Scenarios for using the PAAT

There are four main scenarios when using the Positional Accuracy Assessment Tool (PAAT) tool to assess positional accuracy: using two feature classes, a feature class and a raster, two rasters, and a z-enabled point feature class and digital elevation model (DEM) raster or triangulated irregular network (TIN). The different combinations of layers for evaluation and reference cause the PAAT session to be variable based on what you are using as an evaluation layer versus a reference layer. For example, using two feature classes creates a grid on the reference layer, but if you are comparing a z-enabled point feature class to a DEM, the grid does not appear.

The scenarios and their respective behaviors are as follows:

• Feature class and feature class—Feature classes are used as both the evaluation and reference layers. When one of the feature classes contains points, the point that is digitized in the evaluation layer is automatically associated with a feature in the feature class. You can also configure the session so the point in the evaluation layer automatically snaps to a vertex on the closest feature. This ensures that the points digitized relate to actual features in your feature class.
• Feature class and raster—A feature class is used as one of the layers, and a raster is used as the other. You can use the raster layer as either the reference or evaluation layer in this scenario. When a feature class is used as the reference layer, the features act as control points for the raster you are evaluating. This means you can only review the residual values and the levels of error for the raster layer.
• Raster and raster—Rasters are used as both the evaluation and reference layers. One raster is the reference, and the other is evaluated. When rasters are used for both layers, there is no snapping, since there are no features to snap to. This means you need to visually pick matching points on both layers.
• DEM raster or TIN and a z-enabled point feature class—A z-enabled feature class is used as the evaluation layer, and a raster is used as the reference layer. Rasters can only use z-enabled point feature classes in a PAAT session. The points in the feature class act as control points for the raster because the PAAT evaluates each point in relation to the raster.