ANNUAL BIRD MORTALITY IN THE BITUMEN TAILINGS PONDS IN
NORTHEASTERN ALBERTA, CANADA
KEVIN P. TIMONEY1,3AND ROBERT A. RONCONI2
ABSTRACT.—Open pit bitumen extraction is capable of causing mass mortality events of resident and migratory birds.
We investigated annual avian mortality in the tailings ponds of the Athabasca tar sands region, in northeastern Alberta,
Canada. We analyzed three types of data: government-industry reported mortalities; empirical studies of bird deaths at
tailings ponds; and rates of landing, oiling, and mortality to quantify annual bird mortality due to exposure to tailings ponds.
Ad hoc self-reported data from industry indicate an annual mortality due to tailings pond exposure in northeastern Alberta
of 65 birds. The self-reported data were internally inconsistent and appeared to underestimate actual mortality. Scientific
data indicate an annual mortality in the range of 458 to 5,029 birds, which represents an unknown fraction of true mortality.
Government-overseen monitoring within a statistically valid design, standardized across all facilities, is needed. Systematic
monitoring and accurate, timely reporting would provide data useful to all concerned with bird conservation and
management in the tar sands region. Received 17 November 2009. Accepted 5 May 2010.
Global demand for unconventional energy
sources such as coal bed methane, heavy oil,
and bitumen has grown in recent years. Bitumen
in northern Alberta, Canada, is extracted by two
methods, in situ well-based approaches and truck
and shovel open pit mining. The latter method
produces ‘‘tails’’ during separation of bitumen
from the sand. The tails, a mixture of process-
affected water, residual hydrocarbons, brine, silts
and clays, and metals are discharged into tailings
ponds. The extent of tailings ponds in northeastern
Alberta grew by 422% between 1992 and 2008
(Timoney and Lee 2009). The Athabasca tar sands
development is one of the largest energy projects
in the world. Production of bitumen is predicted to
rise from the current 1.3 million barrels/day to
three million barrels/day by 2018 (Alberta Energy
Water bodies along migration routes attract
many bird species as they afford foraging,
roosting, nesting, and resting opportunities (Ron-
coni 2006). A variety of deterrents have been used
to discourage waterbirds from landing in tailings
ponds such as floating and beach effigies, propane
scare cannons, and sound-producing systems
(Boag and Lewin 1980, Golder Associates Ltd.
2000, Ronconi and St. Clair 2006). Some birds
that land at tailings ponds become oiled and a
proportion of the oiled birds later die. Bird
mortality rates from oiling have not been precisely
measured, but casualties appear to be high for
gregarious species, particularly for diving birds
(Clark 1984). Bird migration is affected by
weather as birds are more likely to land when
they encounter headwinds, low temperatures, and
precipitation (Newton 2007). Storms may increase
the likelihood of bird oiling at tailings ponds
(Ronconi 2006), and inclement weather may
increase the probability of mass mortality events.
Oiled ducks may suffer from reduced insula-
tion, increased metabolic rate, and hypothermia
even from small amounts of oil (Hartung 1967,
McEwan and Koelink 1973). Survival rates of
rehabilitated birds may be as low as 1 to 20% for
some species (Mead 1997). Birds from 43 species
have died due to exposure to tailings ponds in the
area, mostly waterbirds such as dabblers and
divers: Mallard (Anas platyrhynchos), Common
Goldeneye (Bucephala clangula), Northern Shov-
eler (Anas clypeata), Lesser Scaup (Aythya
affinis), American Coot (Fulica americana),
grebes, mergansers, geese, and shorebirds includ-
ing Semipalmated Sandpiper (Calidris pusilla),
Pectoral Sandpiper (C. melanotos), Stilt Sandpiper
(C. himantopus), and Lesser and Greater Yellow-
legs (Tringa flavipes, T. melanoleuca) (Sharp et
al. 1975, Dyke et al. 1976, Gulley 1980, Ronconi
2006). Deaths of birds of prey, gulls, passerines,
and other groups have also been documented.
Mortality rates may be high even at small ponds:
27 dead birds were found at a 0.4-ha tailings pond
lacking deterrents (Dyke et al. 1976). There may
be continual ‘‘incidental take’’ of birds during the
open water season, especially at night when
human observations are impractical. Oiled birds
in tailings ponds have been observed to sink out of
1Treeline Ecological Research, 21551 Township Road
520, Sherwood Park, AB T8E 1E3, Canada.
2Department of Biology, Dalhousie University, 1355
Oxford Street, Halifax, NS B3H 4J1, Canada.
The Wilson Journal of Ornithology 122(3):569–576, 2010
sight (Dyke et al. 1976), minimizing the chance of
Our objective was to provide estimates of
annual bird mortality resulting from bitumen
tailings pond exposure in northeastern Alberta,
Canada through synthesis and analysis of avail-
able data. These data included numbers reported
by industry to government and scientific data on
mortality and landing rates at tailings ponds.
Study Area.—We studied avian mortality in the
Athabasca bitumen (tar) sands region (geographic
center at 57u 039 N, 111u 319 W, Fig. 1) in the
lower Athabasca River watershed north of the city
of Fort McMurray, Alberta, Canada. The
120.6 km2of tailings ponds within the area of
open pit mining, as of March 2008, covered 1.4
times the area of natural water bodies (84.9 km2)
composed of the surface of the Athabasca River
(50.4 km2) and lakes, ponds, and other river
surfaces (34.5 km2) (Timoney and Lee 2009; KPT
and RAR, unpubl. data). The area lies within a
convergence zone of North American waterfowl
flyways; millions of birds migrate through
northeastern Alberta en route to and from local
and distant breeding areas in northern Alberta, the
Peace-Athabasca Delta, Mackenzie River Valley,
and the arctic (Butterworth et al. 2002, Thomas
2002, USDI 2009b). Thirty-five species and
species groups of waterbirds, and 29 other species
have been observed on one lease (Syncrude # 17)
at the natural water body Mildred Lake (Sharp et
al. 1975). More than 16,000 birds were observed
flying over one tailings pond during spring
migration (Ronconi and St. Clair 2006) while
more than 25,000 swans, geese, ducks, Sandhill
Cranes (Grus canadensis), and gulls were ob-
served in daylight during a fall migration at
Syncrude Lease # 17 (McLaren and McLaren
1985). The total number of migratory birds
passing through the lower Athabasca River Valley
Data Collection and Analyses.—Spot censuses
and shoreline searches for dead birds of varying
duration, extent, and frequency at tailings ponds
of known areal extent during the open-water
season (Gulley 1980, Van Meer and Arner 1985)
were used to calculate bird mortalities per km2
from which mortalities were adjusted to the 2008
areal extent of tailings ponds. Mortality data were
collated from three companies with tailings ponds
(Suncor 1990–2008, Syncrude 2000–2007, and
Shell Albian 2000–2008). Data were obtained
from reports produced by the companies (Syn-
crude 2008), and from the Alberta government
(Sustainable Resources Development) under a
freedom of information request (K. P. Timoney,
October 2008). These data, reported by company,
year, and mortality type were analyzed to obtain
mean annual mortality. We estimated the total
number of birds landing and subjected to oiling
during spring migration at the Albian Sands
Muskeg River Mine: landings/hr (from Ronconi
and St. Clair 2006) and oiled birds/day (from
Mortality Rates Estimated by Systematic Sur-
veys.—Systematic surveys for dead birds at
tailings ponds (Table 1), used to calculate mor-
tality per km2(range 7.2 to 145.2 birds), were
extrapolated to estimate total potential mortality
based on 120.6 km2of tailings ponds in 2008. An
estimate based on the lowest observed mortality at
2009). Areas (as of 19 March 2008) undergoing bitumen
extraction are hachured; tailings ponds are black. Tailings
pond names are MLSB 5 Mildred Lake Settling Basin;
ANTP 5 Aurora North tailings pond; SATP 5 Shell Albian
tailings pond; TIP 5 Tar Island Ponds 1 and 1A; 8AEML 5
Suncor Millennium tailings ponds 8A and EML.
Study area (modified from Timoney and Lee
THE WILSON JOURNAL OF ORNITHOLOGY N Vol. 122, No. 3, September 2010
Syncrude’s Mildred Lake Settling Basin (MLSB)
of 7.2 birds/km2in 1985 yielded an annual
mortality of 863 birds. A medium estimate based
on Syncrude’s MLSB (1980–1985) average mor-
tality of 13.38 birds/km2yielded an annual
mortality of 1,614 birds. A high estimate based
on the weighted mean mortality rate for all years
at Syncrude’s MLSB and Suncor’s Tar Island
Ponds 1 and 1A of 41.7 birds/km2yielded an
annual mortality of 5,029 birds.
Industry-based Annual Mortality Reported
to Government.—Annual mortality attributed to
oiling over the period 2000 to 2007 ranged from
17 to 201 birds. The weighted mean (6 SD)
annual mortality was 65 6 59 birds. (Table 2).
Additional annual mortalities attributed to ‘other’
and to ‘unknown’ causes (details in Table 2)
averaged (6 SD) 13 6 9 (max 31 in 2007) and 16
6 9 birds, respectively. Industry data had poor
agreement with the government data released
under the freedom of information request (Ta-
ble 3); the mean difference was 19%.
Annual Bird Mortality Estimated by Landing
and Oiling Rates.—A spring landing rate of
121.44 birds/day in a 3.5 km2tailings pond was
calculated, resulting in an estimated rate of 34.69
landings/km2/day during daylight hours only (low
estimate, Table 4). We calculated 54.93 landings/
day (109.86 landings/km2/day) (high estimate,
Table 4) from observations at a 0.50-km2area
where deterrent testing occurred (RAR, unpubl.
data not previously reported in Ronconi and St.
Clair 2006). Scaling for the total area of tailings
ponds in 2008, about 125,513 to 397,408 birds
may land during a 30-day spring migration period.
Thirteen oiled waterbirds and shorebirds were
found during the same period and at the same site
(Ronconi 2006), most of which were covered in
.50% oil, from which oilings/day and the
proportion of landed birds becoming oiled were
calculated (Table 4). We estimate that 286 to 905
birds may be oiled during spring migration at an
overall oiling rate of 0.2278% for birds that
landed on ponds. About 229 to 815 birds may die
each spring due to oiling if an oiled bird is unable
to land more than once, and 80 to 90% of oiled
birds die (Discussion).
Uncertainties in Mortality Estimates.—Bird
oilings may peak in August and September rather
than in spring (Van Meer and Arner 1985), and it
is reasonable to double the spring mortality to
derive an annual mortality of ,458 to 1,630 birds.
systematic shoreline surveys.
Estimated annual bird mortality/km2/year in the Athabasca tar sands tailings ponds based on spot counts and
scare cannons, human effigies
scare cannons, human effigies
scare cannons, human effigies
human effigies; fresh tailings received 95% of
human effigies; fresh tailings received 100% of
human effigies with artificial lighting at night;
fresh tailings received 92% of days (Apr–Oct)
deterrents?d; fresh tailings received 34% of days
deterrents?d; fresh tailings received 25% of days
deterrents?d; fresh tailings received 0% of days
Pond 11.86 1978 7942.472
Pond 1 1.861979 270145.162
Pond 1A0.56 197743 76.792
Pond 1A 0.56 197831 55.362
Pond 1A 0.56 1979 3358.932
aAreas of MLSB derived from planimetry of airphotographs (1980, AS2165-13; 1984, AS3051-5; 1986, AS3356-280; Ponds 1 and 1A areas derived from Gulley
bDead birds/year 1980–1983 derived mathematically from reported values for 1984 and 1980–1984 (Van Meer and Arner 1985): 1980–1984 average mortality of
170.4 birds/year; total birds dying 1980–1984 5 170.4 3 5 5 852 birds; 1984 mortality of 94 birds; 1980–1983 average mortality 5 (852 2 94)/4 5 189.5 birds/
year, or 16.36 birds/km2; the average weighted mean mortality 1980–1985 5 ((16.36 3 4) + 7.68 + 7.15)/6 5 13.38 birds/km2. The high estimate of annual mortality
is the weighted mean mortality per km2; it is the sum of 500.38 birds/km2for 12 years of data (1980–1983 comprises 4 years of data), 500.38/12 5 41.70 birds/km2.
cReferences: 1 5 Van Meer and Arner (1985); 2 5 Gulley (1980); 3 5 Golder Associates Ltd. (2000).
dQueries sent to Suncor (17 Nov to 12 Dec 2008) regarding deterrents in use on Pond 1A during 1977–1979; no reply received to date (4 May 2010).
Timoney and Ronconi N BIRD MORTALITY AT TAILINGS PONDS
This adjustment to the mortality estimate may be
conservative as it does not include mortalities that
occur before spring, between spring and fall
migration, and after fall migration. Annual
tailings pond mortality estimates derived from
mortality surveys (863 to 5,029 birds) and
landing-oiling rates (458 to 1,630 birds) are
roughly of the same magnitude. Self-reported
oiling mortality data from industry provide the
lowest estimate (65 bird deaths/year) whereas
Wells et al. (2008) provide a high estimate of
8,676 to 156,168 bird deaths/year. Wells et al.
(2008) assumed that all birds that land at tailings
ponds are oiled and that peak landing rates exist
24 hrs/day for 100 days. Our mortality estimates
may be conservative given that 500,000 to one
million birds die annually at oilfield wastewater
ponds in the United States (USDI 2009a). Those
wastewater ponds are similar to bitumen tailings
ponds in their mixture of water, residual oil or
bitumen, and salts.
The presence of extensive tailings ponds
petroleum companies with tailings ponds in northeastern Alberta.a
Bird mortalities attributed to oiling, ‘other’ and ‘unknown’ causes released by the Alberta Government for
Suncor Syncrude Albianb
SuncorSyncrudeAlbianSuncor Syncrude Albian
Recent Mean 6
a‘Other’ includes electrocution, collisions, predation, fights with other birds, and natural causes; ‘Unknown’ includes incidents where company was not able to
identify cause of death and incidents where cause of death was not reported.
bTailings pond at Shell Albian began to fill in 2003; mortality due to oiling not expected prior to 2003. Calculations of recent tailings pond mortalities used years
2000–2007 for Suncor and Syncrude and years 2003–2007 for Shell Albian. Mortality means for ‘other’ and ‘unknown’ use the period 2000–2007. Mortality means
are for each company and mortality type. Average mortalities by year, 2000–2007, oiling 64.8 6 58.7, other 13.1 6 9.3, and unknown 16.5 6 9.1.
cValues for 2008 were ‘year to date’ current to July 2008 with the exception of Syncrude for which the death of 1,606 ducks at the Aurora North tailings pond in
April 2008 was not made public until 2009.
TABLE 3.Annual bird mortality at Syncrudeaas reported by the Alberta government and by Syncrude (2008).
2000200120022003 2004 2005 20062007
aCombined Mildred Lake and Aurora leases.
bDifference 5 (Government 2 Syncrude/Government) 3 100; mean difference without regard to sign 5 19%.
THE WILSON JOURNAL OF ORNITHOLOGY N Vol. 122, No. 3, September 2010
containing bitumen, polycyclic aromatic hydro-
carbons (PAHs), naphthenic acids, brine, heavy
metals, and ammonia along an internationally
significant migratory bird corridor poses long-
term threats to migratory and resident birds
(Schick and Ambrock 1974, Wells et al. 2008).
Tailings ponds may pose the greatest threat in
spring when warm effluent-fed tailings ponds
provide open water at a time when natural water
bodies remain frozen; however, a high risk of
oiling may extend throughout the open water
season (Van Meer and Arner 1985).
There are four aspects of our estimates that
influence their accuracy. First, no nocturnal
observations of migrating or landing birds were
made. Many birds migrate at night (Richardson
1971, Blokpoel 1973, Blokpoel and Burton 1973),
but landing rates during darkness are unknown.
Many birds migrate at night and mortality rates
might be higher if data from night-time observa-
tions were available. There are also no observa-
tions for November through early April, when
natural water bodies are frozen but large areas of
tailings ponds remain unfrozen due to addition of
warm tailings. The frequency of landings then is
unknown, but is presumably greater than zero for
resident birds. Higher rates of nocturnal landings
and landings in non-migratory periods would
increase our estimates.
A second source of estimation error is that
numbers of birds flying over, landing, becoming
oiled, or being found dead are an unknown
fraction of the true parameter values. Some of
these parameters were estimated from the most
recent and systematically collected data available
(e.g., Ronconi 2006, Ronconi and St. Clair 2006);
however, without data from other sites and years,
there is no means to assess variation in rates of
birds landing, becoming oiled, or dying.
Third, we assumed a mortality rate of 80 to
90% for birds that came in contact with oil.
Mortality rates of oiled birds are unknown (Clark
1984), but even very small amounts of oil may kill
birds (Hartung 1967, McEwan and Koelink 1973)
and survival rates of rehabilitated oiled birds may
be as low as 1 to 20% (Mead 1997). Our estimates
assume a small proportion (10–20%) of oiled
birds may survive oiling.
Finally, by scaling our estimates from individ-
ual tailings ponds to the areal extent of tailings
ponds in the region, we assumed that bird use and
associated mortalities are similar across sites. This
assumption remains untested and, for some
species such as shorebirds, the extent of shoreline
contaminated with bitumen may be a better
predictor of mortality than extent of open water.
Individual events may result in large variations
in mortality. A migratory waterfowl mortality
event at the Syncrude Aurora North tailings pond
occurred in April 2008 at which 1,606 dead
waterfowl were later found (CBC 2008, 2010).
Provided that all dead waterfowl were found and
no non-waterfowl died, the single event resulted
in a mortality of 162 birds/km2well in excess of
our highest estimate (Table 1). The frequency of
mass mortality events is unknown.
Inconsistencies and Deficiencies in Reporting
Bird Mortality.—We note three major shortcom-
ings in the data provided by government and
industry. First, mortality estimates based on
mortality surveys and landing/oiling rates are far
higher than those reported by government.
Second, industry-reported mortalities often do
not match mortalities reported by the government
pond during April–May 2003.a
Rates of landing and the proportion of birds that subsequently become oiled at Shell Albian Sands tailings
birds landing Landings/hr Landings/dayc
Oiled birds Oilings/day
% Landed birds
that became oiled
aData compiled from spring migration studies in the 3.5 km2main tailings pond at Shell Albian Sands, Muskeg River Mine (Ronconi 2006, Ronconi and St. Clair
2006) yielding a low estimate of 34.69 landings/km2/day; there were 54.93 landings/day within the 0.50 km2observation area yielding a high estimate of 109.86
landings/km2/day; 47 days of observation for oiled birds.
bLoons, grebes, cranes, herons, cormorants, and coots.
cObservations for daylight hours only; average 15.97 hrs of daylight between 18 April and 29 May, 134.4 hrs of observation (source: www.almanac.com/rise for
Fort McMurray, AB, Canada).
Timoney and Ronconi N BIRD MORTALITY AT TAILINGS PONDS