Raster Modeling in ArcGIS

Part 1: Finding a site for a new school in Stowe, Vermont

 

Tutorial Scenario

The town of Stowe, Vermont, USA, has experienced a substantial increase in population. Demographic data suggests this increase has occurred due to families with children moving to the region, taking advantage of the many recreational facilities located nearby. it has been decided that a new school must be built to take the strain off the existing schools, and as a town planner you have been assigned the task of finding the potential sites.

In this exercise you will find suitable locations for a new school. The four steps to produce such a suitability map are outlined below.

Decide which datasets you need as inputs. The datasets you will use in this exercise are displayed to the right.

Derive datasets. Create data from existing data to derive new information.
 

Reclassify each dataset to a common scale (for example, 1-10) giving higher values to more suitable attributes.
 

Weight datasets that should have more influence in the suitable locations.


Your input datasets in this exercise are Landuse, Elevation, Recreation Sites, and existing Schools. You will derive slope, distance to recreation sites, and distance to existing schools, then reclassify these derived datasets to a common scale from 1-10. You will then weight them according to a percentage influence and combine them to produce a map displaying suitable locations for the new school. The diagram to the right shows the process you will take.

 

Step 1: Setting up the Project

1. Click the Add Data button and add the landuse, rec_sites, elevation and schools layer.
2. Now, we are going to set up the spatial extend for our project to the landuse layer. Click the Spatial dropdown arrow and click Options. Specify a working directory on your local drive in which to place your analysis results (chose My Documents and create a folder with your name). Click the Extent tab. Click the Analysis Extent dropdown arrow and click Same as Layer 'landuse'. Click the Cell Size tab. Click the Analysis Cell Size dropdown arrow and click Same as layer 'elevation' since this is the layer with the lowest resolution of the whole project.
3. Go to File > Map Properties, click on Data Source Options and check Store relative Path names to save files in a relative path structure.
4. Save the project as NewSchool.mxd into your folder.


Step 2: Deriving Datasets

Deriving data from your input datasets is the next step in the suitability model. You will derive the following:
a) Slope from elevation
b) Distance from recreation sites
c) Distance from existing schools

a) Deriving Slope

Since the area is mountainous, you need to find areas of relatively flat land to build on, so you will take into consideration the slope of the land.

1. Click the Spatial Analyst dropdown arrow, point to Surface Analysis, and click Slope.
2. Click the Input surface dropdown arrow and click elevation.
3. Type Slope in the Output raster text box to permanently save your output slope dataset to your folder.
4. Click OK to create the raster. High values (red areas) indicate steeper slopes.

b) Deriving distance from recreation sites

In this model, it is preferable that the school be built near recreational facilities, so you will now calculate the straight-line distance from Recreation Sites.

1. Go to Spatial Analyst > Distance > Straight Line.
2. Choose rec_sites in the Distance to dropdown menu.
3. Save the file as DistRecSites and click OK to create the raster. Values of zero indicate the location of a recreation site, with values (distances) increasing as you move away from each of these sites.
4. Uncheck the box next to schools to turn off this layer so you only see the location of the recreation sites and the distance to them.

c) Deriving distance from schools

You will now derive a dataset of distance from existing schools. It is preferable to locate the new school away from existing schools to spread out their locations through the town.

1. Go to Spatial Analyst > Distance > Straight Line.
2. Choose schools in the Distance to dropdown menu.
3. Save the file as DistSchools and click OK to create the raster.
4. Turn off the recreation sites layer and turn back on the schools layer for further examination.


Step 3: Reclassifying datasets

You now have the required datasets to find the best location for the new school. The next step is to combine them to find out where the potential location can be found. In order to combine the datsets, they must first be set to a common scale. That common scale is how suitable a particular location (each cell) is for building a new school. You will reclassify each dataset to a common scale, within the range 1-10, giving higher values to attributes within each dataset that are more suitable for location the school:
a) Reclassify slope
b) Reclassify Distance to recreation sites
c) Reclassify Distance to schools
d) Reclassify landuse

a) Reclassifying slope

It is preferable that the new school site be located on relatively flat ground. You will reclassify the Slope layer, giving a value of 10 to the most suitable slopes (those with the lowest angle of slope) and 1 to least suitable slopes (those with the steepest angle of slope).

1. Go to Spatial Analyst toolbar and click Reclassify.
2. Choose Slope from the Input raster dropdown arrow.
3. Click Classify.
4. Choose Equal Interval as the classification method. Choose 10 classes. Click OK.
5. You want to reclassify the Slope layer so steep slopes are given low values, as these are least suitable for building on. Click the first New value record in the Reclassify dialog box and change it to a value of 10. Give a value of 9 to the next New value, 8 to the next, and so on. Leave NoData as NoData.
6. Save it as ReclassSlope.
7. Click OK to create the raster. Locations with higher values (less-steep slopes) are more suitable than locations with lower values (steeper slopes).

b) Reclassifying distance to recreation sites

The school should be located near recreational facilities. You will reclasify this dataset, giving a value of 10 to areas closest to recreational sites (the most suitable locations) giving a value of 1 to areas far from recreation sites (the least suitable locations) and ranking the values in between. By doing this you will find out which areas are near and which areas are far from recreation sites.

1. Go to Spatial Analyst > Reclassify.
2. Click the Input raster dropdown arrow and click Distance to rec_sites.
3. Click Classify.
4. Make sure Equal Interval is set up as the classification method and 10 classes are chosen. Click OK.
5. You want to locate the schools near recreational facilities, so you will give higher values to locations close to recreational facilities, as these are the most desirable. As you did when reclassifying the Slope layer, click the first New value record in the dialog box and change it to a value of 10. Give a value of 9 to the next, and so on. Leave NoData as NoData.
6. Save it as ReclassRec.
7. Click OK to create the reclassification raster. It shows locations that are more suitable for locating another school. High values indicate more suitable locations.

c) Reclassifying distance to schools

It is necessary to locate the new school away from existing schools in order to avoid encroaching on their catchment areas. You will reclassify the Distance to schools layer, giving a value of 10 to areas away from existing schools (the most suitable locations) giving a value of 1 to areas near existing schools (least suitable locations) and ranking the values in between. By doing this you will find out which areas are near and which areas are far from existing schools.

1. Go to Spatial Analyst > Reclassify.
2. Click the Input raster dropdown arrow and click Distance to schools.
3. Click Classify.
4. Make sure Equal Interval is set up as the classification method and 10 classes are chosen. Click OK.
5. You want to locate the school away from existing schools, so you will give higher values to locations farther away, as these locations are most desirable. As the default gives high New values (more suitable) to high Old values (locations farther away from existing schools) you do not need to change any values this time!
6. Save it as ReclassSchoo.
7. Click OK to create the reclassification raster. It shows locations that are more suitable for locating another school. Higher values indicate more suitable locations.

d) Reclassifying landuse

At a town planners meeting it was decided that certain landuse types were better for building on than others, taking into consideration the costs involved in building on different landuse types. You will now reclassify landuse. A lower value indicates that a particular landuse type is less suitable for building on. Water and Wetlands will be given NoData as they cannot be built on and should be excluded.

1. Go to Spatial Analyst > Reclassify.
2. Click the Input raster dropdown arrow and click landuse.
3. Choose landuse in the Value field.
4. Type the following values in the New values column:
- Agriculture: 10
- Barren land: 6
- Brush/transitional: 5
- Built up: 3
- Forest: 4
5. You will now remove the Water and Wetland attributes and change their values to NoData. Select the rows for Water and Wetlands (shift key for multiple selections) and click Delete Entries.
6. Check Change missing values to NoData. All values for Water and Wetlands will be changed to NoData.
7. Save it as ReclassLand. Click OK.
8. The output reclassified landuse dataset will be added to your ArcMap session as a new layer. It shows locations that have landuse types that are considered to be better than others for locating the school (higher values indicate more suitable locations).
9. Bring up the Properties of the reclassified landuse layer. Click the Symbology tab and click the Display NoData as dropdown arrow and choose Arctic White to show NoData (Water and Wetlands) in this color. Click OK.


Step 4: Weighting and combining datasets

After applying a common scale to your datasets, where higher values are given to those attributes that are considered more suitable within each dataset, you are ready to combine them to find the most suitable locations. If all datasets were equally important, you could simply combine them at this point. however, you have been informed that it is preferable to locate the new school close to recreational facilities and away from other schools. You will weight all the datasets, giving each a percentage influence. The higher the percentage, the more influence a particular dataset will have in the suitability model. You will give the layers the following percent influence: (Each percentage is divided by 100 to normalize the values):
- Reclass of Distance to rec_sites: 0.5 (50%)
- Reclass of Distance to schools: 0.25 (25%)
- Reclass of landuse: 0.125 (12.5%)
- Reclass of slope: 0.125 (12.5%)

1. Click the Spatial Analyst dropdown arrow and click Raster Calculator.
2. Create following expression:
[Reclass of Distance to rec_sites] * 0.5 + [Reclass of Distance to schools] * 0.25 + [Reclass of landuse] * 0.125 + [Reclass of Slope of elevation] * 0.125 (double-click on the layer names to bring them into the expression box)
3. Click Evaluate to perform the weighting and combining of the datasets.

The output raster dataset shows you how suitable each location is for locating the new school, according to the criteria you set in the suitability model. Higher values indicate locations that are more suitable. The suitable locations are those areas that are close to recreation sites, away from existing schools, on relatively flat land, and on certain types of landuse. The higher weightings set for Distance to schools and Distance to rec_sites have a powerful influence on deciding which areas are more suitable than others.

4. Bring up the Properties for the new layer. Click the Symbology tab and click Classified from the Show list.
5. Setup 10 classes. Scroll to the last three classes, select them, right-click the highlighted classes and click Properties for selected colors. Give them a bright color.
6. Click the Display NoData as dropdown arrow and click the color black. This displays values of NoData (Water and Wetlands) in black. Click OK.

You decide that there are three main potential areas for locating the school. They are labeled in the diagram below.

7. We did not save the Raster Calculation to our folder. To save the file, right-click the new layer in the table of contents and click Make Permanent. Save it as Suitability.
8. Also rename the layer in the table of contents to Suitability.


Step 5: Creating a Layout

Create a layout of the suitability raster (uncheck all other layers to make them invisible) and add a title, your name, scale bar, and other common items you think are appropriate. Export it as NewSchool.pdf and submit it to your Webfolder.
 


Lab created by Eva Grund, February 2005, based on ESRI's ArcGIS 9 - Using ArcGIS Spatial Analyst manual, Chapter 2, Exercise 2.