On the Use of Ground Temperature Data Obtained At Meteorological Stations For Assessment of Permafrost Status

The ground temperature (GT) (to a depth of 320 cm) measured by extraction thermometers at meteorological stations were used to assess the status of permafrost in Russia for 1969 to 1990, which is generally a period of global warming. In all, data for 75 stations located in the zone of discontinuous permafrost and for 44 in the zone of continuous permafrost (perennially frozen ground is observed from the depth of 160-320 cm at of the 33 stations) were analyzed. Linear trends of time series of annual and seasonal GT and surface air temperature (SAT) were used to asses the s ground response to the recent warming. Generally, GT in most of the permafrost zone increases in the cold season, and is explained by the predominant rise of SAT. In summer, however, SAT shows little increase, and the GT does not change in the upper 60 cm layer), but warming is observed below 80 cm. The latter is also attributable to the increase of winter SAT and GT. There are small linear trends of GT in perennially frozen layers and these are insignificant at most of the stations where permafrost exists. Thus, there is no significant warming of perennially frozen ground over most of the permafrost zone of Russia. These results are in accordance with observed energy exchanges in permafrost regions and with the estimates of current permafrost status that have been made using other approaches. Nevertheless, this work shows that some methods used to assess tendencies in times series of meteorological data over large territories are not suitable for those of GT. First, spatial interpolation of GT linear trends to a regular grid was impossible because of the strong variability and irregularity in the distribution of these trends. Second, the use of linear trends is not justified in all cases. For a number of stations, linear trends, even if high and significant, may be false. This can arise due to a low value at the start of the time series and higher ones at the end. Hence, trend estimates made on the basis of such records are misleading. In addition, in some cases, the GT time series for the period 1960s - 1990 include both positive and negative temperature trends, that are lacking in the corresponding SAT time series. The use of linear trends is also inappropriate in these cases. This situation arises due to the effect of fluctuations in the timing, duration, and thickness of snow cover and ground water content on GT. Thus, in comparison with SAT, the analysis of GT data requires their more detailed information at each station. In turn, this will provide fuller understanding of processes occurring in the permafrost zone of Russia.