A RETROSPECTIVE SUMMARY OF RAPTOR MORTALITY IN ONTARIO, CANADA (1991–2014), INCLUDING THE EFFECTS OF WEST NILE VIRUS Open Access

Raptors are atop the trophic hierarchy within many ecosystems and thus contribute to maintenance of biodiversity and ecosystem health (Sergio et al. 2006; Andery et al. 2013). The complex ecological relationship of raptors with their environment underscores the importance of recognizing potential health threats to these birds. For example, in parts of North America, populations of some raptor species, such as the American Kestrel (Falco sparverius), have recently declined in number for unknown reasons (Cadman et al. 2007). Health threats to raptors and other wildlife may be associated with many factors, such as habitat loss, resource depletion, climate change, environmental toxins, and infectious agents; these threats may involve increased interactions with humans and human-altered environments (Gottdenker et al. 2008). Trauma has been identified as the most common cause of mortality in raptors in the mainland US, the state of Hawaii, Brazil, and Spain (Work and Hale 1996; Wendell et al. 2002; Andery et al. 2013; Montesdeoca et al. 2016). In addition, West Nile virus (WNV; family Flaviviridae, genus Flavivirus) has adversely affected raptors in the US and Canada, although population impacts are difficult to document (Gancz et al. 2004; Nemeth et al. 2009).

The province of Ontario, Canada, comprises northern latitudes vulnerable to climatic and landscape changes, in part because of growing human populations. Published data on mortality causes in raptors are sparse for this region, which is home to 25 species of breeding raptors (Cadman et al. 2007). West Nile virus was first detected in Ontario in 2001 (Drebot et al. 2003) and subsequently caused high levels of morbidity and mortality among a large captive population of native northern owl species (Gancz et al. 2004). Long-term data sets (e.g., from wildlife diagnostic centers) allow for more continuous monitoring of free-ranging raptors and other wildlife, increase the awareness of threats to wildlife health, and provide insight into temporal, regional, and taxonomic mortality patterns.

We documented diagnosed causes of raptor mortality in Ontario over a 23-yr period, encompassing the time of the introduction of WNV to the province. Our study objectives were to 1) identify common and uncommon causes of mortality diagnosed among Ontario raptor species submitted to the Canadian Wildlife Health Cooperative (CWHC) from 1991 to 2014; 2) assess demographic patterns in common causes of mortality, including WNV, among all raptors, as well as among the most commonly submitted species, the Red-tailed Hawk (Buteo jamaicensis) and Great Horned Owl (Bubo virginianus); and 3) investigate potential temporal and seasonal patterns for various causes of mortality among all raptors and the most commonly submitted species. Our main goal was to help to identify potential regional risks to raptor health in Ontario, especially those that may be mitigated through natural resource management and public outreach strategies.

Raptor carcasses were received for postmortem examination by the Ontario/Nunavut node of the CWHC between January 1991 and December 2014. Carcasses were submitted by various sources, including rehabilitation centers, veterinary clinics, private citizens, and government agencies (e.g., the Ontario Ministry of Natural Resources and Forestry and Canadian Wildlife Service). Submitters provided the clinical history, date of death, and collection location of the carcass or live bird. Raptors from rehabilitation centers or veterinary clinics were limited to those that died due to presumed natural causes or were euthanatized soon after admission. Based on the date of death, cases were categorized by season: spring (21 March–20 June), summer (21 June–20 September), fall (21 September–20 December), and winter (21 December–20 March). Species, sex, and age (i.e., immature <1 yr old and adult ≥1 yr old as determined by plumage characteristics) were recorded upon postmortem examination, which included gross and microscopic evaluation and ancillary diagnostic testing on a case-by-case basis to arrive at a diagnosis (e.g., immunohistochemistry, histochemical stains or reactions, bacterial or fungal culture, virus isolation, PCR, and gas-chromatography mass spectrometry). Tissue processing for histopathology and ancillary tests was performed by standard protocols developed at the Animal Health Laboratory, an American Association of Veterinary Laboratory Diagnosticians–accredited laboratory at the University of Guelph. A diagnosis of WNV was made based on histopathology in conjunction with positive immunoreactivity by immunohistochemistry or PCR-positive tissue. Aspergillus sp. infections were diagnosed based on fungal morphology and angioinvasion; parasites were evaluated microscopically and generally not identified to species. Toxicoses were diagnosed based on lesions and avian- or species-specific toxin levels in target tissues (e.g., liver, kidney, proventricular or ventricular contents), as determined by laboratory-derived reference values. Mercury toxicosis was diagnosed by renal and hepatic levels in excess of expected environmental levels concurrent with brain and renal lesions (e.g., vascular fibrinoid necrosis and tubular epithelial edema and protein accumulation, respectively).

The primary and contributing causes of mortality and incidental findings were collectively determined based on case history, gross pathology, histopathology, and ancillary diagnostic test results. The primary causes of mortality were broadly differentiated as noninfectious, infectious, or unknown. Noninfectious causes consisted of emaciation, toxicosis, trauma, and unknown. Trauma was further subcategorized as collision (based on surrounding circumstances and corroborating gross lesions such as motor vehicle, building, trees, power line, fence, and wind turbine), electrocution, gunshot, predation, trap, and trauma of unknown cause. Infectious causes were subcategorized as bacterial, fungal, WNV, other viral (i.e., not WNV), parasitic, and suspect infectious (i.e., etiology undetermined or unconfirmed). Culture and PCR were performed to confirm etiologic agents when possible, but in other cases, diagnosis was based on histopathology suggestive of an infectious cause (e.g., cellular composition of inflammation, lesion distribution, intralesional evidence of organisms [e.g., bacteria, fungi, or viral inclusions] with concurrent exclusion of other potential causes of death).

Statistical analyses were conducted using Intercooled STATA14® (StataCorp, College Station, Texas, USA) at a significance level of α=0.05. Univariable logistic regression models were fitted to examine associations between the odds of being diagnosed with select causes of mortality (e.g., all infectious, WNV, trauma, emaciation) and the following temporal, demographic, and taxonomic variables: season, time period before and after the introduction of WNV (1991–2001 and 2002–14, respectively), taxonomic family, age, and sex. Furthermore, using only cases submitted post-WNV introduction, univariable logistic regression models were used to investigate associations between WNV diagnosis and the age and sex of all raptors. Among raptors diagnosed with WNV, exact logistic regression models were fitted to examine for potential associations with season and taxonomic order (i.e., species within the order Accipitriformes versus all others) and age and sex in Red-tailed Hawks and Great Horned Owls. The scoring method was used to calculate P-values (Dohoo et al. 2009). From these models, the odds ratio (OR) and 95% confidence interval (CI) were reported.

Raptor submissions (n=1,448) encompassed 29 species within four orders (Accipitriformes, 12 species; Cathartiformes, 1 species; Falconiformes, 4 species; Strigiformes, 12 species), among which Red-tailed Hawks and Great Horned Owls were the most common (Table 1). Sexes were equally represented, and the majority of birds were adults (Table 2).

The range of the number of raptors received annually by the CWHC-Ontario/Nunavut node was 17 (in 1991) to 109 (in 2002), with high numbers also submitted in 2004 and 2001 (n=105 and 96, respectively), corresponding to the timing of earliest WNV detections and initiation of provincial WNV surveillance (Drebot et al. 2003). Cases were most commonly submitted in the fall (Table 2). Raptors originated from across Ontario, but the number of submissions was highest in southern Ontario (Fig. 1).

Trauma was the most commonly diagnosed primary cause of death among all raptors submitted (Table 1). Among the 716 raptors diagnosed with trauma as the primary mortality cause, the most common subcategory was collision. The majority of collisions were attributed to vehicles and buildings; the remainder involved trees, power lines, fences, or wind turbines, each accounting for <10% of the total collision sources diagnosed (Table 3). The age breakdown for all raptors diagnosed with trauma as a cause of death was 471/716 adults (67%; 95% CI: 62–69%), 189/716 immature birds (26%; 95% CI: 23–30%), and 56/716 of unknown age (8%; 95% CI: 6–10%). Univariable logistic regression models revealed significantly increased odds of adults being diagnosed with trauma versus immature raptors (OR=0.76, 95% CI: 0.61–0.96, P=0.022).

Emaciation was the second most frequently diagnosed primary cause of death among all raptors (Table 1). The age breakdown for these raptors was 150/241 adults (62%; 95% CI: 56–68%), 81/241 immature birds (34%; 95% CI: 30–40%), and 10/241 of unknown age (4%; 95% CI: 2–8%). The odds of being diagnosed with emaciation did not differ by age class (OR=1.21, 95% CI: 0.90–1.63, P=0.213); however, the odds of emaciation being diagnosed as a primary cause of death in males were increased over females (OR=1.62, 95% CI: 1.20–2.20, P=0.002). The odds of being diagnosed with emaciation did not differ between seasons (spring compared to summer: OR=0.73, 95% CI: 0.49–1.09, P=0.127; to fall: OR=0.78, 95% CI: 0.53–1.16, P=0.221; and to winter: OR=0.99, 95% CI: 0.66–1.50, P=0.968), and the seasonal distribution was 67/241 in fall (27.8%; 95% CI: 22.2–33.9%), 56/241 in spring (23.2%; 95% CI: 18.1–29.1%), 61/241 in summer (25.3%; 95% CI: 19.9–31.3%), and 56/241 in winter (23.2%; 95% CI: 18.1–29.1%) with 1/241 of unknown timing (0.4%; 95% CI: 0.0–2.3%). The remaining mortality causes were less commonly diagnosed and included various other noninfectious as well as infectious causes (Table 1).

In total, 214/1,448 (15%; 95% CI: 13–17%) raptors were diagnosed with infectious disease as the primary cause of death, with 69/214 (32%; 95% CI: 26–39%) of these during the pre-WNV introduction period (i.e., 1991–2001) and 145/214 (68%; 95% CI: 61–74%) diagnosed after the introduction of WNV to Ontario (i.e., between 2002 and 2014). The odds of a raptor being diagnosed with an infectious disease were significantly greater during the post-WNV period as compared to the pre-WNV period (OR=2.02, 95% CI: 1.58–2.57, P=0.000) and in summer and fall compared to spring (summer: OR=2.05, 95% CI: 1.45–2.91, P=0.000; fall: OR=1.90, 95% CI: 1.34–1.69, P=0.000). The odds of being diagnosed with an infectious disease were not significantly different in winter compared to spring (P=0.186) or between taxonomic family (P=0.302 for Cathartidae, P=0.335 for Falconidae, P=0.731 for Pandionidae, P=0.205 for Strigidae), age (P=0.115), or sex (P=0.063; Supplementary Table).

West Nile virus was the most commonly diagnosed cause of infectious disease among all raptor species throughout the study period, followed by bacteria, fungi, parasites, other (non-WNV) virus, and suspect infections of unconfirmed etiology (Table 1). Prior to the arrival of WNV, bacterial infections accounted for the majority (32/69, 46%; 95% CI: 34–59%) of infections diagnosed. Bacteria cultured from tissues of 14/52 raptors diagnosed with bacterial infections (27%; 95% CI: 16–41%) were most commonly Escherichia coli (17/52, 33%; 95% CI: 20–47%) and Staphylococcus spp. (10/52, 19%; 10–33%), followed by Pasteurella spp. (6/52, 12%; 95% CI: 4–23), Pseudomonas spp. (4/52, 8%; 95% CI: 2–19), alpha-hemolytic Streptococcus spp. (3/52, 6%; 95% CI: 1–16), Enterococcus spp. (3/52, 6%; 95% CI: 1–16), Actinobacter spp. (2/52, 4%; 95% CI: 1–13%), Clostridium perfringens (2/52, 4%; 95% CI: 1–13%), Klebsiella pneumoniae (2/52, 4%; 95% CI: 1–13%), Proteus spp. (2/52, 4%; 95% CI: 1–13%), Salmonella spp. (2/52, 4%; 95% CI: 1–13%), and beta-hemolytic Streptococcus spp., Chlamydia spp., and Fusobacterium spp. (each with 1/52, 2%; 95% CI: 0–10). Multiple bacteria species were cultured from tissues from some raptors. Culture was not attempted or was nondiagnostic in 19 cases.

The first confirmed positive WNV case in an Ontario raptor was an American Kestrel on 14 August 2002. Since its introduction to Ontario through 2014 (Table 1), individuals from 13/29 species were diagnosed with WNV, accounting for 53% (77/145; 95% CI: 45–61%) of infectious mortalities. Red-tailed Hawks and Great Horned Owls were the two species most commonly diagnosed with WNV, constituting 27% (21/77; 95% CI: 18–39%) and 23% (18/77; 95% CI: 15–34%) of the total WNV cases, respectively. Sharp-shinned Hawks (Accipiter striatus) made up the next highest percentage of WNV cases with 13% (10/77; 95% CI: 6–23%), followed by American Kestrels and Merlins (Falco columbarius), each making up 7% of WNV cases (5/77; 95% CI: 2–15%). The species-specific percentages of WNV-positive cases following the introduction of WNV was highest for Northern Harriers (Circus cyaneus) with 25% (1/4; 95% CI: 1–81%), Goshawks (Accipiter gentilis) with 23% (3/13; 95% CI: 5–54%), Sharp-shinned Hawks with 19% (10/52; 95% CI: 10–33%), Great Horned Owls with 16% (18/116; 95% CI: 10–23%), and Rough-legged Hawks (Buteo lagopus; 1/7; 95% CI: 0–58%), American Kestrels (5/35; 95% CI: 5–30%), Red-tailed Hawks (21/152; 95% CI: 9–20%), and Merlins (5/38; 95% CI: 4–28%), each with 13–15%. The remaining species each accounted for <10% of observed WNV-associated mortalities.

There was no significant difference in WNV prevalence among taxonomic orders (OR=1.3, 95% CI: 0.79–2.16, P=0.326), age class (OR=1.22, 95% CI: 0.73–2.04, P=0.451), or sex (OR=0.93, 95% CI: 0.57–1.52, P=0.769). For Red-tailed Hawks, the odds of being diagnosed with WNV were significantly higher in immature compared to adult raptors (OR=3.82, 95% CI: 1.31–11.8, P=0.012) but not in either sex (OR=0.52, 95% CI: 0.15–1.63, P=0.324). For Great Horned Owls, neither age nor sex was associated with a WNV diagnosis (OR=0.18, 95% CI: 0.00–1.24, P=0.105 and OR=0.69, 95% CI: 0.20–2.16, P=0.651, respectively).

Based on univariable exact logistic regression, there was a significant seasonal difference in WNV diagnoses for all raptors, and the odds of being diagnosed with WNV in spring (OR=0.02, 95% CI: 0.0–0.12, P<0.001) or winter (OR=0.02, 95% CI: 0.0–0.13, P<0.001) were significantly less likely than in fall. However, the odds of a raptor being diagnosed with WNV were not significantly different in summer (OR=0.89, 95% CI: 0.53–1.50, P=0.709) compared to fall.

A number of mortality causes had taxonomic predilections. For example, toxicosis was most commonly diagnosed in Bald Eagles (Haliaeetus leucocephalus; 23/36, 64%; 95% CI: 46–79%), in which it was the most commonly diagnosed cause of death (23/87, 26%; 95% CI: 18–37%). Lead poisoning was the most common toxicosis in Bald Eagles (14/23, 61%; 95% CI: 39–80%), followed by barbiturates such as pentobarbital (4/23, 17%; 95% CI: 5–39%) and mercury (3/23, 13%; 95% CI: 3–34%). The majority of Bald Eagles (12/14, 86%; 95% CI: 57–98%) with detectable lead tissue levels were submitted between November and May, corresponding to hunting seasons for white-tailed deer, wild turkeys, and other game species in Ontario. Mercury toxicosis was diagnosed in two Bald Eagles and was suspected in a third (based on tissue levels) in which autolysis precluded histopathologic assessment. Four-aminopyridine (used for control of nuisance bird populations) toxicity as a cause or contributing cause of death (e.g., with trauma as primary cause) was diagnosed in seven Peregrine Falcons (Falco peregrinus) and one Merlin.

Lesion patterns specific to Great Horned Owls included myocarditis and myofiber degeneration or atrophy in 27/237 cases (11%; 95% CI: 8–16%). The majority of these cases (17/27, 63%; 95% CI: 42–81%) had concurrent protozoan cysts consistent with Sarcocystis spp. within cardiomyocytes. Three Great Horned Owls (3/237, 1%; 95% CI: 0–4%) had intravascular evidence of hematozoan parasites. Three additional Great Horned Owls (3/237, 1%; 95% CI: 0–4%) and one Barred Owl (Strix nebulosa; 1/41, 2%; 95% CI: 0–13%) had multifocal hepatic and splenic necrosis associated with herpesvirus infection. Lymphoma was diagnosed in two Great Horned Owls (2/237, 1%, 95% CI: 0–3.0%) and was multicentric (i.e., kidney, spleen, and liver) in one owl and was limited to the kidneys in the other. One Great Horned Owl had synovial chondromatosis in the scapulo-humeral joint (1/237, 0%; 95% CI: 0–2%). In addition, three Red-tailed Hawks had renal adenoma (3/308, 1%; 95% CI: 0–3%).

The present study provides a broad perspective stemming from a long-term data set of mortality causes diagnosed among free-ranging raptors in the province of Ontario, Canada. Trauma and emaciation were the most commonly diagnosed primary mortality causes, the former accounting for death in approximately half of the birds examined. These findings are consistent with previous studies of raptors in North and South America and Europe (Deem et al. 1998; Komnenou et al. 2005; Andery et al. 2013) and likely reflect not only the high ecological stakes associated with their role as predatory birds, but also the challenges related to coexistence with humans. Trauma in wildlife is often a direct or indirect effect of human activity (e.g., gunshot, or collisions with motor vehicles, buildings, or other manmade structures). In addition, human populations have undergone recent rapid growth and expansion across much of southern Ontario, thereby increasing the likelihood of human contact with raptors and other wildlife. The proportion of trauma in raptors in the present study may be overrepresented, because many of these deaths were more likely to occur in proximity to humans, and, thus, the carcasses were more likely to be submitted for evaluation. Emaciation, the second most common cause of death in raptors in the present study, may be associated with suboptimal hunting skills, such as with lack of experience (Morishita et al. 1997). However, adult raptors in Ontario were more commonly diagnosed with emaciation than immature birds, suggesting that limited resource availability or access were more likely contributors.

Infectious diseases were much less commonly diagnosed than noninfectious causes of mortality in raptors in Ontario. This may be due in part to a lower likelihood of recovering carcasses of wildlife that underwent a more gradual onset of illness and were therefore more likely hidden from view at death. Alternatively, reduced diagnostic test sensitivity of detecting chronic infections and from carcasses in suboptimal condition due to prolonged postmortem intervals may have contributed to fewer diagnoses of infectious diseases. A coupling of transmission dynamics and natural history may also be at play. For example, horizontally transmitted pathogens may be less common than vertically transmitted or vector-borne pathogens in raptors due to their semisolitary lifestyle (Kaleta 1990). Further, global climate change (i.e., increased year-round temperatures and extended warm seasons) may facilitate the establishment and maintenance of vector populations and the associated threat of vector-borne pathogen transmission in northern latitudes, such as in Ontario (Ogden et al. 2006). Our study revealed significantly greater odds of WNV diagnoses in raptors recovered in summer and fall (versus spring), reflecting seasonal mosquito activity. The ongoing threat of unpredictable introductions of exotic vectors and associated pathogens to new regions further underscores the importance of continued diagnostic evaluations and disease surveillance among wildlife (Drebot et al. 2003). Vector-borne disease studies that incorporate the wildlife hosts' natural history traits (e.g., diet and habitat) and vector biology (e.g., mosquito species diversity, density, and host preferences) will more accurately explain long-term eco-epidemiologic patterns (Hamer et al. 2011).

West Nile virus–associated disease syndromes and infection outcomes vary among raptor species observed in a rehabilitation setting (Joyner et al. 2006; Nemeth et al. 2009) and likely in the wild as well. In our study of 29 species of free-ranging raptors, Red-tailed Hawks and Great Horned Owls were most commonly diagnosed with WNV, together constituting approximately half of all WNV cases. This is consistent with a 2002 study of WNV in raptors across 12 states in the US as well as a 2003 study in Virginia (Joyner et al. 2006; Saito et al. 2007). Species differences in WNV infection rates may be in part attributed to foraging strategies (i.e., target prey) and habitat use, which may in turn affect mosquito-raptor contact rates. Great Horned Owls and Red-tailed Hawks are resource (prey) generalists and may utilize a wider array of habitats than some other species (Cadman et al. 2007). Immature Red-tailed Hawks (but not Great Horned Owls) were nearly four times more likely to be diagnosed with WNV than were adults. Age-related differences in WNV-associated susceptibility or mortality have not been documented in free-ranging birds, which may be due in part to difficulty in determining age (Ellis et al. 2007). However, studies in domestic birds (e.g., geese and sparrows) reported young birds as more susceptible to WNV-associated morbidity and mortality than adults (Austin et al. 2004; Nemeth et al. 2008). The higher likelihood of being diagnosed with WNV-associated mortality in immature versus adult Red-tailed Hawks in the present study may reflect increased susceptibility to adverse effects of infection in younger birds or varied age-related oral infection rates based on prey selection (i.e., more likely to consume sick animals or carrion that may have been infected). Oral infection has been demonstrated experimentally in raptors, although mosquito-borne infection is by far the more likely route (Nemeth et al. 2006).

Toxicoses were uncommonly diagnosed and were somewhat species-specific in our study. Toxicosis was the most common diagnosis in Bald Eagles and was usually attributed to heavy metals (lead and, less commonly, mercury). Lead poisoning, primarily through ingestion, remains an important cause of death in wild birds in Ontario and Saskatchewan, Canada, despite legislation mandating the replacement of lead shot with nontoxic shot (Miller et al. 2001; Martin et al. 2008). Similar to our study, relatively high concentrations of lead in Bald Eagle tissues corresponded temporally (i.e., fall and early winter) with hunting seasons (Martin et al. 2008). Four-aminopyridine, an environmental avicide used to repel nuisance birds, was detected in Peregrine Falcons and a Merlin in the present study. These and other raptor species that preferentially feed on avian prey are at an increased risk of avicide ingestion (Preston and Beane 2009; Holroyd and Bird 2012), which may cause acute mortality as well as predispose to traumatic injuries. The latter were commonly observed in raptors diagnosed with four-aminopyridine toxicity in our study.

Reporting of less common causes of mortality is important to raise awareness of differential diagnoses and potential underlying or future health threats. In our study, a small number of Great Horned Owls had myocardial lesions associated with protozoan cysts consistent with Sarcocystis spp. Sarcocystis falcatula–induced encephalitis and Sarcocystis neurona–induced meningoencephalitis have been diagnosed in a Great Horned Owl and Bald Eagle, respectively; both birds also had mild myocarditis (Olson et al. 2007; Wünschmann et al. 2009). In addition, a Barred Owl and three Great Horned Owls were diagnosed with herpesvirus infections, which may have been acquired via consumption of infected pigeons. Hepatic and splenic necrosis in these owls is consistent with the lesions of fatal herpesvirus infection documented in four Great Horned Owls during two separate mortality events (1999 and 2010) in Calgary, Alberta, Canada (Rose et al. 2012).

Spontaneous neoplasia is rarely documented in raptors and other wildlife, in part due to the relatively short lifespans as well as limited access to specimens for evaluation (Forbes 2000). Neoplasia in the present study was limited to two Great Horned Owls with lymphoma and three Red-tailed Hawks with renal adenoma. Lymphoma is a relatively commonly diagnosed neoplasm in a variety of avian species (Nemeth et al. 2016); a recent case of disseminated lymphoma in a Great Horned Owl in California was believed to be the first report in this species (Malka et al. 2008). Renal adenoma is generally considered incidental (as in our study) but may affect mobility and, therefore, survival if associated with hind limb nerve impingement (Lierz 2003).

Although passive surveillance in wildlife is often limited by biases such as public awareness, human population density, and carcass visibility and recovery, the evaluation of submitted carcasses may provide insight into short- and long-term population health trends as well as potential regional or large-scale threats (Ishak et al. 2008; Gould and Higgs 2009). These trends may be affected by single or prolonged events, such as the emergence of infectious agents, spread of pathogen-carrying vectors, environmental contamination with toxins, climate change, habitat fragmentation and alteration, and growing and expanding human populations. For example, results of our study undoubtedly reflected the increased submissions of bird carcasses after the arrival of WNV due to heightened awareness and vigilance aimed at public health surveillance. In turn, negative population pressures on wildlife may arise from more indirect sources, such as decreased fecundity due to poor nutrition (e.g., inadequate prey base), lack of appropriate nesting habitat, and associated low-grade, stress-induced immunosuppression (Cress and Langley 1988). Wildlife diagnostic evaluations will continue to contribute to population health monitoring, anticipate future health threats, and, thus, inform conservation management strategies (Andery et al. 2013).

We are grateful to the many wildlife and natural resource biologists, rehabilitators, veterinarians, and members of the public for the submission of raptor carcasses to the CWHC. CWHC colleagues Lenny Shirose, David Cristo, and Erin Harkness provided logistical support, and Paul Oesterle provided advice on data analysis and manuscript preparation. This study was supported by the Natural Sciences and Engineering Research Council of Canada.

Supplementary material for this article is online at http://dx.doi.org/10.7589/2017-07-157.