Permafrost landscapes in transformation - from local-scale processes to the global model NorESM

In Arctic and sub-Arctic regions of the Earth, vast areas are dominated by permafrost. Like a gigantic freezer, the frozen ground preserves large amounts of organic material from decomposition. Already now, permafrost is thawing in many regions, and climate change will likely continue to weaken the permafrost freezer in the future. Consequently, decomposing organic material can release carbon dioxide and methane in the atmosphere, which can accelerate climate change even further in a process termed the permafrost-carbon feedback. In the Permanor project, we have investigated permafrost landscapes rich in ground ice for which thawing can occur much faster than for ice-poor permafrost. Rapid transformations of the surface are triggered by melting of ice layers in the ground which can for example lead to surface subsidence, sideways erosion or new ponds to form. Some of the most carbon-rich permafrost areas are also extremely rich in ground ice, such as Pleistocene "Yedoma" permafrost in Siberia and Alaska, as well as subarctic permafrost peatlands covering vast areas in Canada, Russia and Scandinavia. With field sites in both cold permafrost in Siberia and much warmer permafrost in Norway, Permanor has focused on a unified understanding of thaw processes in these areas which is key to a better understanding of the permafrost-carbon feedback on the global climate. In Permanor, we have made significant progress to improve Earth System Models (ESMs) and other permafrost models. In a series of scientific papers, we have introduced the concept of "laterally coupled tiling" which for the first time made it possible to simulate how the microtopography in permafrost areas evolves as a result of climate change. Traditionally, tiling is employed in ESMs to represent different landcover types within a large-scale modeling grid cell, such as lakes, forests and arable land. However, these tiles evolve independent of each other in time, for example runoff created in one tile does not affect the soil moisture in the other tiles. Interactive tiles, on the other hand, can exchange lateral fluxes of heat, water and snow, so that the runoff from e.g. an upland tile can increase the soil moisture content in an adjacent lowland tile. As demonstrated in publications by Permanor scientists Kjetil Schanke Aas, Jan Nitzbon and Léo Martin, "laterally coupled tiling" is a universal concept that can be implemented in different model tools and that can reproduce observations permafrost thaw in both in Siberia and Scandinavia. In particular, Jan Nitzbon could show that even extremely cold regions in Northern Siberia will experience thawing in the 21st century if the effects of the ice-rich permafrost is taken into account. Similar model improvements were demonstrated for permafrost peatlands in Northern Scandinavia, based on the excellent field data sets from Permanor field campaigns in Finnmark, Norway. Also here, interactive tiling has proven extremely useful - with the method, it is for the first time possible to simulate under which conditions permafrost peatlands are stable, and by which processes they degrade. A final study is in preparation which investigates future greenhouse gas emissions from thawing ice-rich permafrost with "laterally coupled tiling ". For dissemination, we visited the local highschool in Northern Norway to tell the students about our scientific work near their home, relating it to the global picture of climate change impacts. With visits in Karasjok, Kautokeino, Alta and Lakselv, several hundred students could be reached. The strongly positive response of the students, who will become the future "permafrost stakeholders" in their municipalities, is a strong motivation to continue this form of outreach in the future.

-Multi-year field studies in permafrost peatlands in Northern Norway (Finnmark) have delivered data sets for model validation. The new research sites have attracted the interest of a wide scientific community. -Students at high-schools in the vicinity of the research sites have learned about permafrost, climate change and the PERMANOR activities. -A statistical approach (referred to as "laterally coupled tiling") that can represent rapid permafrost thaw processes (so-called thermokarst) has been developed. -Laterally coupled tiling has been implemented in several land surface models. The same model set-up can reproduce rapid permafrost thaw in regions with very different climates. -When including rapid thaw in models, even today's very cold permafrost in Siberia can thaw in case of a high emission scenario (RCP8.5). -There is emerging evidence that projections of future greenhouse gas emissions from thawing permafrost must be revised when rapid thaw processes are considered.

Permafrost is a highly dynamic element of the Cryosphere that is intimately meshed with the global climate system through complex interactions. A massive release of greenhouse gases from thawing organic-rich permafrost could constitute a significant additional warming potential, so that capturing permafrost thaw is among the key priorities for reducing uncertainty in future projections in Earth System Models (ESMs). PERMANOR will for the first time systematically investigate the representation of permafrost in the Norwegian Earth System Model NorESM by bringing together expertise in observations, process based modeling, model downscaling, and ESM development. PERMANOR will bring the highly dynamic evolution of permafrost landscapes in the focus of ESM development. Many observed forms of permafrost thaw, e.g. the development of thermokarst ponds or the disintegrations of subarctic peat plateaus, cannot be explained without lateral transport of energy, water and snow on spatial scales that are several orders of magnitude smaller than a typical ESM grid cell. To reconcile these different scales, the project will pioneer a two-stage concept by aiming for a statistical representation of small-scale processes, while WRF (Weather Research and Forecasting) modeling will facilitate upscaling of parameterizations to ESM scale. A strong focus will be laid on conceptualizing interactions between microtopography, ground ice content and snow depth, with the goal of developing a statistical representation of permafrost landscape evolution. As a spin-off, climate projections will be downscaled to 1km for key regions in N Scandinavia, Svalbard, and Siberia. Ultimately, PERMANOR will conceptualize process understanding from in-situ studies to develop new model algorithms and pursue their implementation in a coupled ESM framework. The improved representation of permafrost landscape dynamics in ESM will lead to reducing our uncertainty in the predictability of future climate change.