Present and LGM permafrost from climate simulations: contribution of statistical downscaling

Abstract

We quantify the agreement between permafrost distributions from PMIP2 (Paleoclimate Modeling Intercomparison Project) climate models and permafrost data. We evaluate the ability of several climate models to represent permafrost and assess the inter-variation between them. <br><br> Studying an heterogeneous variable such as permafrost implies to conduct analysis at a smaller spatial scale compared with climate models resolution. Our approach consists in applying statistical downscaling methods (SDMs) on large- or regional-scale atmospheric variables provided by climate models, leading to local-scale permafrost modelling. Among the SDMs, we first choose a transfer function approach based on Generalized Additive Models (GAMs) to produce high-resolution climatology of air temperature at the surface. Then, we define permafrost distribution over Eurasia by air temperature conditions. In a first validation step on present climate (CTRL period), this method shows some limitations with non-systemic improvements in comparison with the large-scale fields. <br><br> So, we develop an alternative method of statistical downscaling based on a Multinomial Logistic GAM (ML-GAM), which directly predicts the occurrence probabilities of local-scale permafrost. The obtained permafrost distributions appear in a better agreement with data. In average for the nine PMIP2 models, we measure a global agreement by kappa statistic of 0.80 with CTRL permafrost data, against 0.68 for the GAM method. In both cases, the provided local information reduces the inter-variation between climate models. This also confirms that a simple relationship between permafrost and the air temperature only is not always sufficient to represent local-scale permafrost. <br><br> Finally, we apply each method on a very different climate, the Last Glacial Maximum (LGM) time period, in order to quantify the ability of climate models to represent LGM permafrost. The prediction of the SDMs is not significantly better than large-scale fields with 0.46 (GAM) and 0.49 (ML-GAM) of global agreement with LGM permafrost data. At the LGM, both methods do not reduce the inter-variation between climate models. We show that LGM permafrost distribution from climate models strongly depends on large-scale air temperature at the surface. LGM simulations from climate models lead to larger differences with permafrost data, than in the CTRL period. These differences reduce the contribution of downscaling and depend on several other factors deserving further studies

Postglacial recolonization in a cold climate specialist in western Europe: patterns of genetic diversity in the adder (Vipera berus) support the central–marginal hypothesis

“…This species has particularly low dispersal abilities, and populations separated by only a few kilometres are significantly differentiated (Ursenbacher et al 2009). The common adder is a cold-adapted species with a large boreal distribution (probably the largest for a terrestrial snake; Saint Girons 1980) reaching high latitude (>60°N) and the permafrost seems to limit its distribution (Nilson et al 2005;Levavasseur et al 2011). The range-wide mitochondrial DNA phylogeography of the adder indicated three major clades: one in the Balkan peninsula, one in the Italian peninsula and one with a large distribution from France to Sakhalin Island (Kalyabina-Hauf et al 2004;Ursenbacher et al 2006 Clark et al 2009).…”

Climate control on the evolution of Late Pleistocene alluvial‐fan and aeolian sand‐sheet systems in NW Germany

“…During MIS 2 (Late Pleniglacial, 27-15 ka) the temperatures decreased and arid conditions with strong winds prevailed (Bos et al 2001;Guiter et al 2003). The coldest period with polar desert conditions and scarce vegetation occurred c. 20 ka (Huijzer & Vandenberghe 1998;Guiter et al permafrost (Huijzer & Vandenberghe 1998), although modelling results of Levavasseur et al (2011) point to an absence of permafrost during the Last Glacial Maximum (LGM, c. 21 ka) in central Europe. The end of MIS 2 (Lateglacial, 15-12 ka) was characterized by alternating cold (arid) and warm (humid) periods, as indicated by pollen data and soil formation (Kasse 2002;Guiter et al 2003).…”

Mid‐Holocene frozen ground in China from PMIP3 simulations

“…In comparison, the role of soil depth is second to that of horizontal resolution in influencing the abilities of models. Previous air temperature-based studies indicated that the distribution of permafrost might be moderately refined in heterogeneous mountainous regions by improving the method of downscaling air temperature (Levavasseur et al 2011;Saito et al 2014). Levavasseur et al (2011), who investigated permafrost on the basis of the air temperatures obtained from the PMIP2 models, suggested that a sophisticatedly downscaled temperature field could reduce the biases in the Qilian, Tianshan and Altai mountains and on the Tibetan Plateau.…”

“…G-I are for annual thawing index (cumulative degree days above 0°C within a year; units:°C day). [Colour figure can be viewed at www.boreas.dk] air temperature-based permafrost was strongly dependent on large-scale air temperature at the surface (Levavasseur et al 2011). In other words, downscaling is limited by large-scale errors from climate models in the last glacial maximum period.…”