Lab: Weighted Overlay

A generic problem typically addressed with a Weighted Overlay method is the identification of agricultural land suitable for a particular crop. Below we will be working through a simplified exercise looking e.g. for open, agricultural land with a southern aspect, not overly steep and of course not built-up or covered by forest.

First, though, we want to familiarize ourselves with a sample study area in the Austrian alpine foreland:

Before starting with an overlay analysis based on the above introduced factor layers, we can explore the study area in some more detail and make mental notes where we could be looking for suitable cropping areas:

For reproducing the visuals above in ArcGIS Pro, launch a New Map and Add the layers introduced above: L_WLCBaseVue2013, L_Elev250, L_Slope10, L_Aspect8 and L_Curv3 - and make sure you understand the classification of these layers. This is done easily when connected to https://zgis.maps.arcgis.com as an active portal and selecting these from the group AGI_SpatialAnalysis, or by using the links to the samples quoted above.

For completing the below outlined exercises, though, you would have to access or download and open a > layer package , which also can be found in your connection to https://zgis.maps.arcgis.com as an active portal and selecting this package from the group AGI_SpatialAnalysis.

(Again, you of course are more than welcome to use your own choice of problem and supporting data sets - just make sure these are integer rasters suitably reclassified and with a manageable resolution and extent)

If you have not worked with this toolset before, please read > How Weighted Overlay works , then open the Weighted Overlay tool from Geoprocessing (Spatial Analyst).

This kind of results from a Weighted Overlay (sometimes referred to as 'index overlay') will frequently satisfy the requirements and answer the questions and demands for information products put forward by decision makers. Still, we need to consider some of the constraints of this method:

• Weighted averaging implies a metric scale of rating, a value of 4 means twice the value of 2.
• High grades in one criterion (layer) can compensate lower grades in another criterion.

Sometimes, these points do not correspond with reality and multi-thematic (multi-layer) combinations of values only can be stated on nominal levels - this is the domain of Matrix Overlay. This name refers to a combinatorial matrix of grades given to tuples of values.

Think about and define a realistic multi-factor location or suitability problem and translate it into a Weighted Overlay application. Avoid getting started right away with 'doing it', but spend some time on deciding on the how-and-why:

• which rating scale are you going to employ?
• what is a suitable minimum mapping unit for the decision you're trying to support?
• how do you choose and justify the weighting of layers?
• how do you come with the rating ('points') for categories?
• is your approach actually reasonably implemented in a 'metric' environment, or should you switch to a nominal domain, i.e. matrix overlay?
• if you need a certain amount of farmland - e.g. 500ha aligned with the capacity of a processing factory - how could you come up with the best 500ha if more than that are identified as suitable?
• there might be a minimum operational size of a plot, i.e. you do not really want to farm individual, widely distributed 'pixels' - how do you identify suitable plots with a minimum size? Hint - check out the > Region Group  function.
• ...

Now it is your turn to set up your own case study, define a problem or question and independently establish a spatial priority or suitability selection in the domain of your choice, thereby creating a valuable information product for decision making.