A growing body of simulation research has considered the dynamics of permafrost, which has an important role in the climate system of a warming world. Previous studies have concentrated on the future degradation of permafrost based on global climate models (GCMs) or data from GCMs. An accurate estimation of historical changes in permafrost is required to understand the relations between changes in permafrost and the Earth’s climate and to validate the results from GCMs. Using the Community Land Model 4.5 driven by the Climate Research Unit -National Centers for Environmental Prediction (CRUNCEP) atmospheric data set and observations of changes in soil temperature and active layer thickness and present-day areal extent of permafrost, this study investigated the changes in permafrost in the Northern Hemisphere from 1901 to 2010. The results showed that the model can reproduce the interannual variations in the observed soil temperature and active layer thickness. The simulated area of present-day permafrost fits well with observations, with a bias of 2.02 × 106 km2. The area of permafrost decreased by 0.06 (0.62) × 106 km2 decade1 from 1901 to 2009 (1979 to 2009). A clear decrease in the area of permafrost was found in response to increases in air temperatures during the period from about the 1930s to the 1940s, indicating that permafrost is sensitive to even a temporary increase in temperature. From a regional perspective, high-elevation permafrost decreases at a faster rate than high-latitude permafrost; permafrost in China shows the fastest rate of decrease, followed by Alaska, Russia, and Canada. Discrepancies in the rate of decrease in the extent of permafrost among different regions were mostly linked to the sensitivity of permafrost in the regions to increases in air temperatures rather than to the amplitude of the increase in air temperatures. An increase in the active layer thickness of 0.009 (0.071) m decade1 was shown during the period of 1901–2009 (1979–2009). These results are useful in understanding the response of permafrost to a historical warming climate and for validating the results from GCMs.
Guo Donglin, Wang Huijun, 2017: Simulated historical (1901-2010) changes in the permafrost extent and active layer thickness in the Northern Hemisphere. Journal of Geophysical Research: Atmospheres, 122, 12285–12295.
Figure 1. Comparison of the simulated and observed changes in (a) soil temperature at a depth of 1 m from 1981 to 2009 relative to the time period of 1981–2000 and (b) active layer thickness (ALT) from 1996 to 2007 relative to the time period of 1996–2007. Observations are averages among all stations. Linear trend, correlation coefficient (R), and Nash-Sutcliffe efficiency (NSE) of the simulations and observations are given. The observation stations for measurements of the soil temperature and ALT are shown as red rectangles and circles, respectively in Figure 1c. (c) Comparison of the simulated present-day extent of permafrost (shaded color) for 1981–2000 with observations (areas outlined in blue). The bias in the area between the simulated and observed results is given in the bottom right-hand corner of Figure 1c. The four countries considered here as subregions and the Tibetan Plateau (TP), containing mostly of permafrost, are outlined by gray dashed lines.
Figure 2. Changes in the areal extent of permafrost and air temperatures over (a) the entire region of permafrost in the Northern Hemisphere, (b) high-latitude regions with permafrost, (c) high-elevation regions with permafrost, and regions of permafrost in (d) Russia, (e) Canada, (f) Alaska, and (g) China from 1901 to 2009 relative to 1981–2000. The dark lines present the change smoothed using the 21 year moving average. Linear trends in permafrost area and air temperatures for the period of 1901–2009 are given at the top of each panel with trends for the period of 1979–2009 in parentheses.
Figure 3. Changes in areal extent of permafrost in (a) the 2000s relative to the 1900s and (b) in the decade from 1935 to 1945 (1935s) relative to the 1900s. Four countries and the Tibetan Plateau (TP), consisting mostly of permafrost, are outlined by gray dashed lines.
Figure 5. Relationship between (a) decreasing percentage trend in permafrost area (% decade-1), (b) decreasing percentage sensitivity of permafrost area to increasing temperature (% °C-1) (equivalent to percentage size of permafrost degradation when air temperature increase 1°C), and (c) increasing temperature trend (°C decade-1) in the entire permafrost region (entire region), high-latitude (H-latitude) regions, high-elevation (H-elevation) regions, and in permafrost regions in Russia, Canada, Alaska, and China from 1901 to 2009. The correlation coefficient between the decreasing percentage trend in permafrost area and decreasing percentage sensitivity of permafrost area to increasing temperature is 0.93, whereas the correlation coefficient between the decreasing percentage trend in permafrost area and increasing temperature trend is 0.83. Both of correlation coefficients exceed the significance level of 95%.