30% of the oil is under multiyear ice and will surface in September (Norcor, 1977).

3. Approximately 6°70 of the oil under first year ice and 12070 of the oil under multiyear ice does not reach the surface. It sinks, dissolves, disperses or becomes trapped in ice.

EMISSION FROM ARCTIC CRUDE OIL BURNING 141

4. Thirty-five per cent of the oil which surfaces gathers in pools of sufficient

thickness that burning is feasible (Norcor 1977). 5. Approximateoy 9500 m 3 of oil is subjected to burning in June and 3800 m 3 in

September, yielding 2100 m 3 of unburned residue; and 11200 m 3 of oil is converted

into CO2, CO, unburned hydrocarbon and soot. 6. The oil which is burned has gathered into a large number of pools in which

10% of the oil evaporates prior to burning. Each pool then consists of 1.0 m 3 of oil

of an average thickness 7 mm covering an area of 143 m 2 or diameter 13.5 m. 7. The average pool burns for 10 min, leaving a residue of 0.16 m 3 (Coupal,

1976). This event is used as a unit burn in dispersion calculations. 8. During the burning operation, an average of 9.5 pools are ignited at

approximately the same time. Some pool-to-pool fire spreading occurs. This

'cluster' of pools thus contains 9.5 m 3 of oil and is assumed to burn for 1 h with a maximum of 3 pools burning at any one time.

9. There are thus 9500 pools in 1000 clusters in June and 3800 pools in 400

clusters in September. 10. Since the total length of the strip is 240 km, each of the 1400 clusters occupies

an area 171 m long and 100 m wide.

These assumptions are in many cases arbitrary but are regarded as being reasonable and they provide a basis for calculation of the dispersion behavior.

3. Oil Properties

The important oil characteristics are the S and metal contents and the percentage of

the oil which is emitted as soot. This latter quantity depends on the nature of the combustion process. Since no experimental data could be located for soot

formation, some laboratory work was undertaken to obtain an estimate. The chemical nature of Arctic crude oils is, of course, not known. However, data are available on the S and metal analysis of Norman Wells (NWT) and Swan Hills

(Alberta) crude oils (Norcor, 1977). For the present purposes, it is assumed (pessimistically) that S is present at a concentration of 2%, and metals (principally

Va) are present at a concentration of 10 ppm. No information could be located on the fraction of the metal emitted during combustion, thus some experimental work

on this topic was undertaken. It is assumed that all the S is converted to SO 2. Measurements were made of CO to CO 2 ratio in the experiments.

4. Weather

Comprehensive accounts of climate and weather in the Beaufort Sea region have been provided by Burns (1973), Walker (1975) and Berry et al. (1975) with statistics relevant to the air pollution climatology published by Munn et al. (1970); Shaw et

t42 T, DAY E T A L .

al. (1972); and Portelli (1977a). The Beaufort Sea region suffers from lower mixing heights and ventilation coefficients than do most regions of Canada. Generally, the mean summer and fall maximum (afternoon) mixing heights in the region are from

200 to 500 m. For purposes of this study, a figure of 310 m is assumed, the mean wind speed in June and September is about 20 km h-1 (5.6 m s -1) and is variable in direction, but tending to be from the ESE in June and the WNW in September.

5. Experimental

A small laboratory apparatus, Figure 1, was assembled to study oil burning. It consisted of a 15.2 cm diameter by 46 cm long glass column with a hemispherical bot tom containing a sand bed supporting a circular brass dish 9 cm in diameter and

AIR

OIL

iGNITION SOdRCE ~

ORIFICE METER ~ ,

VALVE Ll~J --

TO FILTER

WATER - - J L ~ VOLTMETER

Fig. 1. Laboratory burning apparatus.

1 cm thick containing oil to a maximum depth of 0.75 cm. The oil was ignited by a nichrome wire and air was supplied by a pump at a known flow rate. Thermocouples were used to measure temperatures in the dish, flame and plume. A suction filter was connected to the top of the apparatus to collect all the soot. The temperature of the dish could be varied by circulating water through a coil molded in the dish base. In a typical test, a known weight of oil was placed in the dish, the temperatures and air flow rate set, the oil ignited, and the temperatures measured. The weight of the residue and soot was measured. Electron photomicrographs were taken of the soot particles to obtain data on soot particle size. Vanadium metal analysis of the oil, residue and soot was by instrumental neutron activation analysis. Samples of the gas

were also analyzed to obtain the CO2 to CO ratio using a gas chromatograph equipped with a thermal conductivity detector.

The laboratory experiments showed that the fraction of oil burned depended strongly on dish temperature, air flow rate and quantity of oil.The colder the oil, the

EMISSION FROM ARCTIC CRUDE OIL BURNING 143

Fig. 2. P h o t o m i c r o g r a p h of soot.

less efficient the burn. For example, at 198°C, 82°7o of the oil is burned, whereas at 31°C, only 34°7o burned. The air flow rate experiments indicated that there is an optimum air flow rate to achieve a maximum combustion efficiency. Increasing the quantity of oil burned, resulted in a greater burning efficiency, probably a result of

the insulating effect of the deeper oil pool. The conditions of the laboratory

experiments are far removed from a full scale burn, but the same trends probably apply.

The amounts of soot formed were generally from 2 to 5°7o of the original oil mass. The ratio of CO 2 to CO varied from 8 to 43 with a mean ratio of 20:1.

An example of an electron photomicrograph of soot is shown in Figure 2. The

soot consists of a large number of small particles of diameter 5 nm to 100 nm. These particles were agglomerated into larger particles often about 1 Hm, but occasionally about 10 Hm in diameter. Examination with a light microscope also showed particles of up to I00 Hm in diameter. Selection of a mean particle size, size distribution, and settling velocity is difficult since the soot particles may be continuously agglomerating in the plume.

Activation analysis for Va in oil showed that the concentrations varied from 2 to 4 ppm. After combustion, analysis of the oil residue and soot showed that approximately 66°70 of the original Va was emitted with 1 to 3°70 of the metal present in soot. These results were confirmed by adding a Co tracer to the oil and repeating the experiments.

144 T. DAY E T A L .

A field study was also undertaken in which 0.204 m 3 of oil were spilled on an ice

surface in a pool 1 cm thick and about 5 m in diameter. The oil was ignited using a rag soaked in kerosene. The burn lasted for 6 min. The air temperature was - 4 ° C and the wind speed was 8 km h - 1.Attempts to measure the temperature of the flame and to sample the plume failed because of the high flame height (5 m) and the

intense heat from the fire which prevented close access to the burn. The oil burned very vigorously and left a residue of only 10070 (weight). The plume rising velocity was about 5 m s - l and it rose about 150 m.

6. Dispersion Calculations

It is assumed that each pool consists of 1.0 m 3 of oil, density 0.85 g cm -3, containing

2°7o S (all of which forms SO2), 6°70 of the original mass is emitted as soot and 6 ppm metal is also emitted with the soot. This corresponds to emission rates of 55 g s -1 of SO2, 85 g s -1 of soot and 8.5 × 10 -3 g s -1 of metal. A C mass balance shows that the CO emissions should be about 2.25 times those of the SO 2. The wind speeds investi- gated were 2, 4, 6, and 8 m s -1 with a mixing layer height of 310 m. Particle sizes of 100, 30, 10, and 1/ tm were used in soot calculations. The settling velocities for the above particles are 50, 5, 0.6, and 0.007 cm s -1 respectively (Ledbetter, 1972).

The plume rise was calculated using the equations of Holland, Concawe, Lucas et al., Singer et aL, and Briggs as reviewed by Portelli (1977b) and Seinfeld (1975).

The results in Table I show considerable disagreement. A mean plume rise was selected for each velocity, but this could be in error by ± 50°7o.

The dispersion equations used are the conventional gaussian distribution

equations as reviewed by Turner (1974). Calculations were made for stability classes B, D, E, and F at these wind speeds and at the center line concentrations calculated

at distances of 1 to 40 km downwind. The equations assume, for SO z and CO, total reflection at the ground and at the

top of the mixing layer. For soot, an equation outlined by Somers (1971) was used which takes into account reflection at the top of the mixing layer and deposition at the ground. The settling velocity of the soot is included by using a tilting plume

model.

TABLE I

Calculated plume rises

Wind speed (m s -1) 2 4 6 8

Holland 315 162 105 79 CONCAWE 407 242 179 144 Lucas e t al. 391 195 130 98 Singer e t al. 127 63 44 32 Briggs 179 142 124 113

Selected Value 310 105 114 88

EMISSION FROM A RCT IC CRU D E OIL BURNING 145

7. Results of Dispersion Calculations

Figures 3 and 4 are typical concentration profiles for SO 2. From a series of such figures, maximum concentrations were determined and are presented in Table II. For a cluster of pools, with a maximum of 3 burning pools, the concentrations are arbitrarily multiplied by a factor of 3. For SO 2, the maximum concentration close to the fire is 300 ]2g m -3 with figures of 90 and 30 ~tg m -3 at 10 and 40 km distance downwind respectively. These concentrations may occur under B stability conditions

Fig. 3.

100

k

o ~

1 i I I 10 20 30 40

DISTANCE (kin}

Center line ground level downwind SO 2 concentration as a function of wind speed (stability D).

100

4 - E

=L

Z O

tw

Z UJ

Z O

WIND SPEED z, m/s STABILITY

B e Dm E V F e

10

ill / - 10 20 3 0

DISTANCE (km} 1,0

Fig. 4. Center line ground level downwind SO 2 concentrations as a function o f stability class (wind speed 4 m s- l ) .

146 T. DAY E T A L .

TABLE II

Sulfur dioxide concentra t ions ~ g m -3) at s tated downwind distance (km). First entry is the m a x i m u m

Wind Speed Stabil i ty Class

m s - I km h -1 B D E F

2 7.2 102 at 2 26 at 20 8 at 35 No m a x i m u m 28 at 10 17 at 10 0.3 at 10 ~ 1 at 10

3.2 a t 4 0 1 9 a t 4 0 8 a t 4 0 0.04 at 40

4 14.4 89 at 1 32 at 7 18 at 15 No m a x i m u m 14 at 10 28 at 10 16 at 10 1 at 10 2 a t 4 0 1 0 a t 4 0 1 2 a t 4 0 5 a t 4 0

6 21.6 - 5 4 a t 4 3 3 a t 7 - 28 at 10 32 at 10

7 a t 4 0 11 a t 4 0

8 57.6 - 7 2 a t 2 5 2 a t 5 - 24 at 10 36 at 10

5 a t 4 0 9 a t 4 0

at a low wind speed of 2 m s -1. Carbon monoxide concentrations in the plume can be

estimated by multiplying SO2 concentrations by a factor of 2.25.

Figure 5 is a typical concentration profile for soot deposition. From a series of

such figures, maximum soot concentrations can be determined and are presented in

Table I l l for 100 and 10/~m size particles respectively. Estimates of soot concentra-

tions f rom the values in the tables indicate that, for 3 pools burning simultaneously,

the maximum values may be 2100, 600, and 45 ~tg m -3 at distances of 1, 10, and 45 km

downwind of the burn site. The metal concentrations and deposition rate can be

estimated by simply dividing the particulate quantities by a factor of 104.

It is necessary to estimate the extent to which plumes may superimpose as a result

of a plume blowing over a burning cluster, thus subjecting an area to multiple concentrations. At a wind speed of 4 m s -1, a plume from a cluster which burns for

1 h will be typically 15 km long and 2 km wide, will cover an area of 30 km 2, and will sweep over an area of 30 km 2 h -1. It will take approximately 3 h to drift to the edge

of the spill area or to be diluted to very low concentrations. Assuming a maximum

burning rate of 10 clusters per hour, there will thus be 30 individual plumes in the

spill area, some of which will inevitably superimpose. A simple statistical analysis leads to the conclusion that approximately one-third

of all fires will be swept over by a plume f rom another fire. The justification is that

if 3 of the 10 fires encounter a plume, there will be formed during each hour, 7 single

plumes and 3 double plumes with a total area of 300 km 2 for the 13 plumes. Since there are 30 plumes in the area, their total area will be 700 km 2 or 30°7o of the total spill area of 2400 km 2. There is thus a 30°70 probabili ty that a fire will occur in the area of a plume. Alternatively, each plume which sweeps an area of 30 km 2 h -1 will

E M I S S I O N F R O M ARCTIC C R U D E OIL B U R N I N G 147

Fig. 5.

1000&

E 100

z 0

<

b- z w o z o 10 o

WIND SPEED L m/s STABILITY D PARTICLE SIZE

• 100 ~m • 30 • 10 AND I

5 10 15 20 25 30 35 DISTANCE (kin)

Center line ground level downwind soot concentration (wind speed, 4 m s -1 and stability class D).

have a probability of (10 × 30/2400) or 12°70 of encountering one of the 10 fires in the total area of 2400 km 2 each hour and thus a 36°7o probability in 3 h. Superposition of a 'double' plume on a fire will occur to 1 fire in 10.

This analysis is necessarily approximate in that the superposition probability is very sensitive to the alignment of the clusters with respect to wind direction. Super- position of the two near-fire concentrations of, for example, 300 pg m -3 of SO 2 will be infrequent. The worst common condition will be superposition of a 10 km

concentration with a near-fire concentration, which will occur in 5 to 10% of the plumes, and more frequently, i.e. in 20 to 25O/o of the plumes, one would expect doubling of the 10 km or 30 km concentrations. These superimposed concentrations are shown in Table IV. These conditions, with approximately one-third superposition, will occur only during the most intense burning period.

8. Discussion

It is interesting to compare, in Table V, these concentrations with the Canadian National Air Quality Objectives and typical levels in urban areas of Canada (Department of Fisheries and Environment 1976, 1975).

For SO2, the concentrations are all below the 1 h desirable concentrations, thus

148 T. DAY E T A L .

T A B L E III

Soo t concen t r a t i ons ~ g m -3) a t s ta ted d o w n w i n d d is tance (km). Firs t en t ry is the m a x i m u m .

W i n d speed Stabil i ty Class Par t ic le s i z e u m m s - I k m h -1 B D E F

100 2 7.2 3 0 4 a t 1 4 8 2 a t 1 106at I 143 at 1 1 at 10 Oat 10 Oat 10 Oat 10

< 1 a t 4 0 0 a t 4 0 0 a t 4 0 0 a t 4 0

100 4 14.4 223 a t 0 . 8 682 at 1 513 at 1 171 at 1 2 a t 10 0 a t 10 0 a t l 0 0 a t l 0

0.1 at 40 0 at 40 0 at 40 0 at 40

100 6 21.6 - 6 4 0 a t I 7 2 2 a t I - - Oat 10 Oat 10 - - 0 a t 4 0 0 a t 4 0 -

100 8 57.6 - 5 5 9 a t 1 7 5 2 a t I - - < 1 a t l 0 Oat 10 - - 0 a t 4 0 0 a t 4 0

10 2 7.2 3 9 a t 2 1 0 a t 2 0 n o m a x i m u m n o m a x i m u m 7 a t 10 7 a t 10 0.1 a t l 0 Oat 10

0.6 a t 40 8 at 40 4 at 40 0.03 at 40

10 4 14.4 67 at 1 25 at 7 14 at 15 no m a x i m u m 4 at 10 21 a t 10 12 at 10 0.5 at 10

0.3 a t 4 0 6 a t 4 0 9 a t 4 0 5 a t 4 0

10 6 21.6 - 41 a t 4 25 a t 7 -

- 21 a t 10 25 at 10 -

- 4 a t 4 0 8 a t 4 0 -

10 8 57.6 - 5 4 a t 2 6 0 a t 5 - - 18 a t 4 0 48 a t 10 - - 3 a t 4 0 13 a t 4 0 -

T A B L E IV

M a x i m u m expected concen t r a t i ons (,ug m -3) du r ing b u r n i n g for a single pool (P), a cluster o f pools (CP), a n d supe r imposed clusters (SCP), the la t ter being es t imated by add ing the 10 k m f igure to the concen t r a t i on close to

the fire, doub l ing the 10 k m a n d 40 k m figures.

SO 2 Soo t C O

P C P S C P P C P S C P P C P S C P

C o n c e n t r a t i o n close to fire 100 300 390

C o n c e n t r a t i o n a t 10 k m 30 90 180

C o n c e n t r a t i o n a t 40 k m 10 30 60

700 2100 2700 225 675 875

200 600 1200 67 200 400

15 45 90 22 67 120

EMISSION FROM ARCTIC CRUDE OIL BURNING

T A B L E V

des i rable a n d ac tua l c o n c e n t r a t i o n s (pg m 3) Accep tab le ,

149

SO 2 Par t i cu la t e ma t t e r C O

M a x i m u m accep tab le 1 h concen t r a t i ons 8 or 24 h

1 yr

M a x i m u m desi rable 1 h concen t r a t i on 8 or 24 h

1 yr

Ac tua l O t t a w a (S la te r /Elg in) A n n u a l S u d b u r y (Ash) C o n c e n t r a t i o n E d m o n t o n (109/98) 1975 To ron to (College)

M o n t r e a l ( D r u m m o n d )

Average

9OO 1 35OO 300 120 15000

60 70 -

450 - 15000 150 - 6000

30 60 -

52 77 3600 73 32 - - 73 1500 39 71 - 96 101 -

65 71 2500

exposure of personnel or communities to SO 2 f rom a 2°7o oil is not regarded as

serious.

The soot concentrations close to the fire and at distances up to 10 km are clearly undesirably high.

There are reasons to suspect that oil smoke will contain higher concentrations of

polynuclear aromatic hydrocarbons (PNA's) than will particulate emissions f rom conventional sources. The direct comparison of suspended particulate concentrations

in the plume and with the acceptable or desirable concentrations or urban area

concentrations may, therefore, be misleading. Since PNA's have been established as

potent carcinogens, it is clearly desirable to be prudent and it is, therefore, recom- mended that human exposure to oil smoke be minimized by avoiding situations

where clean-up crews enter plumes or when prevailing winds carry smoke plumes

directly over communities. This can be accomplished by careful coordination of

igniting activities with short term weather forecasts. The use of aerial ignition techniques also reduces exposure.

Although the metal concentrations are low, and the prevalent metals Va and Ni

are not regarded as being particularly toxic, some prudence is also justified in exposing humans to these metal emissions. There is a possibility that toxic metals

such as Hg or Cr may be present in the oil or gas. The chemical speciation is not

known and an unforeseen combination of circumstances could result in unexpected exposure. This is an additional incentive for avoiding exposure to smoke.

The CO concentrations are an order of magnitude less than the desirable concen- trations, thus no problem is likely.

The principal source of hydrocarbons is evaporation of unburned oil rather than the burned oil. Volatile hydrocarbons are not regarded as noxious at low concentra-

150 T. DAy ETAL.

tions in regions not subject to photochemical smog. The only cause for concern may be the occasional local high concentration in which an explosion hazard exists.

A full report of this work is available from the Department of Fisheries and Environment, Environmental Protection Service, Ottawa (Day et al., 1979).

9. Conclusions and Recommendations

A scenario has been established to describe the time and space distribution of oil burning in the Beaufort Sea. The atmospheric concentrations of SO2, CO, soot and metals have been estimated for locations close to the fire and at distances of 10 km and 40 km. Comparison of these concentrations with Air Quality Objectives and typical levels in urban Canada, shows that the SO2 and CO concentrations will be acceptably low. The principal concern is the high soot concentration in the immediate vicinity of the fires. Human entry into such areas should be avoided. Since there is doubt about the effects of polynuclear aromatic hydrocarbons and metals which will be present in the soot, it is prudent to minimize the exposure of personnel and communities to these substances by careful planning of the burning operations using

short range weather forecasts. It is emphasized that these calculations are exploratory in nature and include a

large number of assumptions about atmospheric conditions, applicability of diffusion and plume rise equations, the time and space distribution of the burning operation,

the quantities of SO2, CO, soot, and metals emitted and the probability of plume superposition. Experimental validation (or otherwise) of the calculations would be desirable by undertaking a controlled oil burn in the Beaufort Sea with aerial plume sampling. This work should not be regarded as an accurate calculation of concentra- tions, but rather as a quantitative assembly of the relevant factors in a reasonably

logical framework to permit calculation of concentrations. The probable accuracy is

no better than a factor of 2 or 3. Although burning as a cleanup measure can be criticized in that it causes air

quality problems, it must be assessed on the overall effect which it can have by ameliorating the environmental damage from oil spills. The emissions are only a small fraction of the amount of oil destroyed. The logistical difficulties and hazards to human life of operating cleanup equipment in ice-infested or ice-covered waters

are such as to cast doubt on the feasibility of recovering a significant fraction of the spilled oil. Experience even in temperate waters has not been encouraging and one must expect poorer performance in the Arctic. The one advantage which ice and cold climate bring to cleanup is that the oil is likely to remain longer in a more burnable form, that is, it will lose volatiles and spread more slowly, thus it appears that burning may be more feasible in the Arctic than in temperate waters. Some problems require further elucidation, for example, the effect of soot in reducing albedo, thus modifying ice melting rates and weather, the issue of whether or not burning increases the emissions of PNAs, and the fate of metals. Subjective consider-

EMISSION FROM ARCTIC CRUDE OIL BURNING 151

a t i on o f the a l t e rna t ives o f t e m p o r a r i l y p o o r air qua l i t y ve rsus l o n g t e r m oi l

c o n t a m i n a t i o n o f wa t e r su r faces a n d shore l ines w h i c h c o u l d resu l t in ex tens ive

d e s t r u c t i o n o f wi ld l i fe , p a r t i c u l a r l y seab i rds a n d m a m m a l s , sugges ts t ha t b u r n i n g

p r o b a b l y has an ove ra l l a m e l i o r a t i n g e f fec t a n d s h o u l d thus be r e g a r d e d as a

p r i m a r y c l e a n u p m e a s u r e fn t he Arc t i c .

I t is r e c o m m e n d e d tha t f u r t h e r r e sea r ch be u n d e r t a k e n on oil b u r n i n g m e c h a n i s m s ,

on N o r t h e r n oil c o m p o s i t i o n s , o n oil e v a p o r a t i o n rates , on s o o t d e p o s i t i o n

charac te r i s t i c s , a n d on the d i spe r s i on charac te r i s t i c s o f the a t m o s p h e r e in the

B e a u f o r t Sea area .

Acknowledgments

This work was supported by a contract under the Arctic Marine Oilspill Program of the Department of Fisheries and Environment Canada, Ottawa, Canada. The authors are grateful to Mr. R Blackall, Dr. R. E. Munn, staff of the Atmospheric Environment Service (Downsview), and BP Oil Limited for their advice and assistance.

References

Berry, M. O., Dutchak, P. M., Lalonde, M. E., McColloch, J. A. W., and Savdie, I.: 1975, 'Weather, Waves and Icing in the Beaufort Sea', Beaufort Sea Project, Technical Report No. 21, Environment Canada, Victoria, B.C.

Burns, B. M.: 1973, 'The Climate of the Mackenzie Valley' Vols. i and 2. Environment CanadaAtrnospheric Environment Service Climatological Studies No. 24.

Coupal, B.: 1976, 'Report on Controlled Combustion Tests Carried out Near Rimouski', Environment Canada Report E P S - 4 - E C - 76-2.

Day, T., Mackay, D., Nadeau, S., and Thurier, R.: 1979, 'Characteristics of Atmospheric Emissions From An in situ Crude Oil Fire', Environment Canada Report. EPS-4-EC-79-1.

Department of Fisheries and Environment, Canada: 1976, Criteria for National Air Quality Objectives. Department of Fisheries and Environment: 1975, National Air Pollution Surveillance, Report 1975, EPS

5-AP-76-16. Environment Canada. Ledbetter, J. O.: 1972, Air Pollution: PartA Analysis, Marcel Dekker, Inc., New York. Logan, W. J., Thornton, D. E., and Ross, S. L.: 1975, 'Oil Spill Countermeasures for the Southern Beaufort

Sea', Beaufort Sea Project Technical Report 31a and 31b (Appendix). Environment Canada, Victoria, B.C. Also issued as Reports EPS-3-EC-77-6 and EPS-3-EC-77-7.

Milne, A.: 1975, 'Offshore Drilling in the Beaufort Sea-A Preliminary Environmental Assessment', Beaufort Sea Project Technical Report No. 39, Environment Canada, Victoria, B.C.

Munn, R. E., Tomalain, J., and Titus, R. L.: 1970, Atmosphere8, 52. Norcor Engineering and Research Ltd.: 1977, 'Probable Behaviour and Fate of Winter Oil Spill in the

Beaufort Sea', Environment Canada Report No. EPS-4-EC-77-5. Norcor Engineering and Research Ltd.: 1975, 'The Interaction of Crude Oil with Arctic Sea Ice, Beaufort

Sea Project Technical Report No. 27, Environment Canada, Victoria, B.C. Pimlott, D., Brown, D., and Sam, K.: 1976, 'Oil Under the Ice', Canadian Arctic Resources Committee,

Ottawa. Portelli, R. V.: 1977a, 'Data on Mixing Heights, Wind Speeds and Ventilation Coefficients for Canada',

InternaIReport No. ARQT-4-76, Atmospheric Environment Service of Environment Canada. Portelli, R. V.: 1977b, 'User's Guide to Three Interactive Point Source Dispersion Models', Personal

Communication. Ross, S. L., Logan, W. J., and Rowland, W.: 1977, 'Oil Spill Countermeasures: The Beaufort Sea and the

Search for Oil', Environment Canada, Ottawa. Seinfeld, J. H.: 1975, Air Pollution: Physicaland ChemicalFundamentals, McGraw-Hill, New York. Shaw, R. W., Hirt, M. S., and Tilley, M. A.: 1972, Atmosphere 10, 33.

152 T. DAY ETAL.

Singer, I. A., Frizzola, J. A., and Smith, M. E.: 1964, APCA J. 14, (AP 23S16041). Somers, E. V.: 1971, Dispersion o f Pollutants Emitted into the Atmosphere, Air Pollution Control, Part I,

Editor, Werner Strauss, Wiley-lnterscience, New York. Turner, D. B.: 1974, Workbook o f AtmosphericDispersionEstimates, USEPA Air Programs Pub. No. AP-26. Walker, E. P.: 1975, 'Oil, Ice and Climate in the Beaufort Sea', Beaufort Sea Project TechnicalReport No. 35,

Environment Canada, Victoria, B. C.