Ground surface temperature variability and permafrost distribution over mountainous terrain in northern Mongolia

Figures & data

Figure 1. Location of the study area, its topography, and the positions of temperature measurement transects across the Sugnugur Valley (indicated by T1–T4; see details in Table 1). The inset shows the PPI (at 1-km resolution), which indicates the likelihood of permafrost existence at a global scale (Obu et al. 2019). The hydroclimatic station (green dot) is situated at 1,193 m.a.s.l. BH = the position of the two boreholes located on the north-facing slope near transect 3

Table 1. Descriptions of temperature measurement sites. South-facing slopes with greater inclination receive higher potential potential solar radiation (PSR)

ID	Transect number^a	Temperature measurement	Slope	Elevation (m)	Vegetation type	Aspect	Inclination (°)	PSR (Wm^−2)
1	1	Surface	South	1,281	Sparse vegetation	S	18	182
2		Surface and air	Valley bottom	1,193	Short grassland	SW	9	175
3		Surface	North	1,281	Sparse forest	NW	21	134
4	2	Surface	South	1,393	Sparse vegetation	SW	35	172
5		Surface and air	Valley bottom	1,356	Bare ground	W	4.4	145
6		Surface	North	1,393	Dense forest	NE	22	111
7	3	Surface	South	1,558	Sparse vegetation	S	35	190
8		Surface and air	Valley bottom	1,485	Short grassland	NE	2.8	161
9		Surface	North	1,558	Dense forest	NE	10	149
10		Surface	North	1,560	Dense forest	NE	10	147
11		Surface	North	1,530	Burned forest	N	7	158
12	4	Surface	Valley bottom	1,583	Shrubland	NW	5	148
13		Surface	North	1,612	Sparse forest	NW	14	140
14		Surface and air	Mountain top	2,020	Rock debris	S	8	196

Note: ^aSee Figure 1.

Table 2. Parameters used in the article and their calculation

No.	Abbreviation	Parameter	Unit	Calculation method
1	FN	Frost number	Unitless	$\frac{\sqrt{\mathrm{F}\mathrm{D}\mathrm{D}\mathrm{a}}}{\sqrt{\mathrm{F}\mathrm{D}\mathrm{D}\mathrm{a}}+\sqrt{\mathrm{T}\mathrm{D}\mathrm{D}\mathrm{a}}}$
2	MAT	Mean annual air temperature	°C	$\mathrm{M}\mathrm{A}\mathrm{T}=\left(\alpha \times \mathrm{P}\mathrm{S}\mathrm{R}+\beta \right)+\mathrm{L}\mathrm{R}$α and β are empirical parameters
3	PSR	Potential solar radiation	Wm^−2	Direct + Diffusive
4	LR	Lapse rate	°C m^−1	Annual air temperature gradient per meter
5	MGST	Mean annual ground surface temperature	°C	(nt × TDDa − nf × FDDa)/365
6	TDDa	Air thawing degree days	°C days	Annual sum of daily air temperatures above 0°C
7	FDDa	Air freezing degree days	°C days	Annual sum of daily air temperatures below 0°C
8	nt	Thawing n-factor	Unitless	TDDs/TDDa
9	nf	Freezing n-factor	Unitless	FDDs/FDDa
10	TDDs	Surface thawing degree days	°C days	Annual sum of daily surface temperatures above 0°C
11	FDDs	Surface freezing degree days	°C days	Annual sum of daily surface temperatures below 0°C
12	SO	Surface offset	°C	MGST − MAT
13	Ta	Ambient air temperature	°C	Observed at 1.5–2 m height
14	Ts	Surface temperature	°C	Observed at 5 cm depth
15	SAVI	Soil adjusted vegetation index	Unitless	$\frac{\mathrm{N}\mathrm{I}\mathrm{R}-\mathrm{R}\mathrm{E}\mathrm{D}}{\left(\mathrm{N}\mathrm{I}\mathrm{R}+\mathrm{R}\mathrm{E}\mathrm{D}+\mathrm{L}\right)}\ast \left(1+\mathrm{L}\right)$L is assumed to be 0.5
16	ASD	Accumulated snow depth	cm	Maximum late winter snow depth
17	LST	Land surface temperature	°C	Derived from Landsat images
18	FT	Discrete Fourier transform	—	Reveals periodicities in input data
19	SBF	Severe burned forest	—	—
20	UBF	Unburned forest	—	—

Figure 2. Flowchart of the methodology

Figure 3. (a) An example of the temperature variability during the study period and (b) the locations of temperature measurements, typical topography, and land cover in the study region as valley transect. The temperature values are five-day averages

Figure 4. Calculated TDD_s and FDD_s at the hydroclimatic station. The study period covers the complete thawing and freezing seasons when the iButton measurements were conducted

Table 3. Mean and maximum snow depth (cm) at the hydroclimatic station site, together with ground surface temperature (T_s, °C), air temperature (T_a, °C), and incoming solar radiation (ISR, Wm⁻²) in winters since 2012

	2012–2013	2013–2014	2014–2015	2015–2016	2016–2017	2017–2018
Mean	25	19	12	15	19	—
Maximum	41	26	25	26	29	—
Ts	−10.4	−8.9	−10	−9.8	−8.3	−7.9
Ta	−20.3	−17.2	−16.7	−19.2	−17.4	−16.3
ISR	87.8	96.1	92.3	86.5	95.6	94.2

Note: T_s, T_a, and ISR are the daily mean values between November and March during the period of persistent snow cover.

Figure 5. Calculated n-factors at each transect during the study period. Values greater than 1.0 indicate lower T_s during the freezing season and higher T_s during the thawing season relative to T_a

Table 4. Observed surface offset and n-factors at the measurement sites

Transect	SO (°C)	nf	nt	ASD (cm)	SAVI	Surface characteristic	Elevation (m)
1	4	0.52	1.2	24	0.37	Short grassland	1,193
2	3.5	0.6	1.05	20	0.3	Poorly vegetated	1,356
3	4	0.53	1.16	25	0.38	Shrubland	1,485
4	1.5	0.52	0.8	NA	0.08	Rock debris (mountain)	2,020

Note: The annual surface offsets at vegetated sites were relatively similar but differed significantly for rock debris surfaces. The values at transect 1 are the mean of the last five years.

Table 5. Pearson’s correlation coefficient (r) matrix between some parameters;

	Elevation	PSR	TDDs	FDDs	Biomass	ASD	LST	MGST
Elevation	1
PSR	0.27	1
TDDs	−0.46	0.60*	1
FDDs	0.04	0.45	0.39	1
Biomass	0.37	−0.65*	−0.8**	−0.21	1
ASD	0.06	−0.34	—	0.17	—	1
LST	−0.69*	0.32	0.78**	0.4	−0.62*	—	1
MGST	−0.35	0.65*	0.94**	0.68*	−0.7*	−0.32	—	1

Note: *p < .05. **p < .01.

Figure 6. (a) Distribution curves of ten-day average potential radiation and observed and adjusted air temperatures at transect 3. After time lag and FT adjustments, the variance between potential radiation and temperature was significantly reduced. (b) The linear regression on the right shows the correlation between the adjusted T_a and potential radiation at transects 2 and 3. The linear equation was later applied for the estimation of MAT for 2016–2017

Figure 7. Calculated MGST in the Sugnugur catchment for 2016–2017. The PPI lines indicate the permafrost probability index from Obu et al. (2019). The inset shows the relationship between the observed MGST at individual sites and the estimated MGST. Red circles mark the outliers described in the text

Obu, J., S. Westermann, A. Bartsch, N. Berdnikov, H. H. Christiansen, A. Dashtseren, R. Delaloy, B. Elberling, B. Etzelmüller, A. Kholodov, et al. 2019. Northern Hemisphere permafrost map based on TTOP modelling for 2000–2016 at 1 km2 scale. Earth-Science Reviews 193:299–316. doi:https://doi.org/10.1016/j.earscirev.2019.04.023.

 Web of Science ®Google Scholar