TIME, year

Fig. 68. Top: Temperature monitoring at 0.25 and 1.25 m depth in shallow borehole Illulissat (Greenland) from 6 November 1968 to 15 June 1982. Data are biweekly averaged. Bottom: Thickness of the snow cover. Data by Olesen (2003).

( The European Union project, Permafrost and Climate Change in Europe (PACE), is collecting data from permafrost boreholes in mountainous regions of Spain, Italy, and Scandinavia. It is expected that this permafrost monitoring network will provide high-resolution data for the permafrost studies under changing climatic conditions both in northern latitudes and at high altitudes of Europe. The goal of the Circumpolar Active Layer Monitoring (CALM) program is to observe the response of the active layer and near-surface permafrost to climate change over multi-decadal timescales ( The CALM program began in 1991. Its observational network contains more than 100 boreholes for measurements of the soil and permafrost temperatures. Most CALM sites are located in Arctic and sub-Arctic lowlands, while 20 boreholes arranged in the mountainous regions of the Northern Hemisphere above 1300m elevation. A new Antarctic borehole network belonging to the CALM currently includes more than 10 sites. The GTN-P was initiated by the International Permafrost Association (IPA) to organize and manage a global database of the permafrost observatories for detecting, monitoring, and predicting climate change. It began in 1992. Some 370 boreholes from 16 countries (the majority from the Northern Hemisphere) have been suggested for the inclusion in the GTN-P borehole temperature monitoring system. For example, in the frames of this experiment temperature measurements were performed in more than 20 boreholes in central East Siberia and in Yamal Peninsula (NW Siberia). The catalog of permafrost measurements from Russia includes 122 boreholes (Melnikov, 1998; Period of observations was primarily from 1980s to early 1990s. Data from this collection were used for the construction of the circum-Arctic map of the permafrost and detection of the ground-ice conditions. It is impossible to enumerate all similar projects; thus, finally we mention only the mountain permafrost temperature monitoring experiment in Mongolia (Sharkhuu and Tumrubaatar, 1998; Sharkhuu, 2003). The broader goal of all these projects is to scrutinize the statement suggested in the earlier works by Lachenbruch and Marshall (1986), Lachenbruch et al. (1988) that the permafrost temperatures represent a reasonably good indicator of climate change. Because the temperature measurements in permafrost are important for quantifying the long-term terrestrial response to the climate change at high latitudes (and/or altitudes), it is therefore crucial that observations and their processing continue over decadal or even longer periods to assess trends and to detect cumulative, long-term climatic variations.

In principle, available measurements have supported the sensitivity of permafrost to climate and the GST-SAT coupling in cold regions. Observations by Isaksen et al. (2001) performed during 1998-2000 years have shown that the mean annual GST measured in borehole Janssonhaugen (near Svalbard, Norway) closely correlates with the nearby airport SAT data, indicating a direct linkage between the atmosphere and ground on the annual scale. However, such coherence occurred because of the presence of bedrock close to the surface and practically the absence of a boundary layer of snow, vegetation, organic material, and the mineral soil. The small GST-SAT differences were also found in the regions of the northeastern Canadian High Arctic where active layer is also thin (Taylor et al., 2006).

In more compound environments observations revealed regional variability of the warming trends and demonstrated that the variations of the permafrost not only depend on the SAT change, but represent a complex response on the ground composition, microrelief, the moisture and depth of active layer, the amount of ice in the ground, etc. The penetration of the surface temperature changes into the ground also can be affected by the insulating effects of vegetation, organic material, or snow cover. As about the vegetation, the Arctic tundra represents treeless land. No true soil is developed in this environment. However, it generally possesses a well-grown surface organic layer of shallow rooted vegetation with most of the biomass concentrated in the roots. This layer is characterized by high porosity and high hydraulic conductivity. When it is dry, it represents very effective insulator (Hinkel et al., 2001). And it is generally dry, because the climate of Arctic tundra is characterized by low precipitation19 coupled with strong, drying winds. Kane et al. (1990) has found that in northern Alaska only 15-17% (10-15% according to Harazono et al., 1995) of the net solar radiation was used for melting and warming of the active layer during warm season. Snow cover provides an insulating layer on the ground surface. It is actually advantageous to plants defending them from the harsh winter cold. The summarized effect of vegetation and snow cover may suppress the surface temperature signal in permafrost and result in seasonal GST-SAT decoupling. Due to high hydraulic conductivity of the tundra surface vegetation layer, when it is wet, it effectively transfers heat by advection. In a dry, unfrozen state, tundra vegetation is an extremely good thermal isolator. When it is water-saturated, advection within this layer can produce rapid changes in the temperature at the organic-mineral surface layers

19Yearly precipitation including melting snow is generally of 15-25 cm.

interface during episodic precipitation events. However, because of the low precipitation, the latter effect likely has no significant influence on the annual and longer scale GST-SAT decoupling as the snow cover. The growing season in the Arctic tundra is extremely short (from 50 to 60 days), and the growth of plant species is slow. Plants more likely reproduce by division and building. If the climate does not vary, such conditions provide plant succession, the long-term stability of the surface layer and, thus, invariable long-term mode of the air permafrost temperature coupling. Because the above-mentioned factors are affected by climate change themselves, an interaction of air and permafrost temperatures implicates composite interdependences.

It should be also mentioned that the effect of strong, surface-based atmospheric temperature inversions put difficulties in the way of direct GST-SAT comparison in the Arctic region. Within the range of ground elevations subjected to temperature inversions, mean temperatures may actually increase rather than decrease with increasing elevation. Strength of the winter inversions may reach ~2K/100m (Bradley et al., 1993); thus, they can successively counteract the usual atmospheric thermal lapse rate. Due to all above-mentioned factors, direct comparison of the GST and SAT temperatures in the high-latitude areas represents a very difficult task often giving misleading results. Thus, Taylor et al. (2006) when comparing the GST inferred from the data of five boreholes located at the northeastern Canadian Arctic Archipelago with the mean annual SAT from the neighboring meteorological stations for the 1950-2000 period have found correlation coefficient values ranging from 0.91 through 0.03 to -0.87. In earlier work by Lachenbruch and Marshall (1986) it was emphasized that the GST relates to the base of the active layer and not to the actual ground surface.

Significant part of this ambiguity can be elucidated by careful long-term precise monitoring of the ground temperature oscillations and related meteorological variables. Figure 68 shows temperature record from the shallow borehole Ilulissat (Greenland) measured from the late 1968 to 1982 (approximately 5000 days). Air temperature in the area varies from slightly below -30°C to about +15°C. Snow definitely protects the ground from freezing, but regardless of this protection temperature below the surface can freeze down as low as — 10°C at 0.25 m or to -8°C at 1.25 m depth. During summer the temperature in the uppermost 1 m can be higher above zero up to depth of 1 m (recorded only in 1969-1972), permafrost never melted below 1.25 m depth (well confirmed for the whole 1969-1982 record), so the surface thickness of alternatively melting and freezing conditions (active layer) is about 1 m. In the areas with continuous permafrost and especially cold climate the thickness of the permafrost can vary between approximately 450-600 m (Barrow and/or Prudhoe Bay, Alaska) to 1500 m (northern Lena and Yana River basins in Siberia). Thus, it is comparable and/or exceeds the depth of the most boreholes. There is a time delay between a change in temperature at the ground surface and the change in permafrost at depth. This lag equals to 2-3 months for the time series monitored at 0.25 and 1.25 m depth (Figure 68). Under typical thermal conditions in the Scandinavian permafrost boreholes the annual wave is delayed by half a year from the surface to approximately 8.5m depth. Zero annual amplitude depth equals to 18m (Isaksen et al., 2001). At the base of thick permafrost this lag may be on the order of hundreds to thousands of years, while for thin permafrost it ranges from years to decades.

Precise soil temperature and moisture monitoring results at Barrow (Alaska) during 1993-1999 period were used to analyze the interactions within atmosphere-snow-active layer-permafrost system (Hinkel et al., 2001). A similar study was performed using data of the ground temperature monitoring near Fairbanks (Alaska) (Kane et al., 2001). Both studies were focused on the detection of the interactions between near-surface ground temperature, the physical properties of this layer, and seasonal variations in the underground heat transfer process. The ground temperature monitoring and soil moisture probe data were supplemented by the SAT temperatures from the neighboring meteorological stations. General patterns of the seasonal GST-SAT coupling measured by Hinkel et al. (2001), Kane et al. (2001) were similar to that described in Sections 2.6.1-2.6.2 (Figures 44-47). The authors have identified four seasonal thermal regimes and demonstrated that seasonal variations are caused primarily by freezing, thawing, and redistribution of water contained in the ground. It was proved that the non-conductive heat transport by water and water vapor plays dominant role in specific periods. The snow-free active layer regime occurs in summer and is characterized by a large daily temperature range and fast response to weather fluctuations. The ground moisture content increases rapidly as the thaw front penetrates downward. Latent heat effects are associated not only with thawing. During this period a significant portion of the absorbed solar energy is expended on evapotranspiration (30-65%), while the ground heat flow is relatively low. The evapotranspiration from the surface and within active layer can significantly lower the mean summer temperature of the near-surface ground; thus, summer represents the period of maximum decoupling of surface and ground temperatures. Effect of the summer precipitation is quite low and incomparable with forest sites in interior Alaska and/or northwest Canada, when soil warming of several degrees can occur in response to precipitation events (Hinkel et al., 1997). In the autumn the active layer refreezes from the top down and from the bottom up. The release of latent heat of fusion slows down the penetration of the freezing front, and the temperature of active layer remains constant during the water-ice conversion. This effect is referred to as the "zero-curtain" (see also Section 2.6.2). When it is at work, the upper permafrost is effectively isolated from the surface temperature variations; thus, the temperature signal is effectively attenuated in the depth. This regime is relatively short in duration (30-40 days). Winter represents the time of freezing regime, "zero-curtain" effect is stopped, and heat is removed more rapidly from the upper layer primarily by conduction. The thickness, thermophysical properties, and duration of snow cover have a strong influence on permafrost. In the winter the ground is warmer than the surface. At sites underlain by permafrost in Alaska, the mean annual GST is generally 3-6K higher than the mean annual SAT. Daily temperature fluctuations are strongly attenuated under the snow cover. However, the GST-SAT tracing is stronger than in the summer, and synoptic events are visible in the underground temperature record. The cessation of this regime occurs in the spring, when during fast short snowmelt infiltration of the melt water results in rapid warming of the upper underground layer. General conclusion by Hinkel et al. (2001) and Kane et al. (2001) was that in high-latitude regions, where the mean annual temperature is below freezing, the surface temperatures anyhow represent effective upper boundary condition for the deeper temperatures. On the other hand, in the areas with extensive seasonal freezing of the active layer the latent heat of fusion prevents the deeper subsurface from seeing the warmth in the summer.

Modeling results by Stieglitz et al. (2003) have specified the role of the show cover in the GST-SAT coupling in the permafrost areas on the longer scales. The authors have revealed two main factors influencing the underground temperature, namely the near-surface air temperature and the temporal variations of snow cover with clear preference of the latter factor. Comparison has shown that on the Northern Slope of Alaska the permafrost temperature changes for the period of 1983 to 1998 are consistent with decadal variability of snow cover. On the other hand, Stieglitz et al. (2003) have also detected that decadal variations in snow cover penetrate to not more than 50-60 m depth and, thus, have little influence on the multi-decadal trends that are archived in borehole temperatures to depth of several hundred meters.

The effect of the penetration depth of surface factors that affect the air-permafrost temperature coupling can be illustrated using above Ilulissat (Greenland) data (Figure 68). Maximum and minimum air temperatures were recorded there simultaneously with the ground temperatures. Correlation of biweekly averaged ground temperature series at 0.25 cm depth with maximum and minimum air temperatures for the 1968-1982 time period equals to 0.11 and 0.17, respectively, while similar correlation with temperature measured at 1.25 m depth already reaches 0.31 and 0.41. The correlation between snow cover and ground temperature at 0.25m depth still equals —0.16, and it falls to zero at the 1.25 m depth. The analysis of more extensive database, namely comparison of the air and permafrost temperatures measured at 1.6 m depth at the Churapcha meteorological station (East Siberia), have shown very weak coupling of both variables on the annual scale (Figure 69). The correlation between both temperatures calculated for the 1957-1992 observational period equals to 0.49, and it is even smaller for shorter averaging intervals (Romanovsky et al., 2000). On the other hand, the correlation of smoothed versions of both temperatures (10-years running means) significantly increases and amounts to 0.87. An improvement of the correlation is more noticeable for the longer averaging intervals. This study supports the ability of permafrost to archive long-term surface air temperature changes and suggests the possibility of exact coupling of the SAT and permafrost temperatures on decadal or longer timescales. According to Romanovsky et al. (2000) decoupling of both temperatures on the annual timescale occurs primarily due to significant inter-annual variability of the snow cover duration and its thickness. This effect will likely be smoothed on the long timescale. Anyhow, it is expected that continuous monitoring of the ground temperatures and related meteorological variables that is carried out at high latitudes/altitudes in the frames of numerous above-mentioned international programs will significantly extend available database and improve our present understanding of the GST-SAT linkage in the permafrost areas.

Permafrost is a thermal condition. It is a product of cold climate that compels it to grow from the surface downward, and its formation, persistence, position of the base, and top strongly depend on climate. The air temperature and the length of freezing season are the most important factors that practically determine the existence and stability of the permafrost. The colder the air temperature, whether due to increasing latitude and/or altitude, the more likely permafrost will occur and the thicker it will tend be. This fact can be easily illustrated by examining of the permafrost boundary maps, where the transition from no permafrost to discontinuous and then to continuous zones exhibits clear south-north trend. The coupling between cold climate and permafrost is so strong that the maps of SAT isotherms averaged for long time intervals are often used to delineate various zones of permafrost occurrence (Kudryavtsev et al., 1980). Permafrost is affected by climate variations and thus can be used for detection of the paleoclimate changes.

Fig. 69. Mean annual air and permafrost temperature measured at 1.6 m depth (Churapcha meteorological station, East Siberia) (dashed line) and its 10-year running mean (solid line). Redrawn from Romanovsky (2001) (see also

Fig. 69. Mean annual air and permafrost temperature measured at 1.6 m depth (Churapcha meteorological station, East Siberia) (dashed line) and its 10-year running mean (solid line). Redrawn from Romanovsky (2001) (see also

Similarly to other temperature logs, inversion of the temperature-depth profiles measured in boreholes drilled in permafrost can provide information on the past GST changes.

Mathematical description of the various phenomena occurring in the frozen underground has been presented in numerous works (Carslaw and Jaeger, 1962; Lunardini, 1991; Osterkamp and Gosink, 1991; Rath and Mottaghy, 2003). Detailed modeling of the freezing/thawing processes including heat and mass transport in the permafrost environment is very complex and its physical theory is not fully understood. The heat conduction equation with the phase change is nonlinear and the coupling of the thermal and hydrological processes with climate is complex. For these reasons simplified models are used for description of the thermal state of cold regions. Thus, a simple but effective method for modeling of freezing/thawing processes in the subsurface was suggested by Mottaghy and Rath (2006). The authors have included their scheme into finite-difference code SHEMAT (Simulation for HEat and MAss Transport; see also Clauser, 2003) that simulates a wide variety of thermal and hydrogeological processes in both the 2- and the 3-D approach. Significant facility of the modeling procedure represents the fact that in permafrost itself the heat transfer is achieved through pure conduction. Advection is insignificant due to frozen state of water, so that once boundary conditions have been specified, temperature pattern within permafrost and the position of its boundary can be calculated from somewhat modified Eq. (4) (see Section 2.2).

The thermal regime of cold regions is strongly affected by the processes of water freezing/melting accompanied by consumption/release of latent heat. This effect is extremely important, since it changes enthalpy by orders of magnitude. The latent heat effect can be included into conventional heat conduction equation by substituting an apparent heat capacity for the volumetric heat capacity of unfrozen stratum. One-dimensional equation of conductive heat transfer that takes into account freezing/thawing in the subsurface can be formulated as a modification of the conventional Eq. (4) (see, e.g. Lunardini, 1991)

Was this article helpful?

0 0

Post a comment