ePIC

Observation of thaw and freeze processes in an arctic tundra landscape with TerraSAR-X and RADARSAT-2 data

Edit Item Edit Item

General Information:

Citation:
Sobiech, J. , Heim, B. , Boike, J. and Dierking, W. (2012): Observation of thaw and freeze processes in an arctic tundra landscape with TerraSAR-X and RADARSAT-2 data , Final User Workshop ESA Data User Element DUE Permafrost, Potsdam, Germany, 15 February 2012 - 17 February 2012 .
Cite this page as:
Contact Email:
Download:

Supplementary Information:

Abstract:

Space-borne imaging radar is a suitable tool for the detection of thaw and freeze processes, as the backscattering of radar waves is highly dependent on the dielectric properties of the surface. The backscattered intensity from frozen and unfrozen wet soils in Siberia differed by 3 to 4 dB using C-band radar. The timing of ice formation and melt on lakes and rivers has successfully been determined from space-borne SAR data. The study site is located in the southern central part of the Lena River Delta, Northern Siberia, Russia, at 72°N, 126°E. The area is a polygonal arctic tundra landscape, located in a zone of continuous permafrost and covered by ponds, lakes (15 m2 to 1.3 km2) and river arms. A time series of twelve TerraSAR-X images (Stripmap mode) and six RADARSAT-2 (Fine quad-pol) images acquired during spring 2011 and nine RADARSAT-2 images recorded during fall 2011 were analyzed for this study. The dates of snow-melt, lake- and river-ice decay and onset, and soil thaw / refreeze obtained from the SAR-based study were compared to meteorological data from Samoylov Island, located in the investigation area, and Tiksi, located about 110 km southeast of the study site. Soil thaw and refreeze was investigated in regions with different soil properties and vegetation cover. Temporal variations of the mean backscattered radar intensity were analyzed for the different land units. The results show that snow-melt events are clearly displayed in the mean backscatter values of the ROIs within the time series. The thaw of the active layer could not be observed in the X-band data, in which the mean backscatter values are mainly the same for both, the frozen and the thawed landscape. The backscatter values of the C-band data are about 3 dB higher during the thaw period than during the freezing period, and vice versa for the refreeze of the soils. Lake- and river-ice decay and refreeze was monitored in the frame of this study as well. Ice-decay could be observed in both, the TerraSAR-X and the RADARSAT-2 data. Classification methods were applied to distinguish between ice and open water surfaces of lakes and river arms. The results show that lake-ice decay is not only dependent on lake size and location, as the timing of lake-ice decay of neighboring lakes of the same size was different in some cases. Thus it is assumed that lake depths are different, but this needs confirmation from other data sources. A very high spatial image resolution is useful to distinguish land surfaces and water bodies. The ice cover on lakes remains in spring several weeks after the active layer thaw has started. In fall, the ice-onset in the water bodies is later than the refreeze of the soils. As snow-melt, refreeze, thaw processes, and ice-decay on water bodies happen quite quickly, it is necessary to acquire an image series with a high temporal resolution.

Further Details:

Imprint
AWI
Policies:
read more
OAI 2.0:
http://epic.awi.de/cgi/oai2
ePIC is powered by:
EPrints 3