GEO242.Lab13.SuitabilityAnalysis

You will assess areas suitable for a new park meeting three criteria using raster overlay. Suppose that new parks should be located in areas that are (a) of high density of population with age under 18 (b) far from existing parks (c) close to school.

1. Prepare data

• Download data. Extract the zip file, and copy all extracted files to a working folder (e.g., E:\GEO242\Lab13)
• Open ESRI Map Document SuitabilityAnalysis.mxd (double-click the file)

2. Make choropleth maps

• Make a thematic map showing population under age 18 from a layer Census blkgrp (block group): to make this map, you first need to add a new field AgeUnder18, and populate the field with [AGE_UNDER5] + [AGE_5_17].
• Make a thematic map showing AgeUnder18 per acre: to make this map, you first need to (a) add a new field ACRE, and populate the field using Calculate Geometry (b) add a new field KidsPerAcr, and populate the field with [AgeUnder18] / [ACRE]. Click here for results.

3. Make a kernel density map

An alternative way to visualize density is Kernel density estimation (KDE). Input data of KDE should be point data, therefore you need to convert census block group (polygon) to centroid.

• Create centroid from census block group (polygon) layer: ArcToolbox > Data Management Tools > Features > Feature To Point > Choose appropriate parameters in Feature to Point dialog box. If you keep getting an error (empty output data), just add an existing data bgctr1 to the map from the working folder.
• Create a Kernel density raster dataset named KidsDensity from census block group centroid weighted by KidsPerAcre: ArcToolbox > Spatial Analyst Tools > Density > Kernel Density > Choose appropriate parameters in Kernel Density dialog box. If you have a warning "Tool Not Licensed", it's because an Extension Spatial Analyst is not activated. To activate Spatial Analyst, Customize > Extensions > Check Spatial Analyst
• Turn off layers bgctr and Census blkgrp to see the result of KDE. The KDE map should look similar to kids population density map. Click here for results.

4. Create distance raster datasets

• Create a raster layer of distance to parks: ArcToolbox > Spatial Analyst Tools > Distance > Euclidean Distance > Choose appropriate parameters in Euclidean Distance dialog box.
• Create a raster layer of distance to schools: ArcToolbox > Spatial Analyst Tools > Distance > Euclidean Distance > Choose appropriate parameters in Euclidean Distance dialog box.
• Save the map as SuitabilityAnalysis_InputCriteria.mxd

5. Screen areas suitable for a new park using binary criteria (or dominance rule)

Suppose that suitable areas for a new park should meet the following requirements: (a) kids density is greater than mean (77) + one standard deviation (74) (a) distance to parks > 3000 feet (b) distance to school < 1000 feet

• Determine areas that meet the criterion (a): ArcToolbox > Spatial Analyst Tools > Map Algebra > Raster Calculator > Enter the logical expression "KidsDensity" > 151 in the Raster Calculator; this will create calculation results added to the Table of Contents. Areas whose kids density is greater than mean plus standard deviation will have value 1.
• Determine areas that meet the criterion (b): ArcToolbox > Spatial Analyst Tools > Map Algebra > Raster Calculator > Enter the logical expression "Dist2Park" > 3000 in the Raster Calculator; Areas whose distance to parks is greater than 3000 feet (far from parks) will have value 1. 
• Determine areas that meet the criterion (c): ArcToolbox > Spatial Analyst Tools > Map Algebra > Raster Calculator > Enter the logical expression "Dist2School" < 1000 in the Raster Calculator; Areas whose distance to school is less than 1000 feet (close to school) will have value 1.
• Determine areas that meet all of three criteria above: ArcToolbox > Spatial Analyst Tools > Overlay > Weighted Overlay > Add three output data created from Raster Calculator above (that is, highkidsdens, farfrompark, and closetoschool), and map 0 to NoData, 1 to 1 in the Weighted Overlay dialog box; Results will show only suitable areas (value 1), that is areas that meet all of three binary criteria. To see results clearly, turn off source layers.
• Save the map as SuitabilityAnalysis_Dominance.mxd

6. Assess suitability for a new park using ordinal criteria (or contributory rule)

Suppose that you want to assign composite suitability scores to any areas given the following criteria: (a) the higher kids density the better (b) the farther from parks the better (c) the closer to schools the better.

Because units are different among datasets (especially density and distance), it is necessary to make units comparable. One way of doing this is to reclassify data values (of density and distance) to ordinal scale like 0 to 10 where higher scores means more suitable.

• Reclassify kids density to ordinal scores in 0-10 scale:  ArcToolbox > Spatial Analyst Tools > Reclass > Reclassify > Reclassify KidsDensity Value into 10 classes using natural break where the lowest class is assigned to 1 (the lower the density, the less suitable), and the highest class is assigned to 10 (the higher the density, the more suitable) in the Reclassify dialog box. Change the symbology of a layer reclass_kids so that the scale from least suitable to the most suitable is color coded to graduated scheme from green to red.
• Reclassify distance to parks to ordinal scores in 0-10 scale: ArcToolbox > Spatial Analyst Tools > Reclass > Classify Dist2Park into 10 classes using natural break where the lowest class is assigned to 1 (the closer to parks, the less suitable), and the highest class is assigned to 10 (the farther from parks, the more suitable) in the Reclassify dialog box. Change the symbology to green-red scheme.
• Reclassify distance to schools to ordinal scores in 0-10 scale: ArcToolbox > Spatial Analyst Tools > Reclass > Classify Dist2School into 10 classes using natural break where the lowest class is assigned to 10 (the closer to schools or low distance, the more suitable), and the highest class is assigned to 1 (the farther from schools, the less suitable) in the Reclassify dialog box. Change the symbology to green-red scheme.

Let's assume that three three criteria carry the same weight, thus the composite suitability scores will be calculated as ordinal score of (a)*.33 + ordinal score of (b)*.33 + ordinal score of (c)*.34 (that is weighted sum).

• Calculate composite suitability scores: ArcToolbox > Spatial Analyst Tools > Overlay > Weighted Sum > Add three reclassified layers above, that is reclass_kids, reclass_parks, and reclass_schls, and set weight 0.33, 0.33, and 0.34, respectively in the Weighted Sum dialog box; Results will show composite suitability scores to all areas resulting from weighted sum).
• Change the symbology of contri_output to graduated color scheme (light brown to dark brown). Click here to see results.
• Save the map as SuitabilityAnalysis_Contributory.mxd

7. Compare results

• Make a map that shows results of dominance rules (screening) and contributory rules (linear combination) in combining criteria
• Add necessary map elements
• Submit the image file of this map to D2L dropbox