FAQ: Why are zonal statistics not calculated or inaccurately calculated for some zones?
Why are zonal statistics not calculated or inaccurately calculated for some zones?
The Zonal Statistics and Zonal Statistics as Table tools calculate statistics for rasters within zones set by a feature dataset. In some instances, the tools either do not produce outputs, or produce inaccurate results for some zones. Since the tools calculate statistics for raster cell values within a zone, such as a polygon boundary, the calculations are based on cells with centroids that fall within the boundary. If the raster cell size is larger than a given boundary, there may be no cells with centroids within the boundary, and hence no statistics or inaccurate statistics are produced for that zone.
The following are best practices when using the Zonal Statistics and Zonal Statistics as Table tools:
- Ensure that the raster and the zones are projected on the same projected coordinate system, and not in a geographic coordinate system. This is to provide uniform cell sizes when using the tools. For more information on doing this in ArcMap, refer to the following Web Help page, ArcMap: An overview of the Projections and Transformations toolset.
- The raster cell size must be smaller than the polygon boundaries. Resample the rasters so that multiple cells of a given raster can be evaluated within each boundary. A trial and error approach may be necessary to determine an optimum cell size, which provides the best level of detail without unnecessarily expending processing time to resample at a very small cell size.
- ArcMap: Resample function
- ArcMap: Choosing an appropriate transformation
- ArcMap: Zonal Statistics
- ArcMap: Zonal Statistics as Table