Assessing glacier features supporting supraglacial trees: A case study of the Miage debris-covered Glacier (Italian Alps)

Abstract

The number of debris-covered glaciers featuring supraglacial trees is increasing worldwide as a response of high mountain environments to climate warming. Generally, their distribution on the glacier surface is not homogeneous, thus suggesting that some glacier parameters influence germination and growth of trees. In this study, we focused our attention on the widest Italian debris-covered glacier, the Miage Glacier (Mont Blanc massif). We analyzed the ablation area in the range from 1730 to 2400 m a.s.l. where continuous debris coverage is present and trees are found. Using data obtained by remote sensing investigations and field surveys, we defined a record of glacier parameters to be analyzed with respect to the presence and abundance of trees. We found that supraglacial trees are present at the Miage Glacier (1) whenever exceeding a debris thickness threshold (⩾19 cm), (2) with a gentle slope (⩽10°), (3) with a low glacier surface velocity (⩽7.0 m/yr), and (4) where the vertical changes due to glacier dynamics are positive (i.e. prevalent increase ranging between +7 and +28 m over 28 years due to both slow debris accumulation and preservation of ice flow inputs). The statistical analysis supports our findings. The analysis of the same parameters might be conducted on other debris-covered glaciers featuring supraglacial trees, in order to evaluate whether such conditions are local ones or whether they are general factors driving germination and growth of trees. By identifying the features supporting the presence and growth of trees in these environments, and their thresholds, a contribution is given for a better understanding of the importance of debris-covered glaciers and, in general, of debris-covered ice, as a refuge for trees during glacial and warm intervals of the Holocene.

Keywords

debris-covered glacier Italian Alps Miage Glacier remote sensing supraglacial debris supraglacial trees

Introduction

Debris-covered glaciers (DCGs) are common features worldwide as they have been observed in Europe, New Zealand, Asia, and South America (e.g. Brook et al., 2013; Diolaiuti et al., 2003; Emmer et al., 2015; Ghosh et al., 2014). The recent literature (see Oerlemans et al., 2009) reports darkening phenomena affecting several mountain glaciers, thus contributing to changing their surface conditions and supporting their transformation into actual DCGs. This leads to a noticeable modification in the alpine landscape, also giving new sites of scientific and cultural interest (Bollati et al., 2014). Moreover, the debris coverage represents a new habitat for living organisms, such as microorganisms (Franzetti et al., 2013), animals (Gobbi et al., 2011), yeasts (Turchetti et al., 2013), and plants (Gentili et al., 2015). Arboreal vegetation may also be present, whenever the glacier tongue is located below the treeline. The distribution of supraglacial trees is generally not homogeneous, thus suggesting that some glacier parameters influence germination and growth of plants. In the recent literature, some authors reported the following as possible driving factors, for both herbaceous and arboreal vegetation: thick debris mantle, fine grain size (i.e. from sand to pebbles), and slow surface glacier velocity; these factors also affect debris stability and then the suitability for supraglacial areas to support vegetation growth (Caccianiga et al., 2011; Leonelli and Pelfini, 2013). Arboreal vegetation is also a precious source of information to describe the behavior of DCGs. The strong sensitivity of arboreal vegetation to changes in site stability and surface velocity was analyzed by Pelfini et al. (2007), who used tree rings cored from supraglacial trees to reconstruct glacier dynamics and evolution. In fact, accelerated ice flow causes growth disturbances in supraglacial trees, which manifest as scars, compression wood, and tree-ring eccentricity.

Thus, years characterized by intense glacier flow can be detected studying tree rings of supraglacial arboreal vegetation.

Despite these previous studies, which underlined that supraglacial tree vegetation reflects peculiar environmental conditions (Pelfini et al., 2007) and gave a list of possible factors driving the colonization of the buried ice (Caccianiga et al., 2011), no research was found dealing with the exact role played by each one of the glacier features in allowing and supporting or preventing tree germination and growth. Thus, dedicated studies on selected locations embracing both the analysis of glacier features and arboreal vegetation characteristics (i.e. abundance of trees) are needed.

Among the existing methods allowing the extraction of geomorphological and glaciological data, for describing the supraglacial environment and detecting the most suitable sites to permit and support tree germination and growth, remote sensing is the most valuable one. In fact, this technique not only allows a wide area to be covered, and in this way to collect distributed data, but it also permits the survey to be repeated over several images and sources, thus giving a multi-temporal analysis.

Aerial and satellite images have been applied in the recent past in order to analyze DCGs, with the aims to detect their boundaries and to estimate volume and area changes (e.g. D’Agata et al., 2005; Diolaiuti et al., 2009; Gjermundsen et al., 2011; Ranzi et al., 2004; Smiraglia et al., 2000), to describe supraglacial debris distribution and its changes over time (e.g. Mihalcea et al., 2008a; Minora et al., 2015; Stokes et al., 2007), to identify changes in the velocity of the glacier (e.g. Luckman et al., 2007), and to detect and model their ablation rate (e.g. Fyffe et al., 2014; Mihalcea et al., 2008b; Nicholson and Benn, 2006; Reid and Brock, 2010). Orthophotos, satellite, and aerial images were also used in order to monitor changes in the distribution of alpine and subalpine vegetation (e.g. Müllerová, 2004; Vescovo and Gianelle, 2008), to analyze the recent ecesis in the glacier forefield (Garavaglia et al., 2010), and to detect variations in the treeline (e.g. Danby and Hik, 2007; Leonelli et al., 2009). A method based on the analysis of color orthophotos to rapidly detect and map supraglacial trees on DCGs has been recently tested (unpublished data), and the discontinuous distribution, low density, and small canopy featured by trees were found representing the main limits in detecting them from remotely sensed data. Moreover, color orthophotos are not always available, in particular at high altitudes, and high-resolution satellite images are often still cost prohibitive.

Since the germination and growth of supraglacial vegetation are controlled not only by climate conditions but also by glacio-related features (Pelfini et al., 2012), in this work we aimed at (1) describing and quantifying the overall glacier features (i.e. debris occurrence and thickness, debris-surface temperature, debris moisture, surface slope, aspect, surface velocity, and ablation rates) dominating a representative alpine DCG where an actual supraglacial forest is found and (2) assessing the role and weight played by each glacier feature in driving tree vegetation presence, growth, and distribution.

For these purposes, we focused our analyses on the most representative Italian DCG, the Miage Glacier (Mont Blanc massif), the unique glacier on the southern side of the Alps featuring an actual supraglacial forest in the lower portion of its ablation tongue. In addition, this glacier underwent a long sequence of both direct and remote sensed investigations over the last decades (see details reported in the ‘Study area’ section), thus offering a robust and wide database to look for relations, if any, among tree occurrence and glacier features. Last but not least, the authors of this contribution are also analyzing vegetation of Alpine glacier forelands (increasing their extent because of the ongoing cryosphere degradation), thus supporting comparisons among tree occurrence, growth, and distribution in these newly exposed zones and on DCGs.

Since supraglacial trees are not just a recent phenomena, this study also provides insights into the role played by supraglacial debris in supporting the establishment and growth of arboreal species in the past. By identifying the features supporting the presence of trees in these environments, and their thresholds, a contribution is given for a better understanding of the importance of DCGs and, in general, of debris-covered ice, as a refuge for high elevation tree species during warm periods occurred in the Holocene, as already suggested for plants (Caccianiga et al., 2011; Fickert et al., 2007; Ravazzi, 1999). Consequently, DCGs could act as refuges for certain tree species in a warmer future.

Study area

The Miage Glacier in the Aosta Valley Autonomous Region (45° 47′N, 6° 52′E) is the widest DCG in the Italian Alps (Figure 1). It is located in the Val Veny, on the southwest slope of the Mont Blanc massif; it features an area of about 10 km² (Diolaiuti et al., 2012), and its ablation tongue is characterized by two main lobes and a smaller one in between. The ablation tongue of the glacier shows quite continuous debris coverage from 1730 up to 2400 m a.s.l., whose thickness varies from a few centimeters up to 2 m: increasing debris thickness is generally detected with decreasing elevation (Mihalcea et al., 2008a). The Miage debris cover is mainly composed of igneous and metamorphic rocks (Brock et al., 2010). The debris grain size largely varies from rock boulders to fine pebbles, sand, and clay. Debris thickness and lithology influence albedo and consequently the ablation rate. In particular, the main lithology is given by gneiss and micaschists followed by granites. These latter are composed of milky white quartz, white plagioclase feldspars, and pink potassium feldspars; moreover, also green biotitic and epidotic rocks are present. Gneiss is made from the same minerals as granites. In addition, chloroschists, amphiboles, and epidotic rocks and schists like ardesia featuring a black color and with intrusions of graphite are present. A red patina is present at the rock surface, and it is constituted of minerals from clays, manganese, and iron oxides, and it derived by deep hydrothermal circulation affecting rock cracks (Franzetti et al., 2013). Ice cliffs and supraglacial lakes are present on the largest part of the glacier ablation tongue, where ablation rate is higher, as a consequence of the exposure of bare ice (Reid and Brock, 2014).

Figure 1.

(a) The Miage debris-covered Glacier (45° 47′N, 6° 52′E) in the Val Veny, Mont Blanc Massif, is characterized in its snout part by the presence of (b, c) vegetation on the supraglacial debris, including trees. The 15 selected plots characterized by the presence of trees selected for this study are reported in Figure 1a.

Scientific research has been conducted on this glacier since the 18th century. It mostly concerned its geomorphologic and glaciological features and their changes over time (e.g. Baretti, 1880; Capello, 1959; Cunietti, 1961; Deline, 1999; Diolaiuti et al., 2009; Fyffe et al., 2014; Smiraglia et al., 2000), but also the calving phenomena and drainage events occurring at its ice-contact lake, the Miage Lake (Deline et al., 2004; Diolaiuti et al., 2005, 2006; Masetti et al., 2010), and its educational and touristic values and the role of the glacier in the frame of geo-heritage (Bollati et al., 2013, 2014).

The characteristics of the Miage Glacier as a new habitat for plants have also been studied. For what concerns trees, the most abundant species in the supraglacial debris are Larix decidua Mill. and Picea abies Karst, but other vascular plants are present. Trees taller than 1 m are mainly located at the southern lobe, and the oldest living trees are about 60 years old as they move down-valley together with the supraglacial debris according to the local surface velocity of the glacier, and once they reach the glacier terminus, they fall (Pelfini et al., 2012). Tree density is higher on the northern lobe, but the oldest and tallest trees are found on the southern lobe. At the Miage Glacier, trees represent a valid source of glaciological and climatic data as they archive much information in tree-ring morphology, and also tree-ring stable isotopes and volatile organic compounds emitted from leaves allow the key factors influencing tree development on supraglacial debris to be investigated (Leonelli et al., 2014).

Materials and methods

Extraction and analysis of the main features characterizing the whole ablation tongue of the Miage Glacier

First, a database including all the main features of the Miage debris-covered ablation tongue (ca. from 1730 to 2400 m a.s.l.) was developed. Then, the data were analyzed in order to describe the average conditions of a wide and representative alpine DCG. In particular, the database included the following glacier parameters:

Debris-surface temperature (°C);

Debris thickness (cm);

Normalized Difference Moisture Index (NDMI, absolute value);

Slope (°);

Ablation rate (m water equivalent/year);

Variation in glacier thickness over 28 years (1975–2003) (m);

Aspect (N, S, E, W);

Surface velocity measured in the past (by Lesca, 1974) and in recent time (by Diolaiuti et al., 2005) on some selected points at the glacier surface (m/yr).

The point surface debris temperature was measured by Mihalcea et al. (2008a) every 5 min by thermistors and data loggers, close to the ablation stakes (see below). The sensor tips were attached to flat rock surface (2 cm thick, 10 cm × 10 cm), 2 cm below the debris surface. The data recorded at this depth are normally considered the indicative of point surface temperature and used within several international protocols to study permafrost and frozen ground (see Guglielmin et al., 2008; Osterkamp, 2003). For describing the mean debris thermal conditions, we analyzed the temperature data recorded on a day with clear sky conditions.

Moreover, we also used surface temperature estimated from satellite images acquired at the same time of the point debris temperatures. We analyzed ASTER kinetic surface temperatures derived from an ASTER image acquired on 1 August 2005 at 10:40 a.m. UTC +1:00. These data feature a resolution of 90 m × 90 m and were fully described by Mihalcea et al. (2008a).

Debris thickness was obtained by Mihalcea et al. (2008a) from the ASTER image acquired on August 2005 on a day with clear sky conditions. Debris thickness variability over the whole debris-covered tongue was estimated through an empirical relation developed by these authors coupling measured debris depth at some selected glacier sites with the surface temperature data of the same sites extracted from the thermal level of the ASTER image. This approach, previously tested by Taschner and Ranzi (2002), was found to be a good method to describe the supraglacial debris of DCGs and was also applied more recently by other authors (Minora et al., 2015) on wide and representative DCG areas.

NDMI on the Miage Glacier was derived from a Landsat image acquired on July 2002 on a day with clear sky conditions (image code: LE71950282002230EDC00). No liquid precipitation occurred in the study area in the 6 days preceding the Landsat image acquisition (according to the meteorological data from the network managed by the Regione Autonoma Valle d’Aosta (RAVA)). NDMI was calculated according to Eq. (1), with NIR being Landsat TM Band 4 and MIR being Landsat T Band 5 (Wilson and Sader, 2002):

NDMI = \frac{NIR - MIR}{NIR + MIR}

(1)

Slope and Aspect were calculated by Diolaiuti et al. (2009) from a DEM (Digital Elevation Model) featuring a 10 m × 10 m spatial resolution. The DEM was obtained from 2003 aerial photos (RAVA flight).

The ablation rate values were obtained from a network of ablation stakes installed in summer 2005 (at the same sites where also point surface temperature was measured) and maintained up to the end of summer 2006. More than 20 ablation stakes were drilled into the ice to evaluate ice melt with varying debris thickness and altitude (i.e. from 1800 to 2400 m a.s.l.).

The stakes were distributed according to one longitudinal and two cross-profiles on the debris-covered area. They were monitored from June to October 2005 and 2006.

Variation in glacier thickness over the period 1975–2003 was calculated by Diolaiuti et al. (2009), who compared DEMs, derived from historical records, in particular maps (1975; scale 1:10,000) and stereo pairs of aerial photos (1991 and 2003; scale 1:15,000).

Annual surface velocity of the Miage Glacier was measured by the Differential Global Positioning System method (Caccianiga et al., 2011; Diolaiuti et al., 2005; Franzetti et al., 2013) in the period 2002–2009. Moreover, historic data were already published, thus permitting an analysis of the variability of glacier flow over time.

Analysis of the glacier features on areas characterized by tree presence

With the aim of identifying the role and weight of environmental parameters driving and supporting tree occurrence and growth at the glacier surface, we analyzed the field data partially published by Pelfini et al. (2012) and collected in the summer seasons 2006 and 2007 in the snout part of the Miage Glacier. More precisely, we selected the field plots where these authors found well-established tree vegetation (i.e. all the plots we selected were characterized by tree abundance exceeding 25 trees/plot), and on these sites we looked for the dominant glaciological features.

The information gained in the field included, in particular, the number of trees (both species Larix decidua Mill. and Picea abies Karst) in each plots.

In this study, 15 plots (plot size: 15 m × 15 m) were considered (they are reported in Figure 1) and a categorical value (25, 50, 75, and 100) was assigned to each plot, depending on the abundance of trees:

25 = number of trees ranging between 1 and 25;

50 = number of trees ranging between 25 and 50;

75 = number of trees ranging between 50 and 75;

100 = number of trees above 75.

For the selected plots, the following parameters were extracted from our glaciological database:

Altitude (m a.s.l.);

Debris-surface temperature (°C);

Debris thickness (cm);

NDMI (absolute value);

Slope (°);

Ablation rate (m water equivalent/yr);

Variation in glacier thickness over 28 years (1975–2003) (m);

Aspect (N, S, E, W);

Surface velocity at the glacier surface (m/yr);

Distance from the closest vegetation source area (m).

Distance from the closest vegetation source area was calculated using the 2005 color orthophotos (RAVA), measuring the distance between each selected plot and the closest forested area located outside the glacier tongue.

Then, we analyzed the environmental conditions dominating the 15 plots we selected to find threshold of the parameters suitable to divide areas with abundant tree vegetation from areas with scarce presence of supraglacial trees. For this purpose, diagrams coupling tree abundance at each plot with the environmental parameter values at the same plot were developed.

Moreover, a one-way ANOVA was performed in order to compare 15 supraglacial plots located above the treeline, where trees are surely absent (tree abundance = 0), against the 15 plots previously selected characterized by the presence of trees (tree abundance = 50, 75, 100) located at altitudes below the local treeline. This statistical analysis was performed to evaluate the parameters (slope, debris thickness, debris-surface temperature, ablation rate, variation in ice thickness over 28 years (1975–2003), aspect, and NDMI) more meaningfully related to tree presence and abundance.

Results

The Miage DCG: Main features and characteristics

The parameters indicating the characteristics of the Miage Glacier are reported in Figure 2.

Figure 2.

Characteristics of the Miage Glacier between 1730 and 2400 m a.s.l., including (a) debris surface temperature, (b) debris thickness, (c) NDMI, (d) slope, (e) ablation rate, (f) variation in ice thickness from 1975 to 2003, and (g) aspect. Altitude ranges are reported on the X axis. Maximum values are indicated with a continuous line, minimum values with a discontinuous line with small segments, and average value with a discontinuous line with larger segments.

Point debris-surface temperature shows a general decreasing trend with increasing altitude, in particular the average surface temperature at the glacier terminus is close to 30°C and at 2400 m a.s.l. it is reduced to 17°C. Peculiar values are observed between 2050 and 2150 m a.s.l., where both the minimum, maximum, and average values increase (Figure 2a). Considering a daily cycle 24 h long, it resulted that ground-measured temperatures of debris cover during a sunny summer day exceed 30°C and are about 1–4°C during night. Daily temperature excursion is therefore 28–33°C. Close to the terminus, debris temperature is positive throughout the vertical profile during almost the entire ablation season, and a continuous ice melt occurs at the bottom.

ASTER kinetic surface temperatures resulted to be similar to ground-measured temperatures: mean surface temperature over areas 90 m × 90 m wide varies between 25°C at 2059 m and 32°C close to the terminus. The Pearson’s correlation coefficient between ASTER and the field surface temperature data is 0.94 over the analyzed area.

Calculated debris thickness and ASTER temperatures are not collinear (r = 0.146) and can therefore be entered simultaneously in statistical analyses. The calculated debris thickness varies between 18 and 55 cm (average value over 90 m × 90 m areas). Overall, debris thickness increases downward, but is thinner in crevassed areas. The lowest thickness was measured at 1956 m a.s.l. in a crevassed area where the glacier divides into three lobes. Thickest debris corresponds to areas close to the terminus (Figure 2b).

NDMI on the Miage Glacier shows a minimum value of 16 (corresponding to maximal moisture) between 1951 and 2050 m a.s.l. and a maximum value of 50 (corresponding to drier environments) above 2350 m a.s.l. (Figure 2c). No big difference in the NDMI values is detected between the glacier terminus and 2050 m a.s.l., while maximum NDMI decreases starting from 1950 to 2250 m a.s.l., where its value passes from 49 to 42, and above 2250 m it rises up to 50.

The minimum value of slope detected is 0°, while the maximum value is 45°. The average varies between 8° and 13°, and it does not seem to follow a specific trend related to altitude (Figure 2d).

The ablation rate is influenced by debris distribution with rates varying from −0.3 m/yr, where debris exceeds 55 cm thickness, to −5.5 m/yr, where the debris is very thin or absent (Figure 2e). The ablation rate generally decreases from the higher to the lower altitudes, with higher values in June–July and lower values in August–September.

The changes in ice thickness over the period 1975–2003 show a general glacier volume loss (−16.640 × 10⁶ m³) from 1975 to 2003. Nevertheless, focusing on the two time sub-windows (i.e. 1975–1991 and 1991–2003), opposite trends are evident: in the period 1975–1991, the volume variation of the Miage Glacier was about +19.25 × 10⁶ m³, while in the period 1991–2003, a volume decrease of about −36.2 × 10⁶ m³ occurred. The thickness changes resulted positive (i.e. depth increase) in the lower glacier sector where debris mantle exceeds the critical value (this latter is the debris thickness driving a buried ice ablation rate equal to the one of bare ice at the same elevation; see also Mattson et al., 1993; once this value is exceeded, ablation rates are found diminishing) with values up to +18 m at 1730 m a.s.l.; the thickness variation was found to be negative (i.e. depth decrease) in the upper glacier zones (from 2250 m a.s.l.) where the debris layer is thinner; here, the changes locally exceed −30 m (Figure 2f).

As regards the aspect, 74% of the Miage Glacier surface ranges between 0° and 180° (predominantly East) from 1730 to 2400 m a.s.l. (Figure 2g).

Annual surface velocity of the Miage Glacier measured by the Differential GPS ranged between 0.3 and 90 m/yr, with a clear vertical gradient (Figure 3): smaller surface velocity values are observed with lower glacier elevations and decreasing slopes. The velocity field is characterized by compressing flow that is important for the debris-cover characteristics. In the snout part of the Miage Glacier, Pelfini et al. (2012) described a thicker debris mantle with compressing flow and thinner and sparse debris with extending flow, thus driving crevasse development and evolution.

Figure 3.

Ice flow at the surface of the Miage Glacier at different altitudes. The arrows represent the velocity vectors from 2006 field measurements (average velocity calculated from July to November), from 1950 to 2250 m a.s.l.

Glacier features on the areas showing tree vegetation presence

The characteristics of the Miage Glacier in the 15 selected plots are reported in Figure 4.

Figure 4.

Characteristics of the Miage Glacier in the 15 selected plots characterized by the presence of supraglacial trees, including (a) debris-surface temperature, (b) debris thickness, (c) NDMI, (d) slope, (e) ablation rate, (f) variation in ice thickness from 1975 to 2003, (g) aspect, and (h) annual surface speed. Tree vegetation abundance is reported in the X axis, respectively, at low, medium, and high density (50, 75, 100). See text for detailed description.

All the plots we selected were characterized by tree abundance exceeding 25 trees/plot.

In the selected plots, debris-surface temperature ranges between 19°C and 33°C where trees are present, and in particular, tree vegetation with lower density is only present where temperature has a value between 29°C and 33°C (Figure 4a).

Trees are present where debris thickness ranges between 19 and 55 cm. In particular, more than 90% of the plots are characterized by thickness ranging between 32 and 55 cm (Figure 4b).

Values of NDMI where arboreal vegetation is present range between 19 and 44 (Figure 4c).

Slope ranges between 2° and 10° where trees are present (Figure 4d).

Ablation rate where tree vegetation is present ranges between −0.6 and −1.8 m/yr (Figure 4e).

The average variation in glacier thickness over 28 years (1975–2003) never shows negative values where trees are present, and it ranges between +7 and +28 m (Figure 4f).

More than 85% of the Miage Glacier is characterized by an aspect ranging between 0° and 180° where tree vegetation is present, and in particular, 60% shows an NE aspect (between 0° and 90°) (Figure 4g).

Annual surface speed in the selected plots ranges between 0.8 and 7.0 m/yr (Figure 4h).

Distance from the closest source area (the nearest trees outside the glacier margins) for the selected plots ranges between 33.8 and 177.4 m. A clear trend is detectable: where the density of tree vegetation is lower, the distance from the source area is higher (between 177.38 and 79.25 m), and it gradually decreases when the density of trees increases (the plots with the highest density are located at a distance ranging between 94.0 and 33.8 m from the closest source area) (Figure 5).

Figure 5.

The relationship between vegetation abundance (selected plots) and the distance between the plots and the closest spot featuring vegetation outside the glacier. Tree abundance increases (100) at decreasing distances between the plot and the vegetation source area.

The results of the statistical analysis highlight that there is a statistically significant correlation at the p < 0.05 level between all the considered parameters and the presence/absence of trees, except for the aspect and NDMI factors that were deemed as not statistically significant. The results of the analysis are reported in Table 1.

Table 1.

Analysis of variance (ANOVA) comparing the considered glacier parameters and the presence of trees.

Parameter	ANOVA test
Ablation rate	F(3, 26) = 28.78; p < 0.000
Debris-surface temperature	F(3, 26) = 8.95; p < 0.000
Variation in ice thickness over 28 years	F(3, 26) = 65.42; p < 0.000
Slope	F(3, 26) = 50.33; p < 0.000
Debris thickness	F(3, 26) = 43.47; p < 0.000
Aspect	F(3, 26) = 1.29; p < 0.29
NDMI	F(3, 26) = 1.91; p < 0.15

NDMI: Normalized Difference Moisture Index.

Discussion

The Miage Glacier is known to be one of the few glaciers worldwide (and the only one in Italy) characterized by the presence of abundant supraglacial vegetation, including well developed trees, that can also be detected using color orthophotos with a pixel size of 0.5 m × 0.5 m. However, supraglacial vegetation can be detected in this way only when its density is very high (unpublished data), thus leading to the identification of vegetation on the Miage Glacier only where altitude ranges between 1730 and 1850 m a.s.l. This altitudinal range only represents 5% of the whole glacier area occupied by continuous supraglacial debris (Figure 6). Nevertheless, the treeline in the Val Veny for some species reaches an altitude of 2250 m a.s.l. (Leonelli and Pelfini, 2013). Although the distance between the 15 plots selected in this study and the closest proglacial forested area is an important factor in the establishment of supraglacial trees (Figure 5), their reduced density at an altitude above 1850 m a.s.l. suggests that one or more glacier parameters influence germination and development of supraglacial trees.

Figure 6.

(a) Area of the Miage Glacier tongue per altitude belts (from glacier terminus at about 1730–2400 m a.s.l.); (b) supraglacial tree distribution over the ablation tongue is only detectable between 1730 and 1850 m a.s.l. using color orthophotos with a pixel size of 0.5 m × 0.5 m, an altitude range representing 5% of the area characterized by the presence of supraglacial debris.

Thus, the selection of glacier parameters here presented has been done in order to describe the main morphological and environmental conditions at the surface of an alpine DCG. The selection was oriented to the features derivable from remote sensing sources to be able (1) to cover the whole debris-covered surface and (2) to ensure repeatability of the methods to other DCGs on the Alps and elsewhere. We analyzed the variability of the same parameters on selected glacier areas where supraglacial arboreal vegetation has been observed. The tree presence and features were detected and described through field surveys, thus giving high-resolution data and assuring that we had selected areas with an actual presence of trees. Even if our sample was restricted (overall, the 15 supraglacial selected plots characterized by the presence of trees featured an area of 3375 m²), this study allowed for the first time the identification of the glacial features, and their thresholds, permitting supraglacial tree germination and growth. In fact, trees are only present in areas featuring higher stability (i.e. slow surface velocity, <7 m/yr), thick debris cover (deeper than 19 cm), gentle slope (⩽10°), and positive changes in ice thickness (ranging between +7 and +28 m over 28 years). These conditions seem depicting a supraglacial stable environment favorable to tree germination and growth.

More precisely, although in the lower portion of the ablation tongue debris thickness ranges between 10 and 55 cm, trees are only present where debris is at least 19 cm thick (observed in only 1 of the 15 considered plots), and more than 90% of the selected plots are located where debris thickness exceeds 30 cm. On a DCG, debris thickness plays a key role in root frost occurrence during summer; in fact, at ice–debris interface the temperature is always at the melting point (Brock et al., 2010); thus, a thinner debris causes cooler (and in some cases also frozen) root conditions, while thicker layer allows warmer and more favorable root conditions. In fact, cold drives stress conditions having a negative impact on forest ecosystems, as underlined by Groffman et al. (2001) who found that more frequent soil freezing events could cause changes in root and microbial mortality and losses of nitrogen. The length of the yearly cold period is also an important factor in determining the stress conditions influencing trees: long winters can cause drought stress and loss of nutrients and, as a consequence, also represent a disturbance to the development of supraglacial trees (Tierney et al., 2001). The duration of the favorable period for growth in the European Alps, characterized by daily mean root-zone temperature of about 7°C, has to be at least of 3 months (Körner and Paulsen, 2004). For this reason, dedicated experiments are needed, in order to define what are the actual root conditions for both supraglacial trees and trees of the same species and age outside the glacier area at the same altitude, on stable lateral moraines (not ice-cored and not showing permafrost occurrence); in this way, it will be possible to describe the microclimatic conditions influencing the roots of supraglacial conifers and in particular during summer, when the conditions in the supraglacial area and outside the glacier are deeply different, thus probably requiring a higher root frost tolerance for the supraglacial trees.

Moreover, stable isotopes in tree rings studied by Leonelli et al. (2014) showed that supraglacial trees are mainly fed by water from liquid precipitation, thus suggesting that tree roots are not so close to buried ice to absorb the derived melting water.

Jones et al. (2005) detected supraglacial vegetation on the Matanuska Glacier (Alaska) only where debris exceeds 25 cm thickness. Our findings and the recent literature seem suggesting a thickness threshold allowing germination and growth of supraglacial trees, that is probably linked to root frost tolerance of tree species.

Debris thickness is linked to glacier surface velocity (Gilcrist et al., 2003), another parameter influencing tree establishment. Thick debris cover presents approximately 3–4 layers: a fine layer at the bottom with melting water along the first centimeters followed by a mixed layer of fine and coarse debris, a layer of coarse debris with clasts of 1–10 cm, and a final layer at the surface with clasts larger than 10 cm. The rock debris layer is found generally thicker than the ‘critical value’ (sensu Mattson et al., 1993). This latter is a depth threshold value which has to be locally evaluated and on the Alps is in the range between 4 and 6 cm (Franzetti et al., 2013, see the Supplementary Material). On the Miage, it was found equal to 3 cm (Mihalcea et al., 2008a). The debris mantle on the lower sectors is thicker than this threshold, and it actually reduces magnitude and rates of buried ice melt (Brock et al., 2010), thus allowing the glacier to maintain its ablation tongue at low altitudes as well.

The 15 selected plots featuring tree vegetation are located where slope does not exceed 10° and glacier surface velocity is lower than 7.0 m/yr, thus highlighting the importance of debris stability in the establishment of supraglacial trees. Glacier surface velocity and slope are key factors also in the establishment of herbaceous vegetation on the Miage Glacier (see Caccianiga et al., 2011).

Another environmental variable influencing tree germination on the glacier surface is vertical disturbance due to ice thickness loss. This disturbance can be evaluated through multiannual ice thickness variations, e.g. by comparing 2003 and 1975 DEMs. On the Miage Glacier tongue, supraglacial debris coverage modulates the magnitude and rates of buried ice ablation (see Diolaiuti et al., 2009): variation in ice thickness from 1730 to 2400 m a.s.l. over 28 years has been observed as ranging from −30 to +44 m, the latter where supraglacial debris exceeds 30 cm thickness, but no tree vegetation was detected on the glacier surface where variation in ice thickness shows negative values over 28 years, thus suggesting that the areas characterized by intense reduction in ice thickness are less favorable to tree germination and growth.

Even if ablation rate does not seem to directly play a role in the establishment of supraglacial trees, several studies, carried out both in the European Alps and in the Himalayan glaciers, show that ablation rate is usually reduced where debris coverage exceeds a few centimeters thickness (D’Agata and Zanutta, 2007; Juen et al., 2014; Pratap et al., 2015). In the snout part of the Miage Glacier, where supraglacial debris coverage is thicker, ablation rate is particularly reduced, thus increasing glacier surface stability and, as a consequence, tree establishment.

Debris-surface temperature, aspect, and NDMI do not seem to directly influence tree colonization on the glacier surface. The values of these parameters in the selected plots reflect the characteristics of the Miage Glacier between 1730 and 1850 m a.s.l. also where tree vegetation is not present. Debris-surface temperature on the Miage Glacier was already identified as not being directly linked with plant colonization (Caccianiga et al., 2011). The values observed in the snout part of the glacier do not seem to either limit or support tree growth, although a very wide day/night temperature range could represent a stress factor for the supraglacial arboreal species. The results concerning the aspect in the 15 selected plots reflect the aspect of the whole glacier terminus, thus suggesting that this environmental variable does not directly influence the establishment of trees. NDMI does not show high variability across altitude range on the glacier, especially in its terminal part. Since this index contrasts the near-infrared band (which is sensitive to albedo of leaf chlorophyll) to the mid-infrared band (which is sensitive to absorbance of leaf moisture), it should be directly linked to the presence of vegetation, both herbaceous and arboreal. We suggest that the resolution of the Landsat images used to calculate NDMI does not allow the detection of differences in moisture related to the discontinuous distribution of trees on the supraglacial debris. This index was already used to study changes in moisture in different environments (e.g. Brom et al., 2012; Lin et al., 2009), but we could not find scientific literature about its use on DCGs. Further analysis of moisture characterizing the supraglacial debris will provide a better understanding of the role of this variable in tree establishment.

The statistical analysis, even if restricted to a limited number of plots both below and above the treeline, supports our observations: all the parameters that seem to play a role in tree establishment and growth on the surface of a DCG actually show a statistically significant correlation with tree presence. Ablation rate and debris-surface temperature also appear to be significant in tree establishment, although we could not identify specific thresholds related to the presence of trees.

The statistical ANOVA analysis was not performed for glacier velocity since it was locally evaluated (through DGPS point measurements), and at some plots above the treeline we have no measurements.

DCGs are suitable sites to observe the growth of conifers in alpine soils with cooler root conditions than trees located outside glaciers, and these conditions may become more frequent in a changing climate. In fact, in a warmer climate, colder soils are expected to occur because of a thinner and less persistent snow winter coverage also outside glacier areas (Groffman et al., 2001), thus negatively affecting plants, especially because fine roots are more easily damaged in a situation of colder soils, and as a consequence nutrient uptake is compromised (Tierney et al., 2001). On the other hand, in a warming climate, earlier snow disappearance causes warmer soil temperature in spring (Luetschg and Haeberli, 2005). For this reason, tree species that are able to tolerate colder soils during winter and higher temperature amplitude cycles in summer, as the ones in glacial environments, will probably be the ones with a major occurrence in a context of warmer climate, while trees showing a low root frost tolerance will have more difficulties in surviving these changes.

The study of tree ecesis and germination year in the glacier foreland of a currently retreating debris-free glacier of the Italian Alps, the Forni Glacier (Stelvio National Park, Italy), determined by means of dendrochronological approach and whorls branch counting, shows an acceleration of the ecesis in the last few years, with an average value of 7 years (unpublished data). The glacier foreland that underwent deglaciation about 50–70 years ago is characterized by a much higher tree density compared with what we observe even in the plots characterized by the highest number of trees on the surface of the Miage Glacier, and since tree colonization on this DCG started about 100 years ago (Deline and Orombelli, 2005), we suppose that, even where the supraglacial conditions are favorable to tree colonization, the presence of ice under the debris and the glacier velocity (also if reduced) limit their establishment.

Of course many other conditions outside the glacier where trees are present need to be better investigated and compared with the ones in the supraglacial environment, in particular substrate temperature and moisture, in order to better understand the differences between the surface of DCGs and the areas surrounding the glacier as habitats for trees.

Conclusion

The originality of this research lies in the comparison between environmental parameters characterizing supraglacial debris and tree location, thus allowing the identification of well-defined intervals for several variables that characterize the spots where arboreal vegetation is present and well established. By knowing what are the values of a set of glacier parameters allowing tree establishment on a DCG, the actual or potential presence of trees on the surface of other DCGs may be predicted.

The methodology here presented represents a possible approach for the investigation of remote glacial areas and for the assessment of supraglacial tree presence at the regional scale. Further studies may then be conducted in the field in order to analyze dendroclimatic and dendroglaciological signals (Coppola et al., 2013; Leonelli et al., 2011; Pelfini, 1999), if morphological and environmental conditions suggest arboreal vegetation presence in the study area.

With this study, the already known link between glacier dynamics and supraglacial trees is even more emphasized, and in particular, our results suggest that glacier surface stability is the main factor influencing tree vegetation establishment on the supraglacial debris, with debris thickness, slope, variation in ice thickness, and glacier velocity being the environmental variables mainly driving tree colonization. Our results suggest that whenever the described parameters would show values supporting tree establishment and growth, DCGs may have acted as refugia for tree species during the warmer periods of the Holocene.

Future investigations will aim at (1) analyzing the same glacier parameters on other DCGs featuring supraglacial trees, in order to evaluate whether such conditions are local to the current study area or whether they are general factors driving establishment of trees; (2) investigating debris lithology where trees are present, in order to assess the role played by the lithological properties of supraglacial debris in driving tree presence; (3) investigating the actual conditions along the debris layer where tree roots develop; and (4) describing tree root physiology on the supraglacial debris and outside (comparative analysis).

Footnotes

Funding

This research has been developed within the PRIN 2010–2011 project (grant number 2010AYKTAB_006; local leader C Smiraglia and national leader C Baroni). The authors thank M Vagliasindi, the Fondazione Montagna Sicura ( ), and the Regione Autonoma Valle d’Aosta (RAVA) for the precious cooperation. The authors would also like to thank the two reviewers, whose comments considerably helped improve this manuscript.

References

Baretti

(1880) Il Ghiacciaio del Miage (versante italiano del Monte Bianco). Memorie della R. Accademia delle Scienze di Torino 32: 269–303.

Bollati

Pelfini

Smiraglia

(2013) Assessment and selection of geomorphosites and itineraries in the Miage glacier area (Western Italian Alps) according to scientific value for tourism and educational purposes. Environmental Management 51: 951–967.

Bollati

Leonelli

Vezzola

. (2014) The ecological support role in geomorphosites value assessment at the debris-covered Miage Glacier (Western Italian Alps) based on a 2.5 centuries-long review of scientific production. Geoheritage. DOI: 10.1007/s12371-014-0111-2.

Brock

Mihalcea

Kirkbride

. (2010) Meteorology and surface energy fluxes in the 2005–2007 ablation seasons at the Miage debris-covered glacier, Mont Blanc Massif, Italian Alps. Journal of Geophysical Research: Atmospheres. DOI: 10.1029/2009JD013224.

Brom

Nedbal

Procházka

. (2012) Changes in vegetation cover, moisture properties and surface temperature of a brown coal dump from 1984 to 2009 using satellite data analysis. Ecological Engineering 43: 45–52.

Brook

Hagg

Winkler

(2013) Debris cover and surface melt at a temperate alpine glacier: Franz Josef Glacier, New Zealand. New Zealand Journal of Geology and Geophysics 56: 27–38.

Caccianiga

Andreis

Diolaiuti

. (2011) Alpine debris-covered glaciers as a habitat for plant life. The Holocene 21: 1011–1020.

Capello

(1959) I laghi glaciali del Miage. L’Universo 21: 685–705.

Coppola

Leonelli

Salvatore

. (2013) Tree-ring based summer mean temperature variations in the Adamello-Presanella Group (Italian Central Alps) 1610–2008 AD. Climate of the past 9: 211–221.

10.

Cunietti

(1961) Rilevamenti di ghiacciai e studi glaciologici in occasione dell’anno geofisico (Ghiacciaio del Miage): rilievo fotogrammetrico. Bollettino del Comitato Glaciologico Italiano (II Serie) 9/1: 59–85.

11.

D’Agata

Zanutta

(2007) Reconstruction of the recent changes of a debris-covered glacier (Brenva Glacier, Mont Blanc Massif, Italy) using indirect sources: Methods, results and validation. Global and Planetary Change 56: 57–68.

12.

D’Agata

Smiraglia

Zanutta

. (2005) The recent variation of a debris covered glacier (Brenva Glacier) in the Italian Alps monitored with the comparison of maps and digital orthophotos. Journal of Glaciology 52: 183–185.

13.

Danby

Hik

(2007) Evidence of recent treeline dynamics in southwest Yukon from aerial photographs. Arctic 60: 411–420.

14.

Deline

(1999) Le variations Holocènes récentes du Glacier du Miage (Val Veny, Val d’Aoste). Quaternaire 10: 5–13.

15.

Deline

Orombelli

(2005) Glacier fluctuations in the western Alps during the Neoglacial, as indicated by the Miage morainic amphitheatre (Mont Blanc massif, Italy). Boreas 34: 456–467.

16.

Deline

Diolaiuti

Kirkbride

. (2004) Drainage of ice-contact Miage Lake (Mont Blanc Massif, Italy) in September 2004. Geografia Fisica e Dinamica Quaternaria 27: 113–119.

17.

Diolaiuti

Pecci

Smiraglia

(2003) Liligo Glacier (Karakoram): Reconstruction of the recent history of a surge-type glacier. Annals of Glaciology 36: 168–172.

18.

Diolaiuti

Bocchiola

Vagliasindi

. (2012) The 1975–2005 glacier changes in Aosta Valley (Italy) and the relations with climate evolution. Progress in Physical Geography 36: 764–785.

19.

Diolaiuti

Citterio

Carnielli

. (2006) Rates, processes and morphology of freshwater calving at Miage Glacier (Italian Alps). Hydrological Processes 20: 2233–2244.

20.

Diolaiuti

D’Agata

Meazza

. (2009) Recent (1975–2003) changes in the Miage debris-covered glacier tongue (Mont Blanc, Italy) from analysis of aerial photos and maps. Geografia Fisica e Dinamica Quaternaria 32: 117–127.

21.

Diolaiuti

Kirkbride

Smiraglia

. (2005) Calving processes and lake evolution at Miage Glacier (Mont Blanc, Italian Alps). Annals of Glaciology 40: 207–214.

22.

Emmer

Loarte

Klimes

. (2015) Recent evolution and degradation of the bent Jatunraju glacier (Cordillera Blanca, Peru). Geomorphology 228: 345–355.

23.

Fickert

Friend

Grüninger

. (2007) Did debris covered glaciers serve as Pleistocene refugia for plants? A new hypothesis derived from observations of recent plant growth on glacier surfaces. Arctic, Antarctic and Alpine Research 39: 245–257.

24.

Franzetti

Tatangelo

Gandolfi

. (2013) Bacterial community structure on two alpine debris-covered glaciers and biogeography of Polaromonas phylotypes. The ISME Journal: Multidisciplinary Journal of Microbial Ecology 7: 1483–1492.

25.

Fyffe

Reid

Brock

. (2014) A distributed energy-balance melt model of an alpine debris-covered glacier. Journal of Glaciology 60: 587–602.

26.

Garavaglia

Pelfini

Bollati

(2010) The influence of climate change on glacier geomorphosites: The case of two Italian glaciers (Miage Glacier, Forni Glacier) investigated through dendrochronology. Géomorphologie: relief, processus, environnement 2: 153–164.

27.

Gentili

Baroni

Caccianiga

. (2015) Potential warm-stage microrefugia for alpine plants: Feedback between geomorphological and biological processes. Ecological Complexity 21: 87–99.

28.

Ghosh

Pandey

Nathawat

(2014) Mapping of debris-covered glaciers in parts of the Greater Himalaya Range, Ladakh, western Himalaya, using remote sensing and GIS. Journal of Applied Remote Sensing. DOI: 10.1117/1.JRS.8.083579.

29.

Gilcrist

Baker

Goetz-Ensminger

SL et al.

(2003) Ice-motion control on super-glacial debris thickness and maturity, Matanuska Glacier, Alaska. In: Geological Society of America ‘Abstracts with Programs’, 37th Annual North-Central Sectional Meeting, Kansas City, MO, 24–25 March.

30.

Gjermundsen

Mathieu

Kääb

. (2011) Assessment of multispectral glacier mapping methods and derivation of glacier area changes, 1978–2002, in the central Southern Alps, New Zealand, from ASTER satellite data, field survey and existing inventory data. Journal of Glaciology 57: 667–683.

31.

Gobbi

Isaia

De Bernardi

(2011) Arthropod colonization of a debris-covered glacier. The Holocene 21: 343–349.

32.

Groffman

Driscoll

Fahey

. (2001) Colder soils in a warmer world: A snow manipulation study in a northern hardwood forest ecosystem. Biogeochemistry 56: 135–150.

33.

Guglielmin

Ellis Evans

Cannone

(2008) Ground thermal regime under different vegetation conditions in permafrost areas and sensitivity to climate change: A case study at Signy Island (maritime Antarctica). Geoderma 144: 73–85.

34.

Jones

Beutel

Ham

N et al.

(2005) Strain net analysis of the western terminus region, Matanuska Glacier, Alaska. In: North-Central Section: 39th Annual Meeting, Minneapolis, MN, 19–20 May.

35.

Juen

Mayer

Lambrecht

. (2014) Impact of varying debris cover thickness on ablation: A case study for Koxkar Glacier in the Tien Shan. The Cryosphere 8: 377–386.

36.

Körner

Paulsen

(2004) A world-wide study of high altitude treeline temperatures. Journal of Biogeography 31: 713–732.

37.

Leonelli

Pelfini

(2013) Past surface instability of Miage debris-covered glacier tongue (Mont Blanc Massif, Italy): A decadal scale tree-ring based reconstruction. Boreas 42: 613–622.

38.

Leonelli

Pelfini

Morra di Cella

(2009) Detecting climatic treelines in the Italian Alps: The influence of geomorphological factors and human impacts. Physical Geography 30: 338–352.

39.

Leonelli

Pelfini

D’Arrigo

. (2011) Non-stationary responses of tree-ring chronologies and glacier mass balance to climate in the European Alps. Arctic, Antarctic and Alpine Research 43: 56–65.

40.

Leonelli

Pelfini

Panseri

. (2014) Tree-ring stable isotopes, growth disturbances and needles volatile organic compounds as environmental stress indicators at the debris covered Miage Glacier (Mont Blanc Massif, European Alps). Geografia Fisica e Dinamica Quaternaria 37: 101–111.

41.

Lesca

(1974) Emploi de la photogrammetrie analytique pour la determination de la vitesse superficielle des glaciers et des profondeurs relatives. Bollettino del Comitato Glaciologico Italiano (II Serie) 22: 169–186.

42.

Lin

Chen

Shih

J et al.

(2009) Using modis-based vegetation and moisture indices for oasis landscape monitoring in an arid environment. In: IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Cape Town, 12–17 July.

43.

Luckman

Quincey

Bevan

(2007) The potential of satellite radar interferometry and feature tracking for monitoring flow rates of Himalayan glaciers. Remote Sensing of Environment 111: 172–181.

44.

Luetschg

Haeberli

(2005) Permafrost evolution in the Swiss Alps in a changing climate and the role of the snow cover. Norsk Geografisk Tidsskrift 59: 78–83.

45.

Masetti

Diolaiuti

D’Agata

. (2010) Hydrological characterization of an ice-contact lake: Miage Lake (Monte Bianco, Italy). Water Resources Management. Epub ahead of print 22 October. DOI: 10.1007/s11269-009-9519-x.

46.

Mattson

Gardner

Young

(1993) Ablation on debris covered glaciers: An example from the Rakhiot Glacier, Punjab, Himalaya. In: Young

(ed.) Snow and Glacier Hydrology. Wallingford: IAHS Press (IAHS Publication 218), pp. 289–296.

47.

Mihalcea

Brock

Diolaiuti

. (2008a) Using ASTER satellite and ground-based surface temperature measurements to derive supraglacial debris cover and thickness patterns on Miage Glacier (Mont Blanc Massif, Italy). Cold Regions Science and Technology 52: 341–354.

48.

Mihalcea

Mayer

Diolaiuti

. (2008b) Spatial distribution of debris thickness and melting from remote-sensing and meteorological data, at debris-covered Baltoro glacier, Karakoram, Pakistan. Journal of Glaciology 48: 49–57.

49.

Minora

Sesese

Bocchiola

. (2015) A simple model to evaluate ice melt over the ablation area of glaciers in the Central Karakoram National Park, Pakistan. Annals of Glaciology 56: 202–216.

50.

Müllerová

(2004) Use of digital aerial photography for sub-alpine vegetation mapping: A case study from the Krkonǒse Mts., Czech Republic. Plant Ecology 175: 259–272.

51.

Nicholson

Benn

(2006) Calculating ice melt beneath a debris layer using meteorological data. Journal of Glaciology 52: 463–470.

52.

Oerlemans

Giesen

Van den Broeke

(2009) Retreating alpine glaciers: Increased melt rates due to accumulation of dust (Vadret da Morteratsch, Switzerland). Journal of Glaciology 55: 729–736.

53.

Osterkamp

(2003) Establishing long-term permafrost observatories for active-layer and permafrost investigations in Alaska: 1997–2002. Permafrost and Periglacial Processes 14: 331–342.

54.

Pelfini

(1999) Dendrogeomorphological study of glacier fluctuations in the Italian Alps during the Little Ice Age. Annals of Glaciology 28: 123–128.

55.

Pelfini

Diolaiuti

Leonelli

. (2012) The influence of glacier surface processes on the short-term evolution of supraglacial tree vegetation: The case study of the Miage Glacier, Italian Alps. The Holocene 22: 847–856.

56.

Pelfini

Santilli

Leonelli

. (2007) Investigating surface movements of debris-covered Miage Glacier, Western Italian Alps, using dendroglaciological analysis. Journal of Glaciology 53: 141–152.

57.

Pratap

Dobhal

Mehta

. (2015) Influence of debris cover and altitude on glacier surface melting: A case study on Dokriani Glacier, central Himalaya, India. Annals of Glaciology. DOI: 10.3189/2015AoG70A971.

58.

Ranzi

Grossi

Iacovelli

L et al.

(2004) Use of multispectral ASTER images for mapping debris-covered glaciers within the GLIMS project. In: IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Anchorage, AK, 20–24 September, vol. 2, pp. 1144–1147. New York: IEEE.

59.

Ravazzi

(1999) Distribuzione ed ecologia di due primule endemiche delle Prealpi Calcaree Meridionali, Primula glaucescens e P. spectabilis, e considerazioni sulla loro corogenesi. Archivio di Geobotanica 3: 125–148.

60.

Reid

Brock

(2010) An energy-balance model for debris-covered glaciers including heat conduction through the debris. Journal of Glaciology 56: 903–916.

61.

Reid

Brock

(2014) Assessing ice-cliff backwasting and its contribution to total ablation of debris-covered Miage Glacier, Mont Blanc massif, Italy. Journal of Glaciology. DOI: 10.3189/2014JoG13J045.

62.

Smiraglia

Diolaiuti

Casati

. (2000) Recent areal and altimetric variations of Miage Glacier (Monte Bianco massif, Italian Alps). In: Nakawo

Raymond

Fountain

(eds) Debris-Covered Glaciers. Wallingford: IAHS Press (IAHS Publication 264), pp. 227–233.

63.

Stokes

Popovnin

Aleynikov

. (2007) Recent glacier retreat in the Caucasus Mountains, Russia, and associated increase in supraglacial debris cover and supra-/proglacial lake development. Annals of Glaciology 46: 195–203.

64.

Taschner

Ranzi

(2002) Comparing the opportunities of LANDSAT-TM and ASTER data for monitoring a debris covered glacier in the Italian Alps within GLIMS project. In: IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Toronto, 24–28 June, vol. 2, pp. 1044–1046. New York: IEEE.

65.

Tierney

Fahey

Groffman

. (2001) Soil freezing alters fine root dynamics in a northern hardwood forest. Biogeochemistry 56: 175–190.

66.

Turchetti

Goretti

Branda

. (2013) Influence of abiotic variables on culturable yeast diversity in two distinct Alpine glaciers. FEMS Microbiology Ecology 86: 327–340.

67.

Vescovo

Gianelle

(2008) Using the MIR bands in vegetation indices for the estimation of grassland biophysical parameters from satellite remote sensing in the Alps region of Trentino (Italy). Advances in Space Research 41: 1764–1772.

68.

Wilson

Sader

(2002) Detection of forest harvest type using multiple dates of Landsat TM imagery. Remote Sensing of Environment 80: 385–396.