Calculating raster class areas inside polygon feature dataset

Question

Hi!

I have raster dataset, grid with cell size 1000 x 1000 m, total area of 54 km2. Raster is reclassified into 6 classes. Also, I have one polygon vector layer (and only one feature) as research area. How can I get areas in meters (or cells) calculated as double or float values of each raster class inside my polygon? I tried with Zonal Toolset (Spatial Analyst) and Thematic Raster Summary Tool (Hawth's Analysis Tools for ArcGIS) but with no luck. I would like to have table like Hawth's Tool has, with six new fields (for six classes) filled with values of exact area in km2 inside polygon. Unfortunately, this tool only counts cells of each class inside polygon and fills attribute table with integer numbers.

Thanks!

Dario

DarioPerkovic · Answer

Thank your very much for your help. I'll try again with Zonal Toolset and maybe GME (http://www.spatialecology.com/gme/), successor of Hawth's Analysis Tools. If I find out something new, I'll let you know.

So as you implied in one your post, Raster resampling helped. Cell size is now changed from 1 km to 1 m and after that new raster has been clipped by feature dataset with Extract by Mask tool (Spatial Analyst Tools). Field in attribute table COUNT is showing much accurate values than before.

DarioPerkovic · Answer

Yes, it's possible, but it won't give you more accurate result.
Your result will be more precise but still will be as accurate as your least accurate data set.
In other words, your result with decimals won't be closer to reality than result without decimals.

Thank your very much for your help. I'll try again with Zonal Toolset and maybe GME (http://www.spatialecology.com/gme/), successor of Hawth's Analysis Tools. If I find out something new, I'll let you know.

MarcinGasior · Answer

Is it possible maybe to convert raster to polygon and then with something else calculate intersection of two polygons with decimals, I mean real km2 values?

Yes, it's possible, but it won't give you more accurate result.
Your result will be more precise but still will be as accurate as your least accurate data set.
In other words, your result with decimals won't be closer to reality than result without decimals.