Help needed calculating glacier volume from DEMs on surface and bedrock topography

[BenRobson]

Hi there,
I'm working on my master thesis at the University of Bergen, Norway. I'm trying to work out the volume of the three parts of Folgefonna glacier, in Norway. I have the bedrock topography beneath the ice which was recorded using ground penetrating radar and ice radar during fieldwork last spring. I also have DEMs of the glacier surface from many years ranging from the 1930s to 2010. In addition this I have shapefiles of the glacier area.

I want to calculate the volume between these two DEMs, I've tried using the Cut Fill tool, but I don't know how to turn the results of this into a volume in cubic metres in order to assess the change in volume.

Can anyone help?

Thanks!

Ben Robson

[EricRice]

Hi Ben,

If your projected coordinate system is in units of meters and the pixel values are in units of meters, then the volumes will be calculated in cubic meters.  What coordinate system are both inputs using and what are the z units?  You are using the DEMs of the glacier surface as the "After Surface", correct?  Essentially you have no cuts -only fill. 

If ultimately you want to visualize these volumes in the form of a multipatch feature class or perform analysis on these volumes (shapes), I would suggest having a look at the Extrude Between tool followed by the Add Z Information tool.  It only works on TIN's though, so you'd have to convert your raster data to TIN.  If they are very large, it may not work because we have to dump everything into memory for that particular tool and you can run out resulting in tool failure.  If it were too large (too many nodes) you can thin the TIN during the conversion or post conversion with Decimate TIN Nodes tool.

Regards,
Eric

[BenRobson]

Hi,
The DEMs were either created on PCI Geomatica or were provided by the water agency of Norway, they all have the same pixel size and co-ordinate system, and the Z factor I assume is the same for each as they agree on heights of the surrounding bedrock. I tried using CutFill where the glacier surface was the After Surface, and the bedrock the Before Surface, and Surface Volume where I calculated below a planar surface above the glacier for both inputs, and then subtracted one from another, but it seems that some of the DEMs give volumes several orders of magnitude out from others. I will try Extrude between to see if I get the same problem.

Do you think that the fact the DEMs were made on a British computer - i.e. it used a . as a decimal point, while my computer which I'm trying to analyse them is Norwegian, so uses a , for a decimal point. Having said that the heights the DEMs give are all quite similar.

[EricRice]

Hi Ben,

The delimiter doesn't sound like the issue for now.  What is the coordinate system of the data?  To check, open the layer properties and go to the source tab.  Is it using meters for the unit?  We need to verify that your X,Y,Z units are all meters.

How large are these datasets?  Are they publicly available for me to download somewhere?

-Eric

[BenRobson]

Hi Eric,
I checked, and a couple of the DEMs have the Spatial Reference Angular Unit as metres, but most have Degree (0,017453292519943295). Is there a precise way to convert these? Or should I make the DEMs again? Some of the datasets are publicly available, but there's a long wait in acquiring them, and some of them I made on PCI Geomatica. I can perhaps try sending them to you though if needed?

Thanks for your help so far!

Ben

[EricRice]

Ben

Ok, so some of your data is using a geographic coordinate system.  This is the problem.  LxWxH can't have mixed units of decimal degrees and meters.  The precise method you inquire about is to project those rasters that have units of DD to a projected coordinate system that uses Meters.  To do this, open the Project Raster geoprocessing tool. Since this is elevation data, please make sure that when projecting the data you change the resampling method to Bilinear.  If you happen to switch datums be sure to set an appropriate geographic transformation for your area.  Once X,Y,Z (I'm assuming Z is already meters) are the same the volumes should make a lot more sense.

-Eric

[BenRobson]

They all have WGS 1984 UTM ZONE 32N as the Spatial Reference, apart from one or two which are Zone 31N, should I adjust this too? Reprojecting them seems to work fine, so now do you suggest I convert them to TINs and use Extrude Between rather than Cut Fill or Surface Volume.

Thanks,

Ben

[BenRobson]

Hi Erik,
I've managed to make the TIN from the extrude between tool, I've looked at it on ArcScene and it all seems fine, but when I try to Add Z Information it gives values for Volume and slope as 0. Any idea why this could be?

Thanks!

Ben

[EricRice]

Hey Ben,

Yes, I do know what that means and it isn't good news.  The only time volume will be returned as zero is when the input multipatch is not closed.  The tool is supposed to create closed multipatches.  I tested 10.1 beta 2 and it's working properly.  I did find a bug where it still produces non closed multipatches but only if the inputs lack spatial reference information.  I know you have defined the coordinates though and even reprojected.  Not until 10.1 do we have the ability to easily close a multipatch - there is a new tool coming called Enclose Multipatch.  If you run Is Closed 3D does it return 'No'?  Did you output a shapefile multipatch or did you send it to a geodatabase?  Also, what service pack do you have installed?  I saw an older bug that said the tool produced non closed when going to shapefile but was working when going to geodatabase.  Cut/Fill can still tell you the volume...if it comes to that.

Regards,
Eric

[BenRobson]

OK, I have version 10.0 Service Pack 2, do you think that updating would fix this error then? And yes I get a "no" for if it's closed, and the output is a multipatch. I can't actually see an option for outputing as a geodatabase, it just gives me an output as a .shp. So you do think updating would fix this?

Thanks for all your help!

Ben