Abstract
(Englisch)
|
Various efforts have been undertaken in recent years to improve spatial modeling of mountain permafrost distribution based on empirical and statistical data. Most of today's models simulating permafrost distribution in alpine regions with complex terrain directly relate documented permafrost occurrence to topoclimatic factors such as altitude, slope, aspect, air temperature, solar radiation, and snow-cover, which can easily be measured or computed. These factors are proxy variables of selected energy-balance factors and reflect a simplified relation between climate and permafrost distribution. Air temperature, solar radiation, and snow-cover are considered the most important factors influencing mountain permafrost distribution.
However, modeling of spatial permafrost distribution under different climatic conditions requires detailed knowledge of the energy-exchange processes at the atmosphere/lithosphere boundary. To improve modeling and simulation of spatial permafrost distribution patterns in different mountain regions and for various climatic scenarios, special emphasis should be given to all vertical energy-exchange processes at the surface and within the active layer and especially to aspects of snow-cover (duration, snow-height), wind and ground thermal characteristics.
To develop an advanced understanding of the energy fluxes involved with permafrost distribution, microclimatic studies and detailed energy-balance measurements were performed on two high-mountain permafrost sites in the Engadin and the Bernese Oberland (both Swiss Alps). The results of these measurements were expected to improve the knowledge about, and the understanding of, the energy-exchange processes at the ground surface. Particular finding can be obtained about advective energy-exchange processes within the active layer: Determination of all vertical incoming and outgoing energy fluxes, and thus the energy balance at the ground surface, will make an approximation of the lateral energy flux possible.
Within the PACE project all vertical energy-exchange fluxes (K¯, K, L¯, L, QH, QLE, QG and QM) were measured over a three-year period (1997-1999) at rock glacier Murtèl-Corvatsch (Engadin) and during one year (1999/2000) at the Schilthorn summit (Bernese Oberland) by the Department of Geography, University of Zurich. The results show the following:
· The radiation components represent the dominating factors whereas QH, QLE, QG and QM (during the snowmelt season) are much smaller.
· An accurate determination of the turbulent heat fluxes QH and QLE remains difficult in alpine topography with its rugged terrain. Further contributions to the difficulties are the rather low wind speed at Murtèl-Corvatsch and the missing information on the humidity of the snow-free ground at Schilthorn.
· The sum of all measured components of the energy-balance at Murtèl-Corvatsch is negative during winter and positive during summer. The deviation between a zero energy-balance and the calculated sum of the energy-balance components, averaged over one year, is around 18 Wm-2 and constant over the three years of the measurement period. The deviation at the Schilthorn site could not be determined due to the different instrumentation of the micrometeorological station.
· A most probable reason for the surplus energy at Murtèl-Corvatsch are the coarse blocks at the surface of the rock glacier which allow air to circulate within the active layer. Besides a system of vertical funnels appearing within the snow-cover on the rock glacier in early winter, the high measured MAGST (5.9 °C in 1997 and 5.0 °C in 1998) is a clear indicator for such a lateral (advective) energy flux within the layer of coarse boulder blocks (cold-air drainage).
· The MAGST measured at Schilthorn permafrost site was 4.4 °C , even though the surface consists of fine-grained material. Different processes to those found at Murtèl-Corvatsch must be responsible for this high temperature difference between the surface and the permafrost itself. Running water on the inclined surface could be a possible factor, it is not yet clear, however, which processes are involved and further investigations have to be carried out.
The computer model PERMEBAL simulating ground thermal conditions based on an energy balance approach has been developed with help of the knowledge gained by these measurements. The model consists of two principal modules, a surface energy-flux module and a thermal-offset module. The energy balance module simulates ground surface temperatures with the help of suitable parameterizations for all important energy fluxes and using simple meteorological data, digital elevation models and information on surface characteristics as input. The thermal-offset module links these ground surface temperatures to thermal conditions at the permafrost table. Such a process-based model does not only show where permafrost has to be expected and where not, but also explains why permafrost is present or absent, that is, which parameters are the critical ones. This enables the modeler to adapt the model to changed environmental conditions and thus to compute various effects from climate change scenarios.
The surface energy-flux module of PERMEBAL was applied to an area of 16km2 around the Piz Corvatsch (Engadin, Switzerland) and of 35km2 in the Schilthorn massif (Bernese Oberland, Switzerland). The model calculations were verified independently by carrying out energy balance measurements at permafrost sites in the investigated areas. Furthermore, the model results were compared to BTS-measurements indicating the actual permafrost distribution. In addition, the snowmelt evolution was found as an independent parameter to test the modeling results: Many of the crucial parameters determining the permafrost distribution (air temperature, solar radiation, wind, etc.) also strongly influence the snowmelt pattern. Furthermore, there is a strong interaction between the seasonal snow cover and ground thermal conditions. It can therefore be assumed that an energy balance model which is able to correctly reproduce the snowmelt evolution will also supply reasonable ground surface temperature simulations.
Reproduction of the radiative energy fluxes K¯, K and L¯ by PERMEBAL gives a high correlation with the measured energy fluxes (r2 between 0.97 and 0.73). The results for the simulated windfield and turbulent heat fluxes are, on the other hand, low quality. The comparison of the resulting surface temperature of both the model calculations and the energy-flux measurements gives high correlation (r2 =0.71).
Comparison of simulated with observed snowmelt evolution showed a good correspondence of the model results with reality, both spatially and temporally: Simulation of the snow-distribution pattern reaches a spatial agreement with observations of more than 78% of the pixels during several days within the snowmelt period. The day of final melt out was determined quite exactly over 14 years (mean overestimation of the persistence of the snow-cover: 5.5 days) without resetting the model. It can, therefore, be concluded that the model enables a realistic reproduction of the energy-exchange processes taking place at the ground/snow-cover/atmosphere interface during winter and spring.
The application areas of PERMEBAL could be divided into three classes of mean annual sums of daily ground-surface temperatures of snow-free pixels, similar to `permafrost probable', `permafrost possible' and `permafrost improbable'. The same scale of classification could be applied to both application areas. A quantitative comparison of measured BTS-points with the classes of mean annual sums of daily ground-surface temperatures of snow-free pixels at Corvatsch-Furtschellas (after application of a thermal offset to the gridpoints with a coarse blocky surface layer) yields an agreement of BTS-values to the assigned class of 43%. 47% of the points were assigned to a neighbouring class and can thus, not be termed `correct' or `incorrect'. The remaining 10% were assigned to an incorrect class. At Schilthorn, the percentage of coincidence between BTS-values and an assigned class of mean annual sum of daily ground-surface temperatures of snow-free pixels is 39%. 55% of the BTS-points were assigned to a neighbouring class. Only 6% were assigned to an incorrect class.
|
