I am using "Summarize Within" (ArcPro 2.7.3) to summarize the area of land use polygons having different land cover types in catchments of lakes. This means the input polygons are lake catchments, the summary features are land cover polygons having different land cover types, and the group field used is the code of the land cover types. The land use data for sure cover the whole lake catchments.
However, the sum of the area of the summary features given in the output grouped table is smaller and not matching the area of the input polygons!
For example, in the screenshot attached to this post, the catchment polygon MSCDLV_1100000 has an area of 1572268.24 squaremeters while the summarized area is only 1570380.6 squaremeters. Moreover, if I clip the land cover data manually using the lake catchment polygon, I can sum up the area of the land cover polygons belonging to the different land cover types, which gives me the correct results but the summed areas for these land cover types given by the Summarize Within tool are slightly smaller.
The only explanation I can think of is the following: The tool converts the summary feature layer to a raster and sums up all raster cells having their center located in the input polygon. This would results in somewhat smaller areas since the raster cells at the border of the input polygons would not be included. However, that would be "suboptimal" since it gives incorrect results. In former times, I had an ArcView Avenue script that calculated summed areas by groups in input polygons. I hope I don't have to switch back to this one ;-).
Does anybody now - especially from the ESRI team - how the tool actually works and if this is the reason for the observed differences?
Many thanks!