5.3 Freeze thaw fronts
Freeze-thaw processes in soils, including changes in FTFs, are very sensitive to warming. However, the latest climate models do not predict changes in FTFs directly. Implementation of the FTF module into CAS-FGOALS was performed by generally following the prescription of Gao et al. (2019). We used the maximum thaw depth in permafrost regions as the active layer depth, where its spatial distribution is shown in Fig. 11(a). Although a direct comparison with observations was difficult because no subgrid-scale permafrost representation was currently used in the model, the spatial distribution of the model simulation was still comparable with the maps of permafrost provided by the International Permafrost Association (IPA; Brown et al. 1998) (Figs. 11(a–b)).
In terms of the FTFs, Fig. 12 shows their spatial distribution for February and August, and the seasonal cycle for two points, one in seasonally frozen ground, and another in a permafrost region. The thaw depth was usually the shallowest in February and deepest in August. The simulated thaw depths in February were thus all near the surface, except for a few places near the borders of seasonally frozen ground and permafrost (Fig. 12(a)). In August, the simulated thaw depths presented a decreasing trend as the latitude increased (Fig. 12(b)). The simulated frost depths, on the other hand, showed an opposite trend as seen for the thaw depths. It was seen that the frost depth extended to the south in February and retreated north in August (Figs. 12(c–d), respectively). Figures 12(e) and 12(f) show the weight-averaged FTF seasonal cycle of two regions (seasonally frozen ground and permafrost, respectively). For the seasonally frozen ground (Fig. 12(e)), the frost depth was deeper in winter and gradually became shallower as time advanced to summer. The thaw depth, on the other hand, starts to melt down in Spring, and gets deeper in May until both the frost and thaw depths meet, then the thaw depth gets shallower again in summer to fall. For the permafrost region (Fig. 12(f)), as the frost depth below is basically frozen all the time, there is only thaw depth that starts to melt in spring and gets deeper through summer, and gets shallower until winter when the thaw depth basically stay on the surface. Overall, the FTFs tended to simulate the seasonal variation of the freeze and thaw process, and the FTF-derived permafrost extent was comparable to that seen in the observations. Thus, the inclusion of FTF was apt for simulations of the seasonal freeze thaw process in seasonally frozen ground and permafrost regions.