Geoscience Reference
In-Depth Information
polygons fell onto areas of very poor habitat that likely would have been avoided
by BTD. While our use of random placement was necessary to satisfy statistical
requirements, the high degree of variation in habitat quality estimated here may
well have exceeded that in habitats actually selected by deer. Decoupling of statisti-
cally sound sampling requirements from the biology of animals, such as behaviors
associated with foraging and use of cover by BTD (both of which, for example,
would result in avoidance of sparsely vegetated habitats such as rock faces), contin-
ues to be a challenge in wildlife-habitat modeling (e.g. Yoccoz
1991
) and certainly
introduce variation into habitat models not present in resource selection decisions
of the animals themselves.
Despite high variation, however, at least 1 model form in each year was reason-
ably close to deer population estimates (Fig.
2.3
) and all models followed the general
pattern of population increases and declines. Whether, as our models suggest, roads
and cover were of less importance during early successional years (Fig.
2.3
)is
unknown although NCC values were most closely associated with deer numbers
(Table
2.3
). Despite a lack of cover immediately after the eruption, there is evidence
that ungulates made significant use of early successional habitats in the blast zone
(Merrill et al.
1986
), likely because limits on human access, including road closures,
mitigated the lack of security cover available. Early use of the area despite a lack
of security cover likely allowed BTD to achieve a high quality diet, which in turn
allowed for a rapid increase in population despite the lack of cover.
Using such general forage biomass estimates ultimately limits any ability to esti-
mate anything beyond general patterns. Given time lags between changing forage
quality and deer population dynamics (e.g. Gill et al.
1996
;Gilbertetal.
2007
)it
is often the forage conditions in preceding seasons which affect population demo-
graphics (Farmer et al.
1982
). The population peak in 1988 was certainly influenced
by vegetation conditions prior to this time. An abundance of Shrub-Seedling resulted
in the estimates of NCC peaking in 1991, yet may have represented an overesti-
mate of forage quantity. By 1991, some Shrub-Seedling stands were surpassing 10
years of age and likely had canopy cover values exceeding 60% and thus the level
of available forage in these stands was already declining (see Taylor and Johnson
1976
).
Even considering high variation around model estimates, closed canopy forests
clearly began to dominate this landscape within a decade of the 1980 eruption. This
successional trajectory resulted in loss of understory forage and declines in BTD
populations in the MSH area. Such patterns in forest succession are also occurring
across the PNW region in response to state and federal management emphasis on
late-successional (old growth) forests. If management to increase BTD numbers is
to be successful in the PNW, then the maintenance or establishment of early succes-
sional habitats, either by natural or man-induced disturbance, needs to be integrated
into forest management plans at far greater levels than at present. Similarly, appre-
ciation of the deleterious effects of maintaining large areas of even-aged closed
canopy forest on deer and other species dependent upon early-successional habitats
needs increased consideration in forest planning across the PNW.
