3.4.2 Ground Temperature

As such, ground temperature is a challenging metric of permafrost change as it often requires the counter-intuitive and difficult interpretation of relating less change to a presumed greater thaw. Ground temperature data is often not shared or shared in non-standardized form and only for individual sites and studies. As such, the current compilation is incomplete and represents a starting point for future work. Work toward data interoperability and standards is underway in NSERC PermafrostNet (https://www.permafrostnet.ca/data/). Ground temperature, along with other relevant permafrost conditions, varies widely over scales of metres to tens of metres. This makes the interpretation of observations difficult even for nearby locations and it also makes interpretation of gridded data in a local context difficult. Ground temperature data can support local planning and engineering design, the quantification and analysis of change, and the testing of computer models.

3.4.2.1 Data sources

Global Terrestrial Network for Permafrost (GTN‐P): http://gtnpdatabase.org/boreholes GTN-P is the primary international program concerned with monitoring permafrost parameters. This database provides metainformation for borehole sites as well as some time-series data.

Permafrost Ground Temperature for the Northern Hemisphere, v3.0 (Westermann et al., 2020): https://catalogue.ceda.ac.uk/uuid/b25d4a6174de4ac78000d034f500a268 Mean annual ground temperatures at the ground surface and 1 m, 2 m, 5 m and 10 m depth are derived from CryoGrid 3. In the data product A, the model is driven by a combination of MODIS Land Surface temperature observations and downscaled ERA near-surface air temperature data. In data product B, the same thermal model is driven by ERA5 near-surface temperature data. CryoGrid 3 includes land-atmosphere coupling and a subsurface heat transfer model. Snowfall and snowmelt are represented by a degree-day based snow model driven by downscaled ERA5 reanalysis. The land cover is expressed as a fractional coverage of seven land cover classes (ESA landcover_cci). Each 1x1 km pixel is using an ensemble approach on land cover fraction to statistically account for natural variability of snow depth. The model has been tested at permafrost sites in Russia and Norway. Uncertainties are introduced in the pre-processing step of the input dataset and through simplified process representation (due to computation cost). The scarcity of validation sites for input parameters, such as stratigraphy, introduces an additional factor of uncertainty.

Borehole and near-surface ground temperatures in northeastern Canada: http://www.cen.ulaval.ca/nordicanad/dpage.aspx?doi=45291SL-34F28A9491014AFD Nordicana-D data are from 11 sets of locations in Nunavik and Nunavut, northeastern Canada. The available datasets cover the period from 1988 to 2019. There are a total number of 46 boreholes with the depth ranging from 0.02 to 20 m. Almost half of the boreholes have near-surface depth (2–5 cm) and 12 boreholes are drilled deeper than 10 m. Among those, 29 boreholes cover Akulivik, Aupaluk, Kangiqsualujjuaq, Puvirnituq, Quaqtaq, Tasiujaq, Umijuaq, Iqaluit, Pangnirtung and Île Bylot regions while other 17 boreholes are located in Salluit region. 14 boreholes have more than 20 years of data while 15 boreholes have data records shorter than 15 years.

A map and summary database for permafrost temperatures in Nunavut:

https://geoscan.nrcan.gc.ca/starweb/geoscan/servlet.starweb?path=geoscan/fulle.web&search1=R=292615. This dataset is from 20 sites in Nunavut where measurements were made in 111 boreholes located within a few tens of metres to a few kilometres of each other. Borehole depth range from 3 to 36 m with 56 boreholes deeper than 20 m and 22 less than 10 m depth. The mine site boreholes account for almost 80% of the boreholes included in the database. Data on ground temperature has been compiled into a MS Excel Spreadsheet (NUGT.xls). Most of the data is available for 2010s.

Air and near-surface ground temperatures Mackenzie Valley Corridor: https://doi.org/10.4095/294835 This dataset included near-surface ground temperatures, indices and summary statistics from 1993 to 2012 for the Mackenzie Valley Corridor, Northwest Territories.

NWT Permafrost Database: Contains ground temperature data collected for a variety of purposes including but not limited to linear and vertical infrastructure, government and academic research sites and community planning. Represents ground temperatures from boreholes with varying depths and captures relevant terrain information useful for interpreting ground temperature data. The database is currently in the testing phase and development of the web portal will begin soon. This will make NWT ground temperature data available for download to the public (Contact Niels Weiss: niels_weiss@gov.nt.ca)

Yukon Permafrost Database: ( https://service.yukon.ca/permafrost/, contact Derek Cronmiller: Derek.Cronmiller@yukon.ca) Contains 80 sites currently with data that includes roads, airstrips, mining projects and academic research sites.

3.4.2.2 Strengths and limitations of datasets

Station data: Ground temperature data associated with station data represents a single point on the landscape and is unlikely to reflect conditions at a regional scale. Relationships between station data and surface and terrain information may be used to extrapolate relations but caution is required as ground temperature is very site-specific. Station data provide accurate information on local conditions.

Gridded data: Ground temperature data is represented by spatially defined grids. It can be derived from interpolated station data, thermal models and satellite data. Limitations include regional biases due to poor spatial coverage of station data and model shortcomings. The spatial resolution of gridded data will depend on the spatial resolution of the sensor.

Reanalyses and Reanalysis-Based Datasets: Ground temperature data available as part of reanalyses has not been systematically evaluated for permafrost areas. ERA5-Land, and likely ERA5, exhibit strong and spatially variable bias (Cao et al., 2021).

References – Ground Temperature

Cao, B., S. Gruber, D. Zheng, and X. Li, 2020: The ERA5-Land Soil Temperature Bias in Permafrost Regions. The Cryosphere, 14(8), 2581-2595, doi.org/10.5194/tc-14-2581-2020.

Westermann, S., A. Bartsch, and T. Strozzi, 2020: Permafrost - Algorithm Theoretical Basis Document. European Space Agency Contract Report: 1–30, Retrieved from https://climate.esa.int/en/projects/permafrost/key-documents/.

Westermann, S., M. Langer, J. Boike, M. Heikenfeld, M. Peter, B. Etzelmüller, and G. Krinner, 2016: Simulating the Thermal Regime and Thaw Processes of Ice-Rich Permafrost Ground with the Land-Surface Model CryoGrid 3. Geoscientific Model Development, 9(2), 523-546, doi.org/10.5194/gmd-9-523-2016.