Significant soil warming across Alaska permafrost and non-permafrost regions from 1997 to 2023

Abstract

Introduction:

Air temperatures in Alaska are increasing at twice the rate of the contiguous United States. Soil temperatures have been shown to be increasing across various landscapes but have overall received less attention. No comprehensive studies have been conducted in Alaska looking at soil temperature trends on a broad scale.

Methods:

In this paper we synthesized soil and air temperature data from 43 weather stations across Alaska spanning a 27-year period (1997-2023). Stations were divided into three regions based on permafrost extent (continuous, discontinuous, and no permafrost). Soil temperature trends were calculated at 5, 20, and 50 cm depth in all three permafrost regions, and additionally at 70, 95, and 120 cm depth for the continuous permafrost region only.

Results:

Annual average soil temperatures increased significantly across all regions with the fastest warming rates occurring at the highest latitudes. Whole profile warming averaged 0.64, 0.37, and 0.35°C dec-1 in the continuous, discontinuous, and no permafrost regions, respectively. Warming was not slowed by increasing soil depth. Air temperatures warmed faster than soil (p < 0.05) in the continuous permafrost region (1.15°C dec-1) but were not significantly different from soil in the discontinuous and no permafrost regions (p > 0.1). Seasonally, soils in the continuous and discontinuous regions warmed fastest in the winter months, whereas in the no permafrost region soils warmed fastest in the summer months.

Discussion:

The relationship between air and soil temperatures was significantly impacted by the presence of snow in the winter months. The correlation between air and soil temperatures was weakened by the presence of snow in each permafrost region, particularly those with deeper snowpacks (discontinuous and no permafrost).

1 Introduction

The higher rate of climate change in the Arctic relative to lower latitudes, referred to as polar amplification, has been the subject of many studies [Aalto et al., 2014; Intergovernmental Panel on Climate Change (IPCC), 2023; Polyakov et al., 2002; Previdi et al., 2021; Stuecker et al., 2018]. The studies have primarily focused on air temperature and precipitation trends. However, soil temperatures are also expected to be affected by polar amplification, and to have a more direct impact than air temperature on ecosystem and hydrologic processes including permafrost thaw, thermokarst formation, soil decomposition, and root and seedling growth. Although air and soil temperature trends are certainly coupled, they are not interchangeable, because soils are likely to be buffered from extreme air temperatures during long winters with consistent snow cover (Lawrence and Slater, 2010; Soong et al., 2020; Lembrechts et al., 2022). Until recently, a paucity of soil temperature monitoring in the Arctic has made it difficult to retrospectively consider how soil temperatures are changing in this region and elsewhere (Lembrechts et al., 2020; Phillips, 2020). The present study helps to fill that gap by providing a state-wide evaluation of soil temperature change across Alaska, spanning Arctic, sub-Arctic, and maritime climate regions.

The Arctic region has experienced some of the strongest climatic changes globally. Estimates of recent air temperatures trends have ranged from approximately 0.50–0.90 °C dec⁻¹ for latitudes 60–90°N (Chylek et al., 2009; Bieniek et al., 2014; Comiso and Hall, 2014; Przybylak and Wyszyński, 2020; Ballinger et al., 2023). Precipitation has also increased across the Arctic throughout the year, and a significant retreat of sea ice extent has been observed (Bengtsson et al., 2004; Serreze and Barry, 2011; Ford and Frauenfeld, 2022; Serreze et al., 2024). In addition to the risk of releasing large amounts of sequestered carbon to the atmosphere, thawing permafrost will have extensive consequences for Arctic infrastructure (Larsen et al., 2008; Ramage et al., 2021; Manos et al., 2025). In Alaska, it is estimated that 169,000 people live on top of permafrost and the number increases to 5 million people in the global Arctic (Ramage et al., 2021). However, those numbers are expected to decrease to 29,000 and 1.7 million, respectively, by 2050 due to the degradation of permafrost. A recent analysis by Manos et al. (2025) estimated that the total costs of the destabilization of infrastructure in Alaska in the continuous and discontinuous permafrost zones will be between $37–51 billion by mid-century.

Climate change, particularly in permafrost regions, has been responsible for large changes in terrestrial ecosystems including losses of soil carbon (Schuur et al., 2008; Vanhala et al., 2008; Grosse et al., 2011), changes in plant growth (Euskirchen et al., 2006; Bunn et al., 2007), biodiversity (Pauchard et al., 2016; Trew and Maclean, 2021; Rubenstein et al., 2023), and hydrologic cycles (Kundzewicz, 2008; Ballinger et al., 2023). The risk of releasing large amounts of sequestered carbon into the atmosphere through the thawing of permafrost has been well documented (Schuur et al., 2008; Jorgenson et al., 2010; Grosse et al., 2011; Zhu et al., 2019; Leal Filho et al., 2023). Tarnocai et al. (2009) estimated that in the northern circumpolar permafrost region there is approximately 1,672 Pg of organic carbon, accounting for about 50% of the estimated belowground organic carbon pool. High latitude regions (north of 60°N) could shift from being a sink to a source of CO₂ by the end of the 21st century (Koven et al., 2011). Methane emissions are also expected to increase due to permafrost thaw (Christensen et al., 2004; Anisimov, 2007; Olefeldt et al., 2013; Neumann et al., 2019).

Existing research of soil temperature trends at high latitudes show that soils are significantly warming with decadal average temperature trends reported between 0.11 and 1.00 °C dec⁻¹ for a range of depths (Oelke et al., 2004; Zhang et al., 2005; Woodbury et al., 2009; Qian et al., 2011; Batir et al., 2017; Reshotkin and Khudyakov, 2019; Chen et al., 2021; Royer et al., 2021; Swanson et al., 2021; Petersen, 2022; Smith et al., 2022; AMAP, 2024). The magnitude of reported trends varies by region, latitude, time period, and methodology (observed vs. modeled). Many of the cited studies have also found that air temperatures are one of the major influences on soil temperatures, often outpacing soil warming, with trends ranging widely from 0.1 to over 1.0 °C dec⁻¹.

The decoupling of soil and air temperatures trends is attributed to the presence of snow in the winter months (Sturm et al., 1997; Zhang, 2005). Though several studies have looked at the effect of snow cover on soil temperatures, the extent to which snow insulates soil varies based on snow depth and timing as well as ecosystem type (Brown et al., 2010; Lawrence and Slater, 2010; Maurer and Bowling, 2014; Ekici et al., 2015; Zhang et al., 2018). In addition to snow cover, there are other factors that can influence soil temperatures that may make air temperature trends a poor predictor of soil temperature trends including elevation, topography, land use type, influence of water bodies, vegetation, soil moisture, soil structure, fire history, and geology (Lembrechts and Nijs, 2020; Lembrechts et al., 2020; Zellweger et al., 2020). Estimates of how soil and air temperature trends, as well as precipitation patterns, are very context dependent and thus there is a need for more observational studies of these parameters to understand how soil and air temperatures are changing on regional scales.

Soil temperature monitoring has been historically limited in Alaska, but has become more common in recent decades. Federal agencies including the USDA NRCS (National Water and Climate Center and the Soil Climate Research Stations), USGS, NPS, and NOAA, as well as scientists at the University of Alaska Fairbanks, established several soil temperature monitoring stations across Alaska in the late 1990s and early 2000s, but to our knowledge there has not yet been a systematic analysis of statewide soil warming trends. Though some of the stations included in this study have been analyzed in other studies (Wang et al., 2018; Smith et al., 2022), many have never been analyzed in this way before. In this paper we evaluated soil and air temperature trends across the state of Alaska and across permafrost regions (continuous, discontinuous, and no permafrost). We also evaluated trends in snow depth to understand how snow cover has influenced soil temperatures. We synthesized soil and air temperature records from 43 stations (snow depth from 36 stations) across a 12° latitude gradient (59–71°N). While there have been studies of changes in soil temperature at individual stations/regions over short timescales (Batir et al., 2017; Swanson et al., 2021), there have been limited statewide studies over longer periods of time. The only other state-wide synthesis, to our knowledge, is the permafrost data set by Wang et al. (2018) which included data from 72 weather stations (some of which were included in this study as well). However, many of the stations included had short timescales and thus were not suitable for trend analyses.

Though the relationship between air and soil temperatures, and the decoupling due to snow presence, is well understood, there is a critical lack of comprehensive, long-term observational baselines across Alaska. The objective of our study is to fill that gap by providing data on air, soil, and snow depth trends from across Alaska, as well as how each of these trends are related. To support these objectives, we have developed the following set of hypotheses. (1) Air, soil, and snow depth trends vary significantly across Alaska’s permafrost regions (continuous, discontinuous, and no permafrost) and by time of year. (2) Changes in air temperature are a major driving force of soil temperature changes in Alaska but the strength of that varies based on permafrost extent. (3) The coupling between air and soil temperatures is influenced by the presence and depth of winter snowpack. A deeper snowpack will provide greater insulation to the soil and lead to a decoupling of the relationship between air and soil temperatures.

2 Methods

2.1 Study location

The study region includes the state of Alaska, USA (latitude 54 to 71°N, longitude 130°W to 173°E). Alaska covers over 1.7 million square kilometers and represents a large range in climates. The climate zones encompass the Arctic at the northernmost latitudes (e.g., Utqiagvik, MAAT of −12.3 °C, MAP 119 mm) to a maritime zone along the southern coast (e.g., Juneau, MAAT of 5.1 °C, MAP 1502 mm) (climate zones based on designations used by the Alaska State Climate Center). Due to the large range of climates, the study area was divided into three zones based on the presence of permafrost (zones of continuous, discontinuous, and no permafrost). Permafrost zones were determined using the USDA Natural Resources Conservation Service Major Land Resource Areas (MLRA) designations for Alaska (United States Department of Agriculture, Natural Resources Conservation Service, 2022). The zone of continuous permafrost primarily consists of the region north of the Arctic Circle (primarily the North Slope), with Koppen Climate Classifications of ET (Tundra), and Dfc (Subarctic). The zone of discontinuous permafrost consists of what is generally referred to as the interior portion of Alaska with Koppen Climate Classifications of Dfc and Dsc (Dry-summer subarctic). The no permafrost zone covers the southern portion of the state with Koppen Climate Classifications of Dsc, Dfc, Cfc (Subpolar oceanic), and Cfb (Oceanic).

2.2 Data analysis

Soil temperature data were collected from weather stations from a variety of sources (Table 1). Stations with at least 10 consecutive years of daily data collection (with few gaps) were included. Stations located on mountain peaks, rocky terrains, and glacial areas were excluded. A full list of stations is provided in Supplementary Table S1. After data cleaning and quality control (Supplementary Figure S1) we ended up with 43 stations, of which 21 were located in the continuous permafrost region, 13 in the discontinuous, and 9 in the no permafrost region (Figure 1). Daily snow depth data was included in 29 of the stations described above (see Supplementary Table S1 for list of stations). An additional 5 stations were included just for snow depth data (Table 1). Data records ranged from 10 to 27 years long, between 1997 and 2023 (Figure 1). Stations reported hourly or daily temperatures. For each station, the reported hourly or daily soil and air temperature data were processed to remove unrealistic values including temperatures outside of the range of −70 to 40 °C, large sudden jumps in temperature, anything that resembled a sensor malfunction, or obvious outliers (assessed visually) (Supplementary Figure S1). Hourly data remaining after plausibility checks were summarized to daily averages (Supplementary Figure S2). Monthly averages (T_AVG) were then calculated from the daily values. Months that had fewer than 25 days of data were excluded. Annual averages were calculated from the monthly averages, and only years that had data from all 12 months were included in analyses evaluating annual average temperatures. Minimum (T_MIN) and maximum (T_MAX) monthly and annual temperatures were calculated as the minimum or maximum of the daily average temperatures. Monthly and annual temperature range (T_RANGE) was calculated by subtracting the minimum temperature from the maximum.

Figure 1

The depths at which soil temperatures were measured varied across the different stations. To accommodate this variation, depths were binned based on which depths were most prevalent across sites. In this study, soil temperature trends are reported at 5, 20, 50, 70, 95, and 120 cm depth. The 5, 20, and 50 cm depths represent data reported in 0–5, 15–20, and 45–50 cm ranges, respectively. If a station reported more than one depth within a bin, the depths were averaged before trends were calculated. The 70, 95, and 120 cm depths were only reported by stations in the continuous permafrost region from the USGS stations and the USDA NRCS Soil Climate Research Stations. These deeper depths represent discrete depths for each station and were not binned across a range of depths.

To evaluate the effect of snow depth and timing on soil temperature, daily snow depth data was also collected, and monthly averages were determined using the same methods as for soil and air temperatures. Number of snow cover days (SCD) were determined by summing the number of days within a year that had a reported snow depth above 0 cm. Onset was determined as the first day that a station reported a snow depth above 0 cm and thaw as the last day that snow was present.

2.3 Trend analysis

The non-parametric Mann-Kendall trend test was used to calculate trends in soil, air, and snow depth(Gilbert, 1995; Kendall, 1975; Mann, 1945) (Supplementary Figure S3). Trends were evaluated at each station and each depth independently. Temperature data were non-normally distributed with histograms containing multiple peaks due to distinct summer and winter regimes. Annual trends were calculated by fitting times series of monthly averages using the Seasonal Mann-Kendall trend test developed by Hirsch et al. (1982) implemented with the “EnvStats” R package (v2.8.1, Millard, 2013). The Mann-Kendall test uses Sen’s slope estimator (Sen, 1968) to determine change in a given variable over time (e.g., °C/year for temperatures or cm/year for snow depth). Slopes were multiplied by 10 to determine decadal rates. All tests were run in R version 4.4.2 (R Core Team, 2024).

We also analyzed year-over-year trends for each month (i.e., monthly trends) to evaluate seasonal changes. We tested for autocorrelation using the Ljung-Box test (Box and Pierce, 1970) in the base R “stats” package. If no autocorrelation existed, monthly trends were calculated using the regular Mann-Kendall Trend test also in the “EnvStats” package. If autocorrelation did exist, we applied a modified version of the Mann-Kendall test where the trend is prewhitened using the Zhang method (Zhang et al., 2000). This modified version was calculated using the “zyp” package (v0.11–1) (Bronaugh and Schoeneberg, 2023). Monthly soil temperature trends were calculated at each station and depth. When evaluating on a seasonal basis, monthly trends within a season are averaged (tests not calculated on a seasonal basis) and significance is based on the result of the Wilcoxon test comparing the population of trends within a season to zero.

Air temperature trends were calculated at each station using the same methods. Additionally, monthly snow depth trends were analyzed in a similar way. Trend significance for each individual test was based on the p-value (p < 0.05). When evaluating whether trends (including temperature and snow depth) were significant across a permafrost region, the Wilcoxon test was used to evaluate whether the median trend across sites in a region differed from zero.

Since stations varied in the number of years measurements were recorded (10–27), we evaluated if trends were significantly affected by the number of years included, by binning stations into three categories based on number of years of data collection (10–15 years, 15–20 years, 20 + years). Comparing slopes between these three categories showed no significant differences (Kruskal-Wallis, p-value > 0.1) for any of the permafrost regions. Since there was no evidence that the calculated slopes (or their significance) were affected by the number of years included in the calculation, stations were not subdivided based on time period for any analyses. A sensitivity analysis for spatial dependence among stations for soil, air, and snow trends was tested for using Moran’s I test in the “ape” (v5.8–1) package in R (Moran, 1950; Paradis and Schliep, 2019).

3 Results

3.1 Regional temperature and snow depth profiles

Here we describe the general features of annual air and soil temperature profiles as well as snow depth and timing across Alaska’s permafrost regions. Across all three regions, mean annual air temperatures were colder than soil temperatures (Table 2; Figure 2). Both T_MIN and T_AVG annual temperatures increased sharply moving from air into the soil. Within soil, annual T_AVG was essentially constant with depth and T_RANGE narrowed with increasing soil depth.

Figure 2

In the continuous permafrost region, air and soil T_AVG remained below freezing throughout the soil profile. Soil T_MIN reached well below freezing, demonstrating that once soils were frozen and had passed the phase change temperature plateau, they continued to cool efficiently. Summer T_MAX was above-freezing near the soil surface, reflecting seasonal thaw that occurred within the active layer of permafrost. Below about 50 cm depth, T_MAX remained near freezing, at the phase change temperature plateau. Annual T_RANGE contracted considerably from 55 °C in air, to 33 °C at 5 cm depth, to only 12 °C at 120 cm depth. Though the continuous permafrost region had the smallest snowpack on average (Table 3), this region had the longest snow season with an average of 251.18 (±20.18) snow cover days (SCD) per year. The median date of snow onset was Sept 29 (middle 50% of dates being between Sept 22 and Oct 5). The median date of snow thaw (last day snow is present) was May 30 (May 24–Jun 8).

Averaging across locations in the discontinuous permafrost region, annual air T_AVG was just below freezing, while annual soil T_AVG was just above freezing. Soils in this region had a narrower annual T_RANGE than permafrost soils. Annual air T_RANGE was similar in the permafrost and discontinuous zones (56 and 57 °C respectively), indicating a smaller ground heat flux in the discontinuous zone. Soil temperature at 5 cm depth ranged only 19 °C annually. In the discontinuous region, soil T_MIN was several degrees below freezing, but was considerably buffered from the colder air T_MIN (−36.67 °C). Soil T_MAX was several degrees above freezing but was more moderately buffered from air T_MAX. The snow season was slightly shorter in the discontinuous region compared to the continuous with an average of 204.98 (±32.62) SCD per year. The median date of snow onset in this region was Oct 12 (Oct 3–Oct 22) and the median thaw date was May 6 (Apr 25–May 17).

Locations in the no permafrost region had air T_AVG close to freezing, and soil T_AVG above freezing. Annual T_RANGE was narrower in air and soil compared to the discontinuous and permafrost regions, likely due to the maritime influence at several sites. The snow season length was similar to that of the discontinuous region with 200.64 (±25.57) SCD per year. This region had the latest median snow onset date of Oct 26 (Oct 15-Nov 4), about two weeks later than the discontinuous region. The median thaw date, however, was slightly later than the discontinuous with an average thaw date of May 17 (May 6–May 24).

3.2 Soil temperature trends

3.2.1 Annual soil temperature trends

Annual soil temperatures significantly warmed throughout Alaska, with the highest rates occurring in the Arctic region and decreasing with latitude (Figure 3). In the continuous permafrost region, annual soil warming was significant across the entire soil profile (Wilcoxon test, p < 0.05, Table 4; Figure 3) and there were no significant differences in warming rates across depths (Kruskal-Wallis, p > 0.05). However, the percentage of stations that showed a significant trend (Mann-Kendall trend test, p < 0.05) increased with depth from 70% of stations at 5 cm (N = 20) to 92% of stations at 120 cm (N = 13, Table 4). Soil T_MIN and T_MAX trends showed a similar pattern to T_AVG trends, averaging 0.70 and 0.43 °C dec⁻¹ respectively, with no significant differences in trends between depths. In the shallowest soil depths (5–20 cm) there were no significant differences between warming trends for T_AVG, T_MIN, and T_MAX (Kruskal-Wallis, p > 0.05). In the deeper soil depths (50–120 cm) rates of warming were significantly faster for T_AVG and T_MIN than for T_MAX. For depths 95–120 cm T_MIN warmed slightly faster than T_AVG (p = 0.08). There was evidence of some spatial autocorrelation among stations for soil trends (T_AVG) for depths 5, 20, 70, and 95 cm (p < 0.1, Moran’s I).

Figure 3

The discontinuous permafrost region had significant soil warming (0.37 °C dec⁻¹) across all depths, with no significant differences in rate across depths (Table 4). The percent of stations with significant warming were 55, 45, and 60% at 5, 20, and 50 cm depth, respectively. Soil T_MIN and T_MAX exhibited similar patterns across the soil depths with an average warming trend of 0.48 and 0.26 °C dec⁻¹, respectively, across the soil profile. T_AVG and T_MAX at 5 cm depth were not significantly warming (p > 0.05). Despite the slower warming trend for T_MAX, there were only slight differences (p = 0.07) between T_MIN and T_MAX at 5 cm depth and no significant differences at 20 or 50 cm depth. There was also evidence of spatial autocorrelation for both the 20 (p < 0.05) and 50 (p < 0.1) cm depth. However, this was primarily due to the fact that there are two stations located near the University of Alaska Fairbanks, though they are in different locations. When one of the two stations is removed the autocorrelation is no longer significant.

In the no permafrost region, soil T_AVG warmed significantly at a rate of 0.35 ± 0.21 °C dec⁻¹ (Table 4) with no significant differences between 5 to 50 cm depths. The percentage of stations with significant warming was 44% at 5 cm depth, and 56% at 20 and 50 cm depths. Soil T_MIN and T_MAX were significantly warming across the soil profile and there were no significant differences between T_AVG, T_MIN, and T_MAX warming rates at any depths. There was no spatial autocorrelation among stations for soil trends in this region (p > 0.1).

3.2.2 Seasonal temperature trends

Seasonal soil warming trends are presented in Figure 4 (results for individual months are available in Supplementary Table S3). In the continuous permafrost region, soil temperatures warmed the fastest in winter (DJF) and spring (MAM) averaging 2.14 (±1.46) and 2.32 (±1.48)°C dec⁻¹, respectively, and warming was significant for all depths (p < 0.05, Wilcox test). Shallow soil warming (5–20 cm) was not significant (p > 0.05) in summer (JJA). At deeper depths (50–120 cm) soil warming was significant in summer (0.44 ± 0.47 °C dec⁻¹), though slower than in winter and spring. In fall (SON) soil warming at all depths was significant but much lower than winter and spring averaging (0.47 ± 1.06 °C dec⁻¹).

Figure 4

The discontinuous permafrost region exhibited a similar pattern to the continuous with soils warming the fastest in winter, averaging 1.07 ± 1.52 °C dec⁻¹ across depths. Warming in spring was non-significant at the shallowest depth (0.31 °C dec⁻¹ at 5 cm), but significant at 20 and 50 cm depth (0.89 ± 1.40 °C dec⁻¹). In summer, soil was not significantly warming at any depths. In fall soils were significantly warming at all depths but at a slower rate than winter and spring, with an average rate of 0.47 ± 0.63 °C dec⁻¹ across depths.

The no permafrost region exhibited a pattern opposite to the continuous and discontinuous permafrost regions with warming occurring fastest in summer at an average rate of 1.04 ± 1.51 °C dec⁻¹ across depths with all depths warming significantly. Warming was significant in other seasons but at much lower rates. Winter, spring, and fall warming rates averaged 0.22, 0.25, and 0.36 °C dec⁻¹, respectively. Air temperature trends.

3.3 Air temperature trends

3.3.1 Annual air temperature trends

Air temperature trends were significantly warming throughout Alaska, but most rapidly in the continuous permafrost region (Table 4). Air T_AVG, T_MIN, and T_MAX were all significantly warming in the continuous permafrost region averaging 1.15, 1.18, and 1.06 °C dec⁻¹, respectively. For all three trends, air temperatures were warming significantly faster than soil temperatures. However, the percentage of stations that had a significant warming trend (Mann-Kendall trend test, p < 0.05) was lower for air temperature trends compared to soil temperature trends with only 53% (n = 19) of stations significantly warming. There was spatial autocorrelation among stations in this region for air temperature trends (T_AVG, p < 0.05, Moran’s I).

In the discontinuous permafrost region, air T_AVG and T_MIN were both significantly warming at 0.45 and 0.71 °C dec⁻¹, respectively. Air T_MAX in this region was not significantly warming. There were no significant differences between air and soil warming rates (p > 0.05, Kruskal-Wallis). Similar to the continuous permafrost region, fewer stations had a significant trend for air temperatures with only 18% of stations having a significant Mann-Kendall trend test result. There was no spatial autocorrelation for air trends in this region (p > 0.1).

In the no permafrost region, air T_AVG and T_MIN were significantly warming at 0.60 and 0.45 °C dec⁻¹. As in the discontinuous permafrost region, air T_MAX was not significantly warming. Warming rates between air and soil were not significantly different for air T_AVG, T_MIN, or T_MAX. Though the average warming rate for stations in this region was significantly greater than 0 (Wilcox test, p < 0.05), none of the individual stations had a significant Mann-Kendall trend test result for air temperature trends in this region. There was no spatial autocorrelation for air trends in the no permafrost region (p > 0.1).

3.3.2 Seasonal air temperature trends

In the continuous permafrost region, air temperatures were warming the fastest in spring at 2.55 (± 2.41)°C dec⁻¹ followed by significant warming trends in winter and fall (1.75 and 1.64 °C dec⁻¹, respectively) (Figure 4). Neither soil nor air temperatures were significantly warming in summer. In fall, air temperatures were warming significantly faster than soil temperatures at all depths (p < 0.05, pairwise Wilcox test). There were no differences in warming rates between air and soil temperatures in winter or spring.

In the discontinuous permafrost region, air temperatures were significantly warming in winter, summer, and fall, but not in spring. Warming was fastest in fall (1.00 ± 1.44 °C dec⁻¹), followed by winter (0.56 ± 1.52 °C dec⁻¹), and summer (0.37 ± 0.79 °C dec⁻¹). There were no significant differences between air and soil warming in any season. Similarly, air temperatures in the no permafrost region were significantly warming in winter, summer, and fall with an average of 0.61, 0.75, and 0.69 °C dec⁻¹, respectively. Air temperatures were not warming in spring (0.19 ± 1.23 °C dec⁻¹). Winter air temperature trends were warming significantly faster than soil temperature trends at all depths.

3.4 Snow depth and timing trends

During the timeline of this study (1997–2023), there were only modest changes in snow depth or timing throughout the state. For the continuous permafrost region snow depth trended toward an increase during winter (DFJ) months and significantly increased in December by an average rate of 4.47 ± 7.46 cm dec⁻¹ (Figure 5C). However, this region did not experience any significant changes in snow timing, including number of snow cover days (SCD), snow onset date, and snow thaw date (Table 3).

Figure 5

Snow depth also increased significantly in the discontinuous permafrost region from December to May with the largest increase from Feb-Apr (16.88 ± 11.65 cm dec⁻¹) (Table 3; Figure 5C). The discontinuous region also exhibited a significant increase in SCD over time (p < 0.05, Wilcox test) with an average increase of 14.51 (± 19.32) days dec⁻¹. This region also showed a significant trend in thaw date with an increase of 8.23 days dec⁻¹ meaning that snow thaw is occurring about a week later per decade. There was little increase in snow depth in the no permafrost region with the only month showing a significant trend being March (7.07 ± 8.60 cm dec⁻¹). There were also no significant changes in snow timing metrics in this region. It should be noted that year to year variations of snow timing and depth are substantial. Standard deviations within a region were high due to large differences in snow depth between stations. Snow depth trends for each individual station can be found in Supplementary Table S4. There was no evidence of spatial autocorrelation among stations for snow trend in any of the three permafrost regions. Relatonship between air and soil temperature.

Simple linear regression analysis of daily air vs. daily soil temperatures showed that soil temperatures were strongly influenced by air temperatures in each region (Figure 6A). The correlation coefficient (Pearson’s) was strongest in the continuous permafrost region (R = 0.86) and was slightly lower in the discontinuous and no permafrost regions (R = 0.79 and 0.75, respectively). The coupling between air and soil temperature was also evaluated under snow free vs. snow conditions (Figures 6B,C). In each permafrost region, the correlation between air and soil decreased when snow was present, however the degree varied by region. In the continuous permafrost region, the correlation between soil and air temperature had a coefficient value of 0.87 when snow was not present. When snow was present, the correlation coefficient decreased to 0.74. In the discontinuous permafrost region the correlation coefficient when snow was not present was 0.76 and decreased to 0.42 when snow was present. In the no permafrost region, the correlation coefficient was 0.70 when snow was not present and 0.34 when snow was present. The insulating effect of snow is also demonstrated in Figure 5. When snow was present (Figure 5A), the offset between shallow soil temperature (5 cm depth) and air temperature was positive, meaning that soil temperatures were warmer than air temperatures (Figure 5B). When snow was not present, the temperature offset was negative or near 0, meaning that air temperatures were warmer than soil.

Figure 6

4 Discussion

4.1 Coupling of air and soil temperatures

Soil and air temperatures warmed significantly throughout Alaska. Polar amplification was evident in soil temperatures as well as air temperatures, with warming occurring fastest in the continuous permafrost region (the Arctic) and decreasing with latitude (Table 4; Figure 3). Soil temperatures were warming across the soil profile (measured from 5 to 120 cm in the Arctic, and 5–50 cm at lower latitudes) and with no significant differences across measured depths within a region. Air temperature warming at high latitudes has been attributed to changes in atmospheric circulation and the retreat of sea ice (Bengtsson et al., 2004; Graversen et al., 2008; Walsh and Brettschneider, 2019). In the continuous permafrost region average annual air temperatures increased at a faster rate than soil temperatures (1.15 °C dec⁻¹ and 0.64 °C dec⁻¹ respectively) indicating that although air temperature was a significant driver of soil temperature, the influence was being moderated through other mechanisms such as snow cover.

The presence and depth of snow had paradoxical effects on soil warming, because snow insulated soils, keeping soil warmer than air during winter, and led to an overall higher annual T_AVG in soil than air (Table 2; Figure 2). However, snow presence also buffered soil from the full effects of air warming (Table 4). In other words, soils were warmer than air, but soils warm more gradually over time than air, where snow and ice occur. The insulating effects of snow and latent heat fluxes resulting from soil freeze–thaw dynamics, limited conductive and convective heat loss to the atmosphere on a seasonal and annual basis. This can be seen in warmer soil than air temperatures in winter (Figure 5B), a commonly observed feature in mid to high-latitudes (Chang, 1958; Beltrami and Mareschal, 1992; Baker and Ruschy, 1993; Soong et al., 2020; Lembrechts et al., 2022). However, these same buffering effects have restricted conductive heat transfer from the atmosphere to soil on a decadal basis, slowing soil warming relative to air warming (Zhang et al., 2005; Bouchard et al., 2024).

The coupling of soil and air temperatures in Alaska fluctuated seasonally, which resulted in differences in how much soils are warming relative to air. Factors influencing this coupling include the presence of snow, snow depth, and soil freezing and thawing (Henry, 2008; Lawrence and Slater, 2010; Maurer and Bowling, 2014; Ekici et al., 2015; Zhang et al., 2018). The difference in degree of snow thermal insulation was likely due to differences in snow depth in the three regions (Table 3; Figure 5A). In the continuous permafrost region where snow depth was shallowest (average depth 25.11, Nov–Mar), soil and air temperatures were highly correlated when snow was not present (r = 0.87) and still highly correlated when snow was present (0.74). In the discontinuous permafrost region where average snow depths were intermediate (42.84 cm. Nov––Mar), soil and air temperatures were highly correlated when snow was not present (r = 0.76), but much less when snow was present (r = 0.42). In the no permafrost region where winter snow depth was greatest (59.49 cm, Nov–Mar), soil and air temperature were correlated when snow is not present (r = 0.70), but only weakly so when snow was present (0.34). These results show that higher snow depths at lower latitudes provided greater thermal insulation to soils compared to those at higher latitudes.

Timing of snow, not just amount, can also have important effects on soil temperatures. In the early winter months, greater snow accumulation could trap more heat which would increase winter soil warming (Maurer and Bowling, 2014). On the other hand, less snow in the early months could result in colder soils as more heat is lost (Bouchard et al., 2024). The smaller snowpack in the fall/early winter in the continuous permafrost region, compared to the other regions, resulted in a lesser degree of snow insulation leading to colder soils. The rate of warming in the fall in the continuous permafrost region was significantly greater for air temperatures compared to soil temperatures in the fall (Figure 4). This was likely due to the loss of heat from soils during these months as the air was cooling, but the snowpack was shallow. Additionally, air temperatures were warming significantly faster than soil in the winter months in the no permafrost region, in this case it was the larger snowpack that buffered the soils from the warming air temperatures.

In spring, earlier thaw dates could make soils warmer in summer as soils thaw and dry out earlier (Seo et al., 2026). Later thaw dates would delay spring/summer warming and could decrease soil warming during those seasons (Maurer and Bowling, 2014). Trends in snow cover days (SCD) and thaw date in the discontinuous region suggest spring soil warming could lag behind air warming. Winter precipitation trends from other studies have found varied results, likely due to the high degree of heterogeneity in snowfall patterns. Some studies have found that snow was thawing earlier in Alaska and other parts of the Arctic (Brown et al., 2010; Swanson, 2017). However, it is projected that with warming temperatures, the hydrological cycle will intensify leading to more precipitation (Ballinger et al., 2023; AMAP, 2024). At northern latitudes (continuous and discontinuous) where temperatures are warming, but still well below freezing in winter, this is leading to increased snowfall in winter (McCrystall et al., 2021; Bigalke and Walsh, 2022). However, at more southern latitudes (no permafrost) this is leading to a decrease in snow and an increase in rainfall. The increased snowfall could lead to a later thaw date if it takes longer for the larger snowpack to fully melt.

Though it was not assessed in this study, snow density also affects thermal insulation of soil. Snow density controls the thermal conductivity of the snowpack with denser snow increasing thermal conductivity which leads to greater heat loss and colder soils (Sturm et al., 1997; Rixen et al., 2008; Cheng et al., 2025). Several factors influence snow density including air temperatures, snow depth, and snow cover duration (Bormann et al., 2013). Warmer climates are likely to increase snow density when winter air temperatures warm above 0 °C. Freeze–thaw events of snowpack increases the snow density as it becomes more compacted (Schaefer et al., 2009; Lawrence and Slater, 2010). Deeper snow depths and longer snow cover durations also compact snow and increase snow density (Che et al., 2025; Cheng et al., 2025). Given the importance of snow density on thermal insulation of soils, it should be taken into consideration when modeling the impact of snow on soil temperatures. In the continuous permafrost region where the snowpack is the smallest, changes in snow density are likely to have a significant effect on soil temperatures. Although snow density increases with snow depth, once the snowpack is deep enough, changes to snow density have minimal impact on the thermal insulation of soil (Ge and Gong, 2010). Therefore, although it is expected that the snowpack in the discontinuous and no permafrost regions would be denser due to the higher snow depth, the thermal insulation of the soil in these regions is greater because the snowpack is deep enough to offset changes in density.

4.2 Additional moderating factors on soil temperature warming

In summer, other factors like vegetation and soil moisture are more likely to influence soil warming than in winter (Kropp et al., 2021; Way and Lapalme, 2021). Ballinger et al. (2023) found that summer precipitation was increasing significantly in the North Slope (continuous permafrost) and interior (discontinuous permafrost) by 2.79 and 1.81% dec⁻¹, respectively, from 1957 to 2021. The increasing summer precipitation moderates soil warming in the summer months at higher latitudes, as soils with high moisture content are slower to warm than dry soils (Harazono et al., 1998; Boike et al., 2008; Ekici et al., 2015; Greiser et al., 2024). Precipitation was not significantly increasing in southcentral (no permafrost) region where most of our stations were located with a trend of just 0.62% dec⁻¹. They also found greater summer air warming along the southern coast of Alaska compared to more northern regions. The lack of an increase in precipitation, warmer air temperatures, and the long summer photoperiod would lead to drier soils that warm more quickly in the no permafrost region. Summer soil warming at lower latitudes in Canada have also been observed (Zhang et al., 2005; Qian et al., 2011).

In the continuous permafrost region, the summer (Jun–Aug) warming pattern, with slightly faster warming at deeper soil depths, may have been due to the longer amount of time it takes for deeper soils to thaw. Latent heat effects can cause a delay in the response of permafrost warming to air warming (Osterkamp, 2007). Many of the sites in this region were in wetland tundra ecosystems, which may also slow warming in the summer since water has a higher heat capacity than organic matter and mineral soil, making wet soils slower to heat (Hinkel et al., 2001; Ekici et al., 2015). Tundra ecosystems also tend to have less plant canopy shading in summer which could lead to warmer soils (Zellweger et al., 2020). Vegetation cover in the discontinuous and no permafrost region was much more varied and was beyond the scope of this study, but the effect of vegetation on soil temperatures and the effect of climate change on shifting vegetation cover is an important consideration when modeling future soil warming scenarios (Kropp et al., 2021; Way and Lapalme, 2021).

Topography also plays a role in warming rates (John et al., 2024). There was significant spatial autocorrelation for both air and soil temperature trends in the continuous permafrost region. This indicates that geographic features within the region may have an effect on warming rates and that it is not warming uniformly. Stations along the Brooks Range mountains (the SNOTEL and some of the NRCS Soil Climate Research Stations) showed soil and air warming that were slower compared to those in the Arctic coastal plains along the northernmost coast (the USGS and ITEX datasets). The four stations that had the slowest warming rates (for both air and soil) were also located at the highest elevations (between 244 and 930 m). The fastest warming rates were associated with stations that were below 200 m. Elevations for each station can be found in Supplementary Table S1.

4.3 Comparison to other high latitude studies

Comparing our results with other studies of observed or modeled soil and air temperature changes at similar latitudes shows that our results are largely consistent with the literature with our trends being in the upper range (Woodbury et al., 2009; Qian et al., 2011; Batir et al., 2017; Wang et al., 2018; Reshotkin and Khudyakov, 2019; Chen et al., 2021; Petersen, 2022). Annual trends of Arctic soil temperatures in other studies have been reported between 0.20 to about 1.00 °C dec⁻¹ across a range of depths from near surface soils to 320 cm depth (Oelke et al., 2004; Park et al., 2014; Batir et al., 2017; Chen et al., 2021; Royer et al., 2021; Smith et al., 2022). In this study the average soil warming trend across the continuous permafrost region soil profile was 0.64 °C dec⁻¹. Soils described as being in regions of discontinuous or no permafrost regions have reported on average lower annual warming trends compared to the continuous permafrost region ranging from 0.11 to 0.33 °C dec⁻¹ (Zhang et al., 2005; Woodbury et al., 2009; Qian et al., 2011; Reshotkin and Khudyakov, 2019; Chen et al., 2021; Petersen, 2022). Our results were consistent with the upper range of these studies with soil trends in the discontinuous and no permafrost regions averaging 0.35 and 0.37 °C dec⁻¹, respectively, across the soil profile.

There have been fewer studies of seasonal soil warming trends at high latitudes. Greater winter warming compared to summer warming has been observed in soil temperatures in this region though the magnitude varies by study (differences in time period, latitude, and region) with winter soil warming temperatures ranging from 0.44–2.37 °C dec⁻¹ and from 0.06–0.75 °C dec⁻¹ in summer (Oelke et al., 2004; Woodbury et al., 2009; Petersen, 2022). Our study also found greater winter warming in the continuous and discontinuous permafrost regions with winter (Nov-Mar) warming averaging 2.12 and 1.09 °C dec⁻¹, respectively. In contrast, in a reanalysis study of pan-Arctic soil temperature trends from 1979 to 2018 found that there was no difference between winter and summer soil warming with both seasons approximately warming by 0.40 °C dec⁻¹ (Royer et al., 2021). Studies of seasonal soil temperatures at lower latitudes (45–60°N) often report greater summer warming (0.30–0.93 °C dec⁻¹) compared to winter warming (0.17–0.80 °C dec⁻¹) which is also consistent with our findings (Zhang et al., 2005; Qian et al., 2011; Reshotkin and Khudyakov, 2019).

Although increased snow fall has been observed in the Arctic, as temperatures continue to warm, and if winter temperatures in the Arctic spend more time above freezing, this increased snowfall will shift to rainfall/freezing rain which could slow soil warming if soils are less protected from air temperatures (Lawrence and Slater, 2010; Maurer and Bowling, 2014) or if winters continue to warm quickly soils may warm faster with increasing air temperatures (Mellander et al., 2007; Henry, 2008).

5 Conclusion

The objective of this study was to provide a comprehensive evaluation of air and soil temperatures, as well as snow depth, trends from across Alaska. We hypothesized that each of these trends would vary significantly across the three permafrost regions. We found that both air and soil temperatures significantly warmed throughout Alaska and warmed the fastest in the continuous permafrost region. Air temperatures in the continuous permafrost region warmed at an average rate of 1.15 °C dec⁻¹, whereas soils warmed at 0.64 °C dec⁻¹ across depths. Rates were lower in the discontinuous and no permafrost regions for both air and soil temperatures but were still significant. There were no significant differences in soil warming by depth (5–120 cm). Trends also varied by time of year. In the continuous and discontinuous permafrost regions soils were warming fastest in the winter whereas in the no permafrost region soils were warming fastest in the summer months. Snow depth trends also varied by region, with only the discontinuous region showing a significant increase in snow depth across the winter months. Both the continuous and no permafrost regions exhibited limited increases in snow depth across the study period. As air temperatures continue to warm, altering precipitation patterns, changes in future snowpack depth could have significant effects on winter soil temperatures.

We also hypothesized that air warming is a major driver of soil warming but that the strength of the coupling between these two trends would vary by permafrost region. The regression analysis (Figure 6) between daily air and soil temperatures demonstrates that there is a significant coupling between and air and soil temperatures. Figure 6A shows that throughout the year air and soil are the most tightly coupled in the continuous permafrost region (slope = 0.51, R = 0.86), and slightly weaker in the discontinuous (slope = 0.35, R = 0.79) and no permafrost (slope = 0.32, R = 0.75) regions.

Our third hypothesis was that snow cover and depth would play a moderating role on soil temperature trends, providing insulation to buffer soils from extreme air temperatures in the winter leading to a decoupling of air and soil temperatures when snow was present. This hypothesis is supported in the regression analysis by comparing the relationship between daily snow and air temperatures when snow is not present (Figure 6B) and when snow is present (Figure 6C). In all three regions, the correlation was stronger when snow was not present than when it was present. Additionally, snow depth affected the extent of thermal insulation of soil from snow cover. Since the continuous permafrost region had the smallest snowpack (average of 25.11 cm, Nov-Mar) on average of the three permafrost regions, thermal insulation was weakest in this region. The larger snowpack in the discontinuous (42.84 cm) and no permafrost (59.49 cm) regions provided a greater buffer for soils from winter air temperatures leading to a decoupling of air and soil temperatures during winter in both regions.

These results provide a comprehensive evaluation of air and soil temperature trends across Alaska and how warming rates vary by region. It also provides an analysis of how snow depth moderates the relationship between air and soil temperatures in each region. This data can be used to inform forecasts of future permafrost thaw and the effect it will have on existing and future infrastructure in Alaska.

References

• 1

  AaltoJ.VenäläinenA.HeikkinenR. K.LuotoM. (2014). Potential for extreme loss in high-latitude earth surface processes due to climate change. Geophys. Res. Lett.41, 3914–3924. doi: 10.1002/2014GL060095

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AMAP (2024). AMAP Arctic Climate Update 2024: Key Trends and Impacts. Tromso, Norway: Arctic Monitoring and Assessment Programme (AMAP) doi: 10.21352/fz60-s852.

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AnisimovO. A. (2007). Potential feedback of thawing permafrost to the global climate system through methane emission. Environ. Res. Lett.2:045016. doi: 10.1088/1748-9326/2/4/045016

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BakerD. G.RuschyD. L. (1993). The recent warming in eastern Minnesota shown by ground temperatures. Geophys. Res. Lett.20, 371–374. doi: 10.1029/92GL02724

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  BallingerT. J.BhattU. S.BieniekP. A.BrettschneiderB.LaderR. T.LittellJ. S.et al. (2023). Alaska terrestrial and marine climate trends, 1957–2021. J. Clim.36, 4375–4391. doi: 10.1175/JCLI-D-22-0434.1

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BatirJ. F.HornbachM. J.BlackwellD. D. (2017). Ten years of measurements and modeling of soil temperature changes and their effects on permafrost in Northwestern Alaska. Glob. Planet. Chang.148, 55–71. doi: 10.1016/j.gloplacha.2016.11.009

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BellJ. E.PaleckiM. A.BakerC. B.CollinsW. G.LawrimoreJ. H.LeeperR. D.et al. (2013). U.S. climate reference network soil moisture and temperature observations. J. Hydrometeorol.14, 977–988. doi: 10.1175/JHM-D-12-0146.1

  • CrossRef
  • Google Scholar

• 8

  BeltramiH.MareschalJ. (1992). Ground temperature histories for central and eastern Canada from geothermal measurements: little ice age signature. Geophys. Res. Lett.19, 689–692. doi: 10.1029/92GL00671

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  BengtssonL.SemenovV. A.JohannessenO. M. (2004). The early twentieth-century warming in the Arctic—a possible mechanism. J. Clim.17, 4045–4057. doi: 10.1175/1520-0442(2004)017<4045:TETWIT>2.0.CO;2

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  BieniekP. A.WalshJ. E.ThomanR. L.BhattU. S. (2014). Using climate divisions to analyze variations and trends in Alaska temperature and precipitation. J. Clim.27, 2800–2818. doi: 10.1175/JCLI-D-13-00342.1

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  BigalkeS.WalshJ. E. (2022). Future changes of snow in Alaska and the Arctic under stabilized global warming scenarios. Atmos.13:541. doi: 10.3390/atmos13040541

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  BoikeJ.WilleC.AbnizovaA. (2008). Climatology and summer energy and water balance of polygonal tundra in the Lena River Delta. Siberia. J. Geophys. Res.113:G03025. doi: 10.1029/2007JG000540

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  BormannK. J.WestraS.EvansJ. P.McCabeM. F. (2013). Spatial and temporal variability in seasonal snow density. J. Hydrol.484, 63–73. doi: 10.1016/j.jhydrol.2013.01.032

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  BouchardB.NadeauD. F.DomineF.AnctilF.JonasT.TremblayÉ. (2024). How does a warm and low-snow winter impact the snow cover dynamics in a humid and discontinuous boreal forest? Insights from observations and modeling in eastern Canada. Hydrol. Earth Syst. Sci.28, 2745–2765. doi: 10.5194/hess-28-2745-2024

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  BoxG. E. P.PierceD. A. (1970). Distribution of residual autocorrelations in autoregressive-integrated moving average time series models. Journal of American Statistical Association65, 1509–1526. doi: 10.2307/2284333

  • CrossRef
  • Google Scholar

• 16

  BronaughD.SchoenebergA. (2023). Zyp: Zhang + Yue-Pilon Trends Package doi: 10.32614/CRAN.package.zyp.

  • CrossRef
  • Google Scholar

• 17

  BrownR.DerksenC.WangL. (2010). A multi-data set analysis of variability and change in Arctic spring snow cover extent, 1967–2008. J. Geophys. Res.115:D16111. doi: 10.1029/2010JD013975

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  BunnA. G.GoetzS. J.KimballJ. S.ZhangK. (2007). Northern high-latitude ecosystems respond to climate change. EoS Transactions88, 333–335. doi: 10.1029/2007EO340001

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  ChangJ. (1958). Ground Temperature. Milton, Massachussetts: Blue Hill Meteorological Observatory, Harvard University.

  • Google Scholar
  • View reference in article

• 20

  CheT.DaiL.LiX. (2025). Spatiotemporal distribution of seasonal snow density in the northern hemisphere based on in situ observation. Research in Cold and Arid Regions17, 137–144. doi: 10.1016/j.rcar.2025.02.004

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  ChenL.AaltoJ.LuotoM. (2021). Significant shallow–depth soil warming over Russia during the past 40 years. Glob. Planet. Chang.197:103394. doi: 10.1016/j.gloplacha.2020.103394

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  ChengY.ChengB.PirazziniR.MacfarlaneA. R.VihmaT.DornW.et al. (2025). Seasonal evolution of snow density and its impact on thermal regime of sea ice during the MOSAiC expedition. Cryosphere19, 6001–6021. doi: 10.5194/tc-19-6001-2025

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  ChristensenT. R.JohanssonT.ÅkermanH. J.MastepanovM.MalmerN.FriborgT.et al. (2004). Thawing sub-arctic permafrost: effects on vegetation and methane emissions. Geophys. Res. Lett.31:L04501. doi: 10.1029/2003GL018680

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  ChylekP.FollandC. K.LesinsG.DubeyM. K.WangM. (2009). Arctic air temperature change amplification and the Atlantic multidecadal oscillation. Geophys. Res. Lett.36:L14801. doi: 10.1029/2009GL038777

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  ComisoJ. C.HallD. K. (2014). Climate trends in the Arctic as observed from space. WIREs Clim. Change5, 389–409. doi: 10.1002/wcc.277

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  DiamondH. J.KarlT. R.PaleckiM. A.BakerC. B.BellJ. E.LeeperR. D.et al. (2013). U.S. climate reference network after one decade of operations: status and assessment. Bull. Am. Meteorol. Soc.94, 485–498. doi: 10.1175/BAMS-D-12-00170.1

  • CrossRef
  • Google Scholar

• 27

  EkiciA.ChadburnS.ChaudharyN.HajduL. H.MarmyA.PengS.et al. (2015). Site-level model intercomparison of high latitude and high altitude soil thermal dynamics in tundra and barren landscapes. Cryosphere9, 1343–1361. doi: 10.5194/tc-9-1343-2015

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  EuskirchenE. S.McGuireA. D.KicklighterD. W.ZhuangQ.CleinJ. S.DargavilleR. J.et al. (2006). Importance of recent shifts in soil thermal dynamics on growing season length, productivity, and carbon sequestration in terrestrial high-latitude ecosystems. Glob. Chang. Biol.12, 731–750. doi: 10.1111/j.1365-2486.2006.01113.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  FlemingS. W.ZukiewiczL.StrobelM. L.HofmanH.GoodbodyA. G. (2023). SNOTEL, the soil climate analysis network, and water supply forecasting at the Natural Resources Conservation Service: past, present, and future. J American Water Resour Assoc59, 585–599. doi: 10.1111/1752-1688.13104

  • CrossRef
  • Google Scholar

• 30

  FordV. L.FrauenfeldO. W. (2022). Arctic precipitation recycling and hydrologic budget changes in response to sea ice loss. Glob. Planet. Chang.209:103752. doi: 10.1016/j.gloplacha.2022.103752

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  GeY.GongG. (2010). Land surface insulation response to snow depth variability. J. Geophys. Res.115:2009JD012798. doi: 10.1029/2009JD012798

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  GilbertR. O. (1995). Statistical Methods for Environmental Pollution Monitoring., Nachdr. New York: Van Nostrand Reinhold.

  • Google Scholar

• 33

  GraversenR. G.MauritsenT.TjernströmM.KällénE.SvenssonG. (2008). Vertical structure of recent Arctic warming. Nature451, 53–56. doi: 10.1038/nature06502

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  GreiserC.HederováL.VicoG.WildJ.MacekM.KopeckýM. (2024). Higher soil moisture increases microclimate temperature buffering in temperate broadleaf forests. Agric. For. Meteorol.345:109828. doi: 10.1016/j.agrformet.2023.109828

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  GrosseG.HardenJ.TuretskyM.McGuireA. D.CamillP.TarnocaiC.et al. (2011). Vulnerability of high-latitude soil organic carbon in North America to disturbance. J. Geophys. Res.116:G00K06. doi: 10.1029/2010JG001507

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  HarazonoY.YoshimotoM.ManoM.VourlitisG. L.OechelW. C. (1998). Characteristics of energy and water budgets over wet sedge and tussock tundra ecosystems at north slope in Alaska. Hydrol. Process.12, 2163–2183. doi: 10.1002/(SICI)1099-1085(19981030)12:13/14<2163::AID-HYP727>3.0.CO;2-Y

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  HenryH. A. L. (2008). Climate change and soil freezing dynamics: historical trends and projected changes. Clim. Chang.87, 421–434. doi: 10.1007/s10584-007-9322-8

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  HillK.SousanesP. (2016). National Park Service Central Alaska Inventory and Monitoring Quality Controlled Climate Dataset. Available at:https://irma.nps.gov/DataStore/Reference/Profile/2240059.

  • Google Scholar

• 39

  HinkelK. M.PaetzoldF.NelsonF. E.BockheimJ. G. (2001). Patterns of soil temperature and moisture in the active layer and upper permafrost at Barrow, Alaska: 1993–1999. Glob. Planet. Chang.29, 293–309. doi: 10.1016/S0921-8181(01)00096-0

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  HirschR. M.SlackJ. R.SmithR. A. (1982). Techniques of trend analysis for monthly water quality data. Water Resour. Res.18, 107–121. doi: 10.1029/WR018i001p00107

  • CrossRef
  • Google Scholar

• 41

  HollisterR.Betway-MayK. (2022). Air and soil temperatures and soil moisture in the international tundra experiment (ITEX) plots at Utqiagvik and Atqasuk. Alaska. doi: 10.18739/A2V40K12R

  • CrossRef
  • Google Scholar

• 42

  Intergovernmental Panel on Climate Change (IPCC). (2023). Climate Change 2021 – The Physical Science Basis: Working Group I Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. 1st Edn. Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press doi: 10.1017/9781009157896.

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  JohnA.OldenJ. D.OldfatherM. F.KlingM. M.AckerlyD. D. (2024). Topography influences diurnal and seasonal microclimate fluctuations in hilly terrain environments of coastal California. PLoS One19:e0300378. doi: 10.1371/journal.pone.0300378

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  JorgensonM. T.RomanovskyV.HardenJ.ShurY.O’DonnellJ.SchuurE. A. G.et al. (2010). Resilience and vulnerability of permafrost to climate change. Can. J. For. Res.40, 1219–1236. doi: 10.1139/X10-060

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  KendallM. (1975). Rank Correlation Methods. 4th Edn. London: Griffin.

  • Google Scholar

• 46

  KovenC. D.RingevalB.FriedlingsteinP.CiaisP.CaduleP.KhvorostyanovD.et al. (2011). Permafrost carbon-climate feedbacks accelerate global warming. Proc. Natl. Acad. Sci. USA108, 14769–14774. doi: 10.1073/pnas.1103910108

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  KroppH.LorantyM. M.NataliS. M.KholodovA. L.RochaA. V.Myers-SmithI.et al. (2021). Shallow soils are warmer under trees and tall shrubs across Arctic and boreal ecosystems. Environ. Res. Lett.16:015001. doi: 10.1088/1748-9326/abc994

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  KundzewiczZ. W. (2008). Climate change impacts on the hydrological cycle. Ecohydrology & Hydrobiology195–203. doi: 10.2478/v10104-009-0015-y

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  LarsenP.GoldsmithS.SmithO.WilsonM.StrzepekK.ChinowskyP.et al. (2008). Estimating future costs for Alaska public infrastructure at risk from climate change. Glob. Environ. Chang.18, 442–457. doi: 10.1016/j.gloenvcha.2008.03.005

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  LawrenceD. M.SlaterA. G. (2010). The contribution of snow condition trends to future ground climate. Clim. Dyn.34, 969–981. doi: 10.1007/s00382-009-0537-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  Leal FilhoW.DinisM. A. P.NagyG. J.FracassiU. (2023). On the (melting) rocks: climate change and the global issue of permafrost depletion. Sci. Total Environ.903:166615. doi: 10.1016/j.scitotenv.2023.166615

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  LembrechtsJ. J.AaltoJ.AshcroftM. B.De FrenneP.KopeckýM.LenoirJ.et al. (2020). SoilTemp: a global database of near-surface temperature. Glob. Chang. Biol.26, 6616–6629. doi: 10.1111/gcb.15123

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  LembrechtsJ. J.NijsI. (2020). Microclimate shifts in a dynamic world. Science368, 711–712. doi: 10.1126/science.abc1245

  • CrossRef
  • Google Scholar
  • View reference in article

• 54

  LembrechtsJ. J.Van Den HoogenJ.AaltoJ.AshcroftM. B.De FrenneP.KemppinenJ.et al. (2022). Global maps of soil temperature. Glob. Chang. Biol.28, 3110–3144. doi: 10.1111/gcb.16060

  • CrossRef
  • Google Scholar
  • View reference in article

• 55

  MannH. B. (1945). Nonparametric tests against trend. Econometrica13:245. doi: 10.2307/1907187

  • CrossRef
  • Google Scholar

• 56

  ManosE.WitharanaC.LiljedahlA. K. (2025). Permafrost thaw-related infrastructure damage costs in Alaska are projected to double under medium and high emission scenarios. Commun. Earth Environ.6:221. doi: 10.1038/s43247-025-02191-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 57

  MaurerG. E.BowlingD. R. (2014). Seasonal snowpack characteristics influence soil temperature and water content at multiple scales in interior western U.S. mountain ecosystems. Water Resour. Res.50, 5216–5234. doi: 10.1002/2013WR014452

  • CrossRef
  • Google Scholar
  • View reference in article

• 58

  McCrystallM. R.StroeveJ.SerrezeM.ForbesB. C.ScreenJ. A. (2021). New climate models reveal faster and larger increases in Arctic precipitation than previously projected. Nat. Commun.12:6765. doi: 10.1038/s41467-021-27031-y

  • CrossRef
  • Google Scholar
  • View reference in article

• 59

  MellanderP.-E.LöfveniusM. O.LaudonH. (2007). Climate change impact on snow and soil temperature in boreal scots pine stands. Clim. Chang.85, 179–193. doi: 10.1007/s10584-007-9254-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 60

  MenneM. J.DurreI.VoseR. S.GleasonB. E.HoustonT. G. (2012). An overview of the global historical climatology network-daily database. J. Atmos. Ocean. Technol.29, 897–910. doi: 10.1175/JTECH-D-11-00103.1

  • CrossRef
  • Google Scholar

• 61

  MillardS. (2013). EnvStats: An R Package for Environmental Statistics. New York: Springer.

  • Google Scholar

• 62

  MoranP. A. P. (1950). Notes on continuous stochastic phenomena. Biometrika37, 17–23.

  • Google Scholar
  • View reference in article

• 63

  NeumannR. B.MoorbergC. J.LundquistJ. D.TurnerJ. C.WaldropM. P.McFarlandJ. W.et al. (2019). Warming effects of spring rainfall increase methane emissions from thawing permafrost. Geophys. Res. Lett.46, 1393–1401. doi: 10.1029/2018GL081274

  • CrossRef
  • Google Scholar
  • View reference in article

• 64

  OelkeC.ZhangT.SerrezeM. C. (2004). Modeling evidence for recent warming of the Arctic soil thermal regime. Geophys. Res. Lett.31:2003GL019300. doi: 10.1029/2003GL019300

  • CrossRef
  • Google Scholar
  • View reference in article

• 65

  OlefeldtD.TuretskyM. R.CrillP. M.McGuireA. D. (2013). Environmental and physical controls on northern terrestrial methane emissions across permafrost zones. Glob. Chang. Biol.19, 589–603. doi: 10.1111/gcb.12071

  • CrossRef
  • Google Scholar
  • View reference in article

• 66

  OsterkampT. E. (2007). Characteristics of the recent warming of permafrost in Alaska. J. Geophys. Res.112:F02S02. doi: 10.1029/2006JF000578

  • CrossRef
  • Google Scholar
  • View reference in article

• 67

  ParadisE.SchliepK. (2019). Ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics35, 526–528. doi: 10.1093/bioinformatics/bty633

  • CrossRef
  • Google Scholar
  • View reference in article

• 68

  ParkH.SherstiukovA. B.FedorovA. N.PolyakovI. V.WalshJ. E. (2014). An observation-based assessment of the influences of air temperature and snow depth on soil temperature in Russia. Environ. Res. Lett.9:F02S02. doi: 10.1088/1748-9326/9/6/064026

  • CrossRef
  • Google Scholar
  • View reference in article

• 69

  PauchardA.MilbauA.AlbihnA.AlexanderJ.BurgessT.DaehlerC.et al. (2016). Non-native and native organisms moving into high elevation and high latitude ecosystems in an era of climate change: new challenges for ecology and conservation. Biol. Invasions18, 345–353. doi: 10.1007/s10530-015-1025-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 70

  PetersenG. N. (2022). Trends in soil temperature in the Icelandic highlands from 1977 to 2019. Int. J. Climatol.42, 2299–2310. doi: 10.1002/joc.7366

  • CrossRef
  • Google Scholar
  • View reference in article

• 71

  PhillipsC. L. (2020). How much will soil warm? JGR. Biogeosciences125:e2020JG005668. doi: 10.1029/2020JG005668

  • CrossRef
  • Google Scholar
  • View reference in article

• 72

  PolyakovI. V.AlekseevG. V.BekryaevR. V.BhattU.ColonyR. L.JohnsonM. A.et al. (2002). Observationally based assessment of polar amplification of global warming. Geophys. Res. Lett.29:1878. doi: 10.1029/2001GL011111

  • CrossRef
  • Google Scholar
  • View reference in article

• 73

  PrevidiM.SmithK. L.PolvaniL. M. (2021). Arctic amplification of climate change: a review of underlying mechanisms. Environ. Res. Lett.16:093003. doi: 10.1088/1748-9326/ac1c29

  • CrossRef
  • Google Scholar
  • View reference in article

• 74

  PrzybylakR.WyszyńskiP. (2020). Air temperature changes in the Arctic in the period 1951–2015 in the light of observational and reanalysis data. Theor. Appl. Climatol.139, 75–94. doi: 10.1007/s00704-019-02952-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 75

  QianB.GregorichE. G.GamedaS.HopkinsD. W.WangX. L. (2011). Observed soil temperature trends associated with climate change in Canada. J. Geophys. Res.116:D02106. doi: 10.1029/2010JD015012

  • CrossRef
  • Google Scholar
  • View reference in article

• 76

  R Core Team. (2024). R: A Language and Environment for Statistical Computing. Available at:https://www.R-project.org/

  • Google Scholar

• 77

  RamageJ.JungsbergL.WangS.WestermannS.LantuitH.HeleniakT. (2021). Population living on permafrost in the Arctic. Popul. Environ.43, 22–38. doi: 10.1007/s11111-020-00370-6

  • CrossRef
  • Google Scholar
  • View reference in article

• 78

  ReshotkinO. V.KhudyakovO. I. (2019). Soil temperature response to modern climate change at four sites of different latitude in the European part of Russia. IOP Conf. Ser.: Earth Environ. Sci368:012040. doi: 10.1088/1755-1315/368/1/012040

  • CrossRef
  • Google Scholar
  • View reference in article

• 79

  RixenC.FreppazM.StoeckliV.HuovinenC.HuovinenK.WipfS. (2008). Altered snow density and chemistry change soil nitrogen mineralization and plant growth. Arct. Antarct. Alp. Res.40, 568–575. doi: 10.1657/1523-0430(07-044)[RIXEN]2.0.CO;2

  • CrossRef
  • Google Scholar

• 80

  RoyerA.PicardG.VargelC.LangloisA.GouttevinI.DumontM. (2021). Improved simulation of Arctic circumpolar land area snow properties and soil temperatures. Front. Earth Sci.9:685140. doi: 10.3389/feart.2021.685140

  • CrossRef
  • Google Scholar

• 81

  RubensteinM. A.WeiskopfS. R.BertrandR.CarterS. L.ComteL.EatonM. J.et al. (2023). Climate change and the global redistribution of biodiversity: substantial variation in empirical support for expected range shifts. Environ. Evid.12. doi: 10.1186/s13750-023-00296-0

  • CrossRef
  • Google Scholar

• 82

  SchaeferK.ZhangT.SlaterA. G.LuL.EtringerA.BakerI. (2009). Improving simulated soil temperatures and soil freeze/thaw at high-latitude regions in the simple biosphere/Carnegie-Ames-Stanford approach model. J. Geophys. Res.114:2008JF001125. doi: 10.1029/2008JF001125

  • CrossRef
  • Google Scholar

• 83

  SchuurE. A. G.BockheimJ.CanadellJ. G.EuskirchenE.FieldC. B.GoryachkinS. V.et al. (2008). Vulnerability of permafrost carbon to climate change: implications for the global carbon cycle. Bioscience58, 701–714. doi: 10.1641/B580807

  • CrossRef
  • Google Scholar

• 84

  SenP. K. (1968). Estimates of the regression coefficient based on Kendall’s tau. Journal of American Statistical Association63, 1379–1389.

  • Google Scholar

• 85

  SeoE.DirmeyerP. A.TakS. (2026). Implementation of a multi-layer snow scheme in the GloSea6 seasonal forecast system: impacts on land–atmosphere interactions and climatological biases. Geosci. Model Dev.19, 1261–1280. doi: 10.5194/gmd-19-1261-2026

  • CrossRef
  • Google Scholar

• 86

  SerrezeM. C.BarryR. G. (2011). Processes and impacts of Arctic amplification: a research synthesis. Glob. Planet. Chang.77, 85–96. doi: 10.1016/j.gloplacha.2011.03.004

  • CrossRef
  • Google Scholar

• 87

  SerrezeM. C.BigalkeS.LaderR.CrawfordA.BallingerT. J. (2024). NOAA Arctic report card 2024: precipitation. NOAA Technical Report. doi: 10.25923/xf7c-p592

  • CrossRef
  • Google Scholar

• 88

  SmithS. L.O’NeillH. B.IsaksenK.NoetzliJ.RomanovskyV. E. (2022). The changing thermal state of permafrost. Nat Rev Earth Environ3, 10–23. doi: 10.1038/s43017-021-00240-1

  • CrossRef
  • Google Scholar

• 89

  Soil Survey Staff, Natural Resources Conservation Service, and United States Department of Agriculture. (2023). Soil Climate Research Station Data. Available: Alaskaat:https://www.nrcs.usda.gov/resources/data-and-reports/soil-climate-research-stations.

  • Google Scholar

• 90

  SoongJ. L.PhillipsC. L.LednaC.KovenC. D.TornM. S. (2020). CMIP5 models predict rapid and deep soil warming over the 21st century. JGR. Biogeosciences125:e2019JG005266. doi: 10.1029/2019JG005266

  • CrossRef
  • Google Scholar

• 91

  StueckerM. F.BitzC. M.ArmourK. C.ProistosescuC.KangS. M.XieS.-P.et al. (2018). Polar amplification dominated by local forcing and feedbacks. Nature Clim Change8, 1076–1081. doi: 10.1038/s41558-018-0339-y

  • CrossRef
  • Google Scholar

• 92

  SturmM.HolmgrenJ.KönigM.MorrisK. (1997). The thermal conductivity of seasonal snow. J. Glaciol.43, 26–41. doi: 10.3189/S0022143000002781

  • CrossRef
  • Google Scholar

• 93

  SwansonD. K. (2017). Trends in greenness and snow cover in Alaska’s Arctic National Parks, 2000–2016. Remote Sens.9:514. doi: 10.3390/rs9060514

  • CrossRef
  • Google Scholar

• 94

  SwansonD. K.SousanesP. J.HillK. (2021). Increased mean annual temperatures in 2014–2019 indicate permafrost thaw in Alaskan national parks. Arct. Antarct. Alp. Res.53, 1–19. doi: 10.1080/15230430.2020.1859435

  • CrossRef
  • Google Scholar

• 95

  TarnocaiC.CanadellJ. G.SchuurE. A. G.KuhryP.MazhitovaG.ZimovS. (2009). Soil organic carbon pools in the northern circumpolar permafrost region. Glob. Biogeochem. Cycles23. doi: 10.1029/2008GB003327

  • CrossRef
  • Google Scholar

• 96

  TrewB. T.MacleanI. M. D. (2021). Vulnerability of global biodiversity hotspots to climate change. Glob. Ecol. Biogeogr.30, 768–783. doi: 10.1111/geb.13272

  • CrossRef
  • Google Scholar

• 97

  UeyamaM.IwataH.HarazonoY. (2024). AmeriFlux BASE US-Uaf University of Alaska Fairbanks, Fairbanks, Ver. 11-5,: AmeriFlux AMP doi: 10.17190/AMF/1480322

  • CrossRef
  • Google Scholar

• 98

  United States Department of Agriculture, Natural Resources Conservation Service (2022). Land Resource Regions and Major Land Resource Areas of the United States, the Caribbean, and the Pacific Basin. Department of Agriculture: U.S.

  • Google Scholar

• 99

  UrbanF. E.ClowG. D. (2018). DOI/GTN-P Climate and Active-Layer Data Acquired in the National Petroleum Reserve-Alaska and the Arctic National Wildlife Refuge, 1998–2019 (ver. 1.2, June 2021): U.S. Geological Survey Data Series 1092https://doi.org/10.3133/ds1092.

  • Google Scholar

• 100

  VanhalaP.KarhuK.TuomiM.BjörklöfK.FritzeH.LiskiJ. (2008). Temperature sensitivity of soil organic matter decomposition in southern and northern areas of the boreal forest zone. Soil Biol. Biochem.40, 1758–1764. doi: 10.1016/j.soilbio.2008.02.021

  • CrossRef
  • Google Scholar

• 101

  WalshJ. E.BrettschneiderB. (2019). Attribution of recent warming in Alaska. Pol. Sci.21, 101–109. doi: 10.1016/j.polar.2018.09.002

  • CrossRef
  • Google Scholar

• 102

  WangK.JafarovE.OvereemI.RomanovskyV.SchaeferK.ClowG.et al. (2018). A synthesis dataset of permafrost-affected soil thermal conditions for Alaska, USA. Earth Syst. Sci. Data10, 2311–2328. doi: 10.5194/essd-10-2311-2018

  • CrossRef
  • Google Scholar

• 103

  WayR. G.LapalmeC. M. (2021). Does tall vegetation warm or cool the ground surface? Constraining the ground thermal impacts of upright vegetation in northern environments. Environ. Res. Lett.16:054077. doi: 10.1088/1748-9326/abef31

  • CrossRef
  • Google Scholar

• 104

  WoodburyA. D.BhuiyanA. K. M. H.HanesiakJ.AkinremiO. O. (2009). Observations of northern latitude ground-surface and surface-air temperatures. Geophys. Res. Lett.36:L07703. doi: 10.1029/2009GL037400

  • CrossRef
  • Google Scholar

• 105

  ZellwegerF.De FrenneP.LenoirJ.VangansbekeP.VerheyenK.Bernhardt-RömermannM.et al. (2020). Forest microclimate dynamics drive plant responses to warming. Science368, 772–775. doi: 10.1126/science.aba6880

  • CrossRef
  • Google Scholar

• 106

  ZhangT. (2005). Influence of the seasonal snow cover on the ground thermal regime: an overview. Rev. Geophys.43:2004RG000157. doi: 10.1029/2004RG000157

  • CrossRef
  • Google Scholar

• 107

  ZhangX.VincentL. A.HoggW. D.NiitsooA. (2000). Temperature and precipitation trends in Canada during the 20th century. Atmosphere-Ocean38, 395–429. doi: 10.1080/07055900.2000.9649654

  • CrossRef
  • Google Scholar

• 108

  ZhangY.ChenW.SmithS. L.RiseboroughD. W.CihlarJ. (2005). Soil temperature in Canada during the twentieth century: complex responses to atmospheric climate change. J. Geophys. Res.110:D03112. doi: 10.1029/2004JD004910

  • CrossRef
  • Google Scholar

• 109

  ZhangY.SherstiukovA. B.QianB.KokeljS. V.LantzT. C. (2018). Impacts of snow on soil temperature observed across the circumpolar north. Environ. Res. Lett.13. doi: 10.1088/1748-9326/aab1e7

  • CrossRef
  • Google Scholar

• 110

  ZhuD.CiaisP.KrinnerG.MaignanF.Jornet PuigA.HugeliusG. (2019). Controls of soil organic matter on soil thermal dynamics in the northern high latitudes. Nat. Commun.10:3172. doi: 10.1038/s41467-019-11103-1

  • CrossRef
  • Google Scholar