Observations from 1,047 meteorological stations from September 1, 2006 to August 31, 2015 revealed regional differences in the freezing and thawing processes of seasonally frozen ground (SFG) across China. SFG generally undergoes a one-way freezing process (i.e., top-down), and the stations with a large freeze depth generally experienced long freeze durations. During the thawing process, soil is generally characterized by two-way thawing (i.e., top-down and bottom-up) in the region north of 35 ' N, ' N, especially north of 30 ' N ' N (except in northeastern China). The onset of thawing from the bottom occurs earlier than that from the top at most stations in the two-way thawing region. The stations exhibiting one-way thawing (i.e., bottom-up) were mainly located on the southern edge of eastern China (east of 110 degrees E) degrees E) and in southern part of Xinjiang and southeast part of the Qinghai-Tibet Plateau. The freezing process lasts several days to more than four months longer than the soil thawing process, and this difference tends to be larger in high-latitude and high-altitude regions. All of the sites experienced a discontinuous freeze-thaw process, the station-average duration of which was less than a quarter of that of the continuous freeze-thaw process. Strong associations of soil freeze depth with air temperature (as characterized by the air freezing index and air thawing index) implied a dominant influence of air temperature on the soil freeze-thaw process. During the freezing process, this relationship was partially modulated by snow cover in snowy regions, such as northeast China, northwest China, and the eastern Tibetan Plateau. This paper provides the first overview of regional differences in the freezing and thawing processes of SFG over China, and the findings improve our understanding of the soil freeze-thaw process and provide important information to support research into regional landscapes, ecosystems, and hydrological processes.

Accurate quantification of the distribution and characteristics of frozen soil is critical for evaluating the impacts of climate change on the ecological and hydrological systems in high-latitude and-altitude regions, such as the Tibetan Plateau (TP). However, field observations have been limited by the harsh climate and complex terrain on the plateau, which greatly restricts our ability to predict the existence of and variations in frozen soils, especially at the regional scale. Here, we present a study relying solely on satellite data to drive process-based simulation of soil freeze-thaw processes. Modifications are made to an existing process-based model (Geomorphology-Based Eco-Hydrological Model, GBEHM) such that the model is fully adaptable to remote sensing inputs. The developed model fed with a combination of MODIS, TRMM and AIRX3STD satellite products is applied in the upper Yellow River Basin (coverage of similar to 2.54 x 10(5) km(2)) in the northeast TP and validated against field observations of freezing and thawing front depths (D-ft) and soil temperature (T-soil) at 54 China Meteorological Administration (CMA) stations, as well as frozen-ground types at 22 boreholes. Results indicate that the developed model performs reasonably well in simulating D-ft (R-2 = 0.69; mean bias = -0.03 m) and T-soil (station averaged R-2 and mean bias range between 0.90-0.96 and -0.51 similar to -0.14 degrees C at eight observational depths, respectively), and outperforms the original GBEHM forced with ground-measured meteorological variables. The frozen-ground types are also (in general) accurately identified by the satellite-based approach, except for a few permafrost boreholes located near the permafrost boundary regions. Additionally, we also demonstrate the importance of considering dynamic soil water content in frozen soil simulation: We find that a static-soil-moisture assumption (as used in previous studies) would lead to biased soil temperature estimates by > 0.5 degrees C. Our study demonstrates the value of using satellite data in frozen-soil simulation over complex landscapes, potentially leading to a greater understanding of soil freeze-thaw processes at the regional scale.