Skip to main content
NCBI home page
Ambio logoLink to Ambio
. 2016 Jan 11;45(4):398–414. doi: 10.1007/s13280-015-0762-5

Local and regional variability in snow conditions in northern Finland: A reindeer herding perspective

Sirpa Rasmus 1,✉, Sonja Kivinen 2, Mathias Bavay 3, Janne Heiskanen 4
PMCID: PMC4824707  PMID: 26754168

Abstract

Weather station measurements were used to force the SNOWPACK snow model and combined with reindeer herders’ experiences to study the local and regional variations in snow conditions in a Finnish reindeer herding area for the 1981–2010 period. Winter conditions varied significantly between the four selected herding districts and between open and forest environments within the districts. The highest snow depths and densities, the thicknesses of ground ice, and the lengths of snow cover period were generally found in the northernmost districts. The snow depths showed the strongest regional coherence, whereas the thicknesses of ground ice were weakly correlated among the districts. The local variation in snow depths was higher than the regional variation and limits for rare or exceptional events varied notably between different districts and environments. The results highlight that forests diversify snow and foraging conditions, e.g., ground ice rarely forms simultaneously in different environments. Sufficient and diverse forest pastures are important during the critical winter season if reindeer herding is pursued on natural grazing grounds also in the future.

Keywords: Snow, Snow modelling, Exceptional snow conditions, Reindeer herding, Rangifer tarandus tarandus, Winter

Introduction

Reindeer herding is an integral part of human life in northern Fennoscandian and north-western Siberian tundra and boreal environments. This traditional land use is strongly dependent on climatic conditions, and particularly winter season is a critical period for reindeer grazing. Snow is one of the most important environmental factors affecting well-being of reindeer (Holleman et al. 1979). Exceptional winters with difficult snow and ice conditions have been reported to have long-term negative impacts on reindeer populations and consequently, economic and cultural viability of reindeer herding livelihood (Lee et al. 2000; Kumpula and Colpaert 2003; Helle and Kojola 2008).

The Finnish reindeer herding area is covered by snow for six to eight months per year (Solantie et al. 1996). In wintertime, semi-domesticated reindeer (Rangifer tarandus tarandus) feed mainly on ground-growing lichens (Cladina spp., Cetraria spp.) that they dig through the snowpack (Kumpula et al. 2004). Arboreal lichens (Bryoria spp., Alectoria spp, Usnea spp.) are important winter fodder especially in the late winter during difficult snow conditions (Kumpula 2001). The availability of winter resources can be defined by the total amount of winter forage in the region and reindeer access to these resources. The amount of winter resource varies in the reindeer herding area due to natural variation in vegetation as well as other land use activities, such as commercial forest management, that can have negative impacts on the quantity and quality of lichen pastures (Kivinen et al. 2010). Reindeer access to lichen resources is also related to snow and ice conditions, which can strongly vary within the winter season and between winters (Kohler and Aanes 2004; Hansen et al. 2011).

Climate characteristics in the Finnish reindeer herding area vary notably along latitudinal and continental–maritime gradients. Southern and central herding districts are located in the northern boreal vegetation zone, whereas vegetation in the northernmost districts is dominated by mountain birch forests and mountain heaths. Reindeer management practices in different districts are adapted to local environmental conditions and annual and seasonal weather variations (Kortesalmi 2007). Warming climate is predicted to cause more variable and unpredictable weather conditions especially in wintertime with changes in the amount and phase of precipitation, snow depth and structure, and snow cover duration (Stocker et al. 2013). This can significantly hamper reindeer winter grazing and practical herding, and increase costs for reindeer husbandry (Tyler et al. 2007).

In order to understand the winter-related challenges for reindeer herding, a comprehensive view of snow conditions and their variations in different biotopes is required. Projected changes in winter climate call for more knowledge especially on rare and exceptional snow-related events and their impacts on reindeer herding. Although the structure of snow cover is highly significant for reindeer and other northern animals, studies on snow structural properties are very limited in the circumpolar area. Further, earlier studies have mainly focused on the impacts of snow and ice on reindeer and caribou populations in treeless high-latitude regions (Miller and Barry 2009; Tyler 2010; Hansen et al. 2011; 2014). In this work, weather station measurements were used to force the SNOWPACK snow model and combined with reindeer herders’ experiences to study the local and regional variations in snow conditions in boreal and subarctic zones in northern Finland over the past 30 years. We studied snow and ground ice characteristics within open and forest environments (local level) in four reindeer herding districts (regional level). Our aims were to:

  1. study local and regional variability and trends in snow and ground ice conditions;

  2. compare limits for rare and exceptional snow events (snow depth and density, ground ice layer thickness) between different districts and environments;

  3. examine local and regional coherence in snow and ground ice conditions.

Materials and methods

Study area

The Finnish reindeer herding area is located approximately between 64–70°N and 21–30°E. Four reindeer herding districts (1) Poikajärvi, (2) Pohjois-Salla, (3) Kyrö, and (4) Käsivarsi were selected to represent the variation in climate, topography, vegetation, and herding practices within the area (Fig. 1; Table 1). Herding districts 1–3 are located in the northern boreal vegetation zone dominated by Scots pine (Pinus sylvestris (L.) and Norway spruce (Picea abies (L.), deciduous trees (mainly Betula spp.) occurring to a lesser extent. The northernmost Käsivarsi reindeer herding district is characterized by subarctic birch forests and treeless heaths. Bogs and mires cover 35–53 % of the reindeer herding area (Sevola 2002). Reindeer herding in the northern districts generally relies on natural pastures, whereas winter feeding and fencing are more common practices in southern districts (Turunen and Vuojala-Magga 2014).

Fig. 1.

Fig. 1

Open in a new tab

a The Finnish reindeer herding area (shaded in gray). b The location of the studied reindeer herding districts (1) Poikajärvi, (2) Pohjois-Salla, (3) Kyrö, and (4) Käsivarsi, and the observation stations A Apukka, B Värriö, C Sodankylä, D Pokka, and E Kilpisjärvi. Land cover classes are based on Corine Land Cover 2006 (Finnish Environment Institute)

Table 1.

Characteristics of the studied reindeer herding districts (Kumpula et al. 2009; Mattila 2012; Pirinen et al. 2012)

Reindeer herding cooperative Poikajärvi Pohjois-Salla Kyrö Käsivarsi
Total surface area (km2) 2522.3 2131.0 1737.6 4845.3
Largest permissible reindeer number 4600 4800 3500 10 000
Number of reindeer owners 104 104 92 174
Land cover
 Coniferous/mixed forest (%) 68.3 64.3 56.7 5.4
 Deciduoud forest (%) 0.3 0.7 1.6 16.2
 Transitional woodland (%) 15.7 14.2 9.5 0.6
 Open mire (%) 8.9 14.7 25.5 21.2
 Mountain heath, rock (%) 0.0 4.9 3.1 51.6
Observation station Apukka Sodankylä Pokka Kilpisjärvi
Elevation of the station (m.a.s.l) 106 179 275 480
Mean annual temperature (°C) 0.4 −0.4 −1.3 −1.9
Mean January temperature (°C) −12.8 −13.5 −14.0 −12.9
Mean July temperature (°C) 15.1 14.5 13.3 11.2
Mean annual precipitation (mm) 556 527 547 487
Mean maximum snow depth (cm) 68 87 101 110
Mean snowmelt date May 2 May 3 May 17 May 26
Open in a new tab

Weather data

Weather data for years 1981–2011 (30 winters) from five synoptic observation stations (Apukka, Sodankylä, Värriö, Pokka, and Kilpisjärvi) run by the Finnish Meteorological Institute (FMI) were used in snow modelling (Fig. 1). Following variables were obtained from the stations with 3–6 h resolution: (1) air temperature (°C), (2) relative humidity (%), (3) wind velocity (m s−1), and (4) wind direction (°). Precipitation was obtained as daily sum (mm); snow depth (cm) data were also daily. In addition to these, the dates of snow cover formation and melt were listed. Temperature and humidity were measured at 2-m height and wind at 10-m height. Daily sums of incoming shortwave radiation (kJ) were estimated from a gridded radiation product based on interpolation of available daily radiation values from different observation stations (FMI). From the Värriö observation station, we included only snow depth data to the analyses. As wind data were missing from this station and time resolution of weather observations was only 8 h until the 1990s, we decided to use weather data from the Sodankylä observation station in snow simulations for the Pohjois-Salla herding district. Winter climate in Sodankylä compares well with the mean conditions in the Pohjois-Salla district.

Reindeer herders’ experiences

The annual management reports of herding districts from the years 1981–2011 included reindeer herders’ experiences on winter and snow conditions (Reindeer Herders’ Association). Remarks of snow conditions experienced as problematic (e.g., deep snow, late snow melt, icy snow, and ground ice) were listed and compared with snow observations and ground ice observations and simulations.

Snow cover simulations

SNOWPACK model

The annual evolution of snow depth, snow density, and thickness of ground ice layers was simulated by the SNOWPACK model (Bartelt and Lehning 2002; Lehning et al. 2002a, b). SNOWPACK is a physically based, energy balance model simulating the temporal evolution of the snow mass and energy balance with an arbitrary number of layers. The ability of the model to reliably simulate the snow mass balance and snow structural properties has been validated, e.g., by Lehning et al. (1998) and Rasmus et al. (2007). A canopy sub-model (Lehning et al. 2006) simulates the impact of vegetation on the upper boundary conditions of the underlying snow cover and models snow depth and structure below forest canopies. The reliability of the sub-model in simulating the radiation transmission through the canopies and snow interception and subsequent effect of these on forest snow mass and energy balance has been validated by Stähli et al. (2009) and Musselman et al. (2012).

Model setup and simulations

As forcing data, SNOWPACK normally requires air temperature, relative humidity, wind velocity, incoming shortwave and longwave radiation, and precipitation or snow height with a temporal resolution of at least a few hours so that it can be resampled to the usual 15 min time steps of SNOWPACK. However, since all the parameters were not available or not at the proper temporal resolution, the MeteoIO preprocessing library (Bavay and Egger 2014) that is used by SNOWPACK was configured to use various parameterizations and temporal interpolation strategies. The incoming short wave radiation daily sum was compared with the potential radiation daily sum to compute a daily short wave atmospheric transmissivity that was applied to the instantaneous potential radiation for each SNOWPACK time steps. The daily precipitation sums were distributed to match the 3–6 h resolution of the rest of the input. The distribution was allocated to the time steps showing the highest probability of precipitation, and precipitation was then corrected for undercatch. The incoming longwave radiation was parametrized according to Dilley and O’Brien (1998) for clear sky conditions, and Unsworth and Monteith (1975) for cloudy sky conditions. As lower boundary condition, a constant geothermal flux of 0.04 W m−2 was applied to the bottom of 16-m of soil (Bartelt and Lehning 2002; Steinkogler et al. 2015).

In each herding district, simulations were performed for (1) open environments, (2) forests of average density, and (3) dense forests (only open environments in the Käsivarsi district) for the winters 1981/1982–2010/2011. Simulations covered the period between 1 September and 30 June annually. Canopy sub-model parameterizations were based on data from the forest areas reported by reindeer herders as regular winter grazing grounds. Representative canopy height for forest environments was derived from tree height maps by the Natural Resources Institute Finland (LUKE; http://kartta.metla.fi/) and leaf area index (LAI) based on countrywide LAI maps (Heiskanen et al. 2011). Canopy openness was estimated from LAI by a function presented by Pomeroy et al. (2002). The outputs of SNOWPACK model included a time series for the mass and energy balance components of the snow cover, and graphical and numerical time series of the snow structure for each studied environment.

Validation data for SNOWPACK simulations

Model validation data for the period 1981–2011 included snow depth observations from the weather stations (FMI) and snow water equivalent (SWE) and snow density measurements from permanent 4-km-long snow transects (Finnish Environment Institute) located close to the Apukka, Pokka, and Sodankylä stations (Perälä and Reuna 1990). Additional snow depth and SWE data as well as above snow cover radiation and snow temperature data were available for shorter periods from the Sodankylä station. The model was validated by Spearman’s rank correlation coefficient calculated between the observed and simulated values, mean bias error (MBE), and root mean square error (RMSE). No measurements of ground ice layers were available for model validation. Instead, we used graphical outputs of the models to identify winters with a thick (>5 cm) ground ice layer and compared these with reported occurrences of icy snow/ground ice in the annual management reports of the herding districts. The model performance seen in the validation should be seen as combined performance of SNOWPACK and the preprocessing by MeteoIO.

Simulated snow depths and SWEs showed a strong correlation with observation datasets (r = 0.78–0.94, p < 0.001) (Fig. 2; Table 2). Simulated annual maximum snow depths were in reasonably good concordance with the observed ones (Fig. 2). Due to open fell environment, the snow cover in the Käsivarsi district is highly variable in time and space. This is seen also in poorer performance of the model in Käsivarsi, especially during the winters with deep snow cover (Fig. 2). Correlations were modest for the snow bulk density (r = 0.42–0.73, p < 0.001) and snow surface and ground surface temperatures (r = 0.59–0.86, p < 0.001). Strong correlations were found between the observed and simulated snow surface radiation balance terms (r = 0.75–0.88, p < 0.001). MBE and RMSE between the observations and simulations varied among the locations (Table 2).

Fig. 2.

Fig. 2

Open in a new tab

Observed and simulated annual maximum snow depth (cm) in open environment in a Käsivarsi, b Kyrö, c Pohjois-Salla, and d Poikajärvi herding districts

Table 2.

Statistics for the SNOWPACK model validation. The Spearman’s correlation coefficient (*** p < 0.001), the mean bias error (MBE), and the root mean square error (RMSE) between the observed and simulated values

Correlation MBE RMSE
Snow depths (cm)
 Apukka (n = 45 657) 0.78*** −3.24 15.59
 Sodankylä (two datasets; n = 274/50 344) 0.90*** −1.71 11.05
 Pokka (n = 52 732) 0.91*** 2.38 12.92
 Kilpisjärvi (n = 55 589) 0.94*** 5.74 15.15
Snow water equivalent (mm)
 Pokka (two datasets; n = 136/148) 0.87*** −4.27 37.70
 Sodankylä (two datasets; n = 164/232) 0.87*** −26.46 42.47
Snow bulk density (kg m−3)
 Apukka (n = 31) 0.60*** −41.44 59.17
 Pokka (two datasets; n = 128/148) 0.73*** −23.39 55.69
 Sodankylä (n = 185) 0.42*** −67.86 107.84
Sodankylä only
 Outgoing SW radiation (W m−2; n = 697) 0.88*** −3.90 59.07
 Albedo (n = 470) 0.75*** −0.22 0.26
 Incoming LW radiation (W m−2; n = 1440) 0.84*** 48.88 124.07
 Snow surface temperature (two datasets; n = 229/320) (°C) 0.74*** 3.08 7.63
 Ground surface temperature (two datasets; n = 19/275) (°C) 0.73*** −1.30 2.43
Open in a new tab

Comparisons between the model outputs and reported snow conditions showed that 75–88 % of the winters with icy conditions and 62–73 % of the winter without icy conditions described in the annual management reports could be distinguished. The proportion of failed detections (ground ice reported but not simulated) was relatively small in all districts with only 1–2 missed events during the study period. The proportion of false occurrences (ground ice simulated but not reported) reflects a bias in the snow classification post-processing: when a layer contains even a minimal amount of liquid water, a special marker gets turned on. As soon as the layer refreezes, it gets associated with a melt/freeze crust with high hardness. More frequent simulated occurrences of ground ice compared to reported occurrences could also reflect the fact that herders’ may not report the icy conditions that occur only in limited geographical areas, or alternative grazing grounds have been used during difficult herding conditions. According to the validation statistics and analysis of icy conditions, the model setup seemed to be reliable and model performance was comparable to earlier studies (Lundy et al. 2001; Rasmus et al. 2007; Vikhamar-Schuler et al. 2013; Rasmus et al. 2014).

Statistical analyses of snow characteristics

The SNOWPACK simulation outputs were used to calculate mean values, standard deviation, and the coefficient of variation for (1) maximum snow depth, (2) mean snow density, and (3) ground ice thickness defined as layers at the bottom of the snow cover with density more than 350 kg m−3 in different environments. Snow characteristics were calculated for early winter (1 October–15 December), mid-winter (16 December–28 February), and late winter (1 March–15 May) over the 30-year study period. Winter periods were defined with the help of reindeer herders to represent the cold season from the reindeer herding perspective.

In order to study local and regional coherence in snow and ground ice conditions within reindeer herding districts and different environments, Spearman’s rank correlation coefficient was calculated for time series of snow and ground ice variables. Regional comparisons between different herding districts included only open environments. The Käsivarsi herding district located in the subarctic zone (no forests) was excluded from local comparisons. Long-term trends in snow characteristics were studied using linear and non-parametric models. The least squares method and non-parametric Sen’s trend estimate were used to calculate the magnitude of the trends. The statistical significance of trends was calculated using the Mann–Kendall trend test. Further, time series were studied for rare and exceptional events. Rare events were defined as those occurring once in 10 winters (three smallest and largest values of snow variables during the study period) and exceptional events as those occurring three times in 100 winters (the smallest and largest values of snow variables during the study period) (IPCC 2013).

Results

Snow conditions in reindeer herding districts

Regional snow characteristics

The highest maximum snow depths were simulated in the northernmost Kyrö and Käsivarsi districts, whereas the southernmost district Poikajärvi was generally characterized by the thinnest snow cover (Table 3a). The simulated mean snow densities did not vary considerably between the reindeer herding districts (Table 3b). Snow densities were the highest in late winter, and variability in snow densities was larger in early winter and mid-winter compared to late winter. The highest thicknesses of ground ice layers were found in late winter (Table 3c). The highest thickness of ground ice layers (open environments) occurred in the Käsivarsi district.

Table 3.

Statistics for average snow characteristics in different herding districts and environments (open, forest of average density, and dense forest) in winters 1981/1982–2010/2011

District Environment Early winter Mid-winter Late winter
Mean SD cvar Mean SD cvar Mean SD cvar
(a) Maximum snow depth (cm)
 Poikajärvi Open 36 12 0.35 69 16 0.22 74 17 0.23
Forest; average 26 9 0.37 57 14 0.24 60 16 0.27
Forest; dense 18 7 0.42 42 12 0.28 43 13 0.30
 Pohjois-Salla Open 44 15 0.33 84 14 0.16 87 16 0.19
Forest; average 25 10 0.41 55 12 0.21 54 16 0.29
Forest; dense 21 9 0.44 46 12 0.25 45 15 0.32
 Kyrö Open 50 15 0.30 86 17 0.20 92 16 0.17
Forest; average 36 13 0.36 68 17 0.24 70 19 0.27
Forest; dense 27 11 0.42 55 17 0.30 57 19 0.33
 Käsivarsi Open 47 15 0.32 85 17 0.20 88 17 0.19
 Total Open 44 81 85
Forest; average 29 60 61
Forest; dense 22 48 48
(b) Mean snow density (kg m−3)
 Poikajärvi Open 162 38 0.23 207 36 0.17 299 33 0.11
Forest; average 161 36 0.23 212 41 0.19 308 37 0.12
Forest; dense 160 30 0.19 217 48 0.22 324 38 0.12
 Pohjois-Salla Open 169 35 0.20 205 21 0.10 294 24 0.08
Forest; average 149 29 0.19 204 42 0.21 288 33 0.11
Forest; dense 149 32 0.21 205 49 0.24 290 38 0.13
 Kyrö Open 175 31 0.18 200 18 0.09 283 23 0.08
Forest; average 164 39 0.24 203 24 0.12 284 28 0.10
Forest; dense 167 42 0.25 216 38 0.18 297 29 0.10
 Käsivarsi Open 166 26 0.15 223 30 0.14 313 42 0.14
 Total Open 168 209 297
Forest; average 158 206 293
Forest; dense 159 213 304
(c) Ground ice thicknesss (cm)
 Poikajärvi Open 0.8 1.6 2.06 3.0 5.8 1.94 13.6 9.7 0.71
Forest; average 0.4 0.8 2.37 3.1 5.2 1.66 12.8 8.5 0.67
Forest; dense 0.2 0.5 2.52 4.3 4.6 1.05 13.1 9.0 0.68
 Pohjois-Salla Open 1.1 1.7 1.56 2.5 3.1 1.28 13.6 7.1 0.52
Forest; average 0.2 0.4 2.00 2.3 3.8 1.69 8.5 6.7 0.79
Forest; dense 0.2 0.4 2.58 2.4 3.5 1.43 7.6 5.8 0.76
 Kyrö Open 0.6 1.3 2.13 1.4 2.3 1.68 11.0 7.1 0.64
Forest; average 0.8 2.1 2.56 1.8 3.5 1.90 9.3 7.2 0.78
Forest; dense 0.8 1.8 2.37 3.3 4.9 1.50 10.9 8.5 0.77
 Käsivarsi Open 0.3 0.8 2.20 4.9 7.1 1.45 21.5 16.8 0.78
 Total Open 0.7 3.0 14.9
Forest; average 0.5 2.4 10.2
Forest; dense 0.4 3.3 10.5
Open in a new tab

Regional limits for rare and exceptional events

The average snow period length varied between 174 and 220 days from south to north in the simulations (Fig. 3). The length of a rarely or exceptionally short snow cover period in the southernmost Poikajärvi district was ≤153 or 137 days, respectively, and in the northernmost Käsivarsi district ≤ 204 or 195 days, respectively. The respective lengths of a rarely or exceptionally long snow cover period were ≥198 or 210 days in the Poikajärvi district and ≥233 or 237 days in the Käsivarsi district.

Fig. 3.

Fig. 3

Open in a new tab

Limits for the rare (once per 10 years) and exceptional (three times per 100 years) lengths of the snow cover period in open environment in different reindeer herding districts

Rare and exceptional snow depths differed notably between the reindeer herding districts (Fig. 4a). The lowest limits were found in the southernmost district Poikajärvi with snow depths ≤19–57 cm and 16–47 cm (early–late winter) considered rarely and exceptionally thin, respectively, and snow depths of ≥49–89 cm and ≥60–102 cm considered rarely and exceptionally deep, respectively. The highest limits for rarely and exceptionally thin snow depths, 32–72 cm and 24–64 cm (early–late winter), respectively, were found in the northernmost districts. Further, the highest limits for rarely and exceptionally deep snow in the study region were ≥61–112 cm and ≥80–120 cm, respectively.

Fig. 4.

Fig. 4

Open in a new tab

Limits for rarely occurring (once per 10 years) a snow depths, b snow densities, and c thicknesses of ground ice layers in open environment in the four studied herding districts during different winter periods based on SNOWPACK simulations. Solid line rarely high, dashed line rarely low (early winter 1 Oct–15 Dec; mid-winter 16 Dec–28 Feb; late winter 1 Mar–15 May)

The limits for rarely and exceptionally low snow densities did not vary notably between the districts. Snow densities considered as rarely and exceptionally high showed more variation (Fig. 4b). The limits for rarely high snow densities varied between 198 and 225 kg m−3 in early winter and between 320 and 368 kg m−3 in late winter; for exceptionally high snow densities, the ranges were 207–233 kg m−3 in early winter and 312–390 kg m−3 in late winter. The highest limits for rarely and exceptionally thick ground ice layers (mid- and late winter) were found in the northernmost Käsivarsi district (17–36 cm and 21–69 cm, respectively).

Snow conditions in different environments

Local snow characteristics

The highest maximum snow depths occurred in open environments, and forests of average density had deeper snow cover than dense forests (Table 3a). In early winter, the highest snow densities were found in open environments, whereas dense forests had the highest snow densities in mid- and late winter (Table 3b). The highest interannual variability in snow depths and densities was found in dense forests. Ground ice layers tended to be thicker in open environments compared to forests in early and late winter (Table 3c). In mid-winter, the thickest ground ice layers were found in dense forests. During 30 winters, ground ice occurred more often in dense forests (on average 20 simulated occurrences per district) compared to forests of average density and open environments (on average 14 and 15 occurrences per district, respectively). Ground ice persisting through the winter was not frequent in the simulations (on average 7 occurrences in dense forests and 5 occurrences in forests of average density and in open environments per district).

Local limits for rare and exceptional events

As an example of limits for rare and exceptional events in different environments, results from the Poikajärvi herding district are presented. Other studied districts showed comparable results. Figure 5a shows the limits for rare and exceptional snow depths in different environments in Poikajärvi. Snow depths considered rare or exceptional were higher in open environments compared to forests. The lowest limits for exceptionally deep and thin snow depths were found in dense forests. Rare and exceptional snow densities did not differ between open and forest environments (Fig. 5b), although the highest limits for rare and exceptional snow densities were found in dense forests in mid-winter and late winter. Furthermore, the highest thickness of ground ice considered as normal tended to be found in dense forests (Fig. 5c).

Fig. 5.

Fig. 5

Open in a new tab

Limits for the rare (once per 10 years) and exceptional (three times per 100 years) a snow depths, b snow densities, and c ground ice thicknesses in open environment, forest of average density (A. forest), and dense forest (D. forest) in the Poikajärvi reindeer herding district during different winter periods based on SNOWPACK simulations (early winter 1 Oct–15 Dec; mid-winter 16 Dec–28 Feb; late winter 1 Mar–15 May)

Trends in snow and ground ice conditions

The mean thickness of ground ice layers increased during the study period of 30 winters in forests of average density in the Poikajärvi district in mid- and late winter (0.1 and 0.3 cm, respectively, p < 0.05). Other snow characteristics showed no statistically significant trends. Further, no statistically significant trends in snow characteristics were found in other herding districts.

Spatial coherence in snow conditions

Regional coherence between the herding districts

Significant correlations (winters 1981/1982–2010/2011) between snow formation dates were found among the herding districts (r = 0.68–0.73, p < 0.001) excluding the southernmost Poikajärvi district (Table 4a). The strength of correlations varied between moderate and strong for snow melt dates. Annual maximum snow depths were strongly correlated among districts excluding the northernmost Käsivarsi district (r = 0.69–0.87, p < 0.001). Snow densities were moderately to strongly correlated particularly in mid-winter, whereas the thickness of ground ice was not generally correlated among herding districts. The strongest correlations in the snow densities and thickness of ground ice were found among the southernmost districts.

Table 4.

Interannual correlations (winters 1981/82–2011/12) in snow conditions among (a) reindeer herding districts (open environments) and (b) different environments within reindeer herding districts (* p < 0.05, ** p < 0.01, *** p < 0.001, n.s. meaning not significant)

Snow variable District Poikajärvi Pohjois-Salla Kyrö
(a) Reindeer herding districts (open environments)
 Formation date Pohjois-Salla n.s. – –
Kyrö n.s. 0.73*** –
Käsivarsi n.s. 0.68*** 0.65***
 Melt date Pohjois-Salla 0.73*** – –
Kyrö 0.59*** 0.79*** –
Käsivarsi 0.37* 0.44* 0.48**
 Max. snow depth Pohjois-Salla 0.87*** – –
Kyrö 0.70*** 0.69*** –
Käsivarsi n.s. n.s. 0.47**
Snow density
 Early winter Pohjois-Salla 0.53** – –
Kyrö n.s. 0.63*** –
Käsivarsi n.s. n.s. 0.57**
 Mid-winter Pohjois-Salla 0.83*** – –
Kyrö 0.43* 0.55** –
Käsivarsi n.s. n.s. 0.37*
 Late winter Pohjois-Salla 0.75*** – –
Kyrö n.s. 0.52** –
Käsivarsi 0.45** n.s. n.s.
Thick. of ground ice
 Early winter Pohjois-Salla 0.38* – –
Kyrö n.s. n.s. –
Käsivarsi n.s. n.s. 0.37*
 Mid-winter Pohjois-Salla n.s. – –
Kyrö n.s. n.s. –
Käsivarsi n.s. n.s. n.s.
 Late winter Pohjois-Salla 0.60*** – –
Kyrö n.s. 0.50** –
Käsivarsi n.s. n.s. n.s.
Snow variable Environment Poikajärvi Pohjois-Salla Kyrö
A. forest D. forest A. forest D. forest A. forest D. forest
(b) Different environments within reindeer herding districts
 Max. snow depth Open 0.86*** 0.69*** 0.77*** 0.69*** 0.91*** 0.90***
A. forest – 0.93*** – 0.96*** – 0.97***
Snow density
 Early winter Open 0.93*** 0.85*** 0.60*** 0.60*** 0.76*** 0.77***
A. forest – 0.89*** – 0.94*** – 0.97***
 Mid-winter Open 0.77*** 0.54** 0.57** 0.52** 0.70*** 0.60***
A. forest – 0.67*** – 0.92*** – 0.83***
 Late winter Open 0.89*** 0.55** 0.73*** 0.46** 0.74*** 0.58***
A. forest – 0.75*** – 0.89*** – 0.88***
Thick. of ground ice
 Early winter Open 0.65*** 0.56** n.s. n.s. n.s. n.s.
A. forest – 0.57** – 0.83*** – 0.77***
 Mid-winter Open 0.59*** n.s. n.s. n.s. n.s. n.s.
A. forest – n.s. – 0.76*** – 0.63***
 Late winter Open 0.78*** n.s. 0.57** 0.37* 0.66*** 0.46**
A. forest – 0.52** – 0.89*** – 0.74***
Open in a new tab

Local coherence between open and forest environments

Within each herding district, maximum snow depths were strongly correlated among different environments with the highest correlations among forest environments (r = 0.93–0.97, p < 0.001). Also correlations between snow densities among forest environments were stronger than forest-open environment correlations. Snow density correlations were the strongest in early winter (average–dense: r = 0.89–0.97; forest–open: r = 0.60–0.93, p < 0.001). The thickness of ground ice was moderately to strongly correlated among all environments in late winter (r = 0.37–0.89, p < 0.05), whereas in early and mid-winter significant correlations were found mainly among forest environments (r = 0.57–0.83, p < 0.01). Ground ice occurred simultaneously in all three environments during eight winters. Ground ice did not occur in any of the studied environment during five winters.

Reindeer herders’ experiences

A review of annual management reports revealed significant differences among the herding districts in the frequency of snow conditions experienced as difficult during the studied 30 winters (Fig. 6). The most often reported difficult snow condition was late snow melt (n = 18). Difficulties related to ground ice were reported most often in Käsivarsi (n = 8), and difficulties related to deep snow in Pohjois-Salla (n = 8). Occurrence of icy layers, mold in pastures, or soft snow in the spring was reported only during 1–2 winters or less per district.

Fig. 6.

Fig. 6

Open in a new tab

a Maximum snow depth, b snow melt day (number of day since the beginning of the year), and c mid-winter thickness of ground ice in the four studied herding districts in winters 1981/82–2011/12 and on average in the same period. Also the winters reported difficult by reindeer herders are shown

Winters with observed high maximum snow depths and late snow melt were often experienced as problematic by reindeer herders, and difficult grazing and herding conditions occurred rather simultaneously in different districts (Fig. 6a–b). Ground ice conditions experienced as problematic most often coincided with simulated occurrences of ground ice (Fig. 6c), although the winters with the highest simulated mid-winter thicknesses of ground ice were not necessarily among the winters experienced as most problematic by herders. This may also reflect the bias of the model in classification of the snow layers.

Discussion

Spatial-temporal variations in snow characteristics in the reindeer herding area

The Finnish reindeer herding area is characterized by heterogeneous snow conditions at regional and local levels. The large latitudinal range and related variations in large-scale climate are the basis for regional variability in snow characteristics. Regional climate is determined by dominating air flow either from the Eurasian continent (high pressure) or from the North Atlantic (low pressure systems). Local variation is induced also by altitudinal differences and large water bodies (Kersalo and Pirinen 2009). Our results showed that both the annual maximum snow depths and the length of snow cover period increased approximately by 20 % from south to north. Further, the highest snow densities and ground ice thickness were found in the northernmost Käsivarsi reindeer herding district, whereas no clear differences in these variables were found between the other districts. The weather in the Käsivarsi district is affected by relatively high elevation (60 % of fells of Finland are found within the district) and the vicinity of the Arctic Ocean. Weather may also change very fast; temperature rise of 33 degrees within a day has been reported (Kersalo and Pirinen 2009).

Higher variation in snow depths was found at the local level than at the regional level. The local variability was not considered in the Käsivarsi district, where snow covers were simulated only in open environments (snow drifting not included in the simulations). This district is, however, topographically different from the other studied districts with open fells, valleys, and ridges. Interaction between topography and wind leads to highly variable snow cover (Kivinen et al. 2012). Snow depths in forests of average density and dense forests were approximately 20–40 % and 40–50 % lower than snow depths in open environments, respectively. Forest pastures seem to offer easier access to lichen pastures during winters with ground ice formation, as ground ice layers tended to be thickest in late winter in open environments (Fig. 4c). Further, based on our simulations, it is not common that ground ice forms simultaneously both in open environment and below the forest canopies of different densities. Hence, if there is diversity in pasture environment, it is probable that reindeer will find a place to forage in different weather conditions. The diversifying effect of forest on the pasture environment is amplified when taking into account the variability of the snow cover within a forest. Several studies have shown that there is wide small-scale spatial variability in snow depth and snow quality in coniferous forests (Rasmus et al. 2011). These spatial patterns result from, e.g., local differences in the canopy interception and the canopy effects on snow surface energy balance fluxes that govern snow metamorphoses and melt (Hedström and Pomeroy 1998; Musselman et al. 2012). Effects of forests on snow characteristics depend mostly on tree species, canopy and trunk density, and heterogeneity of the forest structure (forest gaps, dense and sparse regions within the forest). For example, our results showed higher mid- and late winter snow densities in dense forests compared to other environments. The forest also affects the rate of snow accumulation during the early winter and melting in the spring (McKay and Gray 1981) further bringing more diversity in snow conditions in pastures.

During the study period of 30 years, the annual maximum snow depths were moderately to strongly correlated within the reindeer herding area excluding the northernmost Käsivarsi herding district. Interannual variations in snow depth are strongly related to large-scale climate patterns (Arctic Oscillation, North Atlantic Oscillation) that affect the Northern Fennoscandian temperature and precipitation conditions (Irannezhad et al. 2014a, b). From the reindeer herding perspective this means, that, e.g., deep snow conditions are likely to affect a major part of the reindeer management area in certain years. Relatively strong correlation was found also in snow formation and melt dates particularly between the most closely located districts. The snow densities and the thickness of ground ice layers in particular showed notably weaker spatiotemporal patterns. Processes leading to this type of snow structural changes may be strongly modified by local topography, vegetation, and existence of water bodies. This is also seen in our results on local correlations: annual maximum snow depths and snow densities were more strongly correlated between different environments than the thickness of ground ice layers at local level.

Impacts of snow conditions on reindeer herding

Winter climate can have notable impacts on fitness, mortality, and calf percentage of reindeer (Helle 1980; Helle and Jaakkola 2008). Even during easy snow conditions, reindeer need to consume additional energy in moving in the snow and digging lichens through the snowpack (Kumpula et al. 2004). Fitness of reindeer in the beginning of winter is strongly determined by grazing conditions in preceding summer and autumn. Poor grazing conditions during snow free period may result in lowered fitness and survival of reindeer even during normal winters. Harsh winter conditions may be fatal even to fit animals and result in low number of calves in the following spring (Post and Stenseth 1999; Kumpula 2001; Kumpula and Colpaert 2003; Helle and Kojola 2008).

Late snow melt reported most often as a difficult herding condition means longer digging period for reindeer and late access to green vegetation, which may cause problems particularly during the calving time. Ground ice formation prevents reindeer access to ground-growing lichens and has reported to cause catastrophic declines in populations of semi-domesticated (Helle 1980) and wild reindeer (Kohler and Aanes 2004; Tyler 2010; Hansen et al. 2011, 2014). Mold formation on pastures was often related to snow cover formation on unfrozen ground, unstable weather in early winter and ground ice formations. Difficult snow conditions with regional scale coherence, like late snow melt, were also reported slightly more often in several herding districts during the same winter. More local nature of ground ice formation was also seen in less coherent reporting of this phenomenon, although there were four winters during the study period when ground ice formation was reported widely around the reindeer management area.

Experiencing certain snow condition as problematic depends both on local herding practices and also on pasture environment. Highly variable snow depth in the Käsivarsi district may explain why the herders found the deep snow rarely problematic there, although the greatest mean maximum snow depths were observed within this district (Table 1). Icy conditions were reported most often in the northernmost districts (Käsivarsi). Reporting frequency can be explained both by winter climate (e.g., thick icy layers were more common in Käsivarsi than in other study districts, Fig. 4) and also by quality of pastures in the districts. Forest pastures with arboreal lichen do not exist in the northernmost districts and so reindeer need to dig for forage during the whole snow covered period. In this kind of conditions also, herders pay close attention to the structure of the snow cover. Soft snow during the late winter was seen as problematic in districts where the use of forest pastures with arboreal lichen is an important herding practice during a normal year.

Reindeer herding practices have been historically adapted to local conditions. During the past decades, reindeer herding has needed to adapt also to negative impacts of other land uses on grazing resources. The Poikajärvi herding district is an example of a region where increased infrastructure and intensification of commercial forestry have decreased the area and quality of pastures. Winter feeding in pens has become a common practice in southern herding districts, and has been increasingly practiced besides field feeding also in northern districts particularly during difficult winter conditions. Systematic winter feeding mitigates the impacts of difficult winter conditions on reindeer herding in many districts, although it has been criticized, e.g., due to increased costs, risk of diseases, and harmful impacts on ecosystems (Turunen and Vuojala-Magga 2014). For example, it is notable that remarks on difficult snow conditions are rarer in Poikajärvi than in other studied herding districts. This may be related to supplementary winter feeding of reindeer that partly compensate difficult snow conditions. Availability of diverse pasture environment reduces the need for winter feeding.

Future prospects

We used a novel approach to model snow and ground ice condition in different environments, and analyzed the results in the light of reindeer herders’ experiences on problematic winter conditions. Modeling gives valuable information on snow structural phenomena that is otherwise difficult to gain. Our simulations verified the experiential fact that certain difficult snow condition (e.g., ground-icing) is not equally severe in all parts of the reindeer management area, and that there is also local variability. There are certain challenges in the use of herders’ experiences, especially in the form of annual management reports of the districts. Districts may not report difficult conditions if they prevail only in limited area, or if the difficult situation can be overcome, e.g., by using alternative pastures. Still, the experiences add invaluable knowledge that should not be overlooked.

We observed only few statistically significant trends during the study period (increase of mid- and late winter mean thickness of ground ice layers in forests of average density in the Poikajärvi district). In longer and more comprehensive data sets, e.g., winter air temperatures have showed increasing trends in several locations in Finnish reindeer management area (Irannezhad et al. 2014a, b; Kivinen and Rasmus 2014), while duration of the snow season has declined (Rasmus et al 2014). During warmer winters in the future (Kivinen et al 2012; Stocker et al. 2013), it is probable that the snow season will get shorter, snow amounts will change, and snow structure get icier (Rasmus et al. 2004; Jylhä et al. 2008; Räisänen and Eklund 2012). This means also change in the limits for rare and exceptional events, and reindeer herding need to adapt to new kind of winter conditions.

Conclusions

A significant variability in winter conditions occurs within the whole reindeer herding area as well as locally within the herding districts. Limits of rare or exceptional events vary spatially—what is normal in one district or in one environment may be rare in the other. Reindeer herders experientially place a high value on the forest pastures as a source of arboreal lichen during difficult foraging conditions. Forests also diversify the snow and foraging conditions, which increase the reindeer access to lichen resources. In addition to naturally open environments, such as mountain heaths and open mires, open environments occur in the reindeer management areas as a result of clear-cutting of coniferous forests. These land cover transitions have impacts on snow characteristics and thus, accessibility of lichens in winter grazing grounds. Sufficient and diverse forest pastures are important, if reindeer herding is pursued on natural grazing grounds also in the future.

Acknowledgments

We thank Henna-Reetta Hannula, Leena Leppänen, Pentti Pirinen, and Henriikka Simola from Finnish Meteorological Institute (FMI) for providing meteorological data and Heidi Sjöblom from Finnish Environment Institute (SYKE) for acquiring the snow line data. We thank Michael Lehning, Charlez Fierz, and several others from the WSL Institute for Snow and Avalanche Research (SLF) for continuous support in the SNOWPACK use. Ilmo Kukkonen and Lauri Korhonen from University of Helsinki are thanked for the assistance in the soil and canopy parameterization needed in the snow modeling. Reindeer Herders’ Association personnel were ready to help all through the project with reindeer data, annual management reports and maps, which we warmly acknowledge. Our warmest thanks go to reindeer herders and their families from Kyrö, Käsivarsi, Pohjois-Salla, and Poikajärvi herding districts, especially Olli and Juuso Autto, Tuomas Palojärvi, Ensio Pirttilä, and Veikko Heiskari. Finnish Ministry of Agriculture and Forestry (MAKERA—project 2327/312/2011) and the Nordic Center of Excellence Tundra have supported our work financially. We are grateful to Kirsti Jylhä and Minna Turunen for the valuable comments that have improved the manuscript.

Biographies

Sirpa Rasmus

is a postdoctoral researcher at the Department of Biological and Environmental Science, University of Jyväskylä. Her research interests include snow modelling, snow ecology of boreal and sub-arctic plants and animals, and climate change impacts.

Sonja Kivinen

is a postdoctoral researcher at the Department of Geography and Geology, University of Turku. Her research interests include environmental changes in northern regions, land use/cover changes, and biogeography.

Mathias Bavay

is a scientific staff member at the WSL Institute for Snow and Avalanche Research SLF. His research interests include snow cover numerical modeling and catchment hydrology.

Janne Heiskanen

is a postdoctoral researcher at the Department of Geosciences and Geography, University of Helsinki. His research interests include geoinformatics and remote sensing.

Contributor Information

Sirpa Rasmus, Phone: +358 40 5282585, Email: sirpa.rasmus@alumni.helsinki.fi.

Sonja Kivinen, Email: sonja.kivinen@utu.fi.

Mathias Bavay, Email: bavay@slf.ch.

Janne Heiskanen, Email: janne.heiskanen@helsinki.fi.

References

  1. Bartelt P, Lehning M. A physical SNOWPACK model for the Swiss avalanche warning: Part I. Numerical model. Cold Regions Science and Technology. 2002;35:123–145. doi: 10.1016/S0165-232X(02)00074-5. [DOI] [Google Scholar]
  2. Bavay M, Egger T. MeteoIO 2.4. 2: A preprocessing library for meteorological data. Geoscientific Model Development. 2014;7:3135–3151. doi: 10.5194/gmd-7-3135-2014. [DOI] [Google Scholar]
  3. Dilley AC, O’Brien DM. Estimating downward clear sky long-wave irradiance at the surface from screen temperature and precipitable water. Quarterly Journal of the Royal Meteorological Society. 1998;124:1391–1401. doi: 10.1002/qj.49712454903. [DOI] [Google Scholar]
  4. Hansen BB, Aanes R, Herfindal I, Kohler J, Saether BE. Climate, icing, and wild arctic reindeer: Past relationships and future prospects. Ecology. 2011;92:1917–1923. doi: 10.1890/11-0095.1. [DOI] [PubMed] [Google Scholar]
  5. Hansen BB, Isaksen K, Benestad RE, Kohler J, Pedersen ÅØ, Loe LE, Coulson SJ, Larsen JO, Varpe Ø. Warmer and wetter winters: Characteristics and implications of an extreme weather event in the High Arctic. Environmental Research Letters. 2014 [Google Scholar]
  6. Hedström NR, Pomeroy JW. Measurements and modelling of snow interception in the boreal forest. Hydrological Processes. 1998;12:1611–1625. doi: 10.1002/(SICI)1099-1085(199808/09)12:10/11<1611::AID-HYP684>3.0.CO;2-4. [DOI] [Google Scholar]
  7. Heiskanen, J., M. Rautiainen, L. Korhonen, M. Mõttus, and P. Stenberg. 2011. Generating fine resolution leaf area index maps for boreal forests of Finland. IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2011, Vancouver, BC, Canada, July 24–29, 2011. doi:10.1109/IGARSS.2011.6049675.
  8. Helle T. Laiduntilanteen muutokset ja riskinotto porotaloudessa [Changes in pasture environment and risks in the reindeer herding livelihood, in Finnish] Lapin tutkimusseura: Vuosikirja. 1980;21:13–21. [Google Scholar]
  9. Helle T, Jaakkola LM. Transition in herd management of semi-domesticated reindeer in northern Finland. Annales Zoologi Fennici. 2008;45:81–101. doi: 10.5735/086.045.0201. [DOI] [Google Scholar]
  10. Helle T, Kojola I. Demographics in an alpine reindeer herd: Effects of density and winter weather. Ecography. 2008;31:221–230. doi: 10.1111/j.0906-7590.2008.4912.x. [DOI] [Google Scholar]
  11. Holleman DF, Luick JR, White RG. Lichen intake estimates for reindeer and caribou during winter. Journal of Wildlife Management. 1979;43:192–201. doi: 10.2307/3800651. [DOI] [Google Scholar]
  12. IPCC. 2013. Annex III: Glossary (ed. S. Planton). In Climate change 2013: The physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change, ed. T.F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P.M. Midgley. Cambridge: Cambridge University Press.
  13. Irannezhad M, Chen D, Kløve B. Interannual variations and trends in surface air temperature in Finland in relation to atmospheric circulation patterns, 1961–2011. International Journal of Climatology. 2014 [Google Scholar]
  14. Irannezhad M, Marttila H, Kløve B. Long-term variations and trends in precipitation in Finland. International Journal of Climatology. 2014;34:3139–3153. doi: 10.1002/joc.3902. [DOI] [Google Scholar]
  15. Jylhä K, Fronzek S, Tuomenvirta H, Carter TR, Ruosteenoja K. Changes in frost, snow and Baltic Sea ice by the end of the twenty-first century based on climate model projections for Europe. Climatic Change. 2008;86:441–462. doi: 10.1007/s10584-007-9310-z. [DOI] [Google Scholar]
  16. Kersalo, J., and P. Pirinen. 2009. Suomen maakuntien ilmasto [The climate of Finnish regions, in Finnish]. Finnish Meteorological Institute Reports 2009: 8. Helsinki: Finnish Meteorological Institute.
  17. Kivinen S, Kaarlejärvi E, Jylhä K, Räisänen J. Spatiotemporal distribution of threatened high-latitude snowbed and snow patch habitats in warming climate. Environmental Research Letters. 2012 [Google Scholar]
  18. Kivinen S, Moen J, Berg A, Eriksson Å. Effects of modern forest management on winter grazing resources for reindeer in Sweden. Ambio. 2010;39:269–278. doi: 10.1007/s13280-010-0044-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kivinen S, Rasmus S. Observed cold season changes in a Fennoscandian fell area over the past three decades. Ambio. 2014;44:214–225. doi: 10.1007/s13280-014-0541-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kohler J, Aanes R. Effect of winter snow and ground-icing on a Svalbard Reindeer population: Results of a simple snowpack model. Arctic, Antarctic, and Alpine Research. 2004;36:333–341. doi: 10.1657/1523-0430(2004)036[0333:EOWSAG]2.0.CO;2. [DOI] [Google Scholar]
  21. Kortesalmi, J. 2007. Poronhoidon synty ja kehitys Suomessa [Origin and development of reindeer herding in Finland, in Finnish]. SKS:n toimituksia 1149. Helsinki: Suomalaisen Kirjallisuuden Seura.
  22. Kumpula J. Winter grazing of reindeer in woodland lichen pasture—Effect of lichen availability on the condition of reindeer. Small Ruminant Research. 2001;39:121–130. doi: 10.1016/S0921-4488(00)00179-6. [DOI] [PubMed] [Google Scholar]
  23. Kumpula, J., A. Tanskanen, A. Colpaert, M. Anttonen, H. Törmänen, J. Siitari, and S. Siitari. 2009. Poronhoitoalueen pohjoisosan talvilaitumet vuosina 2005—2008; laidunten tilan muutokset 1990-luvun puolivälin jälkeen [The pastures of the northernmost reindeer herding area during 2005—2008; changes in the state of pastures after the mid 1990s]. Riista- ja kalatalous - Tutkimuksia 3. Finnish Institute for Game and Fisheries Research, Helsinki.
  24. Kumpula J, Colpaert A. Effects of weather and snow conditions on reproduction and survival of semi-domesticated reindeer (R. t. tarandus) Polar Research. 2003;22:225–233. doi: 10.1111/j.1751-8369.2003.tb00109.x. [DOI] [Google Scholar]
  25. Kumpula J, Lefrère S, Nieminen M. The use of woodland lichen pasture by reindeer in winter with easy snow conditions. Arctic. 2004;57:273–278. doi: 10.14430/arctic504. [DOI] [Google Scholar]
  26. Lee SE, Press MC, Lee JA, Ingold T, Kurttila T. Regional effects of climate change on reindeer: A case study of the Muotkatunturi region in Finnish Lapland. Polar Research. 2000;19:99–105. doi: 10.1111/j.1751-8369.2000.tb00333.x. [DOI] [Google Scholar]
  27. Lehning M, Bartelt P, Brown B, Fierz C. A physical SNOWPACK model for the Swiss avalanche warning service. Part III. Meteorological forcing, thin layer formation and evaluation. Cold Regions Science and Technology. 2002;35:169–184. doi: 10.1016/S0165-232X(02)00072-1. [DOI] [Google Scholar]
  28. Lehning M, Bartelt P, Brown B, Fierz C, Satyawali P. A physical SNOWPACK model for the Swiss avalanche warning. Part II. Snow microstructure. Cold Regions Science and Technology. 2002;35:147–167. doi: 10.1016/S0165-232X(02)00073-3. [DOI] [Google Scholar]
  29. Lehning M, Bartelt P, Brown B, Russi T, Stöckli U, Zimmerli M. SNOWPACK model calculations for avalanche warning based upon a network of weather and snow stations. Cold Regions Science and Technology. 1998;30:145–157. doi: 10.1016/S0165-232X(99)00022-1. [DOI] [Google Scholar]
  30. Lehning M, Völksch I, Gustafsson D, Nguyen T, Stähli M, Zappa M. ALPINE3D: A detailed model of mountain surface processes and its application to snow hydrology. Hydrological Processes. 2006;20:2111–2128. doi: 10.1002/hyp.6204. [DOI] [Google Scholar]
  31. Lundy C, Brown RL, Adams EE, Birkeland KW, Lehning M. A statistical validation of the SNOWPACK model in a Montana climate. Cold Regions Science and Technology. 2001;33:237–246. doi: 10.1016/S0165-232X(01)00038-6. [DOI] [Google Scholar]
  32. Mattila, E. 2012. Porojen laitumia koskevia arviointituloksia 1970-luvulta alkaen [Results from reindeer pasture surveys from 1970s onwards, in Finnish]. Metlan työraportteja 238. Finnish Forest Research Institute, Vantaa.
  33. McKay GA, Gray DM. The distribution of snowcover. In: Gray DM, Male DH, editors. Handbook of snow. Canada: Pergamon Press; 1981. pp. 153–190. [Google Scholar]
  34. Miller FL, Barry SJ. Long-term control of Peary caribou numbers by unpredictable, exceptionally severe snow or ice conditions in a non-equilibrium grazing system. Arctic. 2009;62:175–189. doi: 10.14430/arctic130. [DOI] [Google Scholar]
  35. Musselman KN, Molotch NP, Margulis SA, Lehning M, Gustafsson D. Improved snowmelt simulations with a canopy model forced with photo-derived direct beam canopy transmissivity. Water Resources Research. 2012;48:W10509. doi: 10.1029/2012WR012285. [DOI] [Google Scholar]
  36. Perälä, J., and M. Reuna. 1990. Lumen vesiarvon alueellinen ja ajallinen vaihtelu Suomessa. Publications of Water and Environment Research Institute No. A 56. Water and Environment Research Institute, Helsinki.
  37. Pirinen, P., H. Simola, J. Aalto, J.-P. Kaukoranta, P. Karlsson, and R. Ruuhela. 2012. Tilastoja Suomen ilmastosta 1981—2010 [Climatological statistics of Finland 1981—2010, in Finnish]. Reports 2012:1. Finnish Meteorological Institute, Helsinki.
  38. Pomeroy, J.W., D.M. Gray, N.R. Hedström, and J.R. Janowich. 2002. Physically based estimation of seasonal snow accumulation in the boreal forest. In Proceedings of the 59th eastern snow conference, 93–108.
  39. Post E, Stenseth NC. Climate change, plant phenology, and northern ungulates. Ecology. 1999;80:1322–1339. doi: 10.1890/0012-9658(1999)080[1322:CVPPAN]2.0.CO;2. [DOI] [Google Scholar]
  40. Räisänen J, Eklund J. 21st century changes in snow climate in Northern Europe as simulated by regional climate models in the ENSEMBLES project: A high-resolution view from ENSEMBLES regional climate models. Climate Dynamics. 2012;38:2575–2591. doi: 10.1007/s00382-011-1076-3. [DOI] [Google Scholar]
  41. Rasmus S, Grönholm T, Lehning M, Rasmus K, Kulmala M. Validation of the SNOWPACK-model in five different snow zones in Finland. Boreal Environment Research. 2007;12:467–488. [Google Scholar]
  42. Rasmus S, Kumpula J, Siitari J. Can a snow structure model estimate snow characteristics relevant for reindeer husbandry? Rangifer. 2014;34:37–56. doi: 10.7557/2.34.1.2675. [DOI] [Google Scholar]
  43. Rasmus S, Lundell R, Saarinen T. Interactions between snow, canopy and vegetation in a boreal coniferous forest. Plant Ecology and Diversity. 2011;4:55–65. doi: 10.1080/17550874.2011.558126. [DOI] [Google Scholar]
  44. Rasmus S, Räisänen J, Lehning M. Estimating snow conditions in Finland in the late 21st century using the SNOWPACK—model with regional climate scenario data as input. Annals of Glaciology. 2004;38:238–244. doi: 10.3189/172756404781814843. [DOI] [Google Scholar]
  45. Sevola, Y. 2002. Metsävarat. In Metsätilastollinen vuosikirja 2002 [Finnish statistical yearbook of forestry, in Finnish], ed. A. Peltola, 33–76. SVT Maa-, metsä- ja kalatalous 2002:45. Finnish Forest Research Institute, Vantaa.
  46. Solantie, R., A. Drebs, E. Hellsten, and P. Saurio. 1996. Lumipeitteen tulo-, lähtö- ja kestoajoista Suomessa talvina 1960/61-1992/1993 [On snow formation and melt dates and snow cover period lengths in Finland during winters 1960/61-1992/1993, in Finnish]. Meteorologisia julkaisuja 34. Finnish Meteorological Institute, Helsinki.
  47. Stähli M, Jonas T, Gustafsson D. The role of snow interception in winter-time radiation processes of a coniferous sub-alpine forest. Hydrological Processes. 2009;23:2498–2512. doi: 10.1002/hyp.7180. [DOI] [Google Scholar]
  48. Steinkogler, W., M. Bavay, A. Crepaz, and M. Lehning. 2015. Permafrost distribution modeling with the energy balance model Alpine3D in the Dolomites. Permafrost Processes, in review.
  49. Stocker, T. F., D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley (ed.). 2013. Climate Change 2013. Cambridge: Cambridge University Press.
  50. Turunen M, Vuojala-Magga T. Past and present winter feeding of reindeer in Finland: Herders adaptive learning of feeding practices. Arctic. 2014;67:173–188. doi: 10.14430/arctic4385. [DOI] [Google Scholar]
  51. Tyler N. Climate, snow, ice, crashes, and declines in populations of reindeer and caribou (Rangifer tarandus L.) Ecological Monographs. 2010;80:197–219. doi: 10.1890/09-1070.1. [DOI] [Google Scholar]
  52. Tyler N, Turi J, Sundset M, Strøm Bull K, Sara M, Reinert E, Oskal N, Nellemann C, McCarthy J, Mathiesen S, Martello M, Magga O, Hovelsrud G, Hanssen-Bauer I, Eira N, Eira I, Corell R. Saami reindeer pastoralism under climate change: Applying a generalized framework for vulnerability studies to a sub-arctic social-ecological system. Global Environmental Change. 2007;17:191–206. doi: 10.1016/j.gloenvcha.2006.06.001. [DOI] [Google Scholar]
  53. Unsworth MH, Monteith JL. Long-wave radiation at the ground. Quarterly Journal of the Royal Meteorological Society. 1975;101:13–24. doi: 10.1256/smsqj.42703. [DOI] [Google Scholar]
  54. Vikhamar-Schuler D, Hanssen-Bauer I, Schuler TV, Mathiesen SD, Lehning M. Use of a multi-layer snow model to assess grazing conditions for reindeer. Annals of Glaciology. 2013;54:214–226. doi: 10.3189/2013AoG62A306. [DOI] [Google Scholar]

Articles from Ambio are provided here courtesy of Springer

RESOURCES