Geoscience Reference
In-Depth Information
of habitat carrying capacity, since the availability of resources during this season
likely establishes the upper limit of population size, particularly during harsh win-
ters (Scharpf
1985
; Gilbert et al.
2007
). The level of forage available at the end of
the growing season is directly related to forest succession and canopy cover and,
unless covered with snow, early successional stages provide the greatest amount of
forage biomass (Scharpf
1985
).
In addition to forage, habitat use can be limited by the availability of security
cover and degree of disturbance from roads (e.g. Witmer et al.
1985
). Disturbance
associated with roads can reduce habitats and can increase disturbance and mortality
in both hunting and non-hunting seasons (Cole and Smith
1983
;Farmeretal.
1982
).
Road disturbance has been estimated to reduce elk habitat use as far as 1.8 km from
the road (Rowland et al.
2000
), though smaller values have been reported for elk in
the Cascades (Witmer and de Calesta
1985
; McCorquodale
2003
) and for mule deer
(
Odocoileus heminous
) (Rost and Bailey
1979
).
2.2.2 Remote Sensing-GIS Based Habitat Evaluations
Remote sensing has been successfully applied in estimating forest structural
attributes such as stand age or canopy cover (Cohen and Spies
1992
; Cohen et al.
1995
; Collins and Woodcock
1996
; Mirik
2005
), and has been used specifically to
estimate understory forage biomass (Davis
1999
; Peek et al.
2002
;Davis
2005
). The
use of remote sensing allows the mapping of landcover over large extents which can
be an invaluable resource for managers. However, even given estimates of forest
seral stage, forage biomass, cover vegetation, and road density it is unlikely that
simply aggregating overall values will be sufficient to predict animal population
responses (e.g. see Hobbs
2003
). Like many animals, ungulates make habitat selec-
tion decisions at multiple spatial scales (Senft et al.
1987
, Boyce et al.
2003
) and
while management often occurs at landscape scales, for evaluation it is important
to consider scale(s) appropriate to the ecology of foraging animals (i.e. Roloff and
Kernohan
1999
).
Although widely applied, geospatial data modeling can introduce potential error
and uncertainty particularly when multiple data layers are used (see Fassnacht et al.
2006
; Van Niel and Austin
2007
). For managers, it is ultimately the degree of cer-
tainty in output map estimates which is perhaps of most use for decision making.
The integration of GIS and remote sensing tools is essential in the analysis of habi-
tat at multi-spatial and multi-temporal scales, yet issues of error and uncertainty are
often not addressed. This is generally due to a lack of field data but, to the extent
possible, these issues should be considered in geospatial modeling efforts.
2.3 Methods
To develop landcover classifications we used 4 Landsat TM and 2 ETM images from
June 17, 1984, August 31, 1988, July 7, 1991, August 21, 1996, September 7, 1999,
and August 18, 2002, respectively. Image date was based upon growing season,
