Deriving datasets, such as slope or distance to schools, is the first step when building a suitability model. Each cell in your study area now has a value for each input criteria (slope, landuse, distance to recreation sites and distance to schools). You need to combine the derived datasets so you can create your suitability map that will idenify the potential locations for the new school. However it is not possible to combine them in their present form—for example, combining a cell value in which slope equals 15 degrees with a cell value for landuse that equals 7 (forest)—and get a meaningful answer that you can compare to other locations. To combine the datasets, they first need to be set to a common measurement scale, such as 1 to 10. That common measurement scale is what determines how suitable a particular location—each cell—is for building a new school. Higher values indicate more suitable locations for the school.

Using the Weighted Overlay tool, you can weight the values of each dataset, then combine them at one time. However, the inputs for the Weighted Overlay tool must contain discrete, integer values. Landuse is already categorized into discrete values; for example, forest equals a value of 7, so you can simply add this dataset directly into the Weighted Overlay tool and assign each cell a new value on the common measurement scale of 1 to 10 (you’ll do this later in the tutorial). The values in the datasets you derived in previous steps are all floating point, continuous datasets, categorized into ranges, and they must first be reclassified so that each range of values is assigned one discrete, integer value. Potentially, the value given to each range can be any number, provided you note the range that the value corresponds to. This is because you can weight these values within the Weighted Overlay tool—the next step after reclassifying the derived datasets. However, it is easier to weight the cell values for derived datasets while reclassifying. In the Weighted Overlay tool, you can accept the default and leave the scale values the same as the input values.

You will reclassify each derived dataset to a common measurement scale, giving each range a discrete, integer value between 1 and 10. Higher values will be given to attributes within each dataset that are more suitable for locating the school.

It is preferable that the new school site be located on relatively flat ground. You’ll reclassify the slope output, slicing the values into equal intervals. You’ll assign a value of 10 to the most suitable range of slopes (those with the lowest angle of slope), 1 to the least suitable range of slopes (those with the steepest angle of slope), and rank the values in between linearly.

To reclassify the output of the Slope tool that is currently in your model, expand the Reclass toolset, click and drag the Reclassify tool onto the ModelBuilder window, right-click the Reclassify tool element and click Open. Click the Input raster drop-down arrow and click the variable Slope output. Variables are denoted with blue icons in the drop-down list. When you drag data onto a ModelBuilder window or set the data referenced by input or derived data elements within a tool’s dialog box, the elements created on the ModelBuilder window are variables that can be shared between processes.

In this example, you'll use an equal interval classification, dividing the range of slope values in to ten classes. Because sites with low slopes are more attractive, you'll reverse the new values, so steep slopes get a low value, and level areas get a high value.