3.4. Winter ecology of goshawks

Wintering hawks tended to be female (IV). They also tended to be more site tenacious, contrary to the general idea that females are less site tenacious in raptors and owls (Newton 1979). This has also been stated about goshawks in earlier studies (Sulkava 1964, Höglund 1964, Kenward et al. 1981, Widen 1985, Marcström & Kenward 1981, Halley 1996). To my mind it was reasonable that females had been more numerous in winter, because their food base is larger due to the fact males are unable to hunt hares. In a Swedish study goshawks ate almost barely squirrels in winter in boreal forests (Widen 1987). This probably resulted from a peak year of squirrels during the study years (Andren & Lemnell 1992). Red squirrels are suitable winter prey for goshawks and I also found a preference for squirrels when they were abundant in the first study year (IV). Almost all juveniles that were marked stayed wintering. More than half of them, however, died during the winter, some quite soon after marking. In these cases winter came suddenly with very cold periods and snow storms. It may, however, be possible that some hawks suffered from trapping and marking, although no injuries were found in those birds except one adult male that lost part of its head feathers. The mortality rate of juveniles in their first year has been stated to be very high 40-50%, also in more southern latitudes (Ziesemer 1982, Kenward et al. 1991). Therefore the mortality rate I found matches well with those estimates.

About one third of the marked hawks left the study area relatively soon after marking. Furthermore, one third of the hawks (33%) remained nesting in the study area or nearby, and the rest that wintered in the area disappeared quite suddenly just before the breeding season. Disappearance may also have been caused by battery-failures of the tags. Territory defences were clearly seen from March onwards. Once, a radio-marked adult male with a female tried to intrude in the middle of three territories. The attempt was not successful and the intruders were chased away. My results imply therefore that less than half of the goshawks were local birds. Yet I assume that many of those that moved may have had their own territory somewhere else. Three of these birds were in their third year but three were older. In Gotland, Sweden, where juvenile goshawks also mainly stay over winter, the population model has recently been able to build (Kenward et al. 1991, 1999). Juvenile males had a higher mortality rate and they started breeding at a younger age than females, on average. In the second year 82% of males were breeding but only 25% of females (Kenward et al. 1991). In open populations, like mine, the dynamics might be more complicated. Based on controls of ringed hawks in Oulu (n = 10), seven of them were controlled near Oulu (mean 23 km) and three were controlled abroad (mean 974 km) during the study period. This suggests that most juveniles may not disperse very far (see also Sulkava 1964, Halley 1996). It may, on the contrary, be possible that Oulu region attracts hawks based on controls of ringed hawks from 200 km south of Oulu. Hawks controlled as adults were born on average 71 km from their birth places (n = 5).

Wintering hawks in the study area ranged over areas extending from 2223 ha to 17030 ha. Males tended to have larger areas than females defined by the convex popygon method (Kenward 1987) (9894 km², n=4, and 6484 km², n=11, respectively). Goshawks ranged in similar sized areas than in the boreal forest of central Sweden (Widen 1985). Because the goshawk is considered an old forest species one would expect that size ranges would depend on the amount of old forest in the area (Kenward 1982). It might be deduced that the goshawk needs a certain amount of preferred habitat in its range, when the amount of old forest should be independent of the size range. I found, however, a positive correlation between these variables, though not significant (r_s = 0.446, n.s.). On the other hand, the percentage of old forest in the range was inversely depending on the range size (r_s = -0.604, p = 0.021), which witness for the expectation.

What was the preferred habitat? The goshawk, expectedly, preferred old forest stands, above all, spruce forests but also deciduous forests, mainly birch stands. This is also well documented in other studies (Kenward 1982, Widen 1989, Iverson & al. 1996, Bright-Smith & Mannan 1994, Hargris et al. 1994, Drennan & Beier 1997). Goshawks also used much younger forest stands but clearly avoided open areas. Clear-cuts were an exception. They were rather highly ranked. Because in every location point the surroundings were also considered within a circle with 100 m radius, it often included many different habitat types, yet, less than random points. The high rank of clear-cuts might refer to perching at forest edges, when hawk might get access to species living in forests and those living in clear-cut or plantations. It must be remembered that the satellite-image was from 1987 and the last locations were from 1995 when clear-cut areas were already covered by young trees. It seems that goshawks are fairly well adapted to mosaic forest landscape produced by modern forestry provided that there is suitable prey available. However, forestry has reduced grouse numbers (Kurki et al. 1997), which harms goshawks. On the other hand, mountain hares might have benefited from forestry that creates new growth edible for hares. Female goshawks can kill hares and they constituted more than 70% of the diet by biomass. Males are, however, not able to kill hares. So they have, essentially, a narrower food base. The average hare biomass per km² is roughly 32 kg while that of grouse and squirrels is 18kg and 3kg respectively, excluding capercaillie males, that makes more than 10 kg out of the male’s reach (II, V). Thus, females have 63kg prey/km² and males only 21kg! If grouse decrease still further goshawks will encounter poor wintering possibilities for males, which on the other hand is contrary to their tendency to stay in their territories (Sulkava 1964, Kenward et al. 1981, Halley 1996).