Ground temperature and snow depth variability within a subarctic peat plateau landscape

Subarctic permafrost peatlands cover extensive areas and store large amounts of soil organic carbon that can be remobilized as active layer deepening and thermokarst formation increase in a future warmer climate. Better knowledge of ground thermal variability within these ecosystems is important for understanding future landscape development and permafrost carbon feedbacks. In a peat plateau complex in Tavvavuoma, northern Sweden, ground temperatures and snow depth have been monitored in six different landscape units: on a peat plateau, in a depression within a peat plateau, along a peat plateau edge (close to a thermokarst lake), at a thermokarst lake shoreline, in a thermokarst lake and in a fen. Permafrost is present in all three peat plateau landscape units, and mean annual ground temperature (MAGT) in the central parts of the peat plateau is −0.3°C at 2 m depth. In the three low‐lying wetter or saturated landscape units (along the thermokarst lake shoreline, in the lake and the fen) taliks are present and MAGT at 1 m depth is 1.0–2.7°C. Topographical differences between the elevated and low‐lying units affect both local snow depth and soil moisture, and are important for ground thermal patterns in this landscape. Permafrost exists in landscape units with a shallow mean December–April snow depth (<20 cm) whereas snow depths >40 cm mostly result in absence of permafrost.


| INTRODUCTION
Permafrost peatlands are widespread in the sporadic and discontinuous permafrost zones around the circum-Arctic and storẽ 300 × 10 15 g carbon (C), which is nearly 25% of the total soil organic carbon (SOC) in the northern circumpolar permafrost region. 1 Because most permafrost peatlands are located in the southern parts of the permafrost region they are already near thawing and very sensitive to future warmer conditions. 2,3 Extensive landscape changes have already started to take place in subarctic peatlands in recent decades as a result of permafrost thaw. [4][5][6][7][8][9][10] As long as the ground stays frozen, these environments are net carbon sinks through uptake of carbon dioxide (CO 2 ) by plants and peat accumulation. Under changing climatic conditions increased emissions of greenhouse gases (CO 2 , methane [CH 4 ] and nitrous oxide [N 2 O]) can be expected as deepening of the active layer, thermokarst processes and changes in the surface hydrology can cause increased remobilization of previously frozen soil carbon and nitrogen. [11][12][13][14][15] Peatlands cover extensive lowland areas in the northern circumpolar permafrost region (~13%). 1 Despite an overall relatively flat landscape, their small-scale topography can vary considerably. Permafrost peatland ecosystems are often characterized by a complex mosaic of different features, or landscape units, such as dry surface peat plateaus and palsas uplifted by frost heave, wet collapse scar fens or bogs and thermokarst lakes in depressions. In this study, landscape units are defined as a portion of the landscape with similar morphological characteristics, largely uniform in terms of topography, vegetation and water table.
Elevated palsas and peat plateaus typically have a thin snow cover, <30 cm, which allows extensive heat loss from the ground in winter, promoting permafrost development and persistence. [21][22][23] In low-lying landscape units, a thick snow cover can reduce heat flux from the soil, because snow has very low thermal conductivity. 24 Several studies have discussed temporal changes in ground temperature and active layer depth in palsa mires, and the impacts of vegetation and snow depth on ground thermal properties and palsa formation. 17,[19][20][21][22][23]26,27 However, few studies have focused on spatial variation in ground thermal regimes across landscape units in peat plateau ecosystems. The projected global warming and Arctic amplification will continue to cause permafrost thaw and ground subsidence in regions with extensive permafrost peatland coverage. 3 Therefore, a better understanding of climate-permafrost-hydrology interactions, and variability across landscapes and scales is essential for model projections of the future carbon balance in these morphologically heterogeneous environments. The objectives of this study are to (a) increase our knowledge of small-scale spatial ground thermal variability within subarctic peat plateau landscapes, and (b) discuss drivers of ground thermal regimes in different landscape units.

| STUDY AREA
Tavvavuoma is a widespread permafrost peatland area in northern Supporting Information Figure S2). The peat is underlain by glaciofluvial and lacustrine sediments. 29 Peat formation at the site started soon after deglaciation of the Fennoscandian Ice Sheet~10,000 cal yr BP. 30 Permafrost did not develop in these wetlands until during the Little Ice Age~600-100 cal yr BP. 30 Within the peat plateau complex, peat depth varies from around 0.5 to >2 m. The mean SOC content in the organic layer is~114 kg Cm −2 , and in the permafrost the wet-basis gravimetric ice content is >77%. 21,30 At present, the active layer depth in the peat plateau is around 55-60 cm, and the mean annual ground temperature (MAGT) is above −0.3 C and has increased by 0.06 C/yr between 2006 and 2013. 21 At the study site, mean annual air temperature is −2.1 C, and mean monthly temperature is 11.5 C in July and and many of the upper thermistors (e.g., at T1 and T5) have been recording air or lake temperatures, rather than soil temperatures, due to extensive ground collapse as a result of permafrost thaw along the lake shoreline and slow continuous ground subsidence on the peat plateau (Supporting Information Figure S1).  October. 16,17 Therefore the term late-season thaw depth is used in this study instead of active layer depth.

| Snow depth
Snow depth in the different landscape units was monitored by a stationary digital camera overlooking the peat plateau complex. Adjacent to each thermistor cable, snow depth stakes (S1-S9) were mounted The stakes (S1-S9) are installed close to, and numbered in accordance with, the thermistor cables presented in Figure 2 (SX is an extra snow depth stake, originally close to T5 at the lake shoreline, but over time inundated by water due to thermal erosionsee Figure S2). (c) Snow depth variability in different landscape units (S1-S8) during three winter seasons (2010/11-2012/13). The data are derived from pictures recorded approximately every 2 weeks. No data are available for S9 because the stake is partly hidden from the camera [Colour figure can be viewed at wileyonlinelibrary.com] variability and to calculate mean December-April snow depths in different landscape units.

| Statistical analysis
Spearman's rank correlation coefficient was used to determine if there was a statistically significant correlation between snow depth and ground temperature at the thermistor cable sites on the peat plateau and along the thermokarst lake shoreline (T1-T7). Because of the small sample size the p-value was calculated with Monte-Carlo permutation (100,000 permutations), using the statistical software Stata version 15. 33 The thermokarst lake and fen landscape units (T8 and T9) were excluded from this analysis because the thermal properties were probably not just affected by snow depth but also by overlying water/ice.
Moreover, no snow depth data were available for T9.

| Ground temperature
MAGT values just below 0 C were recorded at 1 m depth in the central parts of the peat plateau (T3, T2) and at the peat plateau edge close to the thermokarst lake shoreline (T7, T1) ( Figure 2 and Table 2).
In the small depression within the peat plateau (T4), MAGT was just above 0 C (0.3 C) at 1 m depth. However, at 2 m depth permafrost was present and MAGT was the same at T4, T3 and T2 (−0.3 C).
MAGT values >0 C (permafrost-free conditions) were recorded at 1 m depth at the lake shoreline (1.0 and 1.6 C at T6 and T5, respectively), with a slightly higher MAGT at T5 where the shoreline was actively eroding. Permafrost-free conditions also prevailed in the fen and in the lake sediments (MAGT 1.1 and 2.3 C at 2 m depth at T9 and T8, respectively). In the lake sediments, the minimum recorded ground temperature at 0.3 m depth was below 0 C (−0.03 C), supporting the observation during cable installation that the shallow lake freezes to the bottom in winter.

| Statistical analysis
Snow depth had an important impact on ground temperatures. There was a statistically significant correlation between mean December-April snow depth and MAGT at 1 m depth at T1-T7 (Spearman's rho = 0.9727, p = 0.0016).
In the depression within the peat plateau (T4) where mean December-April snow depth is 60 cm, late-season thaw depth is even greater (100-150 cm) than along the peat plateau edge ( On the thermokarst lake mean December-April snow depth is 41 cm. Similar snow depths can be expected for the fen, which is also located in a low-lying part of the landscape. Both in the lake and in the fen, soils are saturated, which increases the thermal conductivity of the peat. Winter freezing occurs down to~0.3 m depth in the lake sediments and~1 m depth in the fen (Figure 2c). Below these depths taliks are found. At the lake the talik is probably underlain by permafrost as the water level is higher compared to surrounding fens and lakes, whereas a through-talik is suggested beneath the fen based on geophysical methods (ground penetrating radar and electrical resistivity tomography). 33,41 The rather high MAGT values in these landscape units (2.7 C in the lake sediments and 1.3 C in the fen at 1 m depth) can be explained by a combination of a relatively thick insulating snow cover, peat saturation and particularly for the lake also the presence of an overlying water column that slows cooling and freezing in autumn.

| CONCLUSIONS
This study provides insights into ground thermal patterns within peat plateau complexes that can be valuable for understanding and projecting future landscape development in these ecosystems. Permafrost is present in the elevated peat plateau and along the peat plateau edge where mean December-April snow cover is thin (<20 cm). However, along the peat plateau edge a deeper late-season thaw depth suggests that the permafrost has started to degrade. Permafrost also occurs in a small and relatively dry depression in the peat plateau where mean snow cover is 60 cm, but a deeper late-season thaw depth (>100 cm) implies that vertical thaw is more extensive here compared to in the surrounding elevated plateau (thaw depth 55 cm). Taliks are present along lake shorelines where mean snow cover is deep (~80-90 cm), and in other low-lying and saturated landscape units such as thermokarst lakes and fens. Because the small-scale landscape morphology affects both snow depth and soil moisture, it is a key parameter for the ground thermal regime in peat plateau complexes. To improve our understanding of future permafrost peatland thaw, more empirical data on landscape heterogeneity, snow cover variability and duration, soil moisture, thermokarst formation and expansion rates, and connections between local hydrology and permafrost are needed. Ideally, ground thermal variability and landscape dynamics within permafrost peatlands should be incorporated into process-based ecosystem models trying to forecast permafrost thaw and associated carbon emissions.