 | Ranking Model |
The right side of figure 6.1.2-1 shows a Ranking Suitability map of Hugag habitat. In this instance the individual binary maps are simply added together for a count of the number of acceptable locations. Note that the areas of perfectly acceptable habitat (light grey) on both the binary and ranking suitability maps have the same geographic pattern. However, the unacceptable area on the ranking suitability map contains values 0 through 2 indicating how many acceptable factors occur at each location. The zero value for the area in the northeastern portion of the map identifies very bad conditions (0 + 0 + 0= 0). The example location, on the other hand, is nearly good (1 + 1 + 0= 2).
The ability to report the degree of suitability stimulated a lot of research into using additive colors for the manual process. Physics suggests that if blue, green and red are used as semi-opaque colors on the transparencies, the resulting color combinations should be those indicated in the left side of the table in figure 6.1.2-2. However, in practice the technique collapses to an indistinguishable brownish-purple glob of colors when three or more map layers are overlaid
Computer overlay, on the other hand, can accurately differentiate all possible combinations as shown on the right side of the figure. The trick is characterizing the unacceptable areas on each map layer as a binary progression of values1, 2, 4, 8, 16, 32, etc. In the example in the figure, 1 is assigned to areas that are too high, 2 assigned to areas that are too steep and 4 assigned to areas that are too northerly oriented. Adding a binary progression of values each combination of values results in a unique sum. The result is termed a Ranking Combination suitability map.
In the three-map example, all possible combinations are contained in the number range 0 (best) to 7 (bad) as indicated. A value of 4 can only result from a map layer sequence of 0 + 0 + 4 that corresponds to OK Elevation, OK Steepness and Bad Aspect. If a fourth map was added to the stack, its individual binary map value for unacceptable areas would be 8 and the resulting range of values for the four map-stack would be from 0 (best) to 15 (bad) with a unique value for each combination of habitat conditions.
6.1.3 Rating Model
The binary, ranking and ranking combinations approaches to suitability mapping share a common assumptionthat each habitat criterion can be discretely classified as either acceptable or unacceptable. However, suppose Hugags are complex animals and actually perceive a continuum of preference. For example, their distain for high elevations isnt a sharp boundary at 1800 feet. Rather they might strongly prefer low land areas, venture sometimes into transition elevations but absolutely detest the higher altitudes.
Figure 6.1.3-1 shows a Rating Suitability map of Hugag habitat. In this case the three criteria maps are graded on a scale of 1 (bad) to 9 (best). For example, the elevation-based conditions are calibrated as
- 1 (bad) = > 1800 feet
- 3 (marginal) = 1400-1800 feet
- 5 (OK) = 1250-1400 feet
- 7 (better) = 900-1250 feet
- 9 (best) = 0-900 feet
The other two criteria of slope and orientation are similarly graded on the same scale as shown in the figure
This process is analogous to a professor grading student exams for an overall course grade. Each map layer is like a test, each grid cell is like a student and each map value at a location is like the grade for test. To determine the overall grade for a semester a professor averages the individual test scores. An overall habitat score is calculated in a similar mannertake the average of the three calibrated map layers.
The example location in the example is assigned an average habitat value of 5.67 that is somewhere between OK and Better habitat conditions on the 1 (bad) to 9 (best) suitability scale. The rating was derived by averaging the three calibrated values of
9 9 assigned Best elevation condition of between 0 and 900 feet
7 7 assigned Better steepness condition of between 5 and 15 %
1 1 (northwest) assigned Bad aspect condition
17 / 3 = 5.67 average habitat rating ...slightly better than OK
Note the increased information provided by a rating suitability map. All three of the extended techniques (ranking, ranking combination and rating) consistently identify the completely bad area in the northeast and southeast. What changes with the different techniques is the information about subtle differences in the marginal through better areas. The rating suitability map contains the most information as it uses a consistent scale and reports habitat goodness values to the decimal point.
That brings the discussion full circle and reinforces the concept of a map-ematical framework for analyzing spatial relationships. Generally speaking, map analysis procedures such as Suitability Modeling take full advantage of the digital nature of maps to provide more information than traditional manual techniques supporting better decision-making.
6.2 Decision Support Modeling
In the past, electric transmission line siting required thousands of hours around paper maps, sketching hundreds of possible paths, and then assessing their feasibility by eyeballing the best route. The tools of the trade were a straight edge and professional experience. This manual approach capitalizes on expert interpretation and judgment, but it is often criticized as a closed process that lacks a defendable procedure and fails to engage the perspectives of external stakeholders in what constitutes a preferred route.
6.2.1 Routing Procedure
The use of the Least Cost Path (LCP) procedure for identifying an optimal route based on user-defined criteria has been used extensively in GIS applications for siting linear features and corridors. Whether applications involve movement of elk herds, herds of shoppers, or locating highways, pipelines or electric transmission lines, the procedure is fundamentally the same 1) develop a discrete cost surface that indicates the relative preference for routing at every location in a project area, 2) generate an accumulated cost surface characterizing the optimal connectivity from a starting location (point, line or area) to all other locations based on the intervening relative preferences, and 3) identify the path of least resistance (steepest downhill path) from a desired end location along the accumulated surface.
Figure 6.2.1-1 schematically shows a flowchart of the GIS-based routing procedure for a hypothetical example if siting an electric transmission line that avoids areas that have high housing density, far from roads, near or within sensitive areas and have high visual exposure to houses.
These four criteria are shown as rows in the left portion of the flowchart in the figure. The Base Maps are field collected data such as elevation, sensitive areas, roads and houses. Derived Maps use computer processing to calculate information that is too difficult or even impossible to collect, such as visual exposure, proximity and density. The discrete Preference Maps translate this information into decision criteria. The calibration forms maps that are scaled from 1 (most preferredfavor siting, grey areas) to 9 (least preferredavoid siting, red areas) for each of the decision criteria.
The individual cost maps are combined into a single map by averaging the individual layers. For example, if a grid location is rated 1 in each of the four cost maps, its average is 1 indicating an area strongly preferred for siting. As the average increases for other locations it increasingly encourages routing away from them. If there are areas that are impossible or illegal to cross these locations are identified with a null value that instructs the computer to never traverse these locations under any circumstances
