Ninth International Conference on Permafrost ... - IARC Research

Toward a Permafrost Map of Central AsiaSergei MarchenkoGeophysical Institute, University of Alaska Fairbanks, USAN. SharkhuuInstitute of Geoecology, Mongolian Academy of Sciences, Ulaanbaatar, MongoliaXin LiCold and Arid Regions Environmental and Engineering Research Institute, Chinese Academy of Sciences, Lanzhou, ChinaMamoru IshikawaFaculty of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, JapanJerry Brown<strong>International</strong> Permafrost Association, Woods Hole, MA, USAVladimir RomanovskyGeophysical Institute, University of Alaska Fairbanks, USADmitri DrodzovInstitute of the Earth Cryosphere, Moscow, RussiaIntroductionAlthough national permafrost maps exist for China,Kazakhstan, Mongolia, and Russia, there is no consistentcartographic or temperature criteria on which to base aunified permafrost map for the more topographically complexregions that prevail in Central Asia. The <strong>International</strong>Permafrost Association’s (IPA) “Circum-Arctic Map ofPermafrost and Ground-Ice Conditions” (1:10,000,000)employed an international legend developed primarily forcontinental (lowland) permafrost regions (Brown et al.1997). That classification was applied to mountainous andhigh altitude regions with considerable uncertainties.The retreat of glaciers and permafrost degradationin Central Asia in recent years is unprecedented asa consequence of warming. Accelerated warming ofpermafrost in mountainous, highland, and plateau regionsof Asia could result in the disequilibria of the water cycle,increased mass wasting processes, and related sedimenttransport and slope hazards. Without a unified and verifiedregional permafrost map, these processes cannot be assessedadequately. Mapping, modeling, and monitoring strategiesin mountain regions are under development to test and toverify climate-change scenarios and models.In response to the difficulties involved in classifying andmapping of the region’s permafrost, recommendations wereapproved at the IPA-sponsored <strong>International</strong> Symposium onMountain and Arid Land Permafrost, Ulaanbaatar, Mongolia,September 2001 (Brown 2001). These included a request thatan international team of experts prepare a unified permafrostmap of Central Asia.The workshop on the distribution and mapping of thepermafrost distribution of Central and Eastern Asia was heldprior to the Asian <strong>Conference</strong> on Permafrost, in Lanzhou,China, August 5–6, 2006, and was hosted by the Cold andArid Regions Environmental and Engineering ResearchInstitute (CAREERI). The workshop participants agreedthat conventional mapping and modeling approaches ofpermafrost in this region recognize both latitudinal andaltitudinal permafrost zonation, furthermore, and that thecategory of mountain permafrost be recognized as a subsetof altitudinal zonation or, if appropriate, remain as a separateclass. The proposed map should delineate each of thesepermafrost zones and contain actual and calculated groundtemperatures and active layer thickness as point observations.Where available, estimates of ground ice would be included.However, for this next-generation map, classes of ground iceand spatial continuity (percentages) will not be part of theclassification.Mapping of Permafrost in Central AsiaPermafrost classificationsA major difficulty is in reconciling the usage of “continuouspermafrost zonation,” as it is commonly applied tocontinental permafrost (greater than 90% of the land surface;see Heginbottom 2002 for a comparison of classificationschemes). A comparison of the spatial classifications in usein the four countries is presented in Table 1.Modeling approachAn alternative approach of altitudinal permafrost mappingis modeling the ground temperature and permafrostdistribution using the process-based models. Such anapproach allows for spatial and temporal extrapolation ofpermafrost thermal state and distribution and also is wellsuited for studies with respect to permafrost response toclimate change. But the process-based model requires anextensive set of input data such as meteorological data,surface characteristics (vegetation, snow cover), groundthermal properties, and topography. For the modeling ofaltitudinal permafrost within the rugged topography of theAltai Mountains, the basic dataset at 100 m resolution for thedigital elevation model (DEM) was generated. The spatialpermafrost model with a grid box size of 5 km uses griddedfields of monthly air temperature (topographically adjusted203