The effects of climate change on permafrost have been well documented in many studies, whereas the effect of climate change on seasonally frozen ground (SFG) is still poorly understood. We used the observed daily freeze depth of SFG and environmental factors data from the period 2007-2016 to examine the seasonal and inter-annual variation of SFG. We quantitatively evaluated the effects of environmental factors on SFG using a boosted regression tree analysis. The results show that, on a seasonal scale, the lower layer soil frost starts freezing in mid-November, with the maximum freeze depth occurring in late March (209 cm), and then begins to thaw in both the lower and upper layers. We identified four stages of the freeze-thaw cycle: the non-frozen phase, initial freezing, deep freezing, and thawing. Furthermore, the thawing process of SFG mainly took place in the upper layer, but the freezing rate of the lower layer from mid-November to early February was similar to the thawing rate of late April to late June. On the inter-annual scale, the maximum freeze depth showed a significant increasing trend (p < 0.05). However, the freeze-thaw duration declined significantly (p < 0.05), which was correlated with the decrease in the period when surface soil temperature is below 0 degrees C. The mean soil temperature and soil heat flux were the most important environmental indicators affecting seasonal variation of SFG depth, and the cumulative negative air and soil temperatures were the dominant factors affecting inter-annual variation of maximum freeze depth. Our results might provide insight into predicting hydrological and ecological responses to future climate change in frozen-ground regions.

Global climate change has altered soil freeze-thaw (FT) patterns but less is known about the responses of soil microbial diversity, soil multifunctionality, and their relationship to FT events. Daxing'an Mountains in China, located in high-latitude permafrost ecosystems, are one of the most sensitive areas to climate change and FT patterns. Here, simulated FT conditions were used to determine the impact of FT events on soil microbial diversity and multifunctionality as well as to elucidate the relationships between bacterial and fungal diversity and multifunctionality. Community composition, alpha-diversity index, and co-occurrence network complexity of fungi significantly changed during FT events, whereas the same parameters did not exhibit significant alterations for bacteria. Soil fungal communities were more sensitive to FT events than soil bacterial communities. FT events significantly affected soil multifunctionality. A random forest analysis showed that the fungal diversity index was the main predictor of soil multifunctionality. Moreover, changes in soil abiotic factors also affected the relationship between soil microbial diversity and multifunctionality. Soil multifunctionality was also constrained by fungal community network complexity. Structural equation model showed that the FT amplitude and FT cycles exerted different impact paths on soil multifunctionality. The effect of FT cycles on soil multifunctionality (0.289) was greater than that of FT amplitude (0.080). As global climate change is expected to accelerate in the future, extension of the FT period in high-altitude and high-latitude regions may have a severe impact on soil function compared to extreme low temperatures caused by the presence of thin snow cover.