Comparison of in situ ground temperatures and active layer depths with the ESA CCI+ Permafrost Mean Annual Temperature and Active Layer Thickness products

Abstract:

Arctic regions are especially subjected to climate change and studies have shown that permafrost is warming rapidly across the globe. As these regions are mostly remote and difficult to access, remote sensing or simulation studies allow assessing and projecting temperature over large regions and time scales and increase the knowledge on ground temperature change.

Within the ESA CCI+ Permafrost project (cci.esa.int/Permafrost), we collected in situ ground temperature data from various sources like international permafrost monitoring programs such as the Global Terrestrial Network for Permafrost (GTN-P) managed by the International Permafrost Association IPA, and national permafrost monitoring programs such as the Swiss mountain permafrost monitoring program PERMROS and the large scale Russian hydrometerological monitoring program Roshydromet, and from data sets published in data repositories such as PANGAEA, Nordicana D and the National Science Foundation (NSF) Arctic Data Center. Additionally, data on active layer depths are collected from the GTN-P Circumpolar Active Layer Monitoring Network CALM .and further Principal Investigators. These in situ permafrost monitoring data are then compared to the ESA CCI+ Permafrost products and the accuracy of the products assessed with standard statistics like bias and root mean square error. Prior to the assessment, extensive checks of the in situ data where necessary, as especially coordinates were often inaccurate, resulting in permafrost boreholes positioned in e.g. the open sea. From the corrected datasets, standardized depths were extracted to allow for consistent comparisons. In a future step, interpolations between measured depths will help the end-users to extract the exactly needed depths.

In its first version published in 2019 (CRDPv0), the mean annual ground-temperature (MAGT) raster product across all depths (1, 2, 5 and 10 m) and years (2003-2017) already had a good fit compared to the in situ dataset with a bias of ~ 0.58 °C MAGT and an RMSE of 1.41 °C. Here, we will present the evaluation of the CRDPv1 dataset, containing further in situ data and an updated ESA CCI+ MAGT product.