HR: 1340h
AN: H23D-1148    [Abstracts]
TI: Shuttle Topography Radar Mission DEM, ASTER Images and Aerial Photography in Evaluation of Mountain Glacier Area and Volume Changes
AU: * Surazakov, A B
EM: sura7359@uidaho.edu
AF: Department of Geography, College of Science, University of Idaho, PO Box 443025, Moscow, ID 83844-3025 United States
AU: Aizen, V B
EM: aizen@uidaho.edu
AF: Department of Geography, College of Science, University of Idaho, PO Box 443025, Moscow, ID 83844-3025 United States
AU: Kuzmichenok, V A
EM: alex@continent.kg
AF: Institute of Water Problems and Hydro Power, Kyrgyz National Academy of Science, 533 Frunze Str., Bishkek, 720033 Kyrgyzstan
AB: The glacier monitoring on a regional scale have been done traditionally by means of optical space images resulting mainly in observing of changes of area, length and other 2-D information. A lack of texture on the snow fields, steep walls with deep shadows and often cloudiness in high mountains significantly reduce quality and availability of photogrammetrically derived Digital Elevation Models (DEM). Laser altimetry and repeat-pass InSAR DEMs show large potential for glacier volume changes measurements but still have very limited spatial coverage. Recently released 3-arcsecond DEM by NASA-JPL resulted from STRM flown in February 2000 provide unique opportunity for regional-scale glacier change assessment. The method of glacial area and glacier volume changes has been developed over the Akshiirak ice-fields in the Tien Shan Mountains, Central Asia using aerial photography of 1977, topographic maps and RS data of 2000/2003. The datum transformation from WGS-84 used in STRM data to Pulkovo 1942 (Russian) coordinate system was accomplished by 7-parameter Helmert transformation with accuracy at least one order higher than STRM horizontal accuracy (20 m). For vertical validation we compared STRM DEM with DEM constructed from 10 and 5 m additional contour lines digitized from 1:25000 topographic maps on non-glacial relatively flat areas. Though well consistent with 16 m (90%) absolute vertical accuracy, relative accuracy requirement of 6 m (90%) can be easily met only after removing systematic wavy bias in along-track direction. The Akshiirak ice-fields have more than 83% area inclination below $30\deg$. These factors reduce influence of slope-related STRM vertical error to final glacier volume change calculations. Glacier boundaries were manually digitized from an ASTER L1A image acquired on August 18, 2003 that was orthorectified in Orthobase digital photogrammetric package with 9.5 m RMSE of 28 GCPs. For delineating of glaciers in problem areas (debris-covered termini, shadows) we used thermal bands and true hardware-enabled stereo viewing with nadir 3N and backward-looking 3B bands. Accuracy of digitized 2003 glacier boundaries were checked against GPS measurements of 7 glacier termini made in 2002. For surface elevation comparison a second DEM was generated from 10 m contour lines for all glaciers (424 km$^{2}$) using 16 topographic maps of 1:25000 scale created from 1977 aerial photography. The map vertical accuracy is 1/3 of contour interval. Glacier boundaries of 1977 were directly delineated from stereo models. It is revealed that from 1977 till 2003 Akshiirak glaciers have lost 10 km$^{3}$ of ice volume and 35 km$^{2}$ of area.
DE: 1655 Water cycles (1836)
DE: 1699 General or miscellaneous
DE: 1863 Snow and ice (1827)
DE: 1894 Instruments and techniques
DE: 1640 Remote sensing
SC: Hydrology [H]
MN: 2004 AGU Fall Meeting