ISPRS Commission II, Vol.34, Part 3A „Photogrammetric Computer Vision“, Graz, 2002 
  
cm/day, and the maximum difference reached 58%. The r.m.s. 
difference was given as 6.34 cm/day. We understood that such a 
comparison could not serve as an estimate of absolute accuracy, 
but indicated only methodological differences. 
  
(7 —- (C 7m», 
complex- complex- 
valued SAR valued SAR 
— À 
  
  
|4 interferogram I> interferogram 
generation generation 
M rr V 
spectral filtering spectral filtering 
  
  
  
  
(optionally) (optionally) 
topogram topogram 
generation generation 
  
  
V. V. 
  
  
  
  
  
  
  
  
  
fluxogram co-registration 
generation and subtraction 
c.VV.. VF. transferential 
e dq ? referencing € 
velogram referencing, € 
generation geocoding 
  
Figure 4. Principal flowchart of the GINSAR technique 
Absolute tachometric validations were performed during the 
field work undertaken in Novaya Zemlya in September 2001, 
when the frontal velocities of several test tidewater glaciers 
were surveyed using precise geodetic equipment. Preliminary 
measurements of the maximum velocities were made in the lab 
using 4 INSAR models taken over the same glaciers in March 
1996. The geodetic surveys were performed using two 
alternative techniques. A conventional geodetic technique of 
forward intersection is based on multitemporal observations of 
identical points at the glacier front from two different positions 
with a known baseline. The polar idea of a non-traditional 
”touch-and-go” technique is to install the laser line in a 
tangential position at a predefined distance from the glacier 
front and to frequently measure the distance to any point on the 
opposite coast in order to record the instant when the glacier 
will cross the laser line and the measured distance will change 
abruptly (Sharov et al., 2001). This technique made use of a 
LDI-3 laser rangefinder (range up to 19 km without retro- 
reflector) mounted on a Leica T1602 theodolite (angular 
accuracy 0.5 mgon). Typical results of velocity measurements 
are given in Table 1. 
The results obtained indicate that the winter INSAR velocities 
are somewhat lower than those from 'summer' geodetic surveys, 
which corresponds to the observations performed by other 
investigators in other regions (e.g. Rabus & Fatland 2000). 
Apart from the different duration of observations and seasonal 
changes in the glacier motion, this fact can be explained by the 
inherent influence of INSAR undersampling if velocity 
gradients between neighbouring pixels exceed 0.014 m/day. In 
spite of the revealed tachometric differences of up to 4096 and 
more, the spatial correlation between the GINSAR velocities 
and those surveyed in the field was quite high over all glaciers. 
  
  
  
  
; Frontal velocity [m/day] 
Glacier : : 
INSAR | Fwd. intersection | Touch-and-go 
Rykatchova 
No. 88 0.56 0,68+0,14 0,83+0,20 
Mack No.91 0.37 0,61+0,13 0,73+0,15 
  
  
  
  
  
  
Table 1. Glacier velocities measured in the lab and in the field 
S. CONCLUSIONS 
The global and stringent GINSAR algorithm devised for the 
glacier motion estimation does not involve complex process 
artifices and does not require additional topographic reference 
models thus eliminating areal error propagation, and the 
resultant velocity gradient values are quite tolerant of local 
phase errors. Therefore, such an approach remains feasible 
even under significant phase noise. The stage of phase filtering, 
which is usually included in all known phase unwrapping 
algorithms, becomes optional in our case and is used mostly for 
cosmetic reasons. The high metric quality, detail and 
complementary thematic contents of the GINSAR products, 
which are called topogram and fluxogram, show the expedience 
of this technique and its applicability to solving various tasks in 
the area of unsupervised glacier change detection and glacier 
mass balance measurement. The transferential approach based 
on the analysis of fast-ice motion forced by the glacier flow 
provides good reference values for the glacier frontal velocity 
and velocity gradients and can be considered as a 
complementary operation to the GINSAR method. The 
combination of our algorithms provides a unique opportunity to 
reduce the computational load and to mitigate some problems 
related to the operation of interferometric phase unwrapping. 
References 
Bamler, R., Hartl, P., 1998. Synthetic apertur radar interferometry. 
Inverse Problems, 14, R1-R54. 
Fatland, D., Lingle, G., 1998. Analysis of the 1993-95 Bering 
Glacier surge using differential SAR interferometry. J.Glac., 44 
(148), pp. 532-545. 
Forster, R. et al, 1999. Interferometric radar observations of 
Glaciares Europa and Penguin. J. Glac., 45 (150), pp. 325-336. 
Gabriel, A. et al., 1989. Mapping small elevation changes over 
large areas: differential radar interferometry. J. Geophys.Res., 94, 
B7, pp. 9183-9191. 
Joughin, I. et al, 1996. Estimation of ice-sheet motion using 
satellite radar interferometry. J.Glac., 42 (142), pp. 564-575. 
Meyer, F., Hellwich, O., 2001. Estimation of Glacier Topography 
and Velocity Using Correlation Minimization. 
Proc. of DGPF Jahrestagung Sept. 2001, Konstanz, pp. 467-474 
Rabus, B.T. Fatland, D.R., 2000. Comparison of SAR- 
interferometric and surveyed velocities on a mountain glacier: 
Black Rapids Glacier. J. G/ac., 46 (152), pp. 119-128. 
Sandwell, D., Price, E., 1998. Phase gradient approach to stacking 
interferograms. J. Geophys. Res., 103, No. B12, 30183-30204. 
Scambos T., Fahnestock, M., 1998. Improving DEMs over ice 
sheets using AVHRR photoclinometry J.Glac., 44(146), pp.97-103. 
Sharov, A.L., Gutjahr K., 2002. Some methodological enhancements 
to INSAR surveying of polar ice caps. In: Observing our 
environment from space. Proc. of the 21* EARSeL Symp, pp.65-72. 
Sharov, A.I. et al., 2001. Results of the field campaign in Barents 
and Kara seas on board the Hydrologist ship. In: Annual scientific 
report to the INCO Commission, P 9, pp.1-20. 
Wegmüller U., Strozzi T., 1998. Characterization of differential 
interferometry approaches. In: Proc. EUSAR’98, Berlin, p.237-240. 
A - 329