topic DEM values/DEM differencing in ArcGIS Spatial Analyst Questions

DEM values/DEM differencing

LaszloKomaromi2 — Fri, 03 Mar 2017 21:28:11 GMT

Hi everyone,

I would like to compare the elevation and volume of three DEMs. When I export the original DEMs in TIFF format the elevation values change. I do not use compression or any other kind of data transformation. It makes me wonder what the real elevation values are.

Before

After

My second question is that I have calculated the volume difference of my DEMs (not the TIFF versions but the original) and it says that there is an increase in total volume between the DEMs. It is nearly impossible, since these are glacier surfaces that are known to decline in this area (and in most parts of the World). It is unlikely that there is a snow bias and my DEMs are in the same coordinate system and all clipped to the same vector shape.

Any idea what could have gone wrong?

Thanks

Laz

Re: DEM values/DEM differencing

ChrisDonohue__GISP — Fri, 03 Mar 2017 22:13:34 GMT

Any chance the cell size varies between your files? This might account for the volume difference.

Also, when you are doing any processing, is a Snap Raster being set in the Environments setting?

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

Chris Donohue, GISP

Re: DEM values/DEM differencing

DanPatterson_Retired — Fri, 03 Mar 2017 22:15:31 GMT

did you check the cell size and the extents explicitly in your environments prior to export? Any change in cell size and/or extent could be fully or partially to count. Your will have to compare your rasters after conversion to ensure that they are indeed the same

Re: DEM values/DEM differencing

LaszloKomaromi2 — Sat, 04 Mar 2017 15:49:28 GMT

Hi Dan,

Thanks for your reply. I resapmled my DEMs to the same resolution(25x25m) before processing them. I have defined the spatial extent myself, by clipping the DEMs to a vector outline through the Image Analysis menu.

My other problem is that when I turn these three DEMs into contour maps two of them doesn't draw. I have imported the third DEM's Coordinate system, also the 'project' tool and tried several other systems, nothing works, whereas they all show up correctly before turning them into contour maps.

Ideally, all layers should be in the same coordinate system, but somehow it is just not possible in my case. For instance, my vector files are in 1993 Lambert system and my Landsat images are in WGS 1984. They perfectly overlap this way, but if I change my vector to WGS 1984 it won't draw. I have been using ArcMap actively for a year by now, performing different tasks but i still can't figure out why this is happening.

Any idea?

Many thanks,

Laszlo

Re: DEM values/DEM differencing

LaszloKomaromi2 — Sat, 04 Mar 2017 15:53:16 GMT

Hi Chris,

Its not the cells size, they are all in agreement. I'll try the Snap Raster though, however, looking at properties, all my DEMs are identical, from cell size through pixel type and format.

I'll keep on experimenting.

Thanks a lot,

Laszlo

Re: DEM values/DEM differencing

DanPatterson_Retired — Sat, 04 Mar 2017 17:57:47 GMT

You don't 'change' coordinate systems, you 'project' a files, which produces a new file with the desired coordinate system. If you are using the 'define projection' tool anywhere in your process to try and accomplish what you are trying to do, you will end up with files that don't align, overlap and are not in the same coordinate space. The only time that the 'define projection' tool should ever be used is when you get a warning that your file does not have a defined coordinate system. If you got no such warning, then you are not doing what you need to do.

So... get your data into one coordinate system, via using the Project tool for vector data or Project Raster for raster data.

Once everything is in the correct coordinate system, you can begin

Re: DEM values/DEM differencing

LaszloKomaromi2 — Sat, 04 Mar 2017 18:20:46 GMT

That's right, but it doesn't explain why the contour maps don't draw. They are in the same projection.

Also, I tried the above Project Raster tool. Usually I get error messages when performing tasks for no apparent reason. Here is the latest one when trying to project raster:

Re: DEM values/DEM differencing

DanPatterson_Retired — Sat, 04 Mar 2017 18:33:16 GMT

that one is easy... turn off background geoprocessing, it is more trouble that it is worth

Re: DEM values/DEM differencing

DanPatterson_Retired — Sat, 04 Mar 2017 18:43:06 GMT

and if the contours don't appear, make sure they are on top of the list, right-click on the layer, zoom to layer and see if it is in the same coordinate space. If it isn't, there is something wrong with the coordinate systems, like perhaps trying to contour data which are in geographic coordinates

Re: DEM values/DEM differencing

LaszloKomaromi2 — Sat, 04 Mar 2017 18:51:12 GMT

I switched off background geoprocessing. The overall performance seems to have improved but I got the same error message while projecting raster.

Re: DEM values/DEM differencing

DanPatterson_Retired — Sat, 04 Mar 2017 19:36:16 GMT

was the raster defined correctly to begin with?

spaces and other stuff in the output paths

type of raster? ie esri grid, tif??

output coordinate system?

cell size?

extent?

the errors are always in the details... all the details

Re: DEM values/DEM differencing

LaszloKomaromi2 — Sat, 04 Mar 2017 19:39:47 GMT

Very true. I double check the above. Thanks.

Re: DEM values/DEM differencing

LaszloKomaromi2 — Sun, 05 Mar 2017 21:19:59 GMT

Apparently, my contour map can only be projected in WGS 84 UTM Zone 27N. WGS 84 projection is not working.

Other than that, my three DEMs were in different formats and spatial resolution. I resampled them, but changing formats result in different elevation values. It is very important to have them right since I am looking at volumetric differences of glaciers through geodetic mass balance method. According to the volume calculations using 'Surface volume' tool, my 2004 DEM shows an increase in volume compared to the 1986DEM. It is unlikely that the glacier has grown since, as ice is known to be retreating rapidly in that region. It bugs me that when changing formats and/or pixel type the elevation values and volue changes. It makes me wonder what the actual elevation values are for my DEMS.

Re: DEM values/DEM differencing

DanPatterson_Retired — Sun, 05 Mar 2017 23:51:29 GMT

if the glacier area was collected with a GCS with elevations in meters, the GCS data can be projected with an appropriate planar cell size. Geographic coordinates are pretty well only useful for recording location. Calculating area and/or volume, (even using the best algorithms), are difficult at best. Projection of the area to planar coordinates at least gives you a known representation error. Suggestions would be to include some data samples and at least what cell size and extent you used during the projection from GCS to projected data.

Re: DEM values/DEM differencing

LaszloKomaromi2 — Mon, 06 Mar 2017 11:55:53 GMT

This sounds way too complicated. Almost makes it impossible to align DEMs. I started to question the precision of every DEM I ecer used.thanks for your help Dan

Re: DEM values/DEM differencing

DanPatterson_Retired — Wed, 15 Mar 2017 00:30:39 GMT

I will change it to a discussion since there appears to be no resolution that has worked.