The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ The Cryosphere doi:10.5194/tc-5-405-2011 © Author(s) 2011. CC Attribution 3.0 License.

Point observations of liquid water content in wet snow – investigating methodical, spatial and temporal aspects

F. Techel and C. Pielmeier WSL Institute for Snow and Avalanche Research SLF, 7260 Davos, Switzerland Received: 22 September 2010 – Published in The Cryosphere Discuss.: 12 October 2010 Revised: 30 March 2011 – Accepted: 13 May 2011 – Published: 18 May 2011

Abstract. Information about the volume and the spatial 1 Introduction and temporal distribution of liquid water in snow is impor- tant for forecasting wet snow avalanches and for predicting melt-water run-off. The distribution of liquid water in snow The distribution and amount of liquid water in snow af- is commonly estimated from point measurements using a fects surface albedo (Warren and Wiscombe, 1985; Gupta “hand” squeeze test, or a dielectric device such as a “Snow et al., 2005), snow stability (e.g. Armstrong, 1976; Kattel- Fork” or a “Denoth meter”. Here we compare estimates of mann, 1985; Conway and Raymond, 1993), and is important water content in the Swiss Alps made using the hand test to for forecasting the onset of melt-water run-off (Jones et al., those made with a Snow Fork and a Denoth meter. Measure- 1983) and reservoir management (Kattelmann and Dozier, ments were conducted in the Swiss Alps, mostly above tree 1999). line; more than 12 000 measurements were made at 85 loca- In Switzerland, most snowpack information comes from tions over 30 days. Results show that the hand test generally snow profiles measured by researchers, avalanche profes- over estimates the volumetric liquid water content. Estimates sionals, and observers from the Swiss snow observation net- using the Snow Fork are generally 1 % higher than those de- work. Measurements are made at flat study plots as well as rived from the Denoth meter. The measurements were also on potential avalanche slopes with different elevations and used to investigate temporal and small-scale spatial patterns aspects. Typically about 1000 snow profiles (mostly in dry of wetness. Results show that typically a single point mea- snow) are collected each year, and used by the avalanche surement does not characterize the wetness of the surround- warning service as input to the national avalanche hazard ing snow. Large diurnal changes in wetness are common in forecast (Schweizer and Wiesinger, 2001). Liquid water con- the near-surface snow, and associated changes at depth were tent is typically estimated by conducting “hand” tests for also observed. A single vertical profile of measurements is each stratigraphic snow layer according to Swiss and inter- not sufficient to determine whether these changes were a re- national observational guidelines (WSL, 2008; Fierz et al., sult of a spatially homogeneous wetting front or caused by in- 2009, Table 1). In this test, a sample is squeezed by hand, and filtration through pipes. Based on our observations, we sug- the water content is estimated by observing its response (Ta- gest that three measurements at horizontal distances greater ble 1). A magnifying lens can also be used to detect the pres- than 50 cm are needed to adequately characterize the distri- ence/absence of liquid water. Measurements of snow tem- bution of liquid water through a snowpack. Further, we sug- perature can also be used to determine the presence/absence gest a simplified classification scheme that includes five wet- of water; snow is expected to be dry at temperatures less than ness patterns that incorporate both the vertical and horizontal 0 ◦C. However, estimating liquid water content is difficult distribution of liquid water in a snowpack. even for experienced observers (Martinec, 1991b; Fierz and Fohn¨ , 1994), partly because water flow through snow varies both spatially and temporally (e.g. Colbeck, 1979; Marsh, 1988; Conway and Benedict, 1994). Our objectives in this study are threefold: (i) investi- gate the reliability of point observations in relation to tem- Correspondence to: F. Techel poral and small-scale spatial variability; (ii) compare mea- ([email protected]) surements using the hand test with other measurements

Published by Copernicus Publications on behalf of the European Geosciences Union. 406 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

Table 1. Hand test for the qualitative estimation of liquid water content (mWC) and the approximate range of liquid water content (θ). The detailed description is taken from the International Classification of Seasonal Snow on the Ground (Fierz et al., 2009, p. 8). This classification is also used in Swiss observational guidelines (WSL, 2008). Half index classes may also be used. ts – snow temperature.

Wetness Index Description θ Content (mWC) [vol. %] ◦ Dry 1 ts ≤0.0 C. Disaggregated snow grains have little tendency 0 to adhere to each other when pressed together. ◦ Moist 2 ts = 0.0 C. The water is not visible, even at 10× magnification. 0–3 When lightly crushed, the snow has a tendency to stick together. ◦ Wet 3 ts = 0.0 C. The water can be recognized at 10× magnification 3–8 by its meniscus between adjacent snow grains, but water cannot be pressed out by moderately squeezing the snow in the hands. ◦ Very Wet 4 ts = 0.0 C. The water can be pressed out by moderately 8–15 squeezing the snow in the hands, but an appreciable amount of air is confined within the pores. ◦ Soaked 5 ts = 0.0 C. The snow is soaked with water and >15 contains a volume fraction of air from 20 to 40 %.

using dielectric methods; (iii) examine whether wetness ob- morphism, hydraulic conductivity and water flow (Jordan servations over larger areas would help improve wet snow et al., 2008). The amount of liquid water influences the avalanche forecasting. mechanical properties of snow. Relatively small amounts of liquid water can reduce the strength of snow; Techel et al. (2011) found that the strength of temperature-gradient 2 Background snow (such as facets or depth hoar) decreased significantly at low water contents (θ < 3 vol. %), while Colbeck (1982) 2.1 Liquid water in snow described loss in strength of seasonal snow at approximately 8 vol. %. Liquid water is introduced to snow by rain and/or melt. Melt- ing at the surface depends primarily on the incoming flux of 2.2 Estimation and measurement of liquid water in shortwave radiation, which varies with slope aspect and ele- snow vation. The advance of a wetting front into snow is seldom uniform; even under controlled laboratory conditions it is dif- Liquid water content of snow has been measured by centrifu- ficult to achieve a homogeneous distribution (Brun, 1989). gal separation, melting calorimetry, freezing calorimetry, al- Infiltration usually starts through isolated small “flow fin- cohol calorimetry and the dilution method. These methods, gers” (Marsh, 1988; Schneebeli, 1995; Waldner et al., 2004), which are summarized by Stein et al. (1997), are difficult to which enlarge into channels with increased time and flow perform and time-consuming, which makes them impracti- (Kattelmann and Dozier, 1999). The pattern and timing of cal for operational forecasting. Liquid water content has also infiltration depends on snow structure, temperature, slope an- been estimated by measuring the real and imaginary parts gle and the amount of liquid water entering the snowpack of the permittivity of snow. This measurement is diagnostic 0 ≈ (Conway and Benedict, 1994; Fierz and Fohn¨ , 1994). Grav- of liquid water because the permittivity of water (i 86) is 0 ≈ 0 ≈ itational forces usually dominate infiltration, although layer- much higher than those of air (i 1), and ice (i 3.15) parallel infiltration is often observed above impermeable ice (Frolov and Macharet, 1999; Louge et al., 1998). In this layers, or capillary barriers that consist of fine grains layers study, we compare “hand” measurements with measurements of coarse grains (Wankiewicz, 1979; Jordan, 1994; Waldner made with a “Finnish Snow Fork” (SnF, Sihvola and Tiuri, et al., 2004). Introduction of liquid water into snow leads to 1986; Toikka, 2009) and also with a Denoth meter (Dn, De- rapid changes in grain shape (Brun, 1989; Coleou´ and Lesaf- noth, 1994). fre, 1998), grain coarsening (Raymond and Tusima, 1979; The Snow Fork is a two-pronged wave-guide that oper- Brun, 1989; Marsh, 1987) and an increase in bulk density ates at ∼1 GHz and measures both the real and the imaginary (Colbeck, 1997; Marshall et al., 1999; Jordan et al., 2008). parts of the permittivity simultaneously measuring changes Important feedback mechanisms exist between snow meta- in the resonance curve between air and snow (Fig. 1).

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ 10 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

Tables Figures

10 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 10 cm

Fig. 1. Denoth (left) and Snow Fork (right) measurement devices. Tables FiguresF. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 407

Fig.Fig. 2. 2.ResponseResponse of of water water flow flow in a slope. Shown are are 50 50 cm cm wide wide frontalfrontal views views of of the the pit-wall pit-wall facing facing down-slope in in experiments experiments using using dye-tracers.dye-tracers. The The left left photo photo was taken immediately after after cutting cutting a a pit-wall,pit-wall, the the right right picture picture approximately approximately 30 s s later. later. The The size size of of the the wetwet area area at at the the pit-wall pit-wall increased increased rapidly due due to to water water flowing flowing from from laterallateral flow flow channels channels into into the the pit-wall. pit-wall.

10 cm 3 Scope and aim

To address the three objectives of this study, we designed a Fig.Fig. 1. 1.DenothDenoth (left) (left) and Snow Fork (right) (right) measurement measurement devices. devices. series of measurements to investigate the following specific questions: Insertion of the Snow Fork compresses the surrounding 1. Does the “hand” test provide a realistic point estimate snow, which increases snow density by approximately 1– of water content in snow? 2 %, which has a small effect on the estimate of the real part of the permittivity (Sihvola and Tiuri, 1986). The snow fork 2. How do layer characteristics such as hardness, grain that we used provided an estimate of water content on sam- shape or size, influence the results from the “hand” test? ples of area 7.5×2 cm2. The SnF has been used in a variety 3. Can comparable measurements of snowpack wetness be Fig.of studies 2. Response including of water measuring flow in a the slope. spatial Shown wetness are 50 distribu- cm wide frontaltion (Williams views of the et pit-wall al., 1999 facing), snow down-slope characteristics in experiments in Antarc- using achieved by either measuring before digging a snow pit dye-tracers.tica (Kark¨ as¨ The et left al., photo2005), was and taken the immediately wetness of aftersnow cutting in ski a or at the side-wall of a snow profile? pit-wall,tracks (Moldestad the right picture, 2005 approximately). 30 s later. The size of the 4. Is it necessary to consider small-scale spatial variability wet area at the pit-wall increased rapidly due to water flowing from The Denoth meter is a capacitance probe, which measures in water content when interpreting wetness profiles? lateralthe permittivity flow channels of into snow the of pit-wall. area 13 × 13.5 cm2 (Fig. 1). A separate measurement of density is required to solve for the Based on the answers to these questions we developed a ro- imaginary part of the permittivity (Denoth, 1994), which is bust sampling strategy for field measurements. For practical necessary to estimate liquid water content. The Denoth me- purposes, we introduced a simplified snow wetness classifi- ter has also been used in field studies to monitor changes in cation, which incorporates information on vertical and hori- snowpack wetness during the melt period (Martinec, 1991a; zontal wetness distribution. Kattelmann and Dozier, 1999), and to validate snowpack models (Mitterer et al., 2010). The accuracy of measurements made by dielectric meth- 4 Methods and data ods is ±0.5 vol. % (Sihvola and Tiuri, 1986; Fierz and Fohn¨ , 1994). Additional uncertainty can arise if sensors near Most of the data analyzed was collected in winter and the surface are affected by solar radiation (Lundberg et al., spring 2008/2009 and spring 2009/2010 in Alpine terrain in 2008). These methods are destructive to the snow sample and Switzerland in the Fribourg and Western Bernese Alps and also require the excavation of a snow-pit, which can cause Pre-Alps, the region of Davos and in the Lower Engadin local disturbances to water flow (Fig. 2). More recently, non- (Fig. 3). The majority of these observations were carried destructive measurement methods like the Snow Pack Anal- out in potential avalanche terrain (slopes steeper than 30◦) yser (Mess-Systemtechnik, 2010) and ground-penetrating above tree-line. The data-set includes only one day when radar installed upward-looking at the snow-ground interface observations followed a rain-on-snow event (less than 5 mm have been applied to measure snow wetness (Heilig et al., precipitation fell as rain). 2009). In addition, multispectral imaging from satellites has been used successfully to map regions of dry and wet snow surfaces (e.g. Gupta et al., 2005).

www.the-cryosphere.net/5/405/2011/ The Cryosphere, 5, 405–418, 2011 408 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

Observation areas Table 2. Spacing and extent of measurement lay-outs as shown

N in Fig. 4 (concept according to Bloschl¨ , 1999). x-direction corre- 0 50 km sponds to horizontal distance, y-direction is distance between con- secutive measurements as in Fig. 4a, and z-direction is equivalent Davos to snow depth measured vertically (in cm, Fig. 4b, c). 2009 / 2010, 23 days

Fribourg and West. Bernese mode extent spacing Prealps and Alps x y z

2009, 6 days [cm] [cm] [cm] [cm] profile 40 20 – 5 Lower Engadin

2009, 1 day profile 300 10 – 5 horizontal 40 20 5 – vertical 500 50 – 5

Fig. 3. Map showing Switzerland (grey) and the regions (blue), ble 1), the Finnish Snow Fork was used to obtain a more where measurements were conducted in 2009 and 2010. The num- quantitative measure of liquid water content while the De- ber of days is shown, when measurements were carried out. The noth meter was used to allow a comparison between the two majority of measurements were taken in the region surrounding instruments. Davos. 4.1.1 Sampling design for liquid water content measurements in a natural snowpack with the 4.1 Field methods Snow Fork

All measurements were carried out in seasonal snow. Our Liquid water content was measured in one of the following plan was to sample data that were representative of diverse three ways: topographic, snowpack and wetness conditions. Slope se- lection was dictated by safety concerns, as many observa- – horizontal – These measurements preceded all slope tions were carried out during periods of wet snow avalanche profiles. Three contiguous measurements spaced 20 cm activity. apart (across the slope) with horizontal measurement In 2008/2009, we focused on measuring diurnal changes intervals of 5 cm into the snowpack were conducted in wetness and comparing results from hand tests with di- (Fig. 4a, Table 2). electric measurements. Measurements were carried out in the morning and repeated in the afternoon in the same slope – profile – These measurements accompanied all slope at a distance of approximately 3–5 m. Wetness content θ was profiles. Measurements were undertaken on the side- first measured horizontally (Fig. 4a), these were followed by wall of a manual snow profile. Measurements were the excavation of a snow-pit and measurements made on a slope- and layer-parallel, and generally less than 50 cm shaded side-wall of the snow-pit (Fig. 4b). Lastly, a man- from horizontal measurements. Vertical measurement ual snow profile including measurements of layer thickness, intervals were 5 cm and the distance between the three hardness, wetness, grain shape and size was recorded ac- measurement rows was 20 cm (Fig. 4b, Table 2). cording to observational guidelines (WSL, 2008; Fierz et al., – vertical – These measurements were conducted to de- 2009). Snow temperatures were measured using a calibrated, ◦ termine the spatial variability of water content. The ver- digital thermometer with an accuracy of ±0.5 C (Milwau- tical distance between consecutive measurements was kee Stick Thermometer TH310). Snow stability during this 5 cm, the horizontal distance (across the slope) 50 cm period is discussed in detail by Techel and Pielmeier (2009). (Fig. 4c, Table 2). During the 2009/2010 winter we focused on measuring changes in snow wetness over several days and the distri- Applying the concept from Bloschl¨ (1999), measurement bution of water within the vicinity of a snow profile recorded spacing and extent are shown in Table 2. The support, the in- according to Swiss observational guidelines. θ was usually tegrated volume of a measurement device (Bloschl¨ , 1999), is measured by inserting the Snow Fork vertically into the snow approximately 47 cm3 for the Snow Fork (Moldestad, 2005). in cross-sections of up to 5 m width. These measurements The idea behind horizontal measurements was to minimize were followed by a manual snow profile and a snowpack sta- the possibility of water flow in front of the probe. This strat- bility test. A qualitative estimation of liquid water content egy, together with the fast measuring speed of the Snow Fork (mWC) was part of the standard observation procedure in (it typically took less than 2 min to measure one vertical pro- all manual snow-profiles (WSL, 2008; Fierz et al., 2009, Ta- file consisting of 15 single measurements) minimized this

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 13

A B C

5 50 50 50 50 50

5 5

5

F. TechelF. Techel and and C. Pielmeier:C. Pielmeier: Point Point observations observations of of liquid liquid water water content in in wet wet snow snow 409 13

A B profileC

5 horizontal 50 50 50beside50 manual50 example of a vertical before profile observations5 profile observations cross-section over 2.5 m distance 5

5 Fig. 4. Liquid water content measurements: sampling design for horizontal, profile and vertical measurements. A) horizontal measurements profile inhorizontal 5 cm steps to a depth of 75 cm. B) observations adjacent to manual snow profile, vertical distance 5 cm, slope-parallel distance 20 cm. beside manual example of a vertical C)before verticalprofile observations measurementsprofile observations for spatial variabilitycross-section observations,over 2.5 m distance measurement steps 5 cm, horizontal distance 50 cm. Refer also to Tab. 2 for

Fig.Fig. 4. Liquid 4. Liquidextent water water content and content measurements: spacingmeasurements: sampling samplingof measurements. design design for for horizontal, profile profile and and vertical vertical measurements. measurements.(A) A)horizontal horizontal measurements measurements in 5in cm 5 stepscm steps to a to depth a depth of of75 75 cm. cm. B)(B) observationsobservations adjacent adjacent to to manual manual snow snow profile, vertical vertical distance distance 5 5 cm, cm, slope-parallel slope-parallel distance distance 20 cm. 20 cm. C) vertical(C) vertical measurements measurements for forspatial spatial variability variability observations, observations, measurement measurement steps 5 cm, cm, horizontal horizontal distance distance 50 50 cm. cm. Refer Refer also also to Table to Tab.2 for 2 for extentextent and andspacing spacing of measurements. of measurements.

Cross section 5 m LWC [vol. %]

0 Cross section 5 m LWC [vol. %] 10

0 10

8

−20 8 6 −20 −40 4 depth [cm] 6

−60 2 −40

0 4 −80 depth [cm] 0 100 200 300 400 500 a) b) 100 distance [cm] 12 −60 2 Fierz et al. 2009 10 Fig. 5. Contour plot showing cross-section of snow wetness (θ) to ● 75 Martinec, 1991b a depth of 80 cm over 5 m wide areas across the slope. Measure- 8 0 −80 ●

ments were conducted at horizontal intervals of 50 cm (lines) with [vol.%] ● 0 100 200 3006 ● 400 50 500 ◦ F a vertical spacing of 5 cm. 4 April 2009, S aspect, 2660 m, 30 . n a) b) S 4 θ, measured with the Snow Fork, is corrected by -0.8 vol.%distance (this [cm]θ 100 ● 25 12 2 ● corresponds to the median offset in dry snow). ● ● ● Fig. 5. Contour plot showing cross-section of snow wetness (θ) to a depth of0 80 cm over 5 m wide areas across the slope. Measurements Fierz et al. 2009 correctly estimated mWC [%] 0 ◦ 10 were conducted at horizontal intervals of 50 cm (lines) with a vertical spacing of 5 cm. 4 April 2009, S aspect, 2660 m, 30 . θ, measured Martinec, 1991b with theFig. Snow Fork, 5.a) isContour corrected by −0.8 plot vol. %showing (this correspondsb) cross-section to the median offsetdry in drymoist of snow).wet snowv. wet wetnessdry (moistθ) towet v. wet ● 75 8 1.5 ● mWC mWC 8 a depth of 80 cm over● SnF 5 m wide●● areas across the slope. Measure- ● ● Dn ● ● ● ● ● ● 6 ● 1 ● ● ● ●●● ● ●● ● ● ●●● ● ● potential problem. As shown in Fig. 5, there● was●●● ●●● usually●●●●●● ●●● were sampled directly above and below the Denoth meter ● ● ● ●●● ●●●●● ●

ments were conducted● at horizontal intervals of 50 cm (lines) with [vol.%] ● ● ● ● ● ●● ● ● ●●●●● ● 3 6 ● 50 ● ● ● Fig. 7. a) Box-plot comparing estimated liquid water content ● a clear distinction between● wet to dry layers● indicating● ● ● ● ● ● that measurements using a 100 cm density cutter and a digital

●● ● ● ● ●● ●● ● ● ● ●● ● [vol.%] ● ● ◦

● F 4 ● ● 0.5 ● ● (mWC) and water content measured with the Snow Fork (θSnF , ● ● water running ahead of the SnF[vol.%] was not a problem. scale. F ● n a vertical spacing of 5 cm. 4 April 2009, S aspect, 2660 m, 30 . n ●● ●

● θ ● S ● ● ● ● n=318, 4 observers). As comparison the mean for each class is S θ ● 4 ● ● ●● ● θ 2 ● ● 0 ● ● shown based on the conversion given in Fierz et al. (2009). b) shows θ,●●●●● measured● ● with the Snow Fork, is corrected by -0.8 vol.% (this 4.1.2● Comparison● ●●●● ●● of Denoth wetness meter and 4.2 Data ● ●●●●●●●●● ●●●● ●●●●●●●●● 25 ●●●●●●● ●● the frequency that mWC was correctly estimated (light blue bars). ● Snow●●●●●●●● Fork● device 2 ● 0 corresponds to−0.5 the median offset inFor dry comparison, snow). data by a previous study is shown (black dashed ● Snow wetness was measured using the Snow Fork in more ● line, black squares Martinec, 1991b, n=518, 9 observers). θSnF is ● Adjacent−2 measurements0 2 4 from6 the Denoth100 meter200 (2–3300 mea- correctedthan 80 different by -0.8 vol.% locations (this corresponds in a variety to of the measurement median offset de- in dry 0 ρ signs and wetness conditions in natural snow. correctly estimated mWC [%] 0 surements) wereθDn [vol.%] compared with measurements [kg/m3] from the snow). Snow Fork (2–6 measurements). The measurements were Measurements in 2008/2009 targeted small-scale variabil- always undertaken on a sidewall of snowa) pits and paral- ity (within 40 cm), diurnal changesb) in wetness within the dry dry Fig. 6. Comparison of measured liquid water content using the De- moist wet v. wet moist wet v. wet lel to the layer stratigraphy. Horizontal and vertical dis- same slope and compared hand and dielectric estimates of noth (θDn) and the Snow8 Fork (θSnF ) instruments in a) all wetness 1.5 tances between measurements was 5 cm. Snow densities snow wetness. More than 7000 measurements were taken conditions (n=134) and in b) in dry snow as a function of snow● den- mWC mWC ● SnF ●● sity (ρ). Linear regression models are shown in both plots. The ● ● Dn ● median off-set of the θ for the dry snow data is marked by the ● www.the-cryosphere.net/5/405/2011/SnF ● ● ● The Cryosphere, 5, 405–● 418, 2011 6 3 ● 1 ● ● ● ●●● ● dark-blue dot (ρ=250 kg m , θSnF =0.8 vol.%). ●● ● ● ●●● ● ● ●●● ●●●●●●●●● ●●● ● ● ● ●●● ●●●●●● ● ● ● ● ● ● ●● ● ● ●●●●● ● ● ● Fig. 7. a) Box-plot comparing estimated liquid water content ● ● ● ●●● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● [vol.%] ● ● ● 4 ● ● 0.5 ● ● (mWC) and water content measured with the Snow Fork (θSnF , ● ● [vol.%]

F ●

n ●● ●

● θ ● S ● ● ● n=318, 4 observers). As comparison the mean for each class is θ ●● ● ● ● ● 2 ● ● 0 ● ● shown based on the conversion given in Fierz et al. (2009). b) shows ●●●●●●● ● ● ● ●●●● ●● ● ●●●●●●●●●●●●●● ●●●●●●●●●● ●●●●●●● ●● the frequency that mWC was correctly estimated (light blue bars). ●●●●●●●●● ● 0 −0.5 For comparison, data by a previous study is shown (black dashed line, black squares Martinec, 1991b, n=518, 9 observers). θSnF is −2 0 2 4 6 100 200 300 corrected by -0.8 vol.% (this corresponds to the median offset in dry ρ θDn [vol.%] [kg/m3] snow).

Fig. 6. Comparison of measured liquid water content using the De- noth (θDn) and the Snow Fork (θSnF ) instruments in a) all wetness conditions (n=134) and in b) in dry snow as a function of snow den- sity (ρ). Linear regression models are shown in both plots. The median off-set of the θSnF for the dry snow data is marked by the 3 dark-blue dot (ρ=250 kg m , θSnF =0.8 vol.%). F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 13

A B C

5 50 50 50 50 50

5 5

5

profile horizontal beside manual example of a before profile observations profile observations cross-section over 2.5 m distance vertical

Fig. 4. Liquid water content measurements: sampling design for horizontal, profile and vertical measurements. A) horizontal measurements in 5 cm steps to a depth of 75 cm. B) observations adjacent to manual snow profile, vertical distance 5 cm, slope-parallel distance 20 cm. C) vertical measurements for spatial variability observations, measurement steps 5 cm, horizontal distance 50 cm. Refer also to Tab. 2 for extent and spacing of measurements.

Cross section 5 m LWC [vol. %]

0 10

8 −20

6 −40 4 depth [cm]

−60 2

0 −80 0 100 200 300 400 500 a) b) 100 distance [cm] 12 Fierz et al. 2009 10 Fig. 5. Contour plot showing cross-section of snow wetness (θ) to ● 75 Martinec, 1991b a depth of 80 cm over 5 m wide areas across the slope. Measure- 8 ●

ments were conducted at horizontal intervals of 50 cm (lines) with [vol.%] ● 6 ● 50 ◦ F

a vertical spacing of 5 cm. 4 April 2009, S aspect, 2660 m, 30 . n S 4 θ, measured with the Snow Fork, is corrected by -0.8 vol.% (this θ ● 25 2 ● corresponds to the median offset in dry snow). ● 410 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow ● ● 0 correctly estimated mWC [%] 0 dry dry Table 3. Data overview: shown are the number of days (n ), a) b) moist wet v. wet moist wet v. wet day 8 1.5 ● mWC mWC locations (nloc), measurement depths (ndepth) and single measure- ● SnF ●● ● ● Dn ● ● ments (n) for the various measurement modes (Fig. 4) using the ● ● ● ● 6 ● 1 ● ● ● ●●● ● ●● ● ● ●●● ● ● ●●● ●●●●●●●●● ●●● ● ● ● ●●● ●●●●● ● Snow Fork (SnF), and for the comparison between SnF and Denoth ● ● ● ● ● ●● ● ● ●●●●● ● ● ● Fig. 7. a) Box-plot comparing estimated liquid water content ● ● ● ●●● ● ● ● ● ● ● ●● ●● instrument (Dn). ● ● ● ●● ● [vol.%] ● ● ● 4 ● ● 0.5 ● ● (mWC) and water content measured with the Snow Fork (θSnF , ● ● [vol.%]

F ●

n ●● ●

● θ ● S ● ● ● n=318, 4 observers). As comparison the mean for each class is θ ●● ● ● ● ● 2 ● ● 0 ● ● shown based on the conversion given in Fierz et al. (2009). b) shows ●●●●●● ● n n n ● ● ●●●● ●● mode day depth ● ●●●●●●●●● ●●●● ●●●●●●●●● ●●●●●●● ●● the frequency that mWC was correctly estimated (light blue bars). (n ) ●●●●●●●● ● loc 0 −0.5 For comparison, data by a previous study is shown (black dashed horizontal 24 (60) >1300 >3500 line, black squares Martinec, 1991b, n=518, 9 observers). θSnF is −2 0 2 4 6 100 200 300 profile 26 (63) >1200 >3500 corrected by -0.8 vol.% (this corresponds to the median offset in dry θ ρ vertical 10 (26) 330 >3500 Dn [vol.%] [kg/m3] snow). Dn-SnF comparison 8 (11) 134 251–637 Fig.Fig. 6. 6. ComparisonComparison of of measured measured liquid liquid water water content content using using the the De- De- nothnoth ( (θθDnDn) and the the Snow Snow Fork Fork ( (θθSnFSnF) instruments in in a)(a) allall wetness wetness conditionsconditions (n=134) (n = 134) and and in b) in in(b) dryin snow dry snow as a function as a function of snow of den- snow accompanying manual snow profiles using the horizontal sitydensity (ρ). (ρ Linear). Linear regression regression models models are are shown shown in in both both plots. plots. The The (Fig. 4a) and profile measuring modes (Fig. 4b, Table 3). medianmedian off-set off-set of of the theθθSnFSnF forfor the the dry dry snow snow data data is is marked marked by by the the dark-blue dot (ρ = 250 kg m3 3, θ = 0.8 vol. %). These were recorded in more than 2500 different measure- dark-blue dot (ρ=250 kg m , θSnFSnF=0.8 vol.%). ment depths or layers. In spring 2010, the focus was on investigating the wet- plied in the contour plots (as in Fig. 5) and is described in the ness variability at spatial scales of up to 5 m. Vertical mea- R-package graphics (R, 2009). surements (Fig. 4c) were carried out in 25 locations. The comparison of Snow Fork and Denoth instrument is based on measurements in 134 different snow layers made in pro- 5 Results file mode (Table 3). 5.1 Liquid water content measurements using Snow Fork and Denoth wetness instrument 4.3 Data analysis In a variety of snow wetness situations, the liquid water con- The water content measured with the Snow Fork (θSnF) tent measured with the Snow Fork (θSnF, in vol. %) is gener- ranged from 0 to 23.6 vol. %. According to the Snow Fork ally higher than with the Denoth meter (θDn, Fig. 6a), but the manual (Toikka, 2008), measurements with θSnF > 10 vol. % measurements were strongly correlated: may not be accurate. However, as these values often corre- ≈ × + 2 = ≤ sponded to areas of high snowpack wetness, they were not θSnF 1.06 θDn 1.0 (r 0.78,p 0.001) (1) excluded from analysis but considered as 10 vol. %. The cal- Our results are similar to previous studies (Denoth, 1994; culation of the water content using the Denoth instrument Williams et al., 1999; Frolov and Macharet, 1999). The vari- (θ ) requires a measurement of snow density ρ. We used the Dn ability between adjacent measurements was similar regard- mean of two measurements adjacent to the dielectric mea- less which instrument was used. surement. In dry snow (hand test dry and snow temperature ≤ The relationship between estimated water content (wet- −0.5 ◦C), we investigated the effect of snow density (ρ) on ness index, mWC) and measured liquid water content (θ ) SnF the measured water content. The Snow Fork recorded a used the conversion shown in Table 1. If θ was within SnF median θ = 0.8 vol. % (standard deviation σ = 0.2 vol. %, ±0.5 vol. % of a class limit, this is considered as an interme- SnF n = 487). The Denoth wetness device showed lower val- diate (half) index class and not considered as a false clas- SnF ues θ = 0.1 vol. % (σ = 0.17 vol. %, n = 281). Measured sification or measurement. Snow wetness was often non- Dn Dn water content in dry snow was generally 0.65 vol. % higher normally distributed. Therefore, the median and the in- using the Snow Fork than with the Denoth meter (Fig. 6b). terquartile range are considered as robust measures of cen- In dry snow poor positive correlations between θ and ρ were tral tendency and data distribution. If several measurements observed: are available the median θ is used. Linear regression models were derived for θ and the Pearson coefficient of determina- 2 θSnF = 0.0019ρ +0.32 (r = 0.28,p < 0.001) (2) tion r2 was calculated. For categorical variables (mWC), the Spearman correlation rs was used. Data was tested for sig- nificant differences using non-parametric tests (Mann Whit- 2 ney U-test, Kruskal-Wallis H-test, sign-test, Crawley, 2007), θDn = 0.0019ρ −0.33 (r = 0.11,p < 0.05) (3) the level of significance α ≤ 0.05. Linear interpolation is ap-

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 13

A B C

5 50 50 50 50 50

5 5

5

profile horizontal beside manual example of a before profile observations profile observations cross-section over 2.5 m distance vertical

Fig. 4. Liquid water content measurements: sampling design for horizontal, profile and vertical measurements. A) horizontal measurements in 5 cm steps to a depth of 75 cm. B) observations adjacent to manual snow profile, vertical distance 5 cm, slope-parallel distance 20 cm. C) vertical measurements for spatial variability observations, measurement steps 5 cm, horizontal distance 50 cm. Refer also to Tab. 2 for extent and spacing of measurements.

Cross section 5 m LWC [vol. %]

0 10

8 −20

6 −40 4 depth [cm] F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 411 −60 2

0 −80 5.2 Influence0 100 of sampling200 design300 400 500 a) b) 100 distance [cm] 12 Fierz et al. 2009 Liquid water content measurements made before digging a 10 Fig. 5. Contour plot showing cross-section of snow wetness (θ) to ● 75 Martinec, 1991b snow-pit (horizontal mode, Fig. 4a) and following manual ● a depth of 80 cm over 5 m wide areas across the slope. Measure- 8 snow profile observations (profile mode, Fig. 4b) were com- ●

ments were conducted at horizontal intervals of 50 cm (lines) with [vol.%] ● 6 ● 50 ◦ pared in 86 locations. F

a vertical spacing of 5 cm. 4 April 2009, S aspect, 2660 m, 30 . n S 4 θ,No measured significant withdifference the Snow Fork, between is corrected either bymode -0.8 of vol.% measur- (this θ ● 25 2 ● ingcorresponds the water to content the median was offset noted indry in locations snow). with low water ● ● ● content (median θ < 1.3 vol %, dry or barely moist snow). 0 SnF correctly estimated mWC [%] 0 In moist and wet snow (θSnF ≥ 1.3 vol %), horizontal mea- a) b) dry moist wet v. wet dry moist wet v. wet surements8 were wetter than the measurements1.5 at the side- ● mWC mWC wall following snow pit excavation (p ●=SnF0.03). However,●● ● ● Dn ● ● ● ● ● ● 6 ● θ = . 1 ● ● ● ●●● ● the median difference● (● 0 43 vol. %) is within the● ● range●●● ● ● ●●● ●●●●●●●●● ●●● ● ● ● ●●● ●●●●● ● ● ● ● Fig. 7. (a) Box-plot comparing estimated liquid water content ● ● ●● ● ● ●●●●● ● of measurement uncertainty (±0.5 vol. %, Sihvola● and● Tiuri, Fig. 7. a) Box-plot comparing estimated liquid water content ● ● ● ●●● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● (mWC) and water content measured with the Snow Fork (θ , [vol.%] ● ● ● SnF 4 ● ● 0.5 ● ● (mWC) and water content measured with the Snow Fork (θSnF , ● ● 1986; Fierz and Fohn¨ , 1994). [vol.%] F ●

n ●● ● n = 318, 4 observers). As comparison the mean for each class

● θ ● S ● ● ● n=318, 4 observers). As comparison the mean for each class is Inθ our data-set,●● 2 % of the recorded values were higher ● ● ● ● 2 ● ● 0 is shown based on the conversion given in Fierz et al. (2009). ● ● shown based on the conversion given in Fierz et al. (2009). b) shows ●●●●●● ● than 10 vol.● ● %.●●●● ●● These high values were more frequently ● ●●●●●●●●● ●●●● (b) shows the frequency that mWC was correctly estimated (light ●●●●●●●●● ●●●●●●● ●● the frequency that mWC was correctly estimated (light blue bars). observed when●●●●●●●● we● measured across layer boundaries in an 0 −0.5 blueFor comparison, bars). Fordata comparison, by a previous data by study a previous is shown study (black is dashed shown undisturbed snowpack before digging the snow-pit than in (black dashed line, black squares Martinec, 1991b, n = 518, 9 ob- line, black squares Martinec, 1991b, n=518, 9 observers). θSnF is −2 0 2 4 6 100 200 300 layer-parallel measurements taken at the sidewall of a snow servers).correctedθ bySnF -0.8is vol.% corrected (this by corresponds−0.8 vol.% to the(this median corresponds offset in to dry the ρ median offset in dry snow). profile. In horizontalθDn [vol.%] measurements, θ-values greater [kg/m3] than snow). 10 vol. % occur more frequently in layers relatively close to theFig. snow 6. Comparison surface and of measured when neighboring liquid water measurementscontent using the also De- Hardness of snow is influenced by the liquid water con- showednoth (θDn high) and values. the Snow Fork (θSnF ) instruments in a) all wetness conditions (n=134) and in b) in dry snow as a function of snow den- tent. This can be observed by comparing the hardness (hand 5.3sity ( Qualitativeρ). Linear regression snow profile models observations are shown in in both wet plots. snow The hardness test, Fierz et al., 2009) with estimated and mea- median off-set of the θSnF for the dry snow data is marked by the sured water content. A negative correlation exists in layers 3 Manuallydark-blue dot estimated (ρ=250 kg water m , θ contentSnF =0.8 (mWC)vol.%). and liquid wa- classified as MF (rs = −0.70 for mWC, rs = −0.43 for θSnF, ≤ ter content (θSnF) were compared in 314 layers. mWC and p 0.001). Not surprisingly, the transition from dry (frozen) θSnF were strongly correlated (rs = 0.73, p ≤ 0.001). Cor- to wet infers a significant hardness decrease. While no clear rectly estimated water content mWC decreased with increas- trend was observed for TG snow, all layers estimated as be- = ing θSnF (Fig. 7a, b). These results are similar to an earlier ing wet (n 6), or measured as being wet (θSnF > 3 vol. %, study (Martinec, 1991b). Both, Martinec (1991b) and our n = 9), had a hand hardness of 1. This was significantly results, show that dry layers are normally well described, softer than the dry hand hardness (p ≤ 0.01) in TG snow. while wetter layers are often incorrectly classified (30 % in our study for wet layers). It is of note, however, that few 5.4 Temporal changes in snow wetness: morning vs. afternoon measurements were done in snow with θSnF > 8 vol. %. Very few layers were estimated as being very wet. In these layers, Snowpack wetness was compared in 33 locations between mWC was always overestimated. morning and afternoon. Generally, snow wetness increased The wetness in layers consisting of coarse melt-freeze- within the uppermost 10 cm of the snowpack (p < 0.05, particles (MF, snow class MF, Fierz et al., 2009) is more Fig. 8a). This is not surprising, as measurements were car- frequently falsely estimated (33 % of cases) than in layers ried out when day-time warming was expected. Analyz- consisting of fine precipitation particles and snow which has ing each location individually, the median snowpack wet- undergone low-temperature gradient metamorphism (LTG, ness (excluding the uppermost 10 cm) changed significantly snow classes PP, DF, RG, 13 %) or coarse medium to in only nine cases (p < 0.05). Surprisingly, six of these nine high temperature gradient metamorphosed grains (TG, snow wetness profiles showed decreasing values (Fig. 8b). These classes FC, DH, 13 %). No significant correlation was ob- significant changes always involved a transition from dry or served between snow hardness and the correct estimation of barely moist snow (median θ m ≤ 1.3 vol. %, third quartile the water content. The wetness range in MF is much greater SnF q3 ≤ m ≥ than in LTG or TG layers (MF: 0–10 vol. %; LTG/TG: 0– θSnF 1.9 vol. %) to moist or wet snow (θSnF 1.2 vol. %, q3 ≥ 5.5 vol. %). θSnF > 3 vol. % is more frequently observed in θSnF 1.9 vol. %) or vice versa (Fig. 8c), where both the MF (>20 % of cases) and rarely in LTG/TG (<4 % cases). changes in median and third quartile are significant (p ≤ LTG layers estimated as being wet were almost always over- 0.001). estimated. The error rate in snow estimated as being wet is smallest in TG snow (20 %). www.the-cryosphere.net/5/405/2011/ The Cryosphere, 5, 405–418, 2011 14 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

a) b) c) 5 5 5

20 20 20

40 40 40

60 60 60 θ

depth [cm] depth [cm] depth [cm] a) 18 March − S, 32°, 2350 m [vol.%] 0 10

80 80 80 −10 8 −20 100 100 100 −30 6 −2 0 2 4 6 0 2 4 6 8 10 0 2 4 6 8 10 −40 4 depth [cm] ∆θSnF [vol.%] θSnF [vol.%] θSnF [vol.%] −50 2 −60 Fig. 8. a) Difference between morning and afternoon liquid water −70 0 0 100 200 300 400 500

content (θSnF ), measured in 33 locations. Positive values indicate distance [cm] b) 19 March − S, 32°, 2300 m θ [vol.%] an increase in θSnF from morning to afternoon. Depth is given in 0 8 cm below snow surface. The median is the bold line, the shaded area −10

represents the interquartile range. b) and c) example of profiles with −20 6 significant changes in θSnF during the day. The bold line represents −30

−40 4

the morning measurements (median of 3), the shaded area and the depth [cm] arrows indicate the change during the day. −50 14412 F. F. Techel Techel and and C. C. Pielmeier: Pielmeier: Point Point observations observations of liquidof liquid water water content content in wet in wet snow snow 2 −60

−70 0 a) b) c) a) 0 100 200 300 400 500 5 5 5 1 distance [cm] r

20 20 20 t θ

n c) 19 March − SW, 33°, 2350 m [vol.%]

e 0 8 i 0.9 c i

f ●

f ● ● −10 40 40 40 ● ● e ● ● ● ● o ● 6

c −20

0.8

n ● o

60 60 60 i ● −30 t θ 4 a

depth [cm] depth [cm] depth [cm] a) 18 March − S, 32°, 2350 m [vol.%] 0 l 10 −40 e 0.7 ● depth [cm] r 80 80 80 r −10 o −50 c 8 2 −20 0.6 −60 100 100 100 6 −30 −70 0 0 100 200 300 400 500 −2 0 2 4 6 0 2 4 6 8 10 0 2 4 6 8 10 0 100 200 300 400 500 −40 4 depth [cm] ∆θSnF [vol.%] θSnF [vol.%] θSnF [vol.%] −50 lag(x) [cm] distance [cm]

● 2 −60 b) c) d) 20 March − SSE, 27°, 2290 m θ [vol.%]

2 3 ● 0 10 ● ● Fig.Fig. 8. 8.a) (a) DifferenceDifference between between morning morning and and afternoon afternoon liquid liquid water water −70 ● 0 0 100 200 300 400 500 ● ● −10 ● ● ● 8 ● ● ● ● ● ● ● ● ● content (θ ), measured in 33 locations. Positive values indicate ● content (θSnF ), measured in 33 locations. Positive values indicate ● ● ● −20 1.5 ● ● SnF ● ● distance [cm] ● ● ● ● ● x ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● θ ● ● ● ● ● ●● ● ● ● ● [vol.%] ● ● 6 b) 19 March − S, 32°, 2300 m ● ● ● ●● ● ● ● ● ● ● ●● ●● ● ● ● θ ●● ● ● ● ● ● an increase in from morning to afternoon. Depth is given in cm ● ● −30 θ ● ●● ● an increase in from morning to afternoon. Depth is given in ● ● ● SnF ● ●● ● ● ● ● ●● ● ● ● ● 0 ● ● ● ● ● ● ● ● ● ● SnF ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ● ●●●● ● ●● ● ● ● ● ● ●● ●● ●●● ● ●● ● ● ● ● ● ● ● ● ● ●●●●● ● ● ●● ● ● ● ● ● ●●● ●●●●●●● ●●● ● ● ● ● ● ●● ●● ●●● ● ● ● ● ● ● ● ●● ●●●●●● ● ● ● ● ● ●●● ●● ● ●● ● ● ●●●●●●●●●● ●● ● [vol.%] [vol.%] ●●● ● ●●● ● ● ● ● ● ● ●●●● ● ● ● ●●●●●●●●●●●●●●●●● ● ● ● ●●●● ● ● ●●● ● ●● ● ● ● ●●●●●● ●●●●●●● ●●●● ● ● ● ● ●● ●●●● ●● ● ● ● ● ● ● ● ● ●●●●●●●●● ● ● ● ● ● 8 ● ●●●●● ● ● ●● ●●●●●●●● ● ● ● ● ● ● 1 0 ● ● ● ● ● ● ●● ●● ● ● ● −40 ●●●● ● ● ● ● ● ●● ● ● ● ● below snow surface. The median is the bold line, the shaded area ● ● ● ● ● ● cm below snow surface. The median is the bold line, the shaded area ● ● −10 ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● 4 ● ● ● depth [cm] F F ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● n n ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● S S ● ● ● ● ● ● ● ● −50 represents the interquartile range. b) and c) example of profiles with −20 ● ● ● represents the interquartile range. (b) and (c) example of profiles θ θ ● ● ● 6 ● ● ● ●

∆ ∆ 2 ● ● ●

● ● 0.5 ● ● −60 significant changes in θ during the day. The bold line represents −30 ● SnF θ ● with significant changes in during the day. The bold line rep- ● SnF ● ● ● −40 4 −70 0

●

the morning measurements (median of 3), the shaded area and the depth [cm] resents the morning measurements (median of 3), the shaded area ● ● 0 100 200 300 400 500 −50 0 −3 arrowsand the indicate arrows the indicate change the during change the during day. the day. 2 distance [cm] −60 0 100 200 300 400 500 0 2 4 6 8 10 −70 0 lag(x) [cm] Median θ [vol.%] a) 0 100 200 300 400 500SnF ● 1 distance [cm] Fig. 10. State of snow wetness, 18 - 20 March 2010 on southerly

r We have only one explanation for the profiles where over-

t θ aspect slopes at similar altitudes in the beginning of the melt phase. alln water content decreased during the day: the snowpack Fig.Fig. 9. 9. Variabilityc) in 19 measuredMarch − SW, 33°, water 2350 m content as a function [vol.%] of lag- e 0 Variability in measured water content as a function of8 lag- i 0.9

c Contour plots showing cross-sections of snow wetness (θ) to a depth i

f ● distance between measurements and water content. (a) Pearson cor- f ● ● −10 was in the initial● ● part of the melt-phase with spatially and distance between measurements and water content. a) Pearson cor- e ● ● ● ● o ● 6 of up to 70 cm over 5 m wide areas across the slope. All observa-

c −20relation coefficient r for all measurement pairs with the same lag temporally 0.8 highly variable water infiltration patterns where relation coefficient r for all measurement pairs with the same lag n ● tions were observed in shallow snowpack areas at similar elevations o

i ● −30distance lag(x). (b) Difference in liquid water content (1θ ) be- t SnF distance lag(x). b) Difference in liquid water content (∆θSnF4 ) be- botha dry and wet regions exist within the snowpack. In such l −40 (2000 - 2300 m) in southerly aspect slopes (SE, S, SW). θ, measured e 0.7 ●

depth [cm] tween measurement pairs at lag distance (x). Compared are the me- r tween measurement pairs at lag distance (x). Compared are the me- r

situationso it is unclear if morning or afternoon measurements −50 with the Snow Fork, is corrected by -0.8 vol.% (this corresponds to c diandian (bold (bold line) line) and and the the range range between between median median and and the the third third quartile quartile2 were0.6 a better representation of snowpack wetness at the time. −60 the median offset in dry snow). θ-values greater than 10 vol.% are forfor each each lag lag distance distance (shaded (shaded area). area). The The data-set data-set is is split split into into mea- mea- −70 0 Considering0 100 just200 the300 six400 observations,500 where snowpack wet- θ shown as 10 vol.%. surementssurements0 where where100 the the median200 θ ofof300 the the measurements measurements400 taken500 taken over over 5 5m ness decreased during the day, almost 20 % of the measure- lag(x) [cm] mwas was less less than than 1.3 1.3 vol. vol. % (lowerdistance % (lower [cm] values, values, red) and red) more and morethan 1.3 than vol. 1.3 % ments may be a poor representation of snowpack● wetness. b) c) vol.(higher % (higher values, values, blue).d) 20 (c)March blue).Difference − SSE, c) 27°, Difference 2290 m between between median median waterθ [vol.%] content water

2 3 ● 0 10 ● ● ● contentand the and first the and first third and quartile third quartile (always (always measured measured within within 5 m dis- 5 ● ● −10 ● ● ● 8 ● ● ● ● tance, 1θ ). The shaded area highlights the interquartile-range ● ● ● ● 5.5 Spatial variability in water content● distribution SnF ● m distance, ∆θ ). The shaded area highlights the interquartile- ● SnF ● ● −20 1.5 ● ● ● ● ● ● ● ● ● x ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● averaged over 25 measurements. 6 ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ●● ● ● ● −30 ● ●● ● range averaged over 25 measurements. ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ● ●●●● ● ●● ● ● ● ● ● ●● ●● ●●● ● ●● ● ● ● ● ● ● ● ● ● ●●●●● ● ● ●● ● ● ● ● ● ●●● ●●●●●●● ●●● ● ● ● ● ● ●● ●● ●●● ● ● ● ● ● ● ● ●● ●●●●●● ● ● ● ● ● ●●● ●● ● ●● ● ● ●●●●●●●●●● ●● ● [vol.%] [vol.%] ●●● ● ●●● ● ● ● ● ● ● ●●●● ● ● ● ●●●●●●●●●● ●●● ● ●●●●●●●●●● ● ● ● ● ●● ● ● ●●●●● ●●●●●●● ●● ● ● ● ● ● ● ● ●●●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●● ● ● ● ● ●●●●●● ● ● ● ● ●● ●●●●●●●● ● ● ● ● ● ● 1 0 ● ● ● ● ● ● ●● ●● ● ● ● −40 ● ● ●●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● 4 ● θ ● depth [cm] ●

TheF variability of liquid water contentF was investigated ● ● ● ● ● ● ● ● ● ● ● ● SnF ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● n n ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● S S ● ● ● ● ● ● ● ● −50 ● ● ●

θ θ ● ● ● ● at horizontal distances of 10–500 cm. The difference● ● ● be-

∆ ∆ 2 ● ● ●

● ● 0.5 ● ● −60 ●

● ● majority of measurements being within 0.5 vol. % (Table 4, ● tween θSnF● at measurement spacings of 10 and● 20 cm was −70 0

● ●

● Fig. 10a, b). In more than 90 % of the cases the measure- significantly0 less than at 40 cm or−3 greater (p < 0.001). In 0 100 200 300 400 500

ment was within the samedistance [cm] wetness class (as defined in Ta- 50 % of0 the100 cases200 variability300 400 500 was similar0 2 to, or4 less6 than8 10 the ± ble 1) (Fig. 10c). The median water content observed in three measurementlag(x) accuracy [cm] (θ 0.5 vol. %).Median However, θ [vol.%] even at SnF ● measurements θm3 at regular intervals of 50 cm, 100 cm or relatively small horizontal spacings of 20 cm, 20 % of the Fig. 10. State of snow wetness, 18 - 20 March 2010 on2 southerly aspect200 slopes cm was at similar very strongly altitudes correlatedin the beginning to θ5 of m the(r melt> 0 phase..93). θm3 Fig.measurements 9. Variability differed in measured by morewater content than 1 asvol. a function % and 10of lag- % of Contourwas in plots more showing than cross-sections70 % of the cases of snow within wetness 0.5 (θ vol.) to a % depth of θ5 m distancethe measurements between measurements differed by and more water than content. 1.8 a) vol. Pearson %. Gener- cor- ofand up to in 70 more cm over than 5 m 98 wide % within areas across±0.5 the mWC-classes. slope. All observa- The in- relationally the coefficient correlation r for between all measurement measurements pairs with at the various same mea- lag tionsterquartile were observed range in within shallow 5 snowpack m sections areas is at generally similar elevations very well distancesurement lag(x). distances b) Difference was moderate in liquid water to strong content (Fig. (∆θ 9a).) be- Sig- SnF (2000 - 2300 m) in southerly aspect slopes (SE, S, SW). θ, measured tweennificant measurement differences pairs between at lag distance the median (x). Comparedθ in are each the me-verti- represented by the minima and maxima of three values (95 % SnF with the Snow Fork, is corrected by -0.8 vol.% (this corresponds to dian (bold line) and the range between median and the third quartile within ±0.5 mWC-class, Fig. 10d–f shows the results for the cal column existed in five of the twenty-five 500 cm-wide the median offset in dry snow). θ-values greater than 10 vol.% are for each lag distance (shaded area). The data-set is split into mea- measurements at 100 cm distance). With an increase in mea- grids (p ≤ 0.05). Variability in θSnF increased marginally shown as 10 vol.%. surementswith greater where measurement the median θ spacingsof the measurements (Fig. 9b). takenThe variabil- over 5 surement spacing, the devation between θ5 m and θm3 was m was less than 1.3 vol. % (lower values, red) and more than 1.3 ity (expressed as the interquartile range) as a function to smaller and the minima and maxima of the three measure- vol. % (higher values, blue). c) Difference between median water ments tend to fall outside the interquartile range of 5 m cross- contentthe median and the water first contentand third within quartile 5 (always m distance measured (θ5 m within) is much 5 sections. The recorded snow wetness differed less from θ5 m msmaller distance, at ∆θ5θ m <).1. The3 vol. shaded % than area at highlightsθ5 m ≥ 1 the.3 vol. interquartile- % (±0.16 SnF − if θ was used rather than just one measurement (p < 0.05). rangeand ± averaged0.52 vol. over %, 25 respectively, measurements.p < 10 16, Fig. 9b, c). m3 Randomly selecting one single measurement showed 2 strong correlations to the θ5 m (r > 0.83) with the large

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ 14 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

a) b) c) 5 5 5 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 15 20 20 20

F. Techel40 and C. Pielmeier:40 Point observations40 of liquid water content in wet snow 413

60 60 60 θ

depth [cm] a) one measurement depth [cm] b) depth [cm] c) a) 18 March − S, 32°, 2350 m [vol.%] ● ●● ● ● ● 0 10 10 ● ● 200 200 ● ● ● ● 80 ● ● ● 80 80 ● ● ●●● −10 8 ● ●● ● ● ● ●●● ● ● ● ● ●● 8 ● ●●● ● ● ●● ●●● ● ● ● −20 ● ● ●● 6 ● ● ● ● ● ● ● ● ●● ● ●● 100 ●● ●● ● ● 100 100 [vol.%] ● ● ●● ● 100 100 6 ● ● ●● −30 ●● ● θ ● ●●●● [vol.%] F 4 ● ● ●● ● a) 20 March − S, 28°, 2190 m ● ●●● ● n ● ● ● ● ● 0 10 ●● ● S ● ●●● ●● ● ● ● Frequency 0 2 4 6 8 10 Frequency 0 2 4 6 8 10 −2 ● 0● ●● ●●2 ● 4 6 −40 θ ● ● ● ● ● ● 4 ●●● ●● depth [cm] ●●● ● ● ● 2 ●●●●●●● ●● ●●●●●●●●●●●● ● −10 ●●●●●●●●●●● ● ●●●●●●●∆●●θ● ● θ θ ●●●●● ● [vol.%] [vol.%] [vol.%] −50 8 ●●●●● SnF SnF SnF −20 0 0 0 2 −60 −30 6 0 2 4 6 8 10 −4 −2 0 2 4 −2 −1 0 1 2 −70 0 Fig. 8. a) Differenceθ between morning∆θ and afternoon liquid water −40 4 5m [vol.%] SnF [vol.%] ∆ mWC class 0 100 200 300 400 500 depth [cm] d) three mesaurements −50 content (θSnF ), measured in 33 locations.e) Positive valuesf) indicate distance [cm] ground 10 ● 2 ● θ ● b) 19 March − S, 32°, 2300 m [vol.%] −60 an increase in θ ● ●●● from morning to afternoon. Depth is given in SnF● ●● ● ●● 200 200 0 ground 8 ●●● ● ● ●●●● ● ●●●● ● ● −70 0 ●● ●● ● 8 cm below snow● ● surface.●●● The median is the bold line, the shaded area ● ● ● −10 ● ●●●● 6 ● ●●● 0 100 200 300 400 ●● ● ●●●● ● ●●●● ● ●● ● represents[vol.%] the●●● interquartile range. b) and c) example of profiles with −20 ● ● ●●●●● ●● distance [cm] 4 ● ● ● 100 100 6 ● ● ●● ● ●● ●●●

m3 ●●● ●●●●● θ ● ● Frequency Frequency −30 [vol.%] significantθ ●● changes● ●●● ● in θ during the day. The bold line represents b) 24 March − SSE, 28°, 2330 m ●● ● SnF ●●●● ● ●● 0 10 2 ●●●●●● ● ●●●●●●●● ●●●●●●●●●● ● ●●●●●●●●●● 4 ●●●●●●●● ● −40 ●●

the morning● measurements (median of 3), the shaded area and the depth [cm] −10 0 0 0 8 −50 −20 arrows indicate the change during the day. 2 0 2 4 6 8 10 −4 −2 0 2 4 −2 −1 0 1 2 −60 6 −30 θ [vol.%] ∆θ [vol.%] ∆ 5m SnF mWC class −70 0 −40 4 depth [cm] a) 0 100 200 300 400 500 −50 1 distance [cm] 2 −60 Fig.Fig.r 12. 10. ComparisonComparison between between measurement measurement samples samples and and the the me- me-

t θ

n [vol.%] diandian of of a a 5 5 m cross-section cross-section at at a a certain certain depth depthθθ5m.. Upper Upper row row plots plots c) 19 March − SW, 33°, 2350 m −70 0 e 5 m 0 8 i 0.9 c 0 100 200 300 400 500 i f ●

f ● ● −10 (a-c)(a–c) showshow the the● comparison comparison● for for one one randomly randomly selected selected measurement measurement e ● ● ● ● distance [cm] o ● 6

toto c θθ5m,, lower lower row row plots plots (d-f)(d–f) comparecompare the the median median of of three three measure- measure- −20 0.85 m θ n ● c) 03 April − S, 20°, 2050 m [vol.%] o mentsmentsi ( (θθm3)) observed observed at at 1 1 m horizontal horizontal● distance distance to to θθ5m.. Scatter- Scatter- −30 0 7 t m3 5 m 4 a l −40 plotsplotse 0.7 show show● the the absolute absolute values values (a,(a, d), d), the the histograms histograms the the difference difference

−10 6 depth [cm] r r

o −50 −20 5 betweenbetweenc θθ-values-values ( (∆1θθ,, plots plots b(b) andand e)(e) and) andthe difference the difference in wetness in wet- 2 0.6 −60 −30 4 classesness classes (∆ mWC) (1mWC) using using the international the international classification classification (Fierz et (Fierz al., −40 3 2009, plots c, f). −70 0 depth [cm] et al., 20090 ,100 plots200(c, f)300). 400 500 0 100 200 300 400 500 −50 2 lag(x) [cm] distance [cm]

−60 1 ● θ −70 0 b) c) d) 20 March − SSE, 27°, 2290 m [vol.%] 2 3 ● 0 10 Table 4. Deviation between median measured water content● in 5 m ● 0 100 200 300 400 500 ●

● ● −10

θ θ● ● ● distance [cm] cross-sections 5 m: if one measurement is done 1 and to median 8 ● ● ● ● ● ● ● ● ● ● ● ● ● −20 1.5 ● ● ● ● ● ● ● ● ● x ● ● ● ● ● θ ● ● water content if three measurements are done at regular● hori- ● ● ● ● ● ● ● ● m3 ● ● ● ● ● ● ● θ ● ● ● [vol.%] ● ● ● d) 17 April − S, 30°, 2210 m ● ● ● ● ● ●● ● ● ● ● ● ● ● ● 6 ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ●● ● ● ● −30 0 10 ● ●● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ● ●●●● ● ●● ● ● ● ● ● zontal spacings of 50–200 cm. ●● ●● ●●● ● ●● ● ● ● ● ● ● ● ● ● ●●●●● ● ● ●● ● ● ● ● ● ●●● ●●●●●●● ●●● ● ● ● ● ● ●● ●● ●●● ● ● ● ● ● ● ● ●● ●●●●●● ● ● ● ● ● ●●● ●● ● ●● ● ● ●●●●●●●●●● ●● ● [vol.%] [vol.%] ●●● ● ●●● ● ● ● ● ● ● ●●●● ● ● ● ●●●●●●●●●● ●●● ● ●●●●●●●●●● ● ● ● ● ●● ● ● ●●●●● ●●●●●●● ●● ● ● ● ● ● ● ● ●●●●●●●● ●● ● ● ● ● ● ● ● ● ● ● ● ●●●●●●●● ●● ● ● ● ● ●●●●●● ● ● ● ● ●● ●●●●●●●● ● ● ● ● ● ● 1 0 ● ● ● ● ● ● ●● ●● ● ● ● −40 ●●●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● −10 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● 4 ● ● ● depth [cm] F F ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 8 ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 1n - 2- 3- n 4- ● ● 5- ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● S S ● ● ● ● ● ● ● ● −50 ● ● ● −20 θ dry upper part wet transitionalθ fully wet lower● part wet ● ● ● ● ● ●

∆ ∆ 2 6 ● ● ● deviation from θ θ θ –50 cm θ –200● ● cm 0.5 ● 5 m 1 m3 m3 ● −60 ●

● ● ●

−30 ● ● 4 −70 0 depth [cm] ● ●

>0.5 vol. % 33 % 25 % 15 % ● −40 0 −3 0 100 200 300 400 500 2 −50 >1 vol. % 18 % 10 % 5 % distance [cm] ground 0 100 200 300 400 500 0 2 4 6 8 10 ground >2 vol. % 8 % 2 % 2 % −60 0 0 100 200 300 400 500 lag(x) [cm] Median θ [vol.%] SnF ● distance [cm] Fig. 11. 10.StateState of of snow snow wetness, wetness, 18–20 18 - March 20 March 2010 2010 in the on beginning southerly aspect slopes at similar altitudes in the beginning of the melt phase. Fig. 9. Variability in measured water content as a function of lag- of the melt phase. Contour plots showing cross-sections of snow 5.6 Temporal evolution of snowpack wetness wetnessContour ( plotsθ) to showinga depth of cross-sections up to 70 cm over of snow 5 m wide wetness areas (θ across) to a depth the Fig. 11. Evolution of snow wetness during melting from 20 March distance between measurements and water content. a) Pearson cor- slope.of up to All 70 observations cm over 5 mwere wide observed areas across in shallow the slope. snowpack All areasobserva- at till 17 April 2010 on southerly aspect slopes at similar altitudes. relation coefficient r for all measurement pairs with the same lag The evolution of snowpack wetness during spring 2010 for similartions were elevations observed (2000–2300 in shallow m) snowpack in southerly areas aspect at similar slopes elevations (SE, S, Contour plots showing cross-sections of snow wetness (θ) to a depth distance lag(x). b) Difference in liquid water content (∆θ ) be- southerly aspect slopes above tree-line is shown inSnF Fig. 11 SW).(2000θ -, 2300 measured m) in with southerly the Snow aspect Fork, slopes is corrected (SE, S, SW). by −θ0.8, measured vol. % of up to 70 cm over 5 m wide areas across the slope. All observa- tween measurement pairs at lag distance (x). Compared are the me- dry mixed wet dry mixed wet dry mixed wet dry mixed wet dry mixed wet (thiswith corresponds the Snow Fork, to the is median corrected offset by -0.8 in dry vol.% snow). (thisθ-values corresponds greater to tions were observed in shallow snowpack areas at similar elevations dianand (bold Fig. line)12. and The the snowpack range between was median shallow and with the snowthird quartile depth thanthe median 10 vol. % offset are shown in dry snow).as 10 vol.θ-values %. greater than 10 vol.% are (2000 - 2300 m) in southerly aspect slopes (SE, S, SW). θ, measured foroften each less lag thandistance 1 m (shaded and dominated area). The by data-set soft, is coarse-grained split into mea- Fig. 13. Wetness classification incorporating vertical and horizontal shown as 10 vol.%. with the Snow Fork, is corrected by -0.8 vol.% (this corresponds to faceted layers due to the relativelyθ dry winter with sustained surementswater content where distribution. the median The x-directionof the measurements shows the percentage taken over of 5 the median offset in dry snow). θ-values greater than 10 vol.% are cold periods. The snowpack was predominantly dry and cold mthe was snow less which than has 1.3 been vol. wetted, % (lower where values, dry/wet red) is and fully more dry orthan fully 1.3 shown as 10 vol.%. with snow temperatures mostly below −1 ◦C on 18 March wide-spread wet snow avalanching in the region of Davos vol.wet and % (higher mixed consists values, ofblue). both c) dry Difference and wet regions. between The median y-direction water contentshows(Fig. 11 the anda). vertical the Water first distribution and infiltration third quartileof snow was (always wetness. slow the measured Diurnal following changes within day 5 (Fig. 3). A cold period with new snow (Fig. 12c) was fol- moccurtoo distance, (Fig. mostly11∆b,θ withinSnF c).). the On The upper-most 20 shaded March area 10 two highlights - 15 grids cm wereof the the interquartile- measured: snowpack lowed by further melting (Fig. 12d). Subsequent avalanch- rangeandthe are first averaged not at considered higher over 25 elevation in measurements. this classification. showed a relatively dry snow- ing from southerly aspect start zones was minor and related pack with first weak flow channels (Fig. 11d), while the sec- to shallow failures of the surface snow. ond, measured later in the afternoon and at lower elevation, was already moist to the ground (Fig. 12a). Four days later, on 24 March the snowpack was moist or wet throughout (Fig. 12b). This first wetting of the snowpack coincided with

www.the-cryosphere.net/5/405/2011/ The Cryosphere, 5, 405–418, 2011 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 15

a) one measurement b) c) ● ●● ● ● ● 10 ● ● 200 200 ● ● ● ● ● ● ● 414 F. Techel and C. Pielmeier: Point observations● ● ●●● of liquid water content in wet snow 8 ● ●● ● ● ● ●●● ● ● ● ● ●● ● ●●● ● ● ●● ●●● ● ● ● ● ● ●● 6 ● ● ● ● ● ● ● ● ●● ● ●● ●● ●● ● ● [vol.%] ● ● ●● ● 100 100 ● ● ●● ●● ● θ ● ●●●● [vol.%] F 4 ● ● ●● ● a) 20 March − S, 28°, 2190 m ● ●●● ● matesn of water● ● content.● We found that one particularly useful ● ● 0 10 ●● ● S ● ●●● ●● ● ● ● Frequency Frequency ● ● ●● ●● ● θ ● ● ● ● ● ● ●●●● ●● ● 2 ●●●●● ● ● ●●●●●●●●●●●● −10 feature of●●●●●●●●● the● ● Snow● Fork is its long arm allowing measure- ●●●●●●●● ● ●●●●●●●●● ● ● ●●●●● ● 8 ●●●●● −20 ments0 at depth without previously0 disturbing the0 water flow −30 6 by digging0 2 a4 snow-pit.6 8 10 Although−4 −2 the0 Denoth2 4 instrument−2 −1 0 sam-1 2 θ ∆θ −40 4 5m [vol.%] SnF [vol.%] ∆ mWC class depth [cm] ples a 7.5 times larger measurement volume than the Snow −50 d) three mesaurements e) f) ground 10 ● 2 Fork, we did not note● significant effects on the variability of −60 ● ●●● ● ●● ● ●● 200 200 ground 8 ●●● ● ● ● ●●● ●● adjacent measurements.● ● ● Our measurement methods do not ● ● −70 0 ● ●● ●● ● ● ●●●● ●●●● ● ●● ●●●● 6 ●●● ● 0 100 200 300 400 ● ●●●●● allow a conclusive● ●●●● interpretation on the accuracy of the mea- ● ● [vol.%] ● ●●● ●●●●● ●● distance [cm] 4 ● ● ● 100 100 ● ● ●● ● ●● ●●●

m3 ●● sured water●●●●●● content. ● θ ● ● Frequency Frequency b) 24 March − SSE, 28°, 2330 m [vol.%] θ ●●●●● ●●●●● ●●●● ● ●● 0 10 2 ●●●●●● ● ●●●●●●●● ●●●●●●●●●● ● ●●●●●●●●●● We have●●●●●●●● ● not investigated the effect impurities, such as soot ●●● −10 0 0 0 8 or dust, or acid components have on the permittivity of the −20 0 2 4 6 8 10 −4 −2 0 2 4 −2 −1 0 1 2 6 −30 ice. We areθ aware, however, that∆θ such impurities exist and 5m [vol.%] SnF [vol.%] ∆ mWC class −40 4 that these also influence permittivity measurements (e.g. Fu- depth [cm] −50 jita et al., 1992). As our observations were conducted away 2 Fig. 12. Comparison between measurement samples and the me- −60 from industrial areas we can only suspect that these effects dian of a 5 m cross-section at a certain depth θ . Upper row plots −70 0 will be relatively small. In the Swiss Alps, visual5m contamina- 0 100 200 300 400 500 (a-c) show the comparison for one randomly selected measurement distance [cm] tion of the snowpack is most frequently observed following to θ5m, lower row plots (d-f) compare the median of three measure- c) 03 April − S, 20°, 2050 m θ [vol.%] strong southerly airstreams, when dust particles originating 0 7 ments (θm3) observed at 1 m horizontal distance to θ5m. Scatter- from the Sahara desert may color the snow surface reddish, −10 6 plots show the absolute values (a, d), the histograms the difference

−20 5 orbetween whenθ snow-values algae (∆θ, is plots present. b and e) However, and the difference during the in wetness mea-

−30 4 surementclasses (∆ campaignmWC) using we thecould international neither observe classification snow (Fierzalgae etnor al., −40 3 2009, plots c, f). depth [cm] dust layers in the snowpack. −50 2 The water content distribution or the detection of lateral −60 1 flow patterns in a snowpack may be best measured using the −70 0 0 100 200 300 400 500 vertical measurement mode (Fig. 4c). It should be considered distance [cm] that the water content is measured over the length of the sen-

d) 17 April − S, 30°, 2210 m θ [vol.%] 0 10 sor (75 mm) and thinner layer specific observations are not

−10 possible. The horizontal or vertical wetness measuring cause 8 1- 2- 3- 4- 5- −20 onlydry small disturbancesupper part wet in thetransitional snowpack;fully yetwet they arelower stillpart wet 6 −30 destructive to the snow sample. Truly non-destructive meth- 4 depth [cm] −40 ods like the ground penetrating radar looking upward from

2 the snow-soil interface (GPR, Heilig et al., 2009) might be −50 ground ground −60 0 better suited to monitor changes in snow wetness in the same 0 100 200 300 400 500 location. distance [cm] The estimation of liquid water content in the field is dif- ficult, even for experienced observers. When estimating the Fig.Fig. 12. 11.EvolutionEvolution of of snow snow wetness wetness during during melting melting from from 20 20 March March wetness of snow layers by hand test, it is important to observe tilltill 17 17 April April 2010. 2010 onContour southerly plots aspect showing slopes cross-sections at similar ofaltitudes. snow stratigraphic layers always on a shaded side-wall of a snow- wetness (θ) to a depth of up to 70 cm over 5 m wide areas across the Contour plots showing cross-sections of snow wetness (θ) to a depth pit (WSL, 2008). Snow temperature measurements may be slope.of up Allto 70 observations cm over 5 werem wide observed areas across in shallow the slope. snowpack All areasobserva- at similar elevations (2000–2300 m) in southerly aspect slopes (SE, S, andry indicatormixed wet ofdry drymixed snow,wet althoughdry mixed smallwet dry amountsmixed wet ofdry liquidmixed wet tions were observed in shallow snowpack areas at similar elevations SW). θ, measured with the Snow Fork, is corrected by −0.8 vol.% water have been measured in snow in temperatures below (2000 - 2300 m) in southerly aspect slopes (SE, S, SW). θ, measured ◦ (this corresponds to the median offset in dry snow). θ-values greater 0Fig.C( 13.θ

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 415 international guidelines but should be understood as an indi- Table 5. Proposition of interpretation of manually estimated water cation only (Fierz and Fohn¨ , 1994). Expanding on a previous content (hand test as per guidelines WSL, 2008; Fierz et al., 2009). study (Martinec, 1991b), we investigated the effect of layer The proposition from Martinec (1991b)(θMartinec, n = 518, 9 ob- characteristics like hardness or grain shape on snow wetness servers) is compared to our study (θ, n = 314, 4 observers). This estimation. It seems that it is more difficult to correctly es- interpretation may apply only to experienced observers estimating timate the wetness in layers consisting of melt-freeze parti- the liquid water content. θ, measured with the Snow Fork, is cor- cles. The liquid water content of melt-freeze-crusts under- rected by −0.8 vol. %. going melting is particularly hard to estimate. We assume that the reason for this is the larger range of snow wetness Hand test Signature θMartinec θ (θ0−10 vol. %) and hardness. In particular at low water con- (mWC) [vol. %] [vol. %] tent when the ice matrix is still frozen for the most part, wa- Dry 1 <0.5 <0.5 ter cannot be seen using a magnifying lens and the squeeze Moist 2 0.5–2 0.5–2 test is not suitable in such hard layers. While layer hardness, Wet 3 2–4 2–4.5 grain shape and size may unconsciously influence wetness Very Wet 4 4–5 4.5–6 estimates, we were unable to quantify these effects. In cases Slush 5 >5 – where it is necessary to quantitatively interpret the estimated wetness, we propose a rough guide which is based on the study by Martinec (1991b) (Table 5). of note that very wet horizontal layers, as in Fig. 5, were ab- Additionally to methodical aspects, small-scale spatial as- sent during the avalanche cycle in spring 2010. We suspect pects should be considered when recording snow wetness in that this is due to the snowpack structure, which contained point locations. Due to a lack of data at greater spatial dis- few spatially expanded possible capillary barriers, like fine- tances than 5 m, we can only assume that a 5 m wide cross- over-coarse grained layer boundaries. Also, clear patterns of section is a good representation of the snowpack wetness larger vertical flow paths could not be observed. in a given slope and that the median water content and in- terquartile range are robust indicators of snow wetness at a 6.3 Qualitative description of the wetness of a snowpack certain snow depth. In particular in a partially wet snow- pack, measuring (for instance) three wetness profiles at dis- As we have discussed before, the reliable estimation and tances greater than 50 cm will provide a robust estimate of measurement of snow wetness in the field is difficult. Tempo- snow wetness and enhance the quality of wetness recordings ral changes in the amount and distribution of liquid water in using dielectric methods. snow may occur rapidly. Therefore, it might be more practi- cal and sufficient for avalanche forecasting and snow hydrol- 6.2 Monitoring the advancement of the wetting front ogy purposes to use a very general description of the wetness of a snowpack. Based on our observations, we propose five The measurements of snowpack wetness during spring 2010 wetness profile types, which may be based on estimated or were conducted on slopes of all aspects (illustrated for quantitatively measured water content. It could also be based southerly aspects in Fig. 11 and 12). This information was simply on the distinction between dry and not-dry snow. We available for the SLF avalanche warning team during the suggest that surface layers are not included in this classifi- March melt-phase and was considered as very helpful to as- cation as these show significant changes during the diurnal sess the advancement of the wetting front. However, for op- freeze-melt-cycles. The classification incorporates both ver- erational purposes it would be of advantage if continuous, tical and horizontal wetness distribution. real-time information from potential avalanche slopes would The snowpack is dry before the melt-phase (Fig. 13, type be available. In some avalanche forecasting operations, for 1). With continued input of liquid water through melt or instance in the maritime climate of Snoqualmie Pass (USA, rain, snowpack wetness increases. Initially, only a part of Stimberis and Rubin, 2009) and the Milford Road area (New the snowpack is wet while some areas remain dry. The wet- Zealand, Carran et al., 2002) snow temperature and water ting may follow a ”step-and-fill-pattern” (Conway and Bene- outflow measurements are used operationally to assess the dict, 1994) (Fig. 13, type 2). Often, in a snowpack consisting advancement of the wetting front. Currently, in Switzer- predominantly of coarse-grained temperature-gradient snow, land, snow temperature is measured mostly at automatic preferential flow fingers will penetrate the snowpack rela- weather stations in flat study plots. However, it would be tively quickly and water may temporarily flow laterally along very useful for avalanche forecasting purposes if such real- capillary barriers (Marsh, 1988) (Fig. 13, type 3). At this time snowpack information would be available from poten- stage, first water outflow at the base of the snowpack may tial avalanche slopes. be observed. With continued water infiltration the snow- The first significant period of water infiltration into pack will be fully wet and homogenize (Jordan et al., 2008) the lower parts of the snowpack coincided with intense (Fig. 13, type 4). Once drainage channels are well estab- avalanche activity in both the springs of 2009 and 2010. It is lished, water outflow will respond quickly to additional input www.the-cryosphere.net/5/405/2011/ The Cryosphere, 5, 405–418, 2011 416 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

1- 2- 3- 4- 5- mated wetness existed in layers consisting of different grain dry upper part wet transitional fully wet lower part wet shape, however no conclusive results were obtained (research question 2). Dielectric measurement methods are generally a more re- liable indicator of snow wetness at a given point within the snowpack. If the measured liquid water content is low (less than approximately 1.0–1.5 vol. %), the measured wetness should be interpreted with caution. In such cases, we believe, that the snow could be either dry or contain small quantities of liquid water. If the presence of very wet layers is of inter- est and an instrument like the Snow Fork is available, mea- dry mixed wet dry mixed wet dry mixed wet dry mixed wet dry mixed wet surements should be undertaken before excavating a snow pit. Otherwise, only small differences existed between mea- Fig. 13. Wetness classification incorporating vertical and horizontal surements made on the shaded side-wall of a snow-pit or in a water content distribution. The x-direction shows the percentage of previously undisturbed snowpack (research question 3). Ver- the snow which has been wetted, where dry/wet is fully dry or fully tical measurements of snow wetness using the Snow Fork are wet and mixed consists of both dry and wet regions. The y-direction an efficient way to record spatial distribution of snow wet- shows the vertical distribution of snow wetness (50–100 cm). Diur- nal changes occur mostly within the upper-most 10–15 cm of the ness. snowpack and are not considered in this classification. As for dry snow observations, site selection is important to observe representative information. To obtain representative wetness information, slope aspect and inclination, elevation of melt-water (Carran et al., 2002). A special case is the and the distance to rocky areas should be considered. Still, situation that the snowpack begins to refreeze or new snow the small-scale spatially heterogeneous distribution of dry falls on an already wet snowpack after a melt-event (Fig. 13, and wet areas within the snowpack may lead to unrepresenta- type 5). tive results. Our observations showed that approximately ev- We propose this very simplified classification and are ery fifth to tenth wetness profile was a poor representation of aware that more variations will exist. However, such a basic the surrounding snow wetness (research question 4). There- classification can facilitate the description of the snowpack fore we propose to observe several measurements at horizon- wetness, in particular for hydrological and avalanche fore- tal distances greater than 50 cm. Because our measurement casting purposes. One advantage of such a classification is extent was limited to 5 m, we can not give conclusive results that the distinction between dry and not dry snow will likely on spatial correlation of liquid water content. As far as we be more accurate than the estimated wetness classes. Addi- are aware, this is the first study quantifying the variability of tionally, it describes the spatial wetness distribution which liquid water distribution in a snowpack at scales up to 5 m. currently is not included in a snow profile observation. The Based on our observations on spatial variability in snow spatial wetness distribution may be observed when excavat- wetness, we propose a snowpack-wetness classification. We ing a snow pit for avalanche forecasting or snow hydrological see this as a first step towards the development of a wet purposes. snow classification scheme as exists for the assessment of dry snow profiles (Schweizer and Wiesinger, 2001). Incor- porating additional snowpack properties like the state of wet 7 Conclusions snow metamorphism (grain shape), snow layering and snow temperature may improve the value of such a classification Methodical, spatial and temporal aspects must be considered for avalanche forecasting purposes and could also assist in when observing snow wetness in the field and interpreting flood forecasting during snow melt periods. wetness information. Point observations in wet snow provide important infor- The method of estimating the liquid water content using mation on the state of wetting of a snowpack, which in- the hand test can not be regarded as a reliable method to fluences wet snow stability. However, as water infiltration record snow wetness if absolute values are of interest (re- in snow is a fast process, timing of observations is critical. search question 1). If it is necessary to quantitatively inter- Therefore, for operational avalanche forecasting purposes it pret the qualitative wetness recordings, Table 5 may provide is necessary to continuously observe parameters like snow- a rough aid for conversion. The hand test is more suited pack temperatures, liquid water content or water outflow not to record the relative wetness differences within one profile only in flat study plots, but also in potential avalanche slopes. (Fierz and Fohn¨ , 1994) and the difference between dry and This is operationally done in some forecasting operations (for not dry snow. Our study expands previous research (Mar- instance in the US or New Zealand). tinec, 1991b) by incorporating snow layer properties like To our knowledge, there is currently no reliable, econom- hardness, grain shape and size. Differences in correctly esti- ical and practical alternative available to the hand test to

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/ F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow 417 measure snow wetness in the field. Therefore, we propose Frolov, A. and Macharet, Y.: On dielectric properties of dry and that future research should investigate possibilities of de- wet snow, Hydrol. Process., 13, 1755–1760, 1999. veloping a practical, hand-held instrument to quantitatively Fujita, S., Shiraishi, M., and Mae, S.: Measurement on the dielectric measure the liquid water content in snow (similar to a digi- properties of acid-doped ice at 9.7 Ghz, Geosci. Remote Sens., tal thermometer). Further, it would be advantageous to know 30, 799 –803, 1992. the distribution of the liquid water content at the slope-scale Gupta, R.P., Haritashya, U. K., and Singh, P.: Mapping dry/wet snow cover in the Indian Himalayas using IRS multispectral im- to allow better interpretations of point observations in wet agery, Remote Sens. Environ., 97(4), 458–469, 2005. snow. Heilig, A., Schneebeli, M., and Eisen, O.: Upward-looking ground- penetrating radar for monitoring snowpack stratigraphy, Cold Reg. Sci. Techn., 59(2–3), 152–162, 2009. Acknowledgements. We greatly thank Adrian Raz,¨ who assisted in Jones, E., Rango, A., and Howell, S.: Snowpack liquid water distri- the field. Christoph Mitterer shared important field information. butions using freezing calorimetry, Nord. Hydrol., 14, 113–126, He, as well as Thomas Stucki and Stephan Harvey provided 1983. valuable comments which helped improve the manuscript. We Jordan, R.: Effects of capillary discontinuities on water flow thank the editor Andrew Klein, one anonymous reviewer and and water retention in layered snowcovers, in: Proceedings of Howard Conway for their very helpful feedback on this paper. Snowsymp, International Symposium on Snow, Manali, India, 1994, 157–170, 1994. Edited by: A. Klein Jordan, R., Albert, M., and Brun, E.: Physical processes within the snow cover and their parametrization, in: Snow and cli- mate: physical processes, surface energy exchange and model- References ing, edited by: Armstrong, R. and Brun, E., Cambridge Univer- sity Press, 12–69, 2008. Armstrong, R.: Wet snow avalanches, in: Avalanche Research and Kark¨ as,¨ E., Martma, T., and Sonninen, E.: Surface snow properties Snow Characteristics, edited by: Armstrong, R. and Ives, J., San and stratigraphy during the austral summer in western Dronning Juan Mountains, Colorado, University of Colorado INSTAAR, Maud Land, Antarctica, Polar Res., 24, 55–67, 2005. Occasional paper 19. p. 256, Boulder, CO. 67–82, 1976. Kattelmann, R.: Wet Slab Snow Instability, in: Proceedings In- Bloschl,¨ G: Scaling issues in snow hydrology, Hydrol. Proc., 13, ternational Snow Science Workshop, 24-27 Oct 1984, 102–108, 2149–2175, 1999. 1985. Brun, E.: Investigation on wet-snow metamorphism in respect of Kattelmann, R. and Dozier, J.: Observations of snowpack ripening liquid-water content, Ann. Glaciol., 13, 22–26, 1989. in the Sierra Nevada, California, USA, J. Glaciol., 45, 409–416, Carran, W., Hall, S., Kendall, C., Carran, A., and Conway, H.: 1999. Snow temperatures and water outflow during rain and melt; Mil- Louge, M., Foster, R., Jensen, N., and Patterson, R.: A Portable Ca- ford Highway, New Zealand, in: Proceedings International Snow pacitance Snow Sounding Instrument, Cold Reg. Sci. Technol., Science Workshop, Big Sky, Montana, 2002, 173–177, 2002. 28, 73–81, 1998. Colbeck, S.: Water flow through heterogeneous snow, Cold Reg. Lundberg, A., Granlund, N., and Gustafsson, D.: “Ground Trut” Sci. Technol., 1, 37–45, 1979. snow measurements – review of operational and new measure- Colbeck, S.: An Overview of Seasonal Snow Metamorphism, Rev. ment methods for Sweden, Norway, and Finland, in: Proceedings Geophys., 20, 45–61, 1982. 65th Eastern Snow Conference, Fairlee (Lake Morey), Vermont, Colbeck, S.: A review of sintering in seasonal snow, US Army USA, 215–237, 2008. Cold Regions Research and Engineering Laboratory Report, 97– Marsh, P.: Grain growth in a wet Arctic snow cover, Cold Reg. Sci. 10, 1997. Technol., 14, 23–31, 1987. Coleou,´ C. and Lesaffre, B.: Irreducible water saturation in snow: Marsh, P.: Flow fingers and ice columns in a cold snowcover, experimental results in a cold laboratory, Ann. Glaciol., 26, 64– in: Proceedings Western Snow Conference, 18–20 April, 1988, 68, 1998. Kalispell, Montana, 105–112, 1988. Conway, H. and Benedict, R.: Infiltration of water into snow, Water Marshall, H., Conway, H., and Rasmussen, L.: Snow densification Resour. Res., 30, 641–650, 1994. during rain, Cold Reg. Sci. Technol., 30, 35–41, 1999. Conway, H. and Raymond, C.: Snow stability during rain, J. Martinec, J.: Flussiger¨ Wassergehalt in der Schneedecke im Gebiet Glaciol., 39, 635–642, 1993. Davos-Weissfluhjoch, April–Juni 1988 (Liquid water content Crawley, M.: The R book, John Wiley & Sons Ltd., 1 edition, 1997. in the snowpack in the region of Davos-Weissfluhjoch, April– Denoth, A.: An electronic device for long-term snow wetness June 1988), internal report 663 (unpublished), WSL Institute for recording, Ann. Glaciol., 19, 104–106, 1994. Snow and Avalanche Research SLF, Davos, 1999a. Fierz, C. and Fohn,¨ P.: Long-term observation of the water con- Martinec, J.: Schneefeuchtigkeit mit dem Denoth-Gerat¨ im Ver- tent of an Alpine snowpack, in: Proceedings International Snow gleich mit dem Handtest, Messungen 1989, 1990 (Snow wetness Science Workshop 1994 Snowbird, Utah, USA, 117–131, 1994. with the Denoth-meter in comparison to the hand test, measure- Fierz, C., R. L., A., Durand, Y., Etchevers, P., Green, E., McClung, ments 1989, 1990), internal report 667 (unpublished), WSL In- D., Nishimura, K., Satyawali, P., and Sokratov, S.: The Interna- stitute for Snow and Avalanche Research SLF, Davos, 1999b. tional Classification for Seasonal Snow on the Ground, IHP-VII Mess-Systemtechnik, S.: Snow Pack Analyser (SPA) Technical Documents in Hydrology N83, IACS Contribution N1, zur Bestimmung des Schneewasseraquivalents¨ und des UNESCO-IHP, Paris, 2009 . www.the-cryosphere.net/5/405/2011/ The Cryosphere, 5, 405–418, 2011 418 F. Techel and C. Pielmeier: Point observations of liquid water content in wet snow

Flussigwasseranteils,¨ Sommer GmbH & Co. KG, Koblach, Techel, F. and Pielmeier, C.: Wet snow diurnal evolution and sta- Austria, available at: http://www.sommer.at, p. 10, 2010. bility assessment, in: Proceedings International Snow Science Mitterer, C., Hirashima, H., and Schweizer, J.: Wet-snow instabili- Workshop, edited by: Schweizer, J. and Van Herwijnen, A., ties: comparison of measured and modelled liquid water content 27 September to 2 October 2009, Davos, Switzerland., 256–261, and snow stratigraphy, Ann. Glaciol., accepted, 2010. 2009. Moldestad, D.: Characteristics of liquid water content and snow Techel, F., Pielmeier, C., and Schneebeli, M.: Microstructural resis- density in a cross-country race ski track, Bull. Glaciol. Res., 22, tance of snow following first wetting, Cold Reg. Sci. Technol., 39–49, 2005. 65, 382–391, 2001. R: R A Language and Environment for Statistical Computing, R Toikka: Snow Fork – manual, Ins.toimisto Toikka Oy, Espoo, Foundation for Statistical Computing, Vienna, Austria, ISBN Finland, p. 20, 2008 3-900051-07-0, 2009. Toikka: Snow Fork – a portable instrument for measuring the Raymond, C. and Tusima, K.: Grain coarsening of water saturated properties of snow., Ins.toimisto Toikka Oy, Espoo, Finland, snow, J. Glaciol., 22, 83–105, 1979. http://www.toikkaoy.com/, p. 2, 2009 Schneebeli, M.: Development and stability of preferential flow Waldner, P., Schneebeli, M., Zimmermann, U., and Fluhler,¨ H.: Ef- paths in a layered snowpack, IAHS-AIHS Publ. 228, 89–96, fect of snow structure on water flow, Hydrol. Process., 18, 1271– 1995. 1290, 2004. Schweizer, J. and Wiesinger, T.: Snow profile interpretation for Wankiewicz, A.: A review of water movement through snow, stability evaluation, Cold Reg. Sci. Technol., 33, 179–188, 2001. in: Proceedings of the modeling of Snow Cover Runoff, edited Sihvola, A. and Tiuri, M.: Snow fork for field determination of the by: Colbeck, S. and Ray, M., CRREL Special Report, 222–252, density and wetness profiles of a snow pack, IEEE T. Geosci. 1979. Remote, GE-24, 717–721, 1996. Warren, S. G. and Wiscombe, W. J.: Dirty snow after nuclear war, Stein, J., Laberge, G., and Levesque,´ D.: Measuring the dry density Nature, 313, 467–470, 1985. and the liquid water content of snow using time domain reflec- Williams, M., Sommerfeld, R., Massman, S., and Rikkers, M.: tometry, Cold Reg. Sci. Technol., 25, 123–136, 1997. Correlation lengths of meltwater flow through ripe snowpacks, Stimberis, J. and Rubin, C. M.: Glide avalanche response to an ex- Colorado Front Range, USA, Hydrol. Process., 13, 1807–1826, treme rain-on-snow event, Snoqualmie Pass, Washington, USA, 1999. in: Proceedings International Snow Science Workshop 2009, WSL: SLF-Beobachterhandbuch: Regionale Beobachter, Hangpro- Davos, Switzerland, 301–305, 2009. filer, Gelandebeobachter,¨ WSL Institute for Snow and Avalanche Research SLF, Davos, 2008.

The Cryosphere, 5, 405–418, 2011 www.the-cryosphere.net/5/405/2011/