Permafrost is a subsurface phenomenon and cannot be directly observed with satellite data. Yet, monitoring can be done based on indicators and via permafrost models. Indicators are especially thermokarst lake dynamics and surface elevation changes. Those phenomena need to be observed on a local scale. Regional to circumpolar monitoring requires the use of permafrost models. Relevant satellite-observable parameters are land surface temperature (LST), snow extent, snow water equivalent (SWE), vegetation, and soil moisture. Existing services have been integrated into the processing system and adapted to the needs of permafrost modelling. All datasets are freely accessible via PANGAEA and a WebGIS which will tie into the permafrost information system of the International Permafrost Association (IPA).

Land surface temperature | Snow | Land cover | Terrain | Landsurface hydrology


Land surface temperature Land surface temperature

The rate at which permafrost evolves can be determined by studying its thermal regime, which is dependent on surface temperature. Surface temperature is a key parameter as it governs the surface energy budget and the thickness of the permafrost active layer. The LST processing subsystem integrates the LST level 2 products from MODIS and AATSR distributed by NASA and ESA, respectively. Post-processing functions supply UW level-3 weekly and monthly LST products for regional and pan-Arctic scales. The main components of the processing subsystem are: Interpolation to regular grid, Spatial averaging, Temporal averaging and weekly LST. The weekly LSTs are calculated from all satellite overpasses within a seven- day period based on aggregation of daily products. It is available for each day based on a 7-day sliding time window giving most recent observations highest pri-ority following the GlobSnow convention. The monthly LSTs are calculated from all satellite overpasses within a calendar month period. It was noticed that the UW level 3 AATSR product was not consistent over the entire year. The product reached the largest deviation from air temperature and MODIS LST around the month of July at both sites, but performed well for other times of the year.



The amount of snow determines insulation properties. An operational monitoring service for snow extent and SWE has been set up within the ESA DUE project GlobSnow.

Land cover

Vegetation is commonly incorporated into spatial models predicting permafrost distribution. The land cover and the surface texture are affected by the seasonal thawing dynamics of the uppermost permafrost layer (active layer).
The yearly MODIS land cover product with a spatial resolution of 500 m, the GlobCover land cover map, SYNMAP and MODIS VCF (vegetation con-tinuous field) have been combined into one land cover dataset. The datasets consist of four layers, describing the percentage information for each class, with a spatial resolution of 1 km. By summarizing all four layers each pixel ends up with a value of 100 %. The harmonized land cover map was improved by using the Circumpolar Arctic Vegetation Map.
RapidEye Mosaic Lena Delta To represent the seasonal vegetation dynamics on pan-arctic scale the LAI product from GlobCarbon with a spatial resolution of 1 km was utilized. Pan-arctic fire information is presented by using different burned area and active fire products (MODIS, GlobCarbon, Terra Norte, ATSR World Fire Atlas).
Subject of the local land cover analysis were three different test sites: Central Yakutsk and Lena river delta (Siberia) as well as North Slope (Alaska). The land cover classification was done by utilizing an object based classification approach. Object characteristics (shape, spectral properties and information within different hi-erarchical object levels) are used to analyze vegetation class properties and to assign each image object to a thematic class. Land cover was analyzed using RapidEye data.



Seasonal subsidence on North Slope, Alaska Information on surface topography and on change in surface topography is fundamental over permafrost regions. According to the characteristics and possibilities of Earth-Observation (EO) technologies we distinguish within our project between Digital Elevation Models (DEM), on one side, and surface subsidence, on the other side. Ice-rich layers are usually close to the surface and are the first to be melted by increases in downward heat energy flux due to changes in the surface energy balance. The immediate result is sub-sidence. The DEM is compiled at pan-arctic scale from available data sources and derived at local scale from optical stereoscopic pairs and SAR interferometry (InSAR).
The Circum-Arctic DEM was compiled with a 3 arcsec spatial resolution. SAR interferometry turned out to be a reliable tool to detect seasonal surface subsidence due to permafrost thaw on many regions thanks to the short repeat interval of 11 days of TerraSAR-X. The time-series of displacement highlighted that subsidence is oc-curring within a relatively short time period. In our investigations we found coherent annual interferograms only using the low frequency ALOS PALSAR data, but these interferograms were largely contaminated by ionospheric artifacts.


Land surface hydrologyPolar view of soil moisture  from METOP ASCAT data of July/August 2007 and 2008.

Variations in parameters which impact heat conductivity play a role in the reaction of the subsurface frozen ground to changes in the atmosphere. Soil moisture information is one of the key parameter for modelling of permafrost extent. The moisture regime is important for active layer development. Soil moisture together with temperature is also a limiting factor for heterotrophic soil respiration. Water bodies or thaw lakes are an important mechanism of landscape modification in the arctic. Water is a class in all available global and regional land cover maps. The spatial resolution of those existing products ranges between 300 m and 1 km. The majority of lakes within the tundra environment is however much smaller than the spatial resolution of those maps.
The ASCAT Level 2 product including soil moisture data are produced by EUMETSAT in near-real time following the method developed and proto-typed for EUMETSAT by the Institute for Photogrammetry and Remote Sensing of the Vienna University of Technology.
The Surface Soil Moisture parameter represents a relative measure of the soil moisture in the top layer of the soil, scaled between 0 and 100%. The ASCAT SSM DUE Permafrost product is the result of an im-proved SSM retrieval algorithm. The SSM Product is delivered with a weekly temporal resolution and 25km spatial resolution.
For integration into the Per-mafrost Information System (PEO) ASCAT data are resampled to a Discrete Global Grid (DGG). The soil moisture product also includes a quality flag which contains the number of used measurements. Data are masked for frozen ground conditions also based on MetOp ASCAT. The product is provided as weekly averaged images north of 50°N in Geo-TIFF/NetCDF format and EASE Grid projection. The initial freeze/thaw status flag has been implemented initially for the weekly averaged surface soil moisture service. It has been improved in cooperation with the ESA STSE ALANIS-Methane project in order to reflect daily status and also snowmelt conditions.
Regional service for surface soil moisture (1km), open water summer extent (150m) and surface status (1km) has been implemented based on ENVISAT ASAR. The processing has been implemented with the TU Wien SGRT. For application at high latitudes the following adjustments to the algorithm have been implemented: (1) Processing (with NEST) in polar stereographic projection and storage of data. An entirely new gridding system has been setup in order to avoid oversampling and to reduce data storage; (2) A dry reference (representing wilting point) correction algorithm has been implemented in order to account for permanently wet areas; (3) Implementation of a post processing function for the production of weekly composites of surface soil moisture and enhanced masking with respect to signal-to-noise ratio and water fraction. Availability of ASAR Global Mode and Wide Swath data is highly variable in the high latitudes and needs to be accounted for. Auxiliary maps with numbers of measurements have been therefore derived as quality indicator.