Slope maps from ArcMap 10.1 and 10.3 are different

Question

I was examining a slope map that I created in ArcMap 10.1 approximately a year ago. Calculating the same slope map in ArcMap 10.3 gives me slightly different results. Did something change in the way slope is calculated? I used the same 100m DEM one year ago as I have today. Also, I noticed that the 10.1 slope map opens in a stretched B&W symbology while the 10.3 slope map opens in a classified symbology. Screenshots below show the slightly different results including an increase in the maximum slope value.

screenshot.pdf

DanPatterson_Retired · Answer

Ok perhaps next time we can leave out the coordinate system.

If... the raster has the exact same extent (check again) in rows and columns and in UTM coordinates AND it has the same cell size and there is a difference.

Is the difference just random? (a screen grab would really help)

Were you using slope in degrees for both OR slope in percent for both?

ThomasNupp · Answer

Here are the results of the raster calculations. As you can see by the range of values and the mean differences the two slope maps are different. 444, 6791100, 1001.15 MBGRIDGenericfloating point32 Bit-3.40282346639e+038absentabsentNoneBasicPermanent3953822.73375472874.114169517274.1141693885922.73375NAD_1983_UTM_Zone_15NMeter (1.000000)Degree (0.0174532925199433)5000000-930.99960D_North_American_1983skipped columns:1, rows:1, ignored value(s): -8.706651687622077.199062824249268-0.068358556105292541.0596882485126030

DanPatterson_Retired · Answer

It looks like the extent are the same and the cell size.

Now subtract them (either in the raster calculator or by using the Minus key).

You should have effectively 0 within floating point error.

Report back

(could just be a symbology thing... in which case, save the symbology from the old raster, and use it for the new raster in the import symbology from... portion of the symbology pane.)