Evaporation Rates of Liquid Hydrocarbon Spills

On Land and Water

I1ON.4LI1 MACKAY mzd RONALD S . MATSUGU
Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario

The mass and heat transfer processes occurring during the On dQrit les processus de transfert de m a w et de chaleur
evaporation of liquid hydrocarbon spills on land and water are qui ont cours diirant l’evaporation d’hydrocarbures liquides
described and equations developed to enable the liquid tem- renverses sur le sol ou I’eau, ainsi que les equations mises au
perature and evaporation rate to be predicted. Experiments on point pour pouvoir predire la temperature et le taux d’kvapora-
the evaporation of cuniene, water and gasoline are described tion des liquides. O n expose les experiences faites sur I’eva-
and the evaporation mass transfer coefficient correlated with poration du c u m h e , de I’eau et de la gazoline et 6tablit une
the wind-speed, liquid pool size and the vapour phase Schmidt cordlation entre le coefficient d e transfert de masse par
Number. Good agreement is obtained between experimental evaporation et la vitesse du vent, les dimensions de la mare
and computed cuniene temperature and evaporation rates. liquide et le nonibre de Schmidt en phase gazeuse. O n a obtenu
Comparison of the correlation with flat plate mass transfer une boiine concordance entre les taux d’kvaporation et la temp&
correlations shows satisfactory agreement and suggests that rature (exphimentale et calculke) du cumhe. O n a aussi
turbulent transfer occurs, the rate being enhanced by the liquid obtenu une concordance satisfaisante en comparant la dite
surface roughness. ?lie more complex problem of estimating correlation avec celles du transfert de masse sur une plaque unie,
evaporation rates of hydrocarbon mixtures is discussed briefly. et il y a lieu de penser qu’il se produit un transfert turbulent
dont la vitesse augmente avec I’aspbite de la surface liquide.
On discute brievement le probkme plus coniplexe de 1’Cvalua-
tion des taux d’evaporation de mklanges d’hydrocarbures.
A knowledge of the rate of evaporation of spills
Df liquid hydrocarbons to the atmosphere im-
portant, not only as this is a major method of mass
is

loss from oil spills but also since evaporation con- Second, there is transfer from the interface to the
tributes to “weathering” or changes in the physical atmosphere, the vapor phase resistance. I t is assumed
properties of the oil. For example, crude oil on losing that there is no inter-facial resistance. This paper is
the more volatile fractions by evaporation becomes concerned mainly with the vapor phase resistance for
more viscous“’ and its flow and spreading character- which an empirical correlation has been developed. A
istics change considerably. In spill incidents arising description of the liquid phase resistance may or may
from accidents to road and rail tankers containing not be necessary depending on conditions and it is
gasoline and other hydrocarbon fractions it is im- discussed only qualitatively here.
portant to know how long an explosion hazard exists‘*’. A third process of importance is heat transfer since
Evaporation rate estimates are thus invaluable in this determines the oil temperature and hence vapor
determining how t o deal safely with such incidents. pressure which directly influences the mass flux by
Such data are also important in estimating the rate determining the concentration difference driving
of disappearance of oils spilled in the ocean where it force. The heat transfer mechanisms are complex
is well e~tablished‘~’ that large quantities of crude and include direct heat transfer from oil to the at-
oil residues are present. This problem is also of mosphere above and the ground or water below, radia-
particular significance in the Arctic where spills tion received and emitted, and enthalpy losses by
may occur on ice““ or from pipelines on land and evaporation.
where evaporation rates may be very much reduced
as a result of the low temperatures but possibly en- Liquid phase resistance
hanced by the high wind-speeds. When a single component liquid evaporates there
The problem considered here is that of determining is no liquid phase resistance since no concentration
the rate of evaporation of a given oil spill as a func- gradients exist. In multicomponent liquids two limit-
tion of temperature, wind-speed, atmospheric condi- ing conditions may exist. First, assuming there is
tions, solar radiation, ground or sea conditions, (i.e. an infinite diffusion rate (molecular or eddy) in the
roughnessj, the dimensions of the oil spill and the liquid (i.e. perfect mixing), the vapor leaving the
volatility and diffusion characteristics of the oil. This surface will be in equilibrium with the bulk liquid
estimate involves a description of the total mass and will generally have a composition different from
transfer process from the bulk of the liquid t o the that of the liquid, This is equivalent to a batch or
atmosphere. The mass transfer process is conveniently Rayleigh distillation. The liquid composition will be
divided into two parts. First there is the liquid phase constant with depth, but will change with time. The
mass transfer resistance which controls the rate of more volatile material evaporates first and as a re-
transfer of material from the bulk of the liquid to sult the rate of evaporation will fall a s evaporation
the interface and depends on the eddy and molecular proceeds.
diffusivities and the flow conditions in the liquid. The second limiting condition exists when there is

434 The Canadian Journal of Chemical Enginem’ng, Vol. 5 1 , August, 1973

assumed to be no diffusion in the liquid phase (i.e. water is almost always subjected to wave action.
no mixing). The vapor composition will equal t hat of Smith and McIntyre"' have shown that the onset of
the liquid and the liquid composition will remain con- whitecapping and breaking waves greatly increases
stant with time and with depth. The evaporation rate evaporation rates.
will remain constant. Physically this process can be Sivadier and Mikolaj'*'' found a threefold increase
envisaged as successive layers of liquid "peeling off" in evaporation rate as the sea roughened from Beau-
the surface to form vapor. fort 2.5 to 4 although part of this increase may be
In practice a situation intermediate between these due to differences in spreading.
two conditions undoubtedly exists and this will be
controlled by the relative magnitudes of the evapora- Velocity characteristics of the
tion rate and the liquid diffusion rate. This problem lorcer atmosphere
is of some importance, particularly under conditions
of low temperature such as in the Arctic where evap- The velocity profiles of the lower atmosphere have
oration rates may be low. Liquid mixing rates will, received intense study by meteorologists and by aero-
however, also be extremely low as a result of high dynamists in connection with the landing character-
oil viscosities. It. has. for example, been observed by istics of aircraft and the effects of wind gusts on
RamseierC5)that "a type of film forms on the surface buildings and other structures.
of the crude oil which essentially stops evaporation of The velocity profile above the oil spill is critical
the lighter fractions". in controlling the eddy diffusivity. There is no uni-
versally applicable equation relating velocity to height
Vapor.phase resistance since this depends on such factors as ground rough-
ness, the atmospheric temperature profile and the
A vapor phase mass transfer coefficient k, is de- height above the ground. Different equations fit the
fined in Equation (1) which gives the mass flux N data better if the profiles are considered over the first
(mols/m' h r ) as a function of the vapor pressure few feet, few hundred feet or few thousand feet
driving force which is the difference between the above the ground. The two most frequently used pro-
hydrocarbon vapor pressure a t the surface P (atm) files are first the logarithmic velocity profile in
and the hydrocarbon vapor pressure in the bulk of which the velocity ratio a t two heights is propor-
the atmosphere Pa, which is usually zero tional to the logarithm of the height ratio, and
second, the power law velocity profile in which the
N = k, (P - P m ) / R T p . .. . . . . . . . . . . . . . . . . . . . . . . , . (1) velocity ratio is proportional to the height ratio to
some power.
k, is a function of the transport conditions in the J t is convenient here to follow Sutton"' and assume
atmosphere immediately above the spill and is pos- that the wind velocity profile follows a power law
sibly a function of the molecular diffusion character- as shown in Equation (2) in which U is the wind-
istics (i.e. Schmidt number) of the hydrocarbon in speed ( W h r ) , U , is the wind-speed a t a height of
the vapor phase. It is probable, however, that tur- 1 m and 2 is the height ( m ) .
bulent transport is the dominant mechanism. k, will
also depend on the size of the pool. Although "stag- -
n
nant diffusion" conditions apply the Stefan or bulk u = u 1 z 2 - n. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
velocity correction factor'" is usually negligible ex- The exponent n is a function of ground roughness
cept in the evaporation of very volatile hydrocarbons. and temperature profile in the atmosphere and typ-
ical values are from 0.25 to 1.00. For average atmoe-
Heat transfer pheric conditions a value of 0.26 for n is reasonable"'.
An evaporating pool of oil generally has a tem-
It is interesting to note that when n is 0.26 this gives
rise to the 1/7th power law familiar in turbulent flow
perature close to that of the atmosphere and ground. in pipes. It should be noted that the standard height
At high rates of evaporation the oil may be consider-
ably cooler due to evaporative enthalpy losses (the for quoting wind-speed is ten metres.
Sutton'8' has shown that using the Von Karman
wet-bulb effect). On the contrary it may be consid-
erably hotter due to absorption of solar radiation, an similarity principle and this aasumed velocity pro-
file, the turbulent diffusivity can be expressed as a
effect commonly observed on black tarred road or
roof surfaces on sunny days. To permit accurate pre- function of the velocity profile and the mass transfer
dictions of evaporation rates it is essential to quan- coefficient can be deduced as in Equation (3) below
in which C is a constant and X is the pool diameter.
tify these effects and relate pool temperature to local
conditions. The pool temperature is determined by -
2 - n -
-n
the direct heat transfer from a ir above and ground
or water below, the incident radiation, the emitted
k, = C u2 + n X 2 +n
pool radiation, the rate of evaporation, the enthalpy For n = 0.25 this becomes
of evaporation and the depth and temperature history k, = c U0.78 x-0.1'.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
of the pool.
A quantitative model of these heat transfer pro- It is interesting to note that for n = 0 the velocity
cesses was constructed and fitted with good agree- is constant with height and the expression reduces
ment to experimental data. to the familiar unsteady-state penetration diffusion
equation in which the mass transfer coefficient is
Area effects inversely proportional to the square root of the con-
tact time X / U . The mass transfer coefficient is then
A further factor which requires consideration is proportional to the square root of the velocity and
that the surface of oil spilled on land may not be inversely proportional to the square root of the pool
planar since surface waves are formed in liquids of size.
low viscosity a t even low wind-speeds. Oil spilled on The approach taken here was to perform experi-

The Canadian lournal of Chemical Engineering, Vol. 5 1 , August, 1973 435

ments which would lend support to the validity of
Equation (3) and suggest values f o r the constant C.
Cumene was evaporated from a pool over a two day
period to estabiish a value for C.
The exponent on the velocity is difficult to con-
firm since there is no control over the atmospheric
conditions, but fortunately this value is well estab-
lished from other studies of turbulent transport. The
- W AH' APIRT,,]
I
I t is desirable to include with M , the weight of the
exponent on the pool dimension X was investigated evaporating hydrocarbon, the equivalent thermal mass
by evaporating water from pools of different sizes. of the tray which is in intimate thermal contact with
Finally studies were made of the rate of evaporation the hydrocarbon.
of gasoline in which there may be a liquid phase re- This equation cannot be integrated analytically but
sistance. can be integrated numerically by a fourth-order
Runge-Kutta method to give T, as a function of time
provided that data a r e available for T., U and L as a
function of time, t h a t A P can be expressed as a
THEORY function of temperature T, by the Antoine Equation
and numerical values are available for a, e,, e., C , q
Figure 1 shows the terms in the heat budget of an and 1'.
evaporating pool. From Sutton's analysis i t can be Once T, is known, the rate of evaporation and
assumed that Equation (3) expresses the mass trans- cumulative mass losses can be calculated.
fer coefficient in terms of the wind-speed and pool The approach adopted here was to assume a value
size, the constant and exponents being a s yet unde- for q and to determine r by water evaporation. The
termined. From the heat and mass transfer analogy"' value of C was established by using the experimental
the heat and mass transfer coefficients can be re- T, data to generate the observed evaporation rates.
lated as follows: Values of a, el, and e. were obtained by trial and error
to match the experimental and computed T, curves.
Finally, the equation was used to compute simultane-
ously evaporation rate and pool temperature as a
thus function of time. Time increments of 0.10 h r s were
used in the integration.
kt, = k., p , C,, (.Sc/Pr)o.u7 The constant C essentially defines the mass transfer
coefficient and will vary from system to system de-
For cumene evaporation a t 30°C Sc is 2.70 and P r is pending on the diffusivity characteristics a s repre-
0.745'". The rate of heat transfer from the atmosphere sented by the Schmidt number. For two systems the
to 1he pool is thus kh( T,,- T,) A cal/ h r where T. and constants C, and C2 can be related by the equation
T , are the atmospheric and pool temperatures and A below,
is the pool area.
C? = c, (SC,/SC?)0.6'
Solar radiation L, (callm') which includes both in-
cident direct solar radiation and scattered sky radia- EXPERIMENTAL
tion, which reaches the earth's surface is available
from meteorological station data f o r each one hour Studies were made of the evaporation of water,
period. A high proportion of this radiation is reflected, cumene (isopropyl benzene) and gasoline. The gaso-
this being quantified by ;Isurface albedo "a" or frac- line (regular) was supplied by B P Oil Limited, the
tion of radiation reaching the ground which is re- cumene by Gulf Canada Limited. The purity of the
flected. The heat input from solar radiation will thus cumene was sufficient to avoid any significant effect
be (1 - a)LA cal/hr. on vapor pressure. The a i r temperature was varied
It is necessary to include the heat loss by long from +5"C to +30°C and the wind-speed varied from
wavelength emitted radiation which will be e,uTdA zero miles per hour to 15 miles per hour.
(cal/hr) where e, is the emissivity of the pool f o r Two sites were used. First, on the roof of the Chem-
long wavelength radiation and u is the Stefan-Boltz- ical Engineering Department at the University of
mann constant. A considerable portion of this emit- Toronto, water was evaporated. Second, a site was ob-
ted radiation is compensated by counter-radiation tained at Toronto Harbour near the Eastern Channel
from the atmosphere. This can be qualified as e,,aT:A of Toronto Island where the hydrocarbons were evap-
(cal/hr) where the constant e. is dependent on the orated. Meteorological wind-speed data were obtained
water vapor pressure of the atmosphere. from Toronto Island Airport.
Evaporative cooling will cause a heat loss of Wooden evaporation pans were used of 4-ft by 4-ft
and 4-ft by 8-ft by 0.75411 deep painted with epoxy
k., AIY AH" n P I R T cal/hr resin. The rate of evaporation was determined for
the pure liquids by measuring the rate of inflow to
where W is the molecular weight of the evaporating the pan necessary to maintain a constant level. The
material, AH" the enthalpy of vaporization (cal/g), apparatus operated on the "chicken-feeder" principle.
and A P the vapor pressure driving force (atm) With gasoline i t is not possible to use this procedure
where AP is normally the vapor pressure of the due to the change in composition with evaporation.
liquid a t the surface. The volume in the pan was measured periodically.
Combining these heat flux terms with the mass of The volume was measured since waves made depth
evaporating hydrocarbon the differential equation measurement difficult.
describing the temperature variation of the perfectly The effect of pool size was determined by evap-
mixed pool with time can be written, orating water from the 8-ft by 4-ft, and 4-ft by 4-ft

436 The Canadian Iournal of Chemical Engineering, Vol. 5 1 , August, 1973

 atmospheric temperature T,

P
I
pool temperature Tp
pool mass M
pool length X pool area A
Figure 1 -Heat budget terms for an
evaporating pool.
I
I
a2
. ... -
Figure 3 Experimental and computed
evaporation rates and radiation.
1. emitted radiation
DIAMETER (METRES)
2. evaporation
3. direct heat transfer Figure 2 -Effect of pool size on eva-
4. solar and atmospheric radiation poration rate.

pans and a 1.0 x 0.7-ft tray placed in close proximity flow correlations confirm Sutton's suggestion of a
on the roof of the Chemical Engineering Department. value of 0.78 for q.
The pan positions were interchanged to avoid any
effect of position on evaporation rate. The earliest work on evaporation rates from lakes
by Dalton"31 led to the suggestion of a linear relation-
ship ( q = 1) between evaporation rate and wind-
RESULTS AND DISCUSSION speed and this has been used in many subsequent
hydrological equations. A somewhat lower value for
Exponent "r" on pool size q seems appropriate and 0.78 is used in this study.
The water evaporation tests showed that as pre-
dicted the evaporation rate per unit area falls slightly Constant "C )'

as the pool size increases. Probably this can be ex- From the cumene evaporation data the best value
plained as an enhanced evaporation rate at the upwind of C can be determined i.e. the value which will give
or entrance region of the pool. The experimental data equal computed and experimental cumulative mass
are given in Figure 2, a plot of the logarithm of the losses over the two day period. The cumene vapor
evaporation rate per unit area against X the pool pressure was calculated from the experimental pool
equivalent diameter. The slope suggests a value for
r of -0.12 compared with -0.11 suggested by Sutton.
temperatures using the Antoine Equation 'l".
This difference is small when it is considered t h a t a
ten-fold increase in pool size would give a reduction
+
log P ( a h ) = 6.93666 - 146O.793/(Tc 207.777)
in rate by a factor of 0.76 using -0.12 and 0.78 using A value of 0.0150 was found. It should be noted
-0.11. that this value has dimensions dependent on the units
I n hydrological studies of water evaporation from used for the mass transfer coefficient, wind-speed
reservoirs, lakes and experimental pans a "pan fac- and pool size, here mlhr, m/hr and m respectively.
tor" of about 0.7 is used to calculate large scale evap- For systems other than cumene C will be given by
oration rates from pan rates(l0). Assuming that this 0.0150 ( 2 . 7 0 / S ~ ) ~ "or
' 0.0292 SC-"~@".
represents about a thousandfold variation in X the
corresponding value for r is -0.05, i.e. l,000-0.05= 0.70. Some confirmation is available for this value from
water evaporation data which has been correlated by
These results suggest that in scaling up evapora- Fitzgerald"".
tion rates the power on the pool size should be in
the range -0.05 to -0.11 and a value of -0.11 as given
by Sutton and confirmed approximately here, seems
E, = +
(0.40 0.199 W ) ( P ' - PA)
reasonable and is used in this study. Using the equation developed here at a temperature
I n view of the experimental uncertainties and the of 20°C. a 4-ft tray, atmospheric pressure, 50% rela-
requirement for only an approximate estimate of tive humidity, a wind-speed of 10 miles per hour and
evaporation rate the selection of any value in the correcting for the Schmidt number, the evaporation
range quoted is sufficiently accurate. rate of water will be 0.65-in/day. The above correla-
tion predicts a value of 0.82-inlday which is in good
Caution should be used when using this relation- agreement considering the differences between water
ship to extrapolate to very small pool sizes, e.g. 0.1 m and hydrocarbons in enthalpy of vaporization and
or less since the equation may not be valid. tendency to induce natural convection. Other water
evaporation correlations were compared giving rates
Exponent "q" on wind-speed of about 0.4-in/das.

In the cumene and water evaporation tests the Albedo and emissivities
wind-speeds were variable and the effect of wind-
speed was impossible to separate from the other ef- In fitting the experimental and computed tempera-
fects, particularly temperature. The results of ture data the best f i t was obtained using an albedo
Himus"'), Hine"" and the general acceptance of a a of 0.14 and a pool emissivity e, of 0.95. These values
power of about 0.8 on Reynolds number in turbulent are in reasonable agreement with generally accepted

The Cunadian Iournal of Chemical Engineering, Vol. 5 1 , August, 1973 437

 101 I
0 24 48 106
TIME (HOURS)
REYNOLDS NUMBER

-
Figure 4 Experimental and computed Figure 5 - Comparison with sniooth flat
pool teinperatures and air temperature. plate data.
TIME (MINUTES)

values for water surfaces"o"". The term e. for coun-

Figure 6 - Gasoline evaporntioii.
ter-radiation was given a value of 0.75 which is in
agreement with literature data"". The complete tem-
perature and evaporation rate data for the cumene To estimate the vapor pressure as a function of
are given in Figures 3 and 4 together with the at- temperature and percent evaporated, a simulated
mospheric temperature and radiation data. composition was calculated consisting of a mixture
of C5 to C , , n-alkanes of such a composition that it
It should be noted that during periods of high in- would give the same ASTM batch distillation (tem-
cident radiation (around noon) the pool becomes con- perature vs fraction evaporated) curve as the gaso-
siderably warmer than the air. At night, however, line. This simulated mixture could thus be used in
evaporative cooling tends to give the pool a lower the numerical integration of the equations for a batch
temperature. Clearly it is inaccurate to assume that multi-component constant temperature distillation
the pool temperature will equal the air temperature. combined with t,he mass transfer equation developed
For approximate calculations under environmental here. The results of the computed and experimental
conditions it would seem valid to make this assump- evaporation rateR a r e shown in Figure 6. I t can be
tion for night and cloudy day conditions. For sunny seen that the computed rates are higher suggesting
conditions the pool may be 20°C hotter than the at- that there is a liquid phase mass transfer resistance,
mosphere. For spills on water it is probably best to the assumption built into the computed curve being
assume an oil temperature equal to the water tempera- that the pool is perfectly mixed. Less than perfect
ture, particularly if the oil layer is thin. mixing will deplete the surface region of the more
volatile components and thus reduce the vapor pres-
Comparison with smooth flat plate sure and the evaporation rate. This could be quanti-
mass transfer data fied by assuming a finite eddy diffusivity value.
It is interesting to compare the correlation with The approximate agreement in rates is encouraging
those for mass transfer from smooth flat plates"8'. and suggests that with inclusion of a liquid phase
This is conveniently done on a plot of j factor versus resistance it should be possible to estimate rates of
Reynolds number, in which the length term can be evaporation of complex mixtures if data is available
defined either as the distance from the leading edge, on vapor pressure and the liquid phase diffusion char-
giving a local transfer coefficient, or the plate length acteristics.
giving an average transfer coefficient as is consid-
ered here. The proposed correlation when converted CONCLUSION
into j factor, Reynolds number form becomes
Data on the evaporation of water and cumene have
j., = 0.0292 Re-0.22 (XO.11 Q.22 p e . 2 2 ) = 0.0565 K0.22
been used to correlate the gas phase mass transfer
This is plotted with literature data on Figure 5 for coefficient a s a function of wind-speed and pool size
values of X, p and p which apply to the present ex- by the equation,
perimental conditions. In this study Re ranged from k, = 0.0292 uO.78 X - O . l l s ~ - O . ~ 7
7.0 x lo' to 4.6 x 10'. As is expected the correlation
suggests that turbulent transfer conditions prevail If pool temperature and vapor pressure data are
with the rates being about 1.26 times the smooth flat available the rate of evaporation can be predicted.
plate data. This is presumably due to the surface The pool temperature is a complex function of the
waves which will increase mass transfer rates. heat fluxes from direct heat transfer, radiation and
evaporative cooling. The direct atmosphere to pool
Gasoline evaporation heat transfer coefficient can be calculated from k,,
by analogy and if the albedo, emissivities, and inci-
In this case a liquid phase resistance will be pres- dent radiation are known the pool temperature can
ent and the evaporation rate will fall a s the more be calculated by numerical integration of the differ-
volatile components evaporate and the total vapor ential heat balance equation. The values of the con-
pressure fa!ls. This system is thus much more com- stant and the powers are in good agreement with
plex and accurate prediction of evaporation rate is other data on water evaporation.
presently impossible. For hydrocarbon mixtures the evaporation process
I t is interesting, however, to attempt the prediction is more complex, being dependent on the liquid dif-
because the ultimate aim is to predict evaporation fusion charmt.eristics, a liquid phase resistance being
rates for such systems. present. Data on gasoline evaporation when compared

438 The Canadian Journal of Cheniiral Engineering, Vol. 5 1 , August, 1973

with computed rates assuming a composition and per- T. = air temperature O K
fect liquid mixing show f a i r agreement a n d confirm T, = cumene temperature “C
TI, = pool temperature OK
the presence of a liquid phase mass transfer resistance. I! = wind-speed metreslhr
It is hoped t h a t t h i s correlation will be of value in w = wind-speed m.p.h.
calculating evaporation rates under a variety of at- W = rr.olecular weight
mospheric conditions and may lead to better predic- X = ~:ooldiameter m
Z = height m
tions of the weathering of spills of hydrocarbon oil u, = viscosity of vapor g/ni.hr
products and crude oil, these predictions being valu-
able in assessing t h e change in oil properties and t h e H ejerences
nature and duration of f i r e and explosion hazards.
(1) Berridge. S. A.. Dean, R. A.. k‘alluws. R. G. and Fish. A.. “The
I’ronerties of Persistent Oils a t Sea”. .I. Inst. petrol.. vol. 54.
A cknowledgtnents Nib. 639. Nov. 11968).
( 2 ) Wilder I. a n d Brugger. J. E.. “Prenent a n d F u t u r e Technology
Rccluirhnents f o r t h e Containment of Hazardous Material Spills”.
The authtbrs a r e grateful to t h e National Research Council of Nat. Conf. on Control of Hazarduus Material Spills. Houston,
Canada for n Scholarship. t o B P Oil Limited a n d Environment Canada Texas. March .!972.
National Advisory Committee on W a t e r Resources Research for financial (3) Hoult. D. P., Oil on t h e Sea”. Plenitm Press, New York. 1969.
support, to Gulf Canada Limited f o r a g i f t of cumene and to t h e (41 Urown. T.. “Oil on lee”, Sierra Club, New Yurk. 1971.
Toronto Harbour Commission for permission to use their facilities. 151 Hamseier, R. 0..“Oil Pollution in l e e Infeated Watera”. Int.
Symp. on Identification and Meaaurement of Environmental
Pollutants. Ottawa, 1971.
Nomenclature 1 6I ‘I’revbal. R. E.. “Mass Transfer Onerationa”. McGraw-Hill. New
Y d k , 2nd Edn.? 1968.
(71 Smith, C. L. a n d Maclntyre, W. G., “Initial Ageing of Fuel Oil
a =-al‘oedo to incident radiation Films on Sea Water”, A P I on Prevention of Oil Pollution. 1970.
A = area of evaporating pool m? tiage 457.
C = constant in Equation (3) IRI Sutton. 0. C.. “Micrometeorulogy”. McGraw-Hill. New York. 1963.
191 Perry. J. H.. “Chemical E n g i n e e n ’ Handbuok”, McGraw-Hill. New
C,,, = I eat capacity of liquid cal/g “C Yurk. 196:j.
C,,, = heat capacity of vapour cal/g “C IIOI Kazmann. R. G . , “Modern Hyclroliwy’. Harr)er and Row. New
e, = emissivity of pool
York. 1965.
I111 Himua. G . , “Conference o n Vapor Absorption and Adsorption”,
en = constant for atmospheric radiation Institute of Chemical Engineers. 1929.
E , = water evaporation rate in/day 1121 Hine. T. B.. Physics Review, Vol. 24, No. 79 (1924).
A H r = enthalpy of vaporization cal/g 11:: I Dalton. J., “Experimental Essays on the Constitution of Mlxed
(;aaea: on the Force of Steam o r v a p o r from Water a n d other
J = J factor dimensionless Liquids in Different Temperatures. Both in a Torricellian Vacuum
k,, = heat transfer coefficient cal/m2 hr “C and in A i r ; o n EvaporatiDn; a n d on the Expansion of Gasen by
k,. = mass transfer coefficient m/hr Heat”. Mem.. Manchester Lit. Phil. Soc.. Ser. 1. Vol. 5. pt. 2,
pp. 556-609 I1802).
L = incident solar radiation cal/m’ I141 Zwolinski. B. J. a n d Wilhoit. R. C.. “Handbook of Vapor Pressures
A1 = mass of evaporating pool g and Heats of Vaporization of H y d r E a r b o n s and Related Com-
ti = exponent in Equation (2, vounds”. A P I Project 44. 1971.
Wialer. C. 0. and Brater, E G., “Hydrolouy”. Wiley and Sons
A‘ = ntass flux g. mol/mz hr Inc.. New York. 1963.
A P = vapor pressure driving force atm Uudykai. M. I., “The H e a t Balance of t h e Earth’s Surface”.
P = vapor pressure of hydrocarbon at surface a h Department of Commerce, Washington. D.C.. 196H.
Hvttel. H. C. a n d Sarofin, A. F.. “Radiative Transfer.’. McGVaw-
P , = vapor pressure h drocarbon at infinity atm Hill. New York. 1967.
PA = vapor pressure orwater in air in. Hg ( i e i w r , R.. “The Climate near t h e Ground”. Harvard University
P ; = 1 randtlnumber Press, Cambridne. 1966.
Uird. R. 13.. Stewart, W. E. a n d Lightfoot, E. N.. “Transport
P = maximum vapor pressure of water in Hg Phenomena”, p. 410. Wiley a n d Sona Inc.. New York. 1960.
q = exponent on wind-speed Sivadier, H. 0. and Mikolaj. P. G.. Symp. o n the Prevention a n d
= exponent on pool sue Control of Oil Spills. p. 476. American Petroleum Inst.. Washing-
k = pas constant alm m3/g. mole “ K Re = Reynolds number ton. D. C., U.S.A., March 1978.
p, = density of vapor g/m3
Sc = Schmidtnumber
a = Stefan-Boltzmann constant cal/m2 h7 OK4 Manuscript received J a n u a r y 30: accepted May 7. 1973.
Sfn = Stanton number for heat transfer
SI, = Stanton number for mass trnasfer
I = time hr * * *

The Canadian lournal of Chemical Engineering, Vol. 5 1 , August, 1973 439