# Cell size (10m) in Raster properties differs from measured cell size (20m) in multiband (Sentinel) raster?

I've always thought that if raster resolution is 10m, the cell size in the raster is 10m, and therefore the length of pixel is 10m.

Now I have started to work with Sentinel 2 optical imagery. Sentinel 2 data contains multiple resolutions within a multiband raster: 10m, 20m, 30m and 60m. I have created a raster composite from all 13 bands. When I load back individual single band rasters from raster composite, here I see that all single band rasters have cell size 10 m, even when I can see differences in cell size by naked eye; and while measuring the pixel (cell) length, I see that the cell size is actually 20m.

I don't understand why in Raster properties of one band the cell size is 10 m, but true pixel length is 20 m? If the data have been re-sampled to 10 m (which does not make sense for pixel information, but is computationally possible), the pixel size should be 10 m as well? Or am I misunderstand the concept about what the cell size represent in ArcGIS?

I am working in ArcGIS 10.4, GCS: WGS84

• Have you checked that your measured property is indeed only a single 20m pixel and not 4 10m pixels in a 2x2 grid as the result of nearest neighbour resampling? – Kersten Feb 9 '18 at 8:57
• wow, how can I check this? – maycca Feb 9 '18 at 9:00
• Pixel size in raster properties is correct in the coordinate system of the raster. If you measure the pixel size in some other coordinate system than the native one the lengths differ more or less. Pixel size is also expressed in the units of the coordinate system which can be something else than meters, like degrees or feet. It is not enought to know the pixel size, you must know the CRS as well. – user30184 Feb 9 '18 at 9:02
• Cell size is not really meaningful here since your coordinate system is not projected. It won't have a constant linear value (1° doesn't correspond to the same length near the pole and near the equator). – GISGe Feb 9 '18 at 9:02
• yes, the CRS is WGS 84 – maycca Feb 9 '18 at 9:04