EGU2016-6408
Outlet-glacier flow dynamics estimation combining in-situ and spaceborne
measurements
Christoph Rohner, Daniel Henke, David Small, Rémy Mercenier, Martin P. Lüthi, Andreas Vieli
Department of Geography, University of Zurich, Switzerland
Study Area
Statistical Analysis
·
#
GPRI Position
Source: Landsat-8 (USGS)
0
4
2
R = 0.9001
RMSE = 0.4744
0
0
1
2
4
6
ௌ௔௧
ீ௉ோூ )
on a pixel-by-pixel basis. The approach is easily
implemented, while taking into account
information about the reliability of the in-situ
measured data.
4
6
8
150
10
7747
5
7746
0
7745
7744
Landsat-8
Intensity Tracking
100 km
7742
10
0
-25
-20
-15
-10
-5
0
5
7746
10
7744
5
7742
Flow velocity [m/d]
15
10
20
7745
10
7744
7743
7742
Flow Velocity [m/d]
7746
Flow velocity [m/d]
15
532
534
536
UTM22N Easting [km]
Flow Velocities Along Flowline
GPRI (Jun/Jul 2015)
Landsat-8 (Jun/Jul 2015)
Fused Data (June 2015)
Joughin (June 2014)
6
4
5
2
0
0
0
530
532
534
536
UTM22N Easting [km]
Glacier flow velocities based on intensity tracking of
two Landsat-8
scenes (masked to
glacier outline)
Conclusions and challenges
• Fusion of GPRI and space-borne data offers possibility to
overcome intrinsic problems of different data sources
• General agreement to reference data from Joughin et al.
(2010)
• Continuous velocity fields right to calving front allows to
detect small scale variations in flow dynamics
0
534
530
7741
7748
532
-25
528
10
Difference GPRI - LS-8 [m/d]
7747
20
530
-20
7741
8
528
528
-15
50
Fused Data Set
(2015/06/20 14:59:55 - 2015/07/06 15:00:04)
UTM22N Northing [km]
Satellite Data
Radarsat-2
-10
7743
7748
Sentinel-1A
-5
Data Assimilation
8
km
536
UTM22N Easting [km]
GAMMA Portable Radar Interferometer
20
5000
10000
15000
Distance from calving front [m]
Challenges – Outlook
• Not applicable to areas in radar shadows
• Missing high-resolution DHM induces errors
(geocoding/velocity adjustment)
• Application to SAR systems
7748
Interferometry
7747
15
7746
7745
10
7744
7743
5
7742
7741
0
528
530
532
534
UTM22N Easting [km]
536
Flow velocity [m/d]
Following a statistical analysis of the retrieved
flow velocities in areas with overlapping in-situ
and satellite data, the data sets were
assimilated. Therefore, the line-of-sight velocity
information from the GPRI
were adjusted
based on a physical flow direction model. Using
the coherence information retrieved by the
GPRI, both flow velocity estimates were
weighted using the equation
2
200
100
n = 5913
2
Eqip Sermia, a medium-sized ocean terminating outlet glacier in western
Greenland.
UTM22N Northing [km]
Data Fusion
ீ௉ோூ
6
7748
Statistical analysis shows general good agreement between flow velocities derived with spaceborne and
GPRI data. The biggest differences can be found in regions bordering the calving front and the side of the
glacier.
Flowline
UTM22N Northing [km]
By applying a feature (i.e. intensity) tracking
methodology (e.g. Joughin, 2002; Strozzi et al.,
2002) to the spaceborne data, continuous flow
velocities can generally be retrieved with a high
spatio-temporal resolution. Issues exist in the
border regions of the glacier, where there are
discontinuities in movement due to solid ground
(side of glacier) or sea-ice (ocean). Applying the
patchwise normalized cross-correlation approach in these areas results in noisy and
incorrect velocity estimates, while the temporal
and geometric decorrelation of the signal
impedes the use of interferometric approaches.
The movement data from the GAMMA Portable
Radar Interferometer (GPRI) can be used to
overcome the issues in these areas.
ீ௉ோூ
250
Data
Intensity Tracking
௙௨௦௘ௗ
8
300
GPRI Derived Flow Speeds [m/d]
Aim
Using flow velocity data from in-situ measurements and spaceborne systems we aim at:
• Assessing the differences between the measurements methods on overlapping areas
• calculating a spatially continuous velocity
field for the whole glacier
Data
Fit
Confidence Bounds
References
Joughin, I. (2002). Ice-sheet velocity mapping: a combined interferometric and speckle-tracking approach. Annals of
Glaciology, 34, 195–201.
Joughin, I., Smith, B.E., Howat, I.M., Scambos, T.A., & Moon, T. (2010). Greenland flow variability from ice-sheet-wide
velocity mapping. Journal of Glaciology, 56, 415–430.
Strozzi, T., Luckman, A., Murray, T., Wegmuller, U., & Werner, C.L. (2002). Glacier motion estimation using SAR offsettracking procedures. IEEE Transactions on Geoscience and Remote Sensing, 40.
Flow velocity difference [m/d]
Esri, HERE, DeLorme,
MapmyIndia, ' OpenStreetMap
contributors, and the GIS user
community
10
Glacier Flow Speed: GPRI - LS-8
(2015/06/20 14:59:55 - 2015/07/06 15:00:04)
Glacier Flow Speed Difference
(2015/06/20 14:59:55 - 2015/07/06 15:00:04)
UTM22N Northing [km]
Tracking Extent
Glacier Flow Speeds: GPRI vs. LS-8
(2015/06/20 14:59:55 - 2015/07/06 15:00:04)
Frequency
Terminus retreat and flow acceleration changes
of ocean-terminating outlet glaciers contribute
significantly to the current mass loss of the
Greenland Ice Sheet and the global sea level
rise. In order to constrain models of ice
dynamics, detailed knowledge of geometry and
ice-flow velocity of such calving glaciers is
needed. Particularly important are the near
terminus velocities, as these flow fields strongly
affect the glacier’s calving rate. For fast moving
outlet glaciers, these flow fields are difficult to
accurately capture with the current temporal
resolution of spaceborne systems, while in-situ
measurements using ground based radar
interferometers are limited in coverage and
constrained by distance and geometric shading
of the glacier.
Results
Landsat-8 Derived Flow Speeds [m/d]
Introduction