topic Re: Hillshade and Slope Calculation Problem in ArcGIS Spatial Analyst Questions

Hillshade and Slope Calculation Problem

DeborahBowyer — Sat, 30 Oct 2010 05:10:24 GMT

When I calculate slope and hillshade from a dem the results I get look weird... like gridlines. In the case of slope it looks like it is calculating the centre of each cell as flat (zero slope) and making the boundaries between slopes relatively steep. With hillshade the cell centres have a high value and the boundaries between cells relatively low values (see attchments). I'm currently using ArcView 9.3 sp1. I was thinking there might be something wrong with the DEM - but when I tried redoing slope and hillshade for a dem that I had successfully dones these for back in early 2009 I got the same problem.

If anyone out there has had this problem and can shed some light on how to fix it I would be very grateful.

Re: Hillshade and Slope Calculation Problem

DanPatterson_Retired — Sun, 31 Oct 2010 18:03:09 GMT

is the map projected? (ie UTM, state plane?)
are the x, y units the same as the Z values? (ie meters, feet)

Re: Hillshade and Slope Calculation Problem

Anonymous User — Sun, 31 Oct 2010 19:00:57 GMT

Original User: debbow73

Projection is WGS_1984_UTM_Zone_34N, and x, y and z are all in metres.

Re: Hillshade and Slope Calculation Problem

heineklar — Mon, 01 Nov 2010 15:03:46 GMT

hey,
ive got the same problem at the moment.....is there a solution, yet?

Re: Hillshade and Slope Calculation Problem

Anonymous User — Mon, 01 Nov 2010 21:42:15 GMT

Original User: debbow73

I figured out how to fix this... in the Hillshade tool go into Environments > Raster Analysis Settings, and set the cell size to Same As the DEM (I thought I had done this on the Spatial Analyst Toolbar > Options -- but I guess these two processes do different things!?!?). Anyway - I now have a perfect slope raster :D.

Hillshade still looks a bit weird (see attachment) - very "white" compared to hillshades I have made before. Values in the Hillshade range from 37(black) - 127(white). Was wondering if the problem is in how I'm displaying it... but I can't figure out how to make 0 black and 255 white. I'll keep working on it! 😮

Re: Hillshade and Slope Calculation Problem

DanPatterson_Retired — Mon, 01 Nov 2010 21:47:33 GMT

check this blog and see if it applies to your situation
http://blogs.esri.com/dev/blogs/imagery/archive/2010/09/07/why-does-my-hillshade-look-different-when-i-use-a-tool-vs-a-function_3f00_.aspx

Re: Hillshade and Slope Calculation Problem

Anonymous User — Mon, 01 Nov 2010 22:13:45 GMT

Original User: debbow73

😮 Lowering the altitude in the Hillshade Tool seems to have done the trick 😄

Re: Hillshade and Slope Calculation Problem

heineklar — Tue, 02 Nov 2010 13:44:07 GMT

Hm.....in my case it still doesn't work. Any other idea to solute this?

Re: Hillshade and Slope Calculation Problem

Anonymous User — Tue, 02 Nov 2010 17:27:48 GMT

Original User: whuber

When I calculate slope and hillshade from a dem the results I get look weird... like gridlines.

Those are artifacts introduced during the construction or processing of the DEM. They are really in the data, not a bug in the hillshade or slope calculations.

You have two strategies for dealing with this issue:

(1) Accept your data and attempt to "fix" them. This requires preliminary smoothing (such as a focal mean) followed by recalculation of the hillshade and slope. Unfortunately, you will need strong smoothing and the result will depart appreciably from correct values. When you're hillshading to display the DEM, that can be ok, but otherwise it might not be.

(2) Find out where in the processing pipeline the errors occurred, go back before then, and redo the processing. These artifacts can appear as the result of converting floats to integer values at some stage; from changing cell sizes; as artifacts of lossy compression algorithms; and likely in other ways I have not yet encountered. This should be the preferred solution unless it's simply impossible (e.g., you obtained the DEM in this form originally and cannot obtain a better version from any other source).

Re: Hillshade and Slope Calculation Problem

heineklar — Thu, 04 Nov 2010 15:41:07 GMT

hmmmm....the grid problem seems to be solved: it foundet in the merging of two DEMs.

but now the new result doesnt look better....all values are over 89 Degrees. has anyone an idea, what could be the cause for that?

Re: Hillshade and Slope Calculation Problem

Anonymous User — Thu, 04 Nov 2010 16:03:51 GMT

Original User: whuber

You performed the calculation with an unprojected grid. You *must* project a grid to obtain valid hillshade and slope values.