Page 1 of 1

can elmerice compute ice velocity

Posted: 17 Nov 2015, 11:03
by lusaisai
Hello,
I am a graduate student in Tongji University, China. These days I did some research in ice sheet modelling by the software Elmer/Ice. And I am interested in ice velocity modelling in Antarctic.
So I followed the example ”teterousse_2014_Step1--teterousse1.sif” by Prof. Gagliardini, then I got the ice velocity distribution map as below.
modelled velocity
modelled velocity
modelled velocity.jpg (87.16 KiB) Viewed 4656 times
However, there is measured velocity map by Prof. Rignot this region as below.
measured velocity from Prof. Rignot
measured velocity from Prof. Rignot
measured velocity.jpg (88.21 KiB) Viewed 4656 times
Comparing to my result, there is large difference.I am confused about that. I wonder whether I made some mistakes in writing the sif file(if yes, can you give me some suggestions), or actually there will be different between modelled and measured ice velocity. Besides, I also want to know can the elmer/ice make the ice velocity map like the measured one.

I am not familiar with Elmer/Ice, so I am really sorry if I overlooked or misunderstood anything.

I will be very appreciated for any help!
Yours,
Saisai

Re: can elmerice compute ice velocity

Posted: 17 Nov 2015, 11:28
by tzwinger
Dear Saisai,
First of all, we have no clue what the units in the picture from Rignot are (cm/day, m/year?) we should compare your results to - if you used the attached SIF your Elmer/Ice velocities are m/year. Looking at your SIF you have a no-slip condition at the bedrock, which kind of contradicts the basic idea behind an ice stream - for having pure deformation you actually achieved quite high velocities, which is due to the high ice temperature of -1 C (making the ice very soft), which of course will not apply throughout an Antarctic ice stream. The second issue I can spot is the no-flow condition on the sidewalls (including the in- and outlet of your ice -stream). What you have to do is to find out the correct basal sliding parameters that reproduce your surface velocity pattern and get a good guess on your internal temperature (which controls the viscosity). Doing so, usually requests inversion of basal sliding from satellite data. Explaining this in a forum post is not possible - this is far from being trivial. I would recommend you studying the material from last advanced Elmer/Ice course which is to be found in the Elmer/Ice Wiki (courses). To re-emphasize: There is a vast difference between an Antarctic ice stream and a poly-thermal alpine glacier, so you cannot just take the unaltered solver input file of the latter to build a simulation for your problem.

Regards,

Thomas

Re: can elmerice compute ice velocity

Posted: 17 Nov 2015, 13:55
by lusaisai
Hi Thomas, thank you for your reply.

The units in the picture from Rignot are m/year, and I also replace the temperature in the sif according to data from the weather station.
Your answer is helpful for me , I think I should think about the basal sliding if I want to get the velocity map.

Yours Sincerely,
Saisai