Contact:

James D. Carroll
jdcarro2@ncsu.edu

+1-919-559-2894

More info:

Refer to the links below to learn more about the use of different ArcMAP tools and resources.

Links:

- Reclassification Tools

- Weighted Overlay Tool

- Fuzzy Membership Tool

 

ESRI Module:

Using Raster Data for Site Selection


The Training Course

The objective of this course is to learn how to use the ArcGIS analysis geoprocess for successful site selection through the understanding of what quality site selection means, and how to use the variety of available tools and methods to produce the best site location results based upon different project criteria and location.

Part One: Site Selection Analysis

The module begins by introducing the multiple types of site selection methods, and then describes what the standard workflow to be followed should be. This beginning introduction is a 3 minute video.

What is site selection?
Site selection is the analysis of suitability by GIS methods to find the best place or site for something. It can be almost anything; a store, school, garden, park, nature area, etc. Raster or vector data can be used for an anslysis.

Ways to perform site selection:
You can do site selection analysis for vector data with a binary analysis, and for raster data use either binary, weighted, or fuzzy logic analysis methods. Examples are provided showing each method.

Site selection analysis workflow:
There is a standard workflow method for for site selection analysis.
1. Define the problem and set the criteria
2. Gather and chose the required data and then make surfaces
3. Use the correct reclassification method
4. Use the correct weight the layers method
5. Create the final output which shows the potential sites meeting the project criteria.

Part Two: Using Weighted Suitability Analysis for Site Selection

The introduction to these methods is by a 3 minute video.
By using a weighted suitability analysis you are able to take the binary to the next step and rank values, usually by using a common scale ofr a raster image of 1 to 5 or 1 to 9 (low is bad, high is good). This section then covers the methods of Reclassification, Weighting Layers, Determine Weights, and when to use the weighted suitability analysis.

Reclassification:
1. Reclassification is the process of replacing input cell values with new output cell values. Reclassification is often used to simplify or change the interpretation of raster data by changing a single value to a new value, or grouping ranges of values into single values. With the reclassification method a new raster is created with the new reclassified values, and when doing a weighted suitability analysis your data should always be reclassified to achieve a successful result.

Weighting Layers:
2. Weighting layers can be a complex endeavor because of the issues involved, you are able to place more importance on some factors and less on others. The decisions about what has a low or high value should be based on the project criteria. The total of the weights must be 100 because it is a percentage. Interestingly the process of assigning weights is usually done be a panel of experts who use the Delphi process.

Determine Weights:
3. Determine weights is the determination based on project criteria of the relative importance of each layer which will be assigned a weight.

When to use weighted suitability analysis
4. When to use weighted suitability analysis is always a question to be asked. It needs to used whenever you need to solve a multicriteria problem.
*Exercise: Perform weighted suitability analysis*
Perform weighted suitability analysis to find suitable sites for a winery and vineyard.

Original Raster Elevation

*Exercise: Refine your analysis*
Refine the site analysis by elevation criteria.

The final suitable sites for the vineyard are in dark blue.

Part Three: Using Fuzzy Logic for Site Selection

The introduction to the fuzzy logic method is a 4 minute video.
When you have a site with discrete boundaries, you can use a binary or weighted suitability analysis; but if you data is for something more complex which does not have set boundaries, such as animal habitat, then there is uncertainty that the analysis can predict precise results. Instead the fuzzy logic method needs to used because it can model inaccurate data with consideration of the inaccuracies.

Fuzzy logic workflow:
This is a more complex method of site selection analysis, because it uses a "continuum of logical values between 0 (completely false) and 1 (completely true). Instead of simply yes and no, fuzzy logic accepts that conditions can be partly true and partly false at the same time".
1. Define the problem and the criteria
2. Collect or make the necessary data
3. Run the Fuzzy Membership Spatial Analyst tool on all the required layers and assign fuzzy membership values
4. Perform a fuzzy overlay the layers process
5. Verify your results and make analysis choices

Choosing a fuzzy membership type:
The membership values range from 0 to 1, with 0 being unlikely and 1 being most likely that a cell is part of a set. The higher the value, the more suitable is the raster cell should be. The tool reclassifies raster cell values on a 0 to 1 scale using an algorithm specified by the choosen membership type. There are six types of Fuzzzy Membership: Linear, Small, Large, MS Small, MS Large, and Near.

Choosing a fuzzy overlay type:
The next step is to use the Fuzzy Overlay Tool to overlay these surface, which is similar to what weighted suitability when used to do weighted overlay on reclassified surfaces. When you use fuzzy logic for site selection, however, you use the Fuzzy Overlay tool, which provides several overlay methods. You choose the one that best meet your requirements. There are five types of Fuzzy Overlay: And, Or, Product, Sum, and Gamma.

When to use fuzzy logic in site selection:
When you need to model complex phenomena that cannot be done with discrete variables, then fuzzy logic should be used. Data which does not have discrete polygons and boundaries are best analyzed with the fuzzy logic method.
*Exercise: Use fuzzy logic to model bald eagle habitat*
Use fuzzy logic tools to model bald eagle habitat near Big Bear Lake in the San Bernardino National Forest.
"The green cells in the map indicate a full membership for distance to water, distance to human disturbance, and tree cover; whereas red cells indicate non-membership. The areas in yellow are more fuzzy and could have bald eagle nests because eagles don't make decisions based on what a map tells them."

Evaluation

I liked this ESRI course and found the methods and tools to be very useful, I have always found the concept of Fuzzy Logic to be an interesting subject. This analysis method is required to perform geoprocessing for many types of projects because of the unknown and changing states. Because I am an architect, I have always interested in learning more about technology for determining better site suitability for projects. This has always been a problem for many projects... Which site should we choose for a specific criteria and use?