29.3 Methods
pended together into seamless grids or coverages
for the entire study area.
29.3.2 Mapping Intensity of Human Use
Human Population Density
To create a model of human population density, the
number of people living in each 30-m 2 grid cell
was calculated by dividing the number of people
assigned to each of the cover-type regions in each
census block by the number of 30-m2 cells in each
region. In other words, each grid cell in a region
was assigned an average population density for the
entire region. To estimate the broader impact of human residences on surrounding lands, we used the
FOCALSUM function in ARCIINFO (ESRI, 1998)
to calculate the number of people living within a
l-km radius of each 30-m2 cell. The l-km search
radius estimates broader spatial effects; it was selected arbitrarily to describe an inverse relationship
between human use and distance from homes,
roads, or trails. A different search radius could be
used if warranted either by empirical data or different management questions.
The resulting grid was then resampled to 90-m2
resolution to match the terrestrial vertebrate data
with which it was to be compared (described later).
This output provides a relative measure of human
occupancy and use. Use is meant in a general sense;
the output model indicates areas likely to be used
by people, but does not point to different types of
use.
Road and Trail Density
A weighted density of roads and trails was mapped
within a I-km radius around each 30-m 2 cell
using the LINEDENSITY function in ARCIINFO.
Weights were assigned according to feature type (3
for improved roads, 2 for unimproved roads, and 1
for trails) to account for differential ease of access.
Rather than count all roads and trails equally, when
we know that some receive far heavier use, the
3-2-1 weighting was chosen as the simplest, albeit
arbitrary, measure of impact. Density was calculated in units of length per unit area. This output
grid was similarly resampled to 90-m2 resolution.
Human Use Intensity
A model of intensity of human use was then created that combined the resampled population and
road-trail density grids. In this case, they were
weighted equally by simply adding the two together, recognizing that roads, trails, and local residences all have impacts, but lacking more detailed
439
evidence on how to rate these highly variable impacts. Because the units differed between these two
grids (numeric count versus length of lines), each
of the input density grids had to be scaled to the
same range of units, 0 to 100, before they could be
combined (output intensity model = population
grid + road and trail grid). The output grid (Figure
29.3a) shows both local intensity from human settlement and the much broader spatial patterns from
roads and trails; it also clearly identifies lands that
are > 1 km from either.
29.3.3 Mapping Distributions of
Terrestrial and Aquatic
Vertebrates
Species of Special Concern
We use the term species of special concern to mean
taxa that are rare, endemic, disjunct, threatened or
endangered, either throughout their range or just in
Montana, following the definition used by the Montana Natural Heritage Program. These species typically are accorded special management attention,
so they offer a reasonable starting point for assessing human impacts on wildlife. Predicted habitat was modeled at a 90-m2 cell size for each
species as part of the Montana Gap Analysis Project (Redmond et aI., 1998); then the outputs were
overlaid to create a richness map for the Montana
portion of the study area (Figure 29.3b). This was
aggregated to four richness classes (0 plus three
area-based classes) and then intersected with a human use intensity grid, also broken down to the
same four classes (Figure 29.3c). Both species richness and human use intensity were aggregated to
four classes so that individual classes could be distinguished on this two-variable map.
Bull Trout
To create a map of bull trout (Salve linus confluentus) distribution for the study area, we used
three inputs: (1) streams and other hydrographic
features, (2) bull trout distribution data, and (3)
watershed boundaries (6th-code hydrologic units,
or HUCs; Seaber et aI., 1987). Hydrography coverages (1 : 100,000 scale, vector format) were acquired from the USGS, National Hydrography
Dataset, Pacific Northwest StreamNet River Reach
Files (http://www.strearnnet.org/pnwrhome.htrnl) and
clipped to the study area boundary. Stream orders
(Strahler, 1957) were then assigned within ARC/
INFO, and a new coverage was created that contained only higher-order streams (codes 2::2). A database of fish status, some known, some predicted,
pended together into seamless grids or coverages
for the entire study area.
29.3.2 Mapping Intensity of Human Use
Human Population Density
To create a model of human population density, the
number of people living in each 30-m 2 grid cell
was calculated by dividing the number of people
assigned to each of the cover-type regions in each
census block by the number of 30-m2 cells in each
region. In other words, each grid cell in a region
was assigned an average population density for the
entire region. To estimate the broader impact of human residences on surrounding lands, we used the
FOCALSUM function in ARCIINFO (ESRI, 1998)
to calculate the number of people living within a
l-km radius of each 30-m2 cell. The l-km search
radius estimates broader spatial effects; it was selected arbitrarily to describe an inverse relationship
between human use and distance from homes,
roads, or trails. A different search radius could be
used if warranted either by empirical data or different management questions.
The resulting grid was then resampled to 90-m2
resolution to match the terrestrial vertebrate data
with which it was to be compared (described later).
This output provides a relative measure of human
occupancy and use. Use is meant in a general sense;
the output model indicates areas likely to be used
by people, but does not point to different types of
use.
Road and Trail Density
A weighted density of roads and trails was mapped
within a I-km radius around each 30-m 2 cell
using the LINEDENSITY function in ARCIINFO.
Weights were assigned according to feature type (3
for improved roads, 2 for unimproved roads, and 1
for trails) to account for differential ease of access.
Rather than count all roads and trails equally, when
we know that some receive far heavier use, the
3-2-1 weighting was chosen as the simplest, albeit
arbitrary, measure of impact. Density was calculated in units of length per unit area. This output
grid was similarly resampled to 90-m2 resolution.
Human Use Intensity
A model of intensity of human use was then created that combined the resampled population and
road-trail density grids. In this case, they were
weighted equally by simply adding the two together, recognizing that roads, trails, and local residences all have impacts, but lacking more detailed
439
evidence on how to rate these highly variable impacts. Because the units differed between these two
grids (numeric count versus length of lines), each
of the input density grids had to be scaled to the
same range of units, 0 to 100, before they could be
combined (output intensity model = population
grid + road and trail grid). The output grid (Figure
29.3a) shows both local intensity from human settlement and the much broader spatial patterns from
roads and trails; it also clearly identifies lands that
are > 1 km from either.
29.3.3 Mapping Distributions of
Terrestrial and Aquatic
Vertebrates
Species of Special Concern
We use the term species of special concern to mean
taxa that are rare, endemic, disjunct, threatened or
endangered, either throughout their range or just in
Montana, following the definition used by the Montana Natural Heritage Program. These species typically are accorded special management attention,
so they offer a reasonable starting point for assessing human impacts on wildlife. Predicted habitat was modeled at a 90-m2 cell size for each
species as part of the Montana Gap Analysis Project (Redmond et aI., 1998); then the outputs were
overlaid to create a richness map for the Montana
portion of the study area (Figure 29.3b). This was
aggregated to four richness classes (0 plus three
area-based classes) and then intersected with a human use intensity grid, also broken down to the
same four classes (Figure 29.3c). Both species richness and human use intensity were aggregated to
four classes so that individual classes could be distinguished on this two-variable map.
Bull Trout
To create a map of bull trout (Salve linus confluentus) distribution for the study area, we used
three inputs: (1) streams and other hydrographic
features, (2) bull trout distribution data, and (3)
watershed boundaries (6th-code hydrologic units,
or HUCs; Seaber et aI., 1987). Hydrography coverages (1 : 100,000 scale, vector format) were acquired from the USGS, National Hydrography
Dataset, Pacific Northwest StreamNet River Reach
Files (http://www.strearnnet.org/pnwrhome.htrnl) and
clipped to the study area boundary. Stream orders
(Strahler, 1957) were then assigned within ARC/
INFO, and a new coverage was created that contained only higher-order streams (codes 2::2). A database of fish status, some known, some predicted,