Mikhail Zhelezniak，Aleksandr Zhirkov，Anatolii Kirillin

(俄罗斯科学院西伯利亚分院麦尔尼科夫冻土研究所，雅库茨克677010，俄罗斯)

0 Introduction

The Stanovoy Upland is situated in Central Asia，in the southern part of the Siberian Platform.Its permafrost is relatively well studied compared to other mountain regions of the Earth.Permafrost investigations were initiated here in the 1950 s by Sumgin M.I..Detailed studies were undertaken during the 1970 s to 1990s by Nekrasov I.A.and other investigators.These studies provided an understanding of permafrost distribution in the region reflected in schematic geocryological maps.However，not enough is yet known about the subsurface temperature field and permafrost thickness owing to a number of difficulties (need for special borehole preparation，long recovery of the ground thermal equilibrium after drilling，high piezometric groundwater level，etc.).

Seasonal frost and permafrost affect human activities in mountain regions and foothill areas.Frozen ground and related slope processes can strongly increase mining operation costs，as well as threaten environmental safety and infrastructure integrity.

The subsurface temperature field and permafrost depth are mainly controlled by mean annual ground surface temperature，thermal properties of soils and rocks，and geothermal heat flow［1］.In the mountain regions，high elevation and rugged topography exert a significant influence on microclimate，heat exchange at the ground surface and geothermal heat flow redistribution，resulting in spatial variability of permafrost parameters.

1 Results and Discussion

The available data forthe Stanovoy Upland indicate that ground temperatures at the depth of zero annual variation range from +2.0 to－8.0℃，the thermal conductivities of main rock types range from 2.40 to 6.40 W/m·K，and the geothermal heat flux varies from 35 to 65 mW/m2over the region.These parameters in combination dictate the occurrence，distribution and thickness of permafrost.The geocryological data indicate that permafrost is widespread at elevations above 1 300 m.The maximum permafrost thickness obtained by geothermal measurements is 616 m in the Udokan Ridge at an elevation of 1 900 m.The estimated maximum thickness is 1 050 m in the high (3 000 m)mountains of the Kodar Ridge［2-3］.

A geocryological database of the region has been developed at the Melnikov Permafrost Institute which compiles geothermal data from 45 sites(over 1 000 boreholes 20 to 1 200 m in depth).Based on the measured data，a series of geothermal cross-sections have been constructed depicting the subsurface temperature field and permafrost thickness in the region (Fig.1－3).Three hypsometric levels have been distinguished in the Stanovoy Upland and surrounding structures which differ in permafrost distribution.The lower level lies between 200 and 600 m isohypses and has continuous permafrost.The middle level，between 600 and 1 300 m isohypses，is the zone of transition from the crystalline shield to the platform and has discontinuous，frequently thin(50～100 m)permafrost.The upper level lies above 1 300 m where the permafrost distribution is continuous.There is a clear altitudinal trend in permafrost temperature and thickness.Each of these zones has been found to have distinctive permafrost landscapes.

Fig.1 Borehole ground temperatures in the central part of the Stanovoy Upland(Tarynnakh and Cheena sites)，14.525，etc.-the number of exploration borehole

Fig.2 Geothermal cross-section for the permafrost zone in the Olekma-Chara Uplift

Fig.3 Geothermal cross-section for the permafrost zone in the Cheena River headwa1ters，Kalar Ridge

Based on ground temperature data for different land features in the Stanovoy Upland，terrain analysis has made it possible to determine correlations and describe regional relationships between ground temperature at the depth of zero annual amplitude(t0)and the main terrain factors for permafrost mapping purposes.While mapping criteria based on relationships among permafrost and terrain factors are well developed for the plateaus in the central part of the Aldan-Uchur Uplift［4-5］，the issue remains open for the areas of very rugged topography.

From analysis of the measured ground temperatures(t0)in the mountainous areas of the Stanovoy Upland，the basic regional factors controlling t0appear to be altitude(H)which controls climatic and vegetation zonation，slope aspect(φ)and slope angle(α) which affect the surface energy balance，and type and properties of earth material.

Having examined the incoming solar radiation patterns within the latitudes from 58°to 62°N［6-7］for different surface inclinations and slope aspects，we calculated inclination and aspect indices which express a relative numerical measure of solar radiation incident on slopes of different aspects.This allowed for the estimation of effects of slope and aspect on ground temperatures［8］.In the present study，we have performed simple and multiple correlation analyses using data from 70 boreholes located on sites with different slope，aspect，and altitude.Because temperature inversion is widespread in intermontane depressions and valleys located at different altitudes，no consistent positive relationship exists between t0and H.Excluding the foothills and lower slopes which experience t0inversion，the simple correlation analysis has shown a consistent positive correlation of t0with H，φ，and(for similar landforms in the middle and high mountains，with a correlation coefficient of 0.92 to 0.99.Some of the regression equations(R =0.85－0.95)obtained for the slopes of different aspects are given below:

The multi-factor correlation analysis performed on the same principle as the simple correlation analysis has shown the existence of a relationship between t0and H，φ，and α with R=0.98:

In the mountainous regions，more severe permafrost conditions exist in areas of very rugged topography，where even at elevations of 1 100～1 300 m permafrost may be as thick as 300～450 m.Terrain ruggedness commonly increases with increasing elevation in mountains.This causes redistribution of heat flow which plays an important control on permafrost thickness in mountainous regions［9］.Altitudinal permafrost zonality is the main criterion in mountain permafrost mapping.This criterion is more consistent in high mountain areas，whereas in foothill areas this relationship is not always straightforward and identical and in most cases is regionally specific.

Altitude(H)is one of the primary integral parameters that control permafrost temperature and thickness (h)in the mountains and foothills.The correlation coefficient between these parameters is 0.47，suggesting that there is no direct correlation.However，if we differentiate the region by geomorphological conditions，there are clearly marked areas in the middle and high mountains which exhibit altitudinal zonality.

Terrain profiles in the high mountains are morphologically similar，while in the middle mountains they show significant variability which，through the set of interrelated factors，determines the characteristics of permafrost.For example，no direct relationship has been found between permafrost thickness and altitude (R = 0.31)for the sites located in the middle mountains.This is because permafrost thickness in the middle mountains(1 000～1 600 m)is determined，to a larger extent，by available relief and slope aspect.

Fig.4 A scatter diagram of lower altitudinal limit of permafrost(Hp)versus altitude(H)on water divides.

Water divides show similar conditions for the temperature field formation.Excluding the plateaus，a consistent correlation between h and H(R=0.91)has been obtained for the Olekma-Chara Upland which can be described by the regression equation:

For this region where altitudes range from 920 to 2 000 m，the coefficient of correlation between h and H is 0.96 and the regression equation has the form:

Thus，permafrost thickness in the water divides underlain by continuous permafrost increases with altitude following the above equation.Based on this correlation，permafrost may be as thick as 850 m at 2 400 m elevation in the Udokan and Kodar Ridges and 1 100 m at 3 000 m elevation in the Kodar Ridge.

A higher linear correlation(R=0.97)and a lower standard deviation(σ)have been found to exist between the lower altitudinal limit of permafrost(Hp) and H(Fig.4)，which can be described by the regression equation:

Using data from 42 sites located on slopes with different aspect，angle，and altitude，we performed simple and multiple correlation analyses between permafrost thickness and t0.Lower slopes were excluded from the analyses.The simple correlation analysis showed positive associations between h and H，t，φ，and α for identical topographic elements in the middle and high mountains with correlation coefficients of 0.85 to 0.99.The attempt to find correlative relationships for all topographic elements resulted in high standard deviations(σ＞50%)and low correlation coefficients (R=0.47－0.52).

The multi-factor correlation analysis revealed a number of relationships with high correlation coefficients［10］:

The dynamics of permafrost in the southeastern Siberian Platform during the Middle Pleistocene to Holocene was controlled by climatic changes.Ground thermal properties and geothermal heat flow，which remained stable for the geological structures of the region during this period，influenced，along with general climatic fluctuations，the thickness of permafrost as well as its spatial variations.

Quasi-steady permafrost occupies the major part of the Stanovoy Upland composed of sedimentary and crystalline rocks.Disequilibrium permafrost occurs in“Baikal type”depressions(Chara and Upper Tokko).The permafrost in these depressions is thick，because the recent period was preceded by the Sartan cold period.

The period from the present back to the Middle Pleistocene has consisted of major glacial cycles with a periodicity of about 100 000 years［11］.Environmental and climatic changes in the region during the Late Pleistocene to Holocene are derived from floral and faunal remains in sediments［12］.These data are now supplemented by rich information recovered from the ice core from Vostok，Antarctica［13］.These paleoclimate proxies provide evidence of significant climate variability on the Earth，much of which is related to astronomical factors.Following the widely held view on synchronism of climate changes on the planet，the derived harmonics of ice surface temperature variation in the Antarctic can be taken，with some assumptions，as ground temperature variation.Comparing it with current data，subsurface paleotemperature estimates can be obtained for different regions with provisional degree of accuracy［14］.In this study，having the paleoscenario of surface temperature history，current permafrost thickness，ground thermal characteristics，and heat flow data for the geological structures of the southeastern Siberian Platform，we have estimated the dynamics of permafrost for the period from the Middle Pleistocene to present(150 000 years).To reconstruct past changes in permafrost thickness，we solved the following problems depending on lithology:a problem without phase change for the uniform stratum of dense crystalline rock with low degree of fissuring(1.0%～0.5%or less)，a phase-change problem for the uniform stratum of fissured(more than 3%)rock，and a multi-layer medium problem for the layered stratum［8，14］.

Using the above problem formulation，ground tem-perature variation，and geothermal data(q and λ)，we estimated paleotemperature curves for the near-surface lithosphere in different landforms of the study region (Fig.5).This allowed us to model the permafrost thickness variation from 150 000 years to the present.

Fig.5 Paleotemperature profiles on the plateau(H=1 206 m)and the high mountain(H=2 120 m)，Stanovoy Upland

The t0variation in the study region exhibits a pronounced 100 000～110 000 years cycle with sharp rises in the Holocene(MIS 1)and in the Late Pleistocene(MIS 5e).These optimums caused formation of closed and open taliks in most of the study region.Widespread thaw occurred in MIS 5e，from 122 000～127 000 years.During this time，ground temperature was 3～5℃ warmer than now and varied from+5.0 on the plateaus to－4.0℃ at 2 000～2 200 m elevations in the high mountains.Lowest t0occurred 22 000～27 000 years(MIS 2)and 137 000～138 000 years (MIS 6)with values 5～7℃ lower than now［8］.

The permafrost thickness variation during the Middle Pleistocene to Holocene shows some synchronism related to climatic fluctuations.The maximum permafrost thickness，150～350 m thicker than today，occurred 130 000～140 000 years(MIS 6).In areas where the earth materials had high thermal conductivity，thick permafrost might have developed 18 000～22 000 years(MIS 2).

As mentioned above，the thickness of permafrost and the temperature of the near-surface lithosphere are controlled by the ground thermal properties，latent heat effects，and geothermal heat flux(q).These properties explain the abrupt increase in permafrost thickness in the crystalline massifs(Fig.6，curves 1 and 3)and the lack of strong temporal variability in permafrost base in the areas of fissured crystalline and sedimentary rocks(Fig.6，curve 5).A 105 000～110 000 years period between maximum permafrost thicknesses (18 000～22 000 years and 120 000～130 000 years) is apparent on the paleothickness curves.Periods of 17 000～60 000 years and 73 000～110 000 years are also evident in the permafrost dynamics when the permafrost base experienced low-amplitude variations related to relatively stable climatic conditions.The variation of the permafrost base and its amplitude were determined by the depth of permafrost occurrence during this period and the geothermal characteristics of the geological structure.These periods are thought to have been most favorable for the formation of frost shattering zones.Some support to this assumption is provided by the geophysical data from the Apsat and Ukduska mining areas where conductance anomalies were detected in the uniform stratum at corresponding depths.

Fig.6 Paleochanges in permafrost base in different geomorphological regions of the Stanovoy Upland1-water divide in very rugged terrain(H-1 200 m);2-plateau(H-1 200 m);3-creek valley in high mountain area(H-1 600 m);4-water divide(H-2 100 m);5-intermontane basin(Chara site，H-700 m)

Analysis of the geothermal data from deep boreholes and the large volume of available geophysical information，in conjunction with the geomorphological approach and geohistory analysis，allowed us to characterize the current subsurface temperature field and permafrost，as well as to examine the permafrost evolu-tion in the central part of the Stanovoy Upland from the Middle Pleistocene to Holocene.

2 Conclusions

Based on geothermal measurements in deep drill holes，ground temperatures to the depth of 1 000 m were determined and geothermal sections，some extending 700 km in length，were constructed.The results show that ground temperatures at the 1 000 m depth vary from 2.2 to 21.0℃.The highest temperatures at this depth occur in the Mesozoic-Cenozoic depressions (Tokko，Chulman，Chara and others).The lowest ground temperatures are observed in the areas of low heat flow.

Based on modeling，the spatial and temporal variations in permafrost thickness were predicted for different geomorphological regions of the southeastern Siberian Platform from the end of the Middle Pleistocene to the present.The results show that at 125 000 years the ground temperature was 3～5°warmer than today and the permafrost occupied less than 50%of the study region.Ground temperatures were reconstructed to have been lowest and the permafrost thickness was greatest during MIS 2 and MIS 6.During the Sartan thermal minimum(MIS 2)，ground temperatures were 5～7° colder than present，while permafrost was as thick as 800 m at 2 100 m elevation in the high-mountain areas and 550～600 m on the Aldan Plateau.

［1］ Melnikov P I，Tolstikhin N I(eds.).Permafrost Science［M］.Novosibirsk:Nauka，1974.

［2］ Nekrasov I A.Permafrost and Its Evolution in Northeastern and Southern Siberia［M］.Yakutsk:Yakutsk Publishing House， 1976.

［3］ Zhelezniak M N.Geothermal Conditions of Permafrost in the Western Portion of the Aldan Basin［M］.Yakutsk:SB RAS Press，1998.

［4］ Kudryavtsev V A.Southern Yakutia:Permafrost Hydrogeology and Engineering Geology of the Aldan Mining District［M］.Moscow: Moscow State University，1975.

［5］ Ospennikov E N，Trush N I，Chizhov A B.et al.Exogeneous Geological Processes and Phenomena［M］.Moscow:Moscow State University，1980.

［6］ Karausheva A I.Climate and Microclimate in the Kodar-Chara-Udokan Area［M］.Gidrometeoizdat，Leningrad，1977.

［7］ Gavrilova M K.Anticipated climate change and permafrost dynamics［J］.Meteorologiya i Gidrologiya，1984:114-116.

［8］ Zhelezniak M N.The Subsurface Temperature Field and Permafrost in the Southeastern Part of the Siberian Platform［M］.Novosibirsk:Nauka，2005.

［9］ Balobaev V T.Geothermics of the Frozen Zone of the Lithosphere in Northern Asia［M］.Novosibirsk:Nauka，1991.

［10］Zhelezniak M N，Shipitsina L I.Permafrost dynamics in the southeastern part of the Siberian Platform in the Holocene-Middle Pleistocene.In:Proceedings of the Conference“Results of Geocryological Investigations in Yakutia in the XX Century and Prospects of Further Development”［M］.Yakutsk:Melnikov Permafrost Institute Press，2001.

［11］Imbrie J.On the structure and origin of major glaciation cycles［J］.Linear responses to Milankovich forcing.Paleoceanography，1992:701-738.

［12］Enikeev F I，Potemkina V N，Staryshko V E.Stratigraphy and E-volution of Climate and Vegetation in Northern Trans-Baikalia during Late Cenozoic［M］.Novosibirsk:Geo，2013.

［13］Kotlyakov V M，Lorius C.Four climate cycles in the deep ice core from Vostok，Antarctic［J］.Izvestiya RAN.Seria Geograficheskaya，2000:7-19.

［14］ Balobaev V T，Tetelbaum A S，Mordovskoy S D.Dynamics of pore water pressure in subpermafrost stratum under secular climatic variations(two-dimensional mathematical model)［J］.In:Proceedings of the Second Conference of Russian Geocryologists，Moscow State University，Moscow，2001，(2):31-38.