Satellite images and aerial photographs from the Arctic are beautiful to look at. The spatio-temporal changes which become visible when comparing multiple images, however, are not only very exciting visually but convey important scientific information on the stability of permafrost and changes in the northern the carbon cycle. The Siberian and Alaskan key study regions of the PETA-CARB project are geographically wide spread, geologically and climatically different, and cover a wide range of permafrost characteristics. In addition to their remoteness and difficult accessibility, all of them are vulnerable to rapid external disturbances due to the presence of very ice-rich permafrost close to the surface.
In our research, we apply different optical remote sensing methods for continuous monitoring of Arctic permafrost regions using a wide range of space and airborne image records. Spatial, temporal, spectral, and radiometric resolution determine the particular fields of application. Landsat data of medium spatial resolution are used for long-term and broad-scale time series analysis. Since adjacent orbits strongly overlap in northern high latitudes the temporal resolution of these data can be increased in order to derive robust trends of waterlogging, surface reflectance, and vegetation indexes. The very high repeat rate of the RapidEye satellite constellation, providing at the same time high spatial resolution and large areal coverage, allows capturing the strong seasonal variability of the Arctic tundra. Snow melt, lake ice out, and different phenological stages during only three summer months, as well as freeze back are being recorded year after year to obtain an understanding of the seasonal variability and to eventually differentiate long-term changes in the context with other data sets.
Mosaicking of adjacent and complementary scenes enables comprehensive multi-spectral classifications which are then used for mapping of typical landscape units and subsequent quantification of soil organic carbon stored in permafrost soils. Accurate documentation of the conditions on site and the selection of transects for soil and sediment sampling in the field play a central role here in order to capture all relevant landscape units and to achieve a high classification quality. As a ubiquitous element of the landscape, thermokarst lakes cover not only a large part of the surface, but also have a strong influence on the thermal characteristics of the surrounding permafrost because of the high heat storage capacity of water. A comprehensive inventory of thermokarst lakes in ice-rich permafrost and the documentation of current lake expansion, tapping, and drainage phenomena is part of our remote sensing studies.
Precise topographical reference measurements on the ground are also fundamental to exploit the great potential of very high resolution satellite images such as from Kompsat, Ikonos, QuickBird, GeoEye, and WorldView for detailed analyses of land subsidence caused by permafrost thaw and related mass movements. Using digital image processing techniques, stereo images are photogrammetrically analysed to derive digital elevation models (DEM). The DEMs are used in turn to correct for terrain-induced distortions in the images themselves, as additional topographic information for classification purposes, as well as for the quantification of erosion volumes and associated sediment and carbon fluxes. Geodetic repeat measurements on instrumented monitoring grids in our key investigation areas provide control data sets for remote sensing time series of elevation change, caused by permafrost thaw-related processes. Therefore, constant improvements of the data basis as well as of the applied methods allow a more precise picture of the rapid ongoing changes in the Arctic permafrost regions.