Two types of criteria were used in MCE to find the relative suitability of each region. Constraints are hard decisions that limit the analysis to a particular physical area. They are comprised of Boolean images that state whether an area IS or IS NOT suitable for analysis. Conversely, factors define the degree of suitability for the whole study area. MCE results create a range of suitability on a scale of 0 (unsuitable) to 255 (perfectly suitable). This project examined 3 constraints and 6 factors that helped defined areas of highest suitability for a new hospital development in Vancouver.
Land-use in the GVRD dictates what type of development is allowed to
take place in a given area. In order to partake in this evaluation,
an area needed to be situated within the Vancouver boundaries, as well
as outside unsuitable land-use areas. For the purpose of this project
unsuitable land-use included recreation and protected areas, lakes and
water bodies, industrial areas, agricultural areas, as well as transport
routes. Additionally, the new hospital could not be located on top of
major roads, and these were excluded from the analysis.
Total
Population Density: A new hospital would benefit a larger group
of people if it were located in an area of high population density.
Density
of Persons Considered to be “At Risk”: People under
the age of 5, and over the age of 55 have a higher rate of admission to
hospitals than other population groups. The new hospital should therefore
be located near an area where the “At Risk” population is dense.
Distance
from road networks: The most suitable location for a hospital
has to be close to a road network, in order to make emergency access for
patients and ambulances quick and uncomplicated.
Distance
from the edge of Vancouver’s boundary: In order to limit
overlap with municipalities surrounding Vancouver, and maximize the buffer
zone around the hospital, it should be located away from the city’s
edges.
Average
household income: Areas of lower income are often linked to increased
usage of emergency and hospital care. For this reason, the new hospital
should be located in regions of lower average household income.
Distance
from existing hospitals: In order to distribute hospital care
amongst the whole Vancouver population, the new location for a hospital
should be far away from hospitals that currently exist in Vancouver. The
new hospital should not be located within a 1000m radius to an existing
hospital.
Distance
from Land-use: Using the RECLASS module in Idrisi Kilimanjaro,
new values were given to land-use types in order to create a Boolean representation
of the suitable land-use. Areas with the value of 1 were considered suitable,
while areas with the value of 0 were considered unsuitable, creating a
suitable_land layer.
Study
Area:
In order to limit the analysis to the study area, the RECLASS module was
again used to create a Boolean image from the original land-use file.
The resulting image shows the outline of the city of Vancouver (land_bul).
Not
Roads:
In order to exclude the major road networks from analysis, a Boolean image
(from roads_rast) was created using the RECLASS module. This raster layer
was named not_roads_bul. Before this operation could be performed, the
CONCAT function was used enlarge the study area of the roads_rast file.
Distance
from Road Networks:
A Boolean image of the road network was used as an input for the DISTANCE
module, creating a distance surface. Before this operation could be performed,
the CONCAT function was used enlarge the study area of the roads_rast
file. The DISTANCE output (roads_distance) was placed into the Decision
Wizard.
Distance
from Existing Hospitals:
An ASSIGN module was used to create a Boolean image of hospitals in Vancouver.
Values of 0-1 were given a value of 0, while values 1-14 were given a
value of 1. Using the CONCAT module, the study area was enlarged to match
other raster layers. The DISTANCE module was applied to the hospital points.
The ASSIGN module was used to only include values of 1000m or more away
from an existing hospitals.
