1. INTRODUCTION

Permafrost results from interactions between the land and atmosphere, so its spatial distribution is mainly determined by climatic conditions (Ferrians and Hobson 1973; Harris, 1981; Cheng and Dramis 1992; Riseborough et al. 2008). However, the influence of local environmental conditions such as slope, aspect, vegetation, snow cover, and soil conditions may outweigh the climatic background, resulting in heterogeneous permafrost conditions (ground temperature, active layer and permafrost thickness) at the local scale (Brown 1973; Williams and Smith 1989; Camill and Clark 1998; Camill, 2000; Cheng 2004; Heggem et al. 2006; Lin et al. 2019; Luo et al. 2019). Local conditions may influence the ground thermal regime of permafrost by controlling, for example, incoming solar radiation, heat convection and conduction processes, and ground ice conditions (Cheng 2003).
The Qinghai-Tibet Plateau (QTP) is one of the highest plateaus in the world, and represents the largest area of high-elevation permafrost area on Earth (Zhou et al. 2000). The occurrence of high-elevation permafrost is mainly controlled by climate, topography, and surface conditions (Cheng 1983; Harris 1986; Gorbunov 1988; Cheng and Dramis 1992). At the global scale, latitude and atmospheric circulation generally control the distribution of permafrost, while local factors such as topography and surface conditions strongly regulate site scale ground thermal conditions (Zhang et al. 2000).
The QTP includes mountainous topography, and the differences in slope aspect are significant. Due to the high elevation, low latitude, and clear air, significant solar radiation reaches the ground surface on QTP (Xu and Chen 2006). As a result, the absorption or reflection of solar radiation differs greatly on slopes depending on aspect (Lin et al. 2015a; Wang et al. 2016), significantly affecting permafrost conditions (e.g., Gorbunov 1978; King 1986).
The influences of slope direction on permafrost have been reported in several areas. For example, in southeast Yukon, Canada, vegetation growth and active layer thickness differed at four sites with similar altitude and geological conditions but different slope directions (Price 1971). At Tianshan Mountain, China, the average annual ground temperature difference between sunny and shady slopes at the same elevation can reach 4.6 ℃ (Cheng 2003). The difference in active layer thickness on north and south facing slopes of Kunlun Mountain, China, was ~1.0 m, and permafrost temperatures differed by 0.5 ℃ (Lin et al. 2015a; Luo et al. 2019). The permafrost thickness on the southwest slope of Fenghuo Mountain was about 70 m, and about 120-145 m on the northeast slope. A recent report from discontinuous permafrost regions in northern Mongolia indicated mean ground surface temperature (MGST) differences of 3-4°C between north- and south-facing slopes over short horizontal distances (Munkhjargal et al. 2020). In addition, several studies have focused on the effect of embankment slope aspect on subgrade engineering by examining asymmetrical subsidence along infrastructure in permafrost regions (Hu et al. 2002; Lai et al. 2004; Chou et al. 2008; Niu et al. 2011; 2015; Zhang et al. 2017). Although these studies reported on differences in ground temperature and subsidence on opposing slopes, there has not been an in-depth explanation of the mechanisms responsible for the differences with supporting field data. This study more fully examines differences in air and ground temperature, moisture content, radiation, and soil texture and SOM content at two sloping sites with opposing aspects in an area with warm and ice-rich permafrost on QTP. The aim of the study is to elucidate the climate-permafrost relationship in this mountainous area, and attempt to understand the impacts of permafrost degradation due to slope effect.