Skip booklet index and go to page content

COSEWIC Assessment and Update Status on the Northern Goshawk (2000)

General Biology


Goshawks form a monogamous pair bond and show strong mate (Detrich and Woodbridge 1994) and nesting area fidelity between years (Crocker-Bedford 1990a; Reynolds and Joy 1998). However, about 25% of breeding females breed in different areas one year to the next (P. Kennedy pers. comm.). Although individual females can breed as yearlings, most breeders in a given population are >2 years old (Squires and Reynolds 1997).

Goshawk territorial behaviour has been observed as early as February in the mid-Atlantic states (Speiser and Bosakowski 1991); mid-March in British Columbia (Beebe 1974). On the Queen Charlotte Islands, courtship displays of the Queen Charlotte Goshawk were recorded as early as 6 March (Chytyk et al. 1998) and are thought to commence during the last week of February (Chytyk and Dhanwant 1999). Some pairs may remain near the nest year-round (Doyle and Smith 1994). Recent telemetry work on Vancouver Island suggested that most male Queen Charlotte Goshawks remain on or near the nesting territory year round, while females generally disperse further from the nest site during the winter months (McClaren 1999).

Eggs are generally laid between mid-April and late May (McGowan 1975; Reynolds and Wight 1978; Bull and Hohmann 1994; Iverson et al. 1996); but as early as 7 April has been reported in British Columbia (Campbell et al. 1990). Egg-laying may be delayed during cold, wet springs and at higher elevations (Squires and Reynolds 1997). On Vancouver Island and the Queen Charlotte Islands it is thought that egg laying occurs during late April and early May (E. McClaren pers. comm.; Chytyk and Dhanwant 1999).

The incubation period ranges from 28 to 32 days per egg (Beebe 1974; McGowan 1975; Reynolds and Wight 1978), and begins with the first or second egg laid (Beebe 1974; Squires and Reynolds 1997). Incubation is performed primarily by the female. During this period, males hunt and deliver food to the female (Brown and Amadon 1968), but occasionally incubate (Lee 1981, E. McClaren pers. comm.). Hatching and fledging dates are variable, ranging from 13 May to 25 June, and 25 June to 28 July, respectively (McGowan 1975; Reynolds and Wight 1978; Bull and Hohmann 1994). In British Columbia, the earliest recorded fledging date is 25 June (Dease Lake), and the latest was calculated to be in the last week of August (Campbell et al. 1990).

On Vancouver Island, seven broods of Queen Charlotte Goshawks fledged from early to mid-July 1997 (McClaren 1997). Fledging generally occurred during the first two weeks of July on the Queen Charlotte Islands (Chytyk and Dhanwant 1999). Goshawks have a relatively long post-fledging dependency period that can be several weeks in duration (Kennedy et al. 1994). In southeast Alaska, all juvenile Queen Charlotte Goshawks appeared to disperse from natal areas before 5 September (Titus et al. 1995). Natal dispersal on Vancouver Island and the Queen Charlotte Islands generally occurred during late August (E. McClaren pers. comm.; Chytyk and Dhanwant 1999).

In North America, clutch sizes usually range between two and four eggs with a mean clutch size of 2.7 eggs (Squires and Reynolds 1997). In British Columbia, clutch sizes also range from two to four eggs (Campbell et al. 1990), but three eggs is the norm (Beebe 1974). Nesting success is highly variable but, in most studies, 80-94% of nest attempts produced at least one fledgling (Reynolds and Wight 1978; Bull and Hohmann 1994; Squires and Reynolds 1997). In the Yukon, the average number of fledglings varied from 0/nest to 3.9/nest (Doyle and Smith 1994). On the Olympic Peninsula of Washington, 2.3 young fledged/breeding attempt in 1996, and 2.0 in 1997 (Finn et al. 1998). See Appendix B for additional productivity data.

On Vancouver Island, productivity for 56 nesting attempts between 1994-1998 averaged 1.7 fledglings/nest; with annual rates varying between 1.4 and 2.2 fledglings/nest (McClaren 1999). On the Queen Charlotte Islands, productivity was variable: 1 nest with 2.0 fledglings in 1995; 2 nests averaging 1.5 fledglings/nest in 1996; 1 nest with 1.0 fledgling in 1997; and 1 nest with 0 fledglings in 1998 (Chytyk and Dhanwant 1999). Nest failure followed by a replacement clutch, has been observed (Johnsgard 1990), but is likely rare, as goshawks require the full spring and summer season to nest successfully (Squires and Reynolds 1997).

Variation in goshawk productivity is associated mainly with prey abundance (McGowan 1975; Crocker-Bedford 1990a; Doyle and Smith 1994) and habitat structure for accessibility to prey (Widen 1989; Crocker-Bedford 1990a; Beier and Drennan 1997). There is evidence that nestling survival is directly dependent on food supply due to starvation or siblicide (Estes et al. 1999), but an alternative explanation is that higher food abundance allows adults to remain longer within the nest area, thus decreasing predation of nestlings (Ward and Kennedy 1996). Initiation of breeding is generally dependent on prey availability, the presence of a suitable mate, and the availability of unoccupied suitable nesting habitat (McGowan 1975; Hennessy 1978; Reynolds and Wight 1978; Doyle and Smith 1994; Iverson et al. 1996; Finn 1997). On Vancouver Island, low productivity for Queen Charlotte Goshawks in 1995 was thought to be related to low abundance of Red Squirrel (Tamias hudsonicus), a primary prey species (Ethier In prep).

Weather, age of breeders, nestling predation, adult mortality, disease and human disturbance may also affect productivity (Squires and Reynolds 1997). Siblicide occurs during food shortages (Squires and Reynolds 1997; Estes et al. 1999). Fishers (Martes pennanti) were taking eggs, nestlings, and adult females in a Wisconsin population and an increasing Fisher population was thought to be one cause of reduced goshawk productivity (Erdman et al. 1998). Other predators of goshawk nestlings include Great Horned Owl, and mammals such as Wolverine (Gulo gulo) (Doyle 1995). On the Queen Charlotte Islands, Raccoon (Procyon lotor) was recently suspected of preying upon a Queen Charlotte Goshawk nest (Chytyk and Dhanwant 1999). 


Estimates of mortality rates of adults and fledged juveniles are difficult to make and few data exist for North American populations. Some specific causes of adult mortality are starvation, predation, disease, and direct and indirect killing by humans (Beebe 1974; Snyder and Wiley 1976 in Palmer 1988; Newton 1979; Duncan and Kirk 1995; Squires and Reynolds 1997). 

Long-term survival rates of juveniles and recruitment rates into the breeding population are unknown. European research indicates that mortality is highest in the first year (58-64%) and decreases with age (Newton 1979). In Arizona, estimated annual survival was 87% for females >1 year old and 69% for males >1 year old (Kennedy 1997; Squires and Reynolds 1997). In southeast Alaska, annual survival of adult Queen Charlotte Goshawks was 76%; late winter and early spring was the period of highest mortality (Titus et al. 1995). During the winter of 1998 on Vancouver Island, 7 of 20 radio tagged adult Queen Charlotte Goshawks (4 females and 3 males) died (McClaren 1999). The majority of recovered bodies were badly emaciated and necropsies indicated starvation as the primary cause of death.

Palmer (1988) suggests a maximum life span of 20 years in North America, but provides no supporting evidence. Goshawks in Europe have been known to live up to 19 years, both in the wild and in captivity (Newton 1979).


Throughout most parts of its range, in particular the southern portions, goshawks tend to nest on north-facing slopes that provide cool micro-environments (Reynolds et al. 1982; Squires and Reynolds 1997). On the Queen Charlotte Islands, 5 active Queen Charlotte Goshawk nests were found on warmer, southwest aspects that ranged between 200° and 245° (Chytyk and Dhanwant 1999; Chytyk et al. 1999). This contrasts with nests on Vancouver Island where 9 were on aspects between 1°-90°, 4 were on aspects between 91°-180°, 10 were on aspects between 181°-270°, and 8 were on aspects between 271°-360° (McClaren 1998, 1999). Elsewhere, there is no clear pattern for aspect of goshawk nests: northeast aspects in Kispiox Forest District,northwest British Columbia (Mahon and Franklin 1997) andsoutheast Alaska (Titus et al. 1994); southeast aspects in the Lakes Forest District, northwest British Columbia (Mahon and Doyle 1999); or southerly aspectsin southeast Alaska (McGowan 1975).

On the Queen Charlotte Islands, 4 of 5 active Queen Charlotte Goshawk were found in dead western hemlocks. This contrasts with other nest trees across its range which tend to be live trees (Beebe 1974, Reynolds et al. 1982, Crocker-Bedford and Chaney 1988, Titus et al. 1994, Squires and Reynolds 1997). On Vancouver Island, only 3 of 32 known Queen Charlotte Goshawk nests were located in dead trees (E. McClaren pers. comm.). On the Queen Charlotte Islands, Queen Charlotte Goshawks may select dead trees because they provide direct exposure to solar heat that may help to regulate nest temperatures, since summer months there are generally wet and relatively cool (Chytyk and Dhanwant 1999). Nesting in exposed dead trees may also be explained by the absence of Great Horned Owls breeding on the archipelago (Campbell et al. 1990). Great Horned Owls are known to prey on young and adult goshawks and other raptors (Squires and Reynolds 1997).

Heavy rainfall is thought to impact reproductive success of goshawks (Penteriani 1997), and may be an important factor in coastal British Columbia given its relatively high rainfall. During a ten year study in the Mediterranean, Penteriani (1997) found that cold, wet springs delayed nest initiation, and heavy levels of annual rainfall between 60-120 mm corresponded with years that had the highest incidence of nest failure. April and May were suggested as the most critical months when weather could affect nest productivity. Other studies also stress the negative impact that precipitation levels, particularly in spring, have on nest productivity for goshawks and other raptor species (Kostrzewa and Kostrzewa 1990).

On the Queen Charlotte Islands, the average annual precipitation is 1359 mm at Sandspit (ne corner Moresby Island) (Environment Canada 2000), a full magnitude greater than the threshold limit in the Mediterranean. In Sandspit, March, April and May have on average 104 mm, 95 mm and 62 mm of precipitation per month respectively. On Vancouver Island, average annual precipitation levels vary from 857 mm in Victoria to 3295 mm in Tofino (Environment Canada 2000). These relatively high precipitation levels, when compared to other regions such as interior North America or the Mediterranean, suggest the Queen Charlotte Goshawk has adapted to wetter environments. However, high rainfall in coastal British Columbia may have an adverse affect on Queen Charlotte Goshawk productivity.


Goshawks are a nomadic species but are probably resident year-round in most years throughout most of its range. Migration, when it occurs, is linked to food shortages (Squires and Reynolds 1997). Residency appeared to be typical for the Queen Charlotte Goshawk in southeast Alaska (Crocker-Bedford 1994; Titus et al. 1994; ADFG 1996). Adults dispersed variable distances from their nesting areas after breeding; some used overlapping summer and winter ranges whereas others dispersed as far as 90 km from their nest area for the winter (ADFG 1996). On Vancouver Island and Queen Charlotte Islands, the Queen Charlotte Goshawk is almost certainly resident (Taverner 1940; Beebe 1974; Campbell et al. 1990). For example, on Vancouver Island, recent data from 17 radio-tagged birds showed a maximum movement from nest sites of >100 km from July through March (E. McClaren unpubl. data). In addition, several radio-tagged females were “lost” during the winter months, suggesting that they may have dispersed much further during winter.

Juvenile residency was well-documented in southeast Alaska when 23 of 27 independent juveniles dispersed an average of 60 km from their natal site; then remained fairly consistently in winter-use areas (ADFG 1996). However, juveniles usually must disperse from their natal areas to find unoccupied habitat and, therefore, tend to disperse greater distances than adults. In another study in southeast Alaska, 14 radio-tagged juvenile Queen Charlotte Goshawks dispersed from nest areas 5 to 7 weeks (August 5 to September 5) after fledging, and were tracked to distances ranging from 16 to 151 km from their natal sites (Titus et al. 1994). 


Goshawks prey on a wide range of small to medium-sized mammals and birds depending on season and region. Geography and variation in prey fauna available in different forest types explain much of the variation in local goshawk diets (R. Reynolds pers. comm.). During the nesting season, mammals were taken more often in interior Alaska, Arizona, Nevada, Utah, and Yukon; birds were taken more often in southeast Alaska, California, New Mexico, and Oregon (Titus et al. 1994; Squires and Reynolds 1997). In Washington, coastal goshawks took more birds (53%) than mammals, compared to interior goshawks (47%), with squirrels, grouse, and Snowshoe Hares (Lepus americanus), being the main prey (Watson et al. 1998). Common prey species include tree squirrels, ground squirrels, rabbits, Snowshoe Hare, woodpeckers, grouse, corvids, and various large songbirds (Squires and Reynolds 1997). Goshawks occasionally use carrion (Squires 1995).

In southeast Alaska, the most common prey of Queen Charlotte Goshawks were Steller’s Jay (Cyanocitta stelleri), Blue Grouse, Spruce Grouse (Dendragapus canadensis), Varied Thrush (Ixoreus naevius), Red Squirrels, and woodpeckers. Smaller numbers of Sharp-shinned Hawk (Accipiter striatus), alcids, yellowlegs, ptarmigan, and Northwestern Crow (Corvus caurinus) were taken (Titus et al. 1994). In southeastern Alaska, of the ten most common prey species for Queen Charlotte Goshawks, none were expected to benefit from clearcut logging and most were likely to decline (Iverson et al. 1996); although Blue Grouse were likely to increase during the earliest seral stages.

On the British Columbia coast, Beebe (1974) thought the Queen Charlotte Goshawk preyed mainly on Steller’s Jays and Varied Thrushes on Vancouver Island; on the Queen Charlotte Islands, he thought they take mainly Northwestern Crows. However, Chytyk and Dhanwant (1997) found Red Squirrel, Red-breasted Sapsucker (Sphyrapicus ruber), Blue Grouse, Varied Thrush, and Hermit Thrush (Catharus guttatus) to be important prey for pairs nesting in inland parts of the Queen Charlotte Islands, and found no evidence of predation on crows. A possible explanation may be that the Northwestern Crow occurs most abundantly near marine coasts and can become rare even a few kilometres inland (Campbell et al. 1997), and Beebe spent most of his field time along the coast. An analysis of 44 pellets collected from the base of nest trees on the Queen Charlotte Islands in 1996 showed that Red Squirrel (44%) and various songbirds (47%) were the major prey of nesting pairs during the breeding season (Roberts 1997).

Recent evidence on Vancouver Island, suggested Red Squirrels may be the most important prey species, at least during early parts of the breeding season (McClaren 1997; T. Ethier pers. comm.); other prey included Varied Thrush, Northern Flicker (Colaptes auratus), Red-breasted Sapsucker, Marbled Murrelets (Brachyramphus marmoratus) (T. Ethier pers. comm.), and bats (J. Deal pers. comm.).


Over the past century, population declines of some subspecies of goshawks have been reported, with habitat loss through logging identified as the primary cause (Widen 1997). However, if taller, older forests are not available, goshawks are capable, to some degree, of using younger and denser forests (Doyle and Smith 1994; Bosakowski and Vaughn 1996; Bosakowski et al. 1999). But there are limits to this adaptability; for example, the minimum age of stands used for nesting was 40 years in the Cascade Mountains of western Washington (Bosakowski et al. 1999), and 75 years in an Idaho Douglas-fir forest (Lilieholm et al. 1994). These younger aged forests are likely the lower age limit of forests that can provide trees with the structural capability to hold goshawk nests.

On Vancouver Island, the Queen Charlotte Goshawk is known to nest in second-growth stands that are >50 years old (E. McClaren pers. comm.). Nests in second-growth stands had similar productivity (1.9 fledglings/active nest for 14 nests) to nests in old-growth forests (1.7 fledglings/active nest for 35 nests) (McClaren 1999). Reoccupancy rates were similar between old-growth and second-growth stands (E. McClaren pers. comm.).

Goshawks are, at times, sensitive to disturbance at or near the nest, and may abandon a nest during incubation or the nestling period, if disturbed by industrial activity, or other human presence (Speiser and Bosakowski 1987; Reynolds 1989; Speiser 1992; Boal and Mannan 1994; Squires and Reynolds 1997). Data from eastern North America, showed that goshawks nest farther from human habitations and paved roads than random sites (Bosakowski and Speiser 1994). However, goshawks can habituate to routine human activities at some distance from the nest. For example, in Arizona, noise from logging trucks that passed by approximately 500 m from 2 active nests elicited no discernible response from a brooding adult female or a lone juvenile (Grubb et al. 1998). In the Cariboo region of central British Columbia, one nest successfully fledged young in 1996, even though logging and road-building occurred nearby during the nesting season (Bosakowski and Rithaler 1997). On Vancouver Island, 2 active Queen Charlotte Goshawk nests occurred within 200 m of an active logging mainline road. Another nest produced 2 fledglings even when tree felling and yarding occurred within 75 m of the nest during the early to mid nesting season (D. Doyle pers. comm.), and a nest within 200 m of a heli-logging operation during the late incubation and nestling phase fledged 1 young (E. McClaren pers. comm.).

It is likely that the risk of human activity that causes nest abandonment or affects productivity is a function of timing in breeding chronology (declining risk as the nesting season progresses), distance (increasing risk as distance shortens), and intensity of activity (increasing risk with increasing intensity).