Skip to content

Post-processing and analysing GNSS records of glacier motion: the potential of unsupervised clustering algorithms in comparison with conventional methods - University of Oslo (UiO)

Notifications You must be signed in to change notification settings

AdrienWehrle/GNSS_post_processing

Repository files navigation

Development Environment

  • These methods have been implemented and developed using a conda virtual environment that can be identically recreated. To this end, create a new environment using ./GNSS_post_processing.yml as below:
    conda env create -f GNSS_post_processing.yml
        

    ResolvePackageNotFound error can be raised. In that case, run conda env export --no-builds > GNSS_post_processing.yml instead.
    Then, run conda activate GNSS_post_processing to activate this new virtual environment.

  • The repository can also be run interactively on Binder, without any download:

    https://mybinder.org/badge_logo.svg

Abstract

Glacier dynamics remain a major source of uncertainty in sea level rise predictions (Stocker et al., 2013) and therefore still need to be better understood in places like Svalbard, known as one of the most climatically sensitive regions in the world (Rogers et al., 2005). In this study, the potential of long-term GNSS records at Holtedahfonna ice field and Etonbreen glacier is investigated to study short-term velocity events. Three automatic methods were developed in order to post-process the data as objectively as possible by getting rid of thresholds fixing. To this end, the data behaviour has been studied with regard to a range of parameterisations and the help of the elbow method (Satopaa et al., 2011). The comparison of the resulted daily velocities showed slight but unexpected variations depending on the method used. These methods have then been compared to others used in previous studies (Helbing (2005); Sugiyama et al. (2015)) which brought into light a high sensitivity of the data depending on the post-processing. Thereafter, a temporal resolution increase revealed a first change in the data behaviour of Holtedahlfonna ice field at an eight-hour and twelve-hour resolution (for respectively 2011 and 2012) while studying the developed methods independently but also among themselves. Finally, a glaciological interpretation consisting in the comparison a mean daily velocity to surface mass balance model outputs showed that even at this scale, various processes associated with a potential hydrometeorological forcing can ben explained. It is therefore highly encouraging for further temporal resolution investigations and thus, for a better understanding of short-term glacier dynamics.

References

Helbing, J.: Glacier dynamics of Unteraargletscher: verifying theoretical concepts through flow modeling, Ph.D. thesis, ETH Zurich, 2005.

Midgley, P. M., et al.: Climate change 2013: The physical science basis, 2013.

Rogers, J. C., Yang, L., and Li, L.: The role of Fram Strait winter cyclones on sea ice flux and on Spitsbergen air temperatures, Geophysical Research Letters, 32, 2005.

Satopaa, V., Albrecht, J., Irwin, D., and Raghavan, B.: Finding a ” kneedle” in a haystack: Detecting knee points in system behavior, in: 2011 31st international conference on distributed computing systems workshops, pp.166-171, IEEE, 2011.

Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V.,Midgley, P. M., et al.: Climate change 2013: The physical science basis, 2013.

Sugiyama, S., Sakakibara, D., Tsutaki, S., Maruyama, M., and Sawagaki, T.: Glacier dynamics near the calving front of Bowdoin Glacier, northwestern Greenland, Journal of glaciology, 61, 223-232, 2015.

About

Post-processing and analysing GNSS records of glacier motion: the potential of unsupervised clustering algorithms in comparison with conventional methods - University of Oslo (UiO)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages