Air temperature thresholds to evaluate snow melting at the surface of Alpine glaciers by T-index models : the case study of Forni Glacier ( Italy )

Introduction Conclusions References


Introduction and study aims
Melting processes at glaciers occur when surface temperature is equal to 273.15 K and the surface energy budget is positive (e.g.Hock, 2005;Senese et al., 2012a).It is however very difficult assessing these conditions, in particular whenever a supraglacial Au-Introduction

Conclusions References
Tables Figures

Back Close
Full tomatic Weather Station (AWS) is not located and running at the glacier surface; thus simple models are generally adopted which assume empirical relationships between air temperature and snow or ice melting rates (i.e.T-index or Degree Days models e.g.Braithwaite, 1985;Cazorzi and Dalla Fontana, 1996;Hock, 1999;Pellicciotti et al., 2005).The degree days amount during the ablation period have naturally to be calculated for any point of the glacier surface.The estimation of distributed degree days amount from air temperature data acquired by meteorological stations located outside a glacier is not always simple and the use of an appropriate lapse rate is very important especially if the used meteorological station features a strong elevation difference with respect to the studied glacier.Unfortunately this condition often occurs as meteorological stations are mainly located close to urbanized areas which are at bottom of mountain valleys and only a small number of glaciers are equipped with AWSs.The availability of a supraglacial AWS is also very useful in order to assess presence and persistence of snow cover.These information are particularly useful in the case of late spring/early summer time (April-June) at the glacier ablation zone in order to establish in detail if snow is at the melting point and how long this phenomenon occurs.
A further very important issue in the application of T-index models is the air temperature threshold adopted for melting conditions: generally a daily average value of 273.15K is used (Braithwaite, 1985;Hock, 2003;Bocchiola et al., 2010).This assumption neglects the fact that melting does not necessarily occur only when daily average temperature exceeds 273.15 K, and there can be days featuring a negative average temperature value but experiencing a fraction of the time (from a few to several hours) with positive air temperature (thus driving melt).Moreover the surface energy balance can be positive also with negative air temperature (Kuhn, 1987).Then the choice of an appropriate air temperature threshold is fundamental to detect the actual melting days and to evaluate the melt amount.
Within this context the aim of this paper is to investigate the relationship between degree days amount and melting processes considering a 7 year period of measures on the Forni Glacier which is the largest Italian valley glacier (Fig. 1).It is widely debris-Introduction

Conclusions References
Tables Figures

Back Close
Full free, even if darkening phenomena are shown (Diolaiuti and Smiraglia, 2010;Diolaiuti et al., 2012).The measures were performed both by means of a supraglacial AWS (named AWS1 Forni, see Citterio et al., 2007;Senese et al., 2012a, b) and by field surveys (i.e.snow pits and ablation stakes).The goal of the research is to contribute better understanding the use of T-index models focusing on the air temperature threshold.
Moreover we aim at evaluating the performances that can be achieved using meteorological stations located outside the glacier to reconstruct supraglacial air temperature conditions.

Data and methods
The analyses presented in this paper are based on data collected in the 2005-2012 period: 1. daily average air temperature records (T AWS ) collected by a supraglacial AWS, the AWS1 Forni at 2631 m a.s.l., and by a meteorological station located at Bormio, a village at 1225 m a.s.l. and about 20 km far from the glacier terminus (Fig. 1); the air temperature data acquired by this second weather station were shifted to the AWS1 Forni elevation by applying the mean tropospheric lapse rate (−6.5 K km −1 , the resulting temperature values are indicated T B from here).
2. Hourly air temperature, relative humidity, wind speed, air pressure, solar and infrared radiation (both incoming and outgoing) data collected by the AWS1 Forni.with very few gaps (3.05 % of the total period).In this study the time frame from 1 October 2005 to 31 December 2012 is analysed.The Bormio meteorological station is managed by the Lombardy Regional Agency for Environmental Protection (ARPA Lombardia), which takes care data validation and disseminates the data through a dedicated web site.The snow pit data have been provided by personnel from the Centro Nivometeorologico di Bormio of ARPA Lombardia according to the AINEVA protocol.The pits were annually dug at the end of the accumulation period (April-May), only in 2007 no surveys were performed.
The snow and ice melting is assessed from surface energy budget (Melt from Energy Budget, M EB , calculated from AWS1 Forni data).The net energy (R S ) available for heating the surface and melting snow and/or ice is calculated following Senese et al. (2012a): where SW net and LW net correspond to the net radiation (shortwave and longwave, respectively), SH and LE to the sensible and latent heat fluxes.All the fluxes (W m −2 ) were defined positive when directed towards the surface.The conductive heat flux at the surface was neglected since no temperature sensors were located in the snowpack and in the ice surface layer.The melt amount (M EB , kg m −2 or mm w.e.) was obtained from: where L m is the latent heat of melting (3.34 × 10 5 J kg −1 ).Whenever the surface temperature (T S derived from outgoing LW data acquired by the AWS1 Forni) is at 273.15 K and R S is positive, the melt amount (M EB ) is quantified.The obtained results are considered reliable and indicative of the glacier actual melt since in a previous paper (Senese et al., 2012a) the melt amount derived from energy budget computations has been Introduction

Conclusions References
Tables Figures

Back Close
Full compared to the melt amount measured on the field at a selection of ablation stakes located nearby the AWS1 Forni.These two dataset resulted strongly correlated (less than 3 % of difference between modelled cumulative melt and measured cumulative one) thus supporting the application of energy budget computation to assess the actual melt amount.Besides to M EB we calculated also other two mass loss amounts: (i) melt occurred with 3 driving Conditions (M 3C ) which corresponds to the snow ablation during days featuring positive energy balance, surface temperature of 273.15 K, and albedo (α, from here) > 0.4: (ii) Melt evaluated only considering the Positive Energy Budget (M PEB ) which is estimated neglecting glacier surface temperature and conditions: The snow melting is also assessed by a T-index model (M T-INDEX ) following Braithwaite (1985): where T t corresponds to the air temperature threshold (K) adopted by the model and DDF to the degree-day factor (mm K −1 d −1 ).The applied temperature data were T B .
The snow Degree-Day Factor (sDDF) was found considering the degree days amount and the Snow Water Equivalent (SWE) values estimated from snow pits performed nearby the AWS1 Forni.The presence of snow or bare ice was deducted from albedo data (from AWS1 Forni) and then the length of the snow coverage period.In Introduction

Conclusions References
Tables Figures

Back Close
Full fact, the SWE was considered completely melted when the albedo becomes lower than 0.4.Finally the sDDF was calculated as: where DD glacier is the sum of Degree Days (from T B data) in the time frame between a snow pit survey and the occurrence of ice albedo.Moreover we also considered different air temperature thresholds.
We chose Bormio temperature record since this village is less affected by thermal inversion than other stations located nearby and inserted in the ARPA Lombardia meteorological network.By comparing T B data with data actually measured by the AWS1 Forni (Fig. 2) a high correlation value results as well (r = 0.91), with a root mean square error of the modelled temperature slightly over 3 K.Moreover the slope coefficient of the linear regression between measured and modelled temperatures at the AWS1 Forni site turns out to be very close to 1 (see Fig. 2).
In order to detect the most suitable daily temperature threshold (T t ) to adopt in the T-index model for quantifying glacier melting in the April-June period, we considered hourly M EB values (obtained from AWS1 Forni data) and studied how long ablation occurred in each day (number of hours per day).Then we sorted these data according to temporal length classes (0, 4, 6, 12 and 24 melting hours per day).For each class we calculated the mean, maximum and minimum values of daily average air temperature from T AWS and T B data.These temperature data represent possible thresholds to be applied to calculate degree days driving snow melt.We performed several attempts of running the T-index model by applying the different temperature threshold values and the obtained melt amounts (M T-INDEX ) were compared with the ones from energy budget computation (M EB ) thus permitting to select the most suitable and performing threshold values.Introduction

Conclusions References
Tables Figures

Back Close
Full

Melting from energy balance
Melting at the AWS1 Forni site estimated from the energy balance is shown in Fig. 4. M EB was obtained considering only hours characterized by null (273.15K) surface temperature and positive energy budget.The time window of our analysis is scheduled according to the hydrological year (i.e. at Mid-latitudes from 1 October of the year x to 30 September of the year x + 1).It results a total M EB of −37.5 m w.e. over a 7 year period.Supraglacial albedo data permit to detect snow cover presence (α > 0.4) during days featuring both positive energy balance and surface temperature of 273.15 K.The occurrence of these three criteria (ranging from 69 to 108 days per year) allows calculation of M 3C (light grey bars in Fig. 3) which is found not negligible with a melt amount from 17.6 % to 29.2 % of the total annual value.In general the ice melting period (dark grey bars in Fig. 3) results longer (with an average value of 100 days per hydrological year) than the M 3C time frame (89 days per year mean value) with the unique exception of the hydrological year 2009/2010 (88 and 101 days for ice and snow, respectively).
In order to assess the importance of the condition concerning glacier surface temperature, we also performed a test calculating ablation (M PEB ) only from positive energy balance data during April-June period (91 days long) and not considering glacier surface temperature.The results were compared with M EB values (Table 1).Totally a M EB for the April-June period over the time window 2006-2012 of −12.72 m w.e. is calculated.This value differs from the M PEB which results equal to −13.43 m w.e.
The quite wide range of variability of our results suggests further analyses.In fact, the April-June M EB corresponds to ca. 1/3 of the total M EB (Table 1).Focusing on this period we found the most intense ablation occurred in 2011 (−2.26 m w.e.contributing for 43.5 % with respect to the total M EB for that year) and the minimum was found in 2008 (−1.55 m w.e., 31.0 % of the annual value).

Conclusions References
Tables Figures

Back Close
Full To choose the most suitable daily air temperature threshold (T t ) to detect melting days in the April-June period and to evaluate the corresponding melting amount by T-index model we analysed T AWS and T B data (see Table 2).
Days without melting (0 M EB hours per day, 3rd column in Table 2) resulted occurring over 4.9 % of the total time frame (April-June over a 7 year long period).On the other hand, days featuring continuous and uninterrupted melting (24 M EB hours per day, 8th column in Table 2) cover 3.3 % of the total time.The cumulative melt occurred over these 21 days is equal to −0.77 m w.e. and it represents only the 6.0 % of the total loss.These days are characterized by daily air temperature at AWS1 Forni always positive with a daily mean value of 278.6 K, the maximum of daily average data equal to 282.1 K and the minimum of daily average values of 275.5 K. T B data over the same time frames are characterized by a mean daily average of 279.8 K, a maximum of daily average of 284.6 K and a minimum daily temperature of 274.7 K.
Days with at least 12 M EB hours (7th column in Table 2) occurred on the 47.3 % of the total analyzed time and they represent the 73.3 % of the whole ablation (cumulative M EB : −9.30 m w.e.).
Days with at least 6 M EB hours (6th column in Table 2) occurred on the 83.8 % of the total analyzed time and they represent the 97.7 % of the whole ablation (cumulative Then the largest part of the melt resulted occurring on days with at least 6 M EB hours thus suggesting to consider temperature data calculated for this class as suitable T t .In particular the minimum value of the average daily temperature calculated with at least 6 M EB hours per day (268.6K and 268.1 K considering the T AWS and T B data, respectively) were applied to detect the actual melting days.The first threshold was applied to T AWS data and permitted to select 586 melting days describing the 99.4 % (−12.64 m

Conclusions References
Tables Figures

Back Close
Full w.e.) of the total M EB .Considering T B data, the threshold of 268.1 K permitted to detect 601 melting days giving a cumulative M EB of −12.66 m w.e.equal to 99.5 % of the total.
From our results it is also clear that the use of the 273.15K as temperature threshold drives an underestimation of ablation.In fact, this value applied to the AWS1 Forni temperature data allows to detect 421 melting days (66.1 % with respect to the total time frame) equal to a cumulative M EB of −11.39 m w.e.(89.5 % with respect to the total melt).The same threshold applied to the T B data allowed to select 481 melting days giving a M EB of −12.00 m w.e.(94.3 % with respect to the total value).
In both cases the application of 273.15K threshold does not represent the most suitable and exhaustive solution and an underestimation up to 10 % of the total M EB in the April-June period can occur.

Melt calculation from T-index model
Besides the calculation of snow and ice melting by means of the energy balance, the snow daily melting rate at the AWS1 Forni site was also calculated by applying the T-index model (M T-INDEX ).T B data represent the input values to calculate the degree days sum driving snow ablation.We focused on melting in the April-June period from 2006 to 2012.
Different temperature thresholds (T t ) estimated from T B data were used: the most common and applied value (273.15K), the minimum of daily average temperature calculated for days featuring at least 6 M EB hours (268.1 K) and the minimum daily average temperature calculated for days featuring 24 M EB hours (274.7 K).By comparing the cumulative M T-INDEX curves (Fig. 4) a similar trend is evident but the total amount of melting over the 7 year long period results quite different depending on the applied T t value: from −6.81 m w.e.(threshold 274.7 K) to −8.23 m w.e.(threshold 268.1 K).
The cumulative M 3C in the same period results −8.23 m w.e.thus underlying that the temperature threshold that better explains magnitude and variability of snow melting corresponds to the minimum of daily average temperature calculated for days featuring at least 6 M EB hours (268.1 K).Introduction

Conclusions References
Tables Figures

Back Close
Full

Discussions
During April-June at the AWS1 Forni site all the snow coverage has completely melted; in fact, from 2006 to 2012 the M 3C over the April-June time frame is found −8.23 m w.e.corresponding to the 97 % of the annual M 3C and to the 65 % of the April-June M EB (see Table 1).This latter results equal to −12.72 m w.e., thus underlying that also a not negligible amount of ice has melted.On the one hand, the knowledge of both T S and α allows the calculation of M 3C amount thus permitting to distinguish between ice and snow melt and evaluating their respective contribution to the whole loss at the glacier tongue.In fact, when we computed the April-June mass loss neglecting T S and α, the ablation we found (M PEB ) results equal to −13.43 m w.e.thus underlining an overestimation (+5.6 % with respect to the M EB over the same period).Then at least the knowledge of glacier surface temperature is really important to avoid misestimation of the melt amount and whenever also α is known the computation of ice and snow contribution to the total melt is possible.On the other hand T S and α are available only on a small number of glaciers where supraglacial AWSs have been running thus suggesting to look for different strategies in order to assess the snow melt amount.The most diffuse and simple method is the T-index approach based on data acquired outside the studied glacier.Then in this study to compute the degree days amount driving glacier melt we used air temperature values measured at Bormio (1225 m a.s.l.) and shifted to the AWS1 Forni site elevation (2631 m a.s.l.) applying the mean tropospheric lapse rate (−6.5 K km −1 ).
To validate this temperature reconstruction we made a comparison among modelled air temperature values (T B ) with air temperature actually measured at the AWS1 Forni (T AWS ).It resulted a strong agreement between the two records of data (r = 0.91) suggesting that it is possible to perform a reasonable reconstruction of supraglacial daily average air temperature starting from meteorological data acquired down valley also using a mean tropospheric vertical gradient as lapse rate.Nevertheless this result could depend on the small distance (ca.20 km) between Bormio village and Forni

Conclusions References
Tables Figures

Back Close
Full Glacier.But on other glaciers the assessment of a local daily vertical gradient could be needed to reconstruct supraglacial air temperature and degree days amount from data acquired by weather stations located far and farer from the glacier site.
From our analysis results that the major uncertainty in computing the degree days amount at the glacier surface and then in assessing snow melt is given by the choice of an appropriate air temperature threshold (T t ).The modelled snow melt largely varies: from −6.81 m w.e.applying a threshold of 274.7 K to −8.23 m w.e. using a value of 268.1 K. Then T-index models show a high sensitivity to T t changes.
To assess the most suitable temperature threshold, we firstly analyzed hourly M EB values (obtained from AWS1 Forni data) in the April-June period to detect if ablation occurs and how long this phenomenon takes (number of hours per day).Then we sorted these data according to temporal length classes (0, 4, 6, 12 and 24 melting hours per day).For each class we calculated the mean, maximum and minimum values of daily average air temperature from T AWS and T B data.Moreover we also evaluated the cumulative M EB for each temporal length class to evaluate its weight with respect to the total ablation.The largest part of the melting (97.7 % of the whole ablation) resulted occurring on days featuring at least 6 M EB hours thus suggesting to consider the minimum average daily temperature value calculated for this class as a suitable temperature threshold (268.6K and 268.1 K considering the T AWS and T B data respectively).These data were applied to detect the actual melting days.The first threshold was applied to air temperature data from AWS1 Forni and permitted to select 586 melting days describing the 99.4 % (−12.64 m w.e.) of the total M EB .Considering T B data, the threshold of 268.1 K permitted to detect 601 melting days giving a cumulative M EB of −12.66 m w.e.equal to 99.5 % of the total.
Then we ran a simple T-index model using as input data T B and varying the temper- also obtained applying as threshold 268.11K to T B data.These results suggest to use as T t the minimum of daily average temperature calculated for days featuring at least 6 M EB hours.This value permits to better explain magnitude and variability of snow melting (Fig. 4).
As regards the temperature threshold largely reported in the literature dealing with T-index model (273.15K e.g.Braithwaite, 1985;Hock, 2003), our tests indicate that it drives an underestimation of snow melt (ca.17 % with respect to the value obtained from M 3C which represents the actual snow melt).Moreover in the case the main purpose should be only the detection of the melting days and not the assessing of the melt amount, the application of the 273.15K threshold drives an underestimation as well.With this threshold 481 melting days are detected against 601 melting days selected by applying the 268.1 K threshold to T B data.
Our findings results in agreement with van den Broeke et al. ( 2010) who modelled the ablation at the Greenland ice sheet.They found an about 5 K lower threshold value by observing the cumulative distribution of daily average temperature for days with melt at three AWSs.From their study this value resulted to allow also an appropriate calculation of snow and ice degree-day factors.In fact applying the common 273.15K threshold they obtained non-sense degree-day factor values.
Some attention has to be spent on the snow DDF (see Eq. 6) values.The snow DDFs we found, ranging from 2.0 to 5.8 mm d −1 K −1 , are in agreement with findings in other studies (see Table 3) With the 268.1 K threshold the snow degree-day factor (Table 3) results lower than the ones derived applying the other two thresholds.This value (2.0 mm d −1 K −1 ) resulted comparable in magnitude with the ones reported by van de Broeke et al. ( 2010) applying a threshold of 268 K to Greenland data and also these authors found smaller snow DDF value occurring with a T t 5 K lower than the 273.15K value.

Conclusions
The most suitable way to evaluate the actual snow melt amount is to calculate M 3C , the melt from energy budget computations considering both null surface temperature and albedo major than 0.4.Nevertheless this approach is applicable only whenever a supraglacial AWS is running thus giving information on T s and α.Otherwise the snow melt can be estimated by applying T-index methods.The main limits in such approaches are due to the need of reconstructing air temperature at the glacier surface from data acquired outside the glacier and choosing a suitable air temperature threshold for computing the DD amount.Then we performed further investigations to contribute to this topic.
Firstly we started evaluating the actual glacier melt.We applied a point energy balance model (following Senese et al., 2012a) at the AWS1 Forni site on the glacier ablation tongue and we found that a not negligible annual snow melt (M 3C ) occurs.In the 7 year period 2005-2012 it ranges from 17.6 % to 29.2 % of the total annual (ice and snow) ablation (M EB ).
Our data also indicate that during April-June at the AWS1 Forni site all the snow coverage disappears due to melting processes: from 2006 to 2012 the April-June M 3C is found corresponding to the 66.59 % of the April-June M EB .
Secondly in order to apply a T-index model we reconstructed air temperature conditions at the glacier surface starting from data measured at Bormio (1225 m a.s.l.) and shifted to the AWS1 Forni site elevation (2631 m a.s.l.) applying the mean tropospheric lapse rate (−6.5 K km −1 ).This approach resulted a reasonable reconstruc-Introduction

Conclusions References
Tables Figures

Back Close
Full tion of supraglacial daily average air temperature starting from meteorological data acquired down valley also using a mean tropospheric vertical gradient as lapse rate.
Then our analysis indicates that the major uncertainty in computing the degree days amount at the glacier surface and then in assessing snow melt is given by the choice of an appropriate air temperature threshold (T t ).In fact, ablation does not occur only with air temperature higher than 273.15K since it is determined by the surface energy balance and this latter is only indirectly affected by air temperature (Kuhn, 1987).
Thirdly, to assess the most suitable temperature threshold, we analyzed hourly M EB values (obtained from AWS1 Forni data) in the April-June period to detect if ablation occurs and how long this phenomenon takes (number of melting hours per day: 0, 4, 6, 12 and 24).The largest part of the melting (97.7 % of the whole ablation) resulted occurring on days featuring at least 6 M EB hours thus suggesting to consider the minimum average daily temperature value calculated for this class as a suitable temperature threshold (268.1 K).This value was applied to detect the actual melting days permitting to select 601 days giving a cumulative M EB of −12.66 m w.e.equal to 99.5 % of the total.
Finally we ran a simple T-index model using as input data the temperatures measured at Bormio and shifted to the glacier elevation and varying the temperature threshold (273.15 K, 268.1 K and 274.7 K).The cumulative M 3C (which witnesses actual snow melt) in the same period results −8.23 m w.e. and this value is also obtained applying as threshold 268.11K.These results indicate as the most suitable T t the minimum of daily average temperature calculated for days featuring at least 6 M EB hours.This value (in agreement with findings by van den Broeke et al., 2010 who studied the Greenland ice sheet) permits to better explain magnitude and variability of snow melting and avoids most of the sampling problem in the application of the positive T-index models, such as the calculation of snow and ice degree-day factors.Then probably the choice of a 268 K value as air temperature threshold for computing degree days amount could be generalized and applied not only on Greenland glaciers but also on Mid latitude and Alpine ones like the Forni Glacier.Introduction

Conclusions References
Tables Figures

Back Close
Full  Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

3.
Snow pits annually performed nearby the AWS1 Forni.The AWS1 Forni (Fig. 1) was set up on the glacier melting tongue on 26 September 2005.It was developed in the framework of the SHARE (Stations at High Altitude for Research on the Environment) program and data are quality checked and validated according to the SHARE protocol and available to all the scientific community upon request.This dataset is quite uninterrupted (from 26 September 2005 up to now) Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 3 Results Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 3.2 Daily air temperature thresholds witnessing glacier melting in the April-June period Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ature threshold (273.15K, 268.1 K and 274.7 K).The cumulative M T-INDEX curves we obtained in the April-June period (Fig. 4) show a similar trend but the total amount of snow melting over the 7 year long period results quite different.The cumulative M 3C (witnessing only snow melt) in the same period results −8.23 m w.e. and this values is Discussion Paper | Discussion Paper | Discussion Paper | : Hock (2003) reported snow DDF values ranging from 2.5 mm d −1 K −1(Clyde, 1931) to 11.6 mm d −1 K −1(Kayastha et al., 2000) andBraithwaite (2008) suggested values from 3.5 to 4.6 mm d −1 K −1 .Moreover the snow DDFs we found result always minor than the ice DDFs we evaluated (they resulted 5.6, 3.6 and 6.7 mm d −1 K −1 with T t 273.15 K, 268.1 K, 274.7 K, respectively) and this is in agreement with the literature dealing with the application of T-index models for computing ice and snow melt (see the overview by Braithwaite, 2011 and the considerations on the different magnitude of ice and snow DDFs reported by van de Broeke et al., 2010).Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Diolaiuti, G. and Smiraglia, C.: Changing glaciers in a changing climate: how vanishing geomorphosites have been driving deep changes in mountain landscapes and environments, Geomorphologie, 2, 131-152, 2010.Diolaiuti, G., Bocchiola, D., D'Agata, C., and Smiraglia, C.: Evidence of climate change impact upon glaciers' recession within the Italian Alps: the case of Lombardy glaciers, Theor.Appl.Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 1 .
Fig. 1.On the top the location of Bormio (a village at 1225 m a.s.l.) and Forni Glacier (Google Earth © raster base) and in the lower picture the location of the AWS1 Forni (black dot, WGS84 coordinates: 46 • 23 56.0 N and 10 • 35 25.2 E, geodetic elevation: 2631 m a.s.l.).In the map the light grey areas (1) are used to mark supraglacial debris coverage and the dark grey zones (2) indicate rock exposures and nunataks.

Table 1 .
For each hydrological year the following data are shown: cumulative annual M EB value (2nd column), cumulative M EB during April-June (3rd column), cumulative M 3C during the snow melting period (5th column), cumulative M PEB during April June (7th column) and percentages with respect to the annual M EB value (4th, 6th and 8th column respectively).All melting amounts are in m w.e.