3 Results

3.1 Air temperature and Ground surface temperature (Ts)

Results presented in this paper are from 1 September 2016 to 31 August 2019 (Figure 6a). The annual mean air temperature and derived values are presented in Table 1. There were no significant differences in air temperature (Ta ) between the sunny and shady slope. The annual mean air temperatures (\(\overset{\overline{}}{T}\)a) for 2016-17 and 2017-18 were -2.90 and -2.56 °C at the sunny slope and -2.63 and -2.36 °C at the shady slope. The differences between the sites in the two years were 0.27 and 0.20 °C. Although the annual mean air temperature in 2018-19 was more than 1 °C lower than the previous two years (-4.02 °C at the sunny slope and -3.82 °C at the shady slope), the air temperature difference between both sites was still only 0.2 °C, near the sensor accuracy.
There was little difference in annual freezing degree days (FDD) and thawing degree days (TDD) for the three years at both sites, particularly in the thawing season where the difference was ~20 degree days (Table 1). The difference in calculated air frost number (F) was only 0.01 (0.62 and 0.63), indicating the air temperature conditions at the two sites are very similar.
Daily variations in Ts are presented in Figure 6b. Compared to air temperature, Ts values were significantly different between the sites during the monitoring period. The mean annual Ts for the three years were 0.28 ±0.40 °C at the sunny slope and -1.02 ±0.40 °C at the shady slope, a difference of ~1.3 °C.

3.2 Ground temperatures (Tns, Tps, Tg)

Daily variations in ground temperature at 5 cm (Tns), 2.5 m (Tps), and 5 m (Tg) depth are shown in Figure 7. Variations in Tns over the three years were similar to Ts because the thermistors were only 3-5 cm apart. The daily mean Tns for the three years at the sunny slope was always higher than at the shady slope, but the difference in Tns was not as great as for Ts. The average daily temperature gradient between Ts and Tns was lower at the shady slope (0.02 ±0.06 °C) than that at the sunny slope (0.10 ±0.04 °C). The difference in daily mean Tns (Tns at the sunny slope minus Tns at the shady slope) was 0.1-3.6 °C, with a mean value of 1.43 °C. And the annual mean temperature difference ranged 1.3 to 1.6 °C.
The difference in Tps between the sites was significant up to 2.5 m depth (Figure 7b). The daily mean Tpsfluctuated about 0 °C at the sunny slope, but was stable below 0 °C at the shady slope. The daily mean Tps for the three years at the sunny slope was also always higher than at the shady slope, with differences between 0.4-3.4 °C, and a mean value of 1.44 °C. The annual mean Tps values at the sunny slope in 2016-2019 were 0.24, 0.30, and -0.12 °C, and -1.21, -1.15, and -1.53 °C at the shady slope, respectively. The annual temperature difference was ~1.4 °C.
At 5.0 m depth (Figure 7c), the daily mean Tg was <0 °C at both sites all year. Daily mean Tg at the sunny slope was always ~-0.1°C over the three years. Because the precision of the sensor is ±0.21 °C, we don’t think this value is very accurate. However, it seems that local environmental factors have little effect on Tg, and conclude that there is a strong temperature control effect at this depth. The daily mean Tg at the shady slope fluctuated between -1 to -2 °C over the study period, with a mean annual value of -1.4 ±0.02 °C. The 1.4 °C difference in Tg between the sunny and shady slopes indicate that the thermal regime is strongly affected by slope aspect.

3.3 Depth of seasonal thawing

The ground surface begins to thaw when Ts rises above 0 °C with the increase of Ta, and the thaw depth usually reaches its annual maximum at the end of August on QTP (Luo et al., 2019). The maximum seasonal thawing depth at both sites in 2016 was approximated during the drilling campaign in July-August, and the observed results were 2.70±0.1 m among nine boreholes at the sunny slope and 1.74±0.1 m at the shady slope. A nearly 1.0 m difference of the maximum depth of seasonal thawing between sites is likely to result from the large difference in Ts between sites. In 2016-19, the mean annual Tps measured at 2.5 m depth was above 0 °C (~0.14 °C) at the sunny slope, indicating that the maximum depth of seasonal thawing was >2.5 m. At the shady slope, the site mean Tps values and accompanying maximum and minimum values remained below 0 °C (~-1.30 °C), indicating that the maximum depth of seasonal thawing was <2.5 m. The difference in depth of seasonal thawing between the sites is related to the difference in thawing period duration. The thawing period was about 30-40 days shorter at the shady slope than at the sunny slope in 2016-19, resulting in a shallower depth of seasonal thawing at the site.

3.4 Soil moisture content

Daily variations in soil moisture content at four different depths within the depth of seasonal thawing (0.25, 0.5, 1.0, and 1.5m) are presented in Figure 8. The difference in moisture content between sites was significant, with the ground at the shady slope always wetter than at the sunny slope. The difference was maintained despite frequent summer precipitation events affecting both sites. The near-surface ground at the shady slope remained relatively moist during the thawing periods. The mean soil moisture content at 0.25 m depth at the shady slope was 0.374±0.003 m3/m3 during the three thawing periods, and was 0.250±0.002 m3/m3 at the sunny slope (Figure 8a). The difference in moisture content between sites occurred up to 1.5 m depth (Figure 8d). The soil moisture content decreased dramatically after ground freezing commencing at the end of October. The residual moisture content may reflect unfrozen water content, but as the sensors are not calibrated for measurement below 0 °C, unfrozen water is not discussed further.

3.5Radiation

The four components of the radiation budget all showed variation due to seasonal changes in solar altitude (Figure 9). The seasonal variation in shortwave downward radiation (DR) is evident at both sites (Figure 9a). The maximum value of daily mean DR reached ~400 W·m-2 in May-July, then gradually decreased to the minimum value of ~100 W·m-2 in mid-December. In contrast with DR, although shortwave upward radiation (UR) exhibited relatively little seasonal variation, there were many peaks caused by snowfall increasing the surface albedo, especially in March-May (the gray area in Figure 9b). The daily mean UR at the sunny slope was much higher than at the shady slope, and during the cold period, the monthly mean UR was 40% higher (~12.0 W·m-2) (Table 2).
The daily mean downward long-wave radiation (DLR) showed a similar seasonal pattern as DR, but variations in DLR were smaller and less scattered over the year. The DLR lagged behind DR by a few weeks (Figure 9c) due to the thermal inertia of the Earth system. Seasonal fluctuation in upward long-wave radiation (ULR) was well correlated with DLR. The daily mean ULR was higher at the sunny slope than that at the shady slope in most months, except in March-May of each year (Table 2). The resulting net long-wave radiation (Rl) was always ~7 W·m-2 lower at the sunny slope than at the shady slope during the warm season and ~5.5 W/m2 lower during the cold season.
The magnitudes of variation in net radiation (Rn) for both sites were different in each month. However, the daily mean Rn in most months was lower at the sunny slope than that at the shady slope (Figure 10 and Table 2). The mean daily Rn for the whole study period (n=969 days) was 12.5±0.65 W·m-2 lower at the sunny slope than at the shady slope. This may be related to differences in surface albedo and surface moisture conditions at the sites. In the warm season, the maximum daily Rn was >850 W·m-2 (Figure 10h), and <350 W·m-2 in the cold season (Figure 10k).

3.6 Surface albedo

Variations in daily mean surface albedo were consistent with variations in UR (Figures 9b, 11), and similar at both sites. However, the daily mean surface albedo was slightly higher at the sunny slope (n=1095, 0.182±0.003) than at the shady slope (n=969, 0.176±0.003). As a result, most of the monthly mean albedo values, and the annual mean albedo was higher are the sunny slope (Table 2). The surface albedo at the two sites fluctuated greatly over the year, with snowfall in winter causing a sharp increase up to a maximum of 0.99. The surface albedo was not significantly higher over the whole cold season (0.175±0.005 and 0.165±0.005) than the warm season (0.178±0.002 and 0.180±0.003) indicating that the snowfall did not persist on the ground for long at the sites. This was confirmed by field observations in winter (Jan. 10-15, 2017, Dec. 9-22, 2018, and Dec. 16-25, 2019).