CLIMATOLOGY OF AN EXPERIMENTAL MOUNTAINOUS … · 2011. 11. 7. · exhaustive study on the set of the 12 years unit now. It seemed thus interesting to lead a climatological study

CLIMATOLOGY OF AN EXPERIMENTAL MOUNTAINOUSLOCATION FOR STUDIES ON SNOWDRIFT.

G. Guyomarc'h*\ Y. Durand\ L. Merindol\ F. Naaim-Bouvef1 Meteo-France CEN (Snow Study Centre)

2 Cemagref Etna (Torrential Erosion,snow and avalanches research unit)

KEYWORDS: Climatology, snowdrift, experimental measurements.

2. INTERESTS

The main points justifying the interestssuch a study are summarized below:

1 - Constitution of a dataaccessible to the partners of the project, nothe Safety and Snow service of "Alpa d'Hu.resort (Sata) and to the scientific communifY:

the thresholds of wind allowing the beginning atthe transport, according to the snow type at thesurface of the snow-cover. Models weredeveloped and are now operational. For thesemulti-field searches, different types of sensorswere recording parameters connected to u.:studied phenomenon during the durationcontracts. Recordings were made from October.'to the end of April from 1988-1989 to 1999-2000winter seasons. These parameters were usedpunctually for every study (generally one orseasons), but had not been the object of artexhaustive study on the set of the 12 years unitnow. It seemed thus interesting to lead aclimatological study to get a tool accessible faall.

296

* Corresponding author address :Gilbert Guyomarc'h, Meteo-France, CEN, 1441rue de la piscine 38406 St Martin d'Heres CedexFRANCEE-mail: [email protected]

The purpose of this work is theclimatological study of various parametersrecorded at an experimental site of high altitude,and the connection with episodes of blowingsnow. For 12 years, several research projectshave succeeded one another on the site of the"Col du Lac Blanc". All these studies had incommon the study of the effects of the snowdriftin the distribution of the snow-pack and in theincreasing of avalanches risk. Indeed, thepresence of wind during or after snowfalls isoften materialized by the constitution of slab.These slabs are often very unstable, and are atthe origin of about 80 % of accidents byavalanche in mountain.

These research projects allowed toimprove the knowledge on the mechanisms ofsnow transport, and to determine in particular

1. INTRODUCTION

ABSTRACT: Over the last 10 years, 2 French laboratories specialized on snow and avalancheresearch have joined their resources to investigate snowdrift effects and their consequences on theincreasing of avalanche hazard. The experimental site is situated at 2 800 meters high in the French

.Alps, at a place where the wind is canalized in 2 directions.The working method was based :

• In a first time, on data measurements separated in automatic measurements (windvelocity, wind direction, snow depth, water equivalent of precipitation, ...) and in a set offield measurements (snow pits, taking of snow samples, vertical profile of snow density,stereo-photogrammetry, ...)

• In a second time, on development of applications for the forecast of blowing snow eventsand the modeling of snow distribution.

• The current project aims at testing an integrated chain of models which can determine thesnow distribution by taking into account the local topography.

This poster will present the results of a climatological study made by using the series of data and theresults of the modeling.

This data base will be useful for the futureresearches and could be completed by results ofnumerical simulations (Safran - Crocus [Brunand others, 1992]).

2 - Valuation of data recorded on a sitewhich equipment was funded by contracts with"region Rhone-Alpes" and "Pole Grenoblois" andpublication.

3- Synthesis of knowledg~ on theexperimental site, critical exam of data,improvement and validation of alreadydeveloped or current models.

4 - exhaustive documentation of theperiods of blowing snow on the set of twelvewinter seasons.

3. DESCRIPTION OF THE EXPERIMENTALSITE

At this experimental location severalresearch programs have succeeded. Thegeneral aim was to study the effect of windduring snowfalls and on the deposited snow.Numbers of sensors have been set up at thislocation, so we have series of data over a periodof twelve winter seasons (in the broad sense!From October to April). This site has --beendescribed with details in previous Issw's papers[Guyomarc'h and others, 1992-1996-1998] andin other publications [Guyomarc'h and Merindol,1997], [Durand and others, 2000], [Michaux andothers, 2000]. We can summarize thedescription as follows :.:. Large pass where the wind is channeled

according to a northern-southern axis (figure1).

•:. The site is situated at an altitude of 2 800m.It assures conditions often favourable to theobservation of the snowdrift phenomenon.Several sensors have been set up for thehourly record of data.

297

years, two other zones were added andwere equipped with sensors in order todetect blowing snow events and to measurethe erosion and the accumulation of snow.

Figure 2 : one point of measurements ("Col duLac Blanc'J (snow depth, snow waterequivalent, wind velocity and direction, airtemperature). At this site we have 12 seasons ofdata.

Figure 3 : the second point of windmeasurements ("Dome des Petites Rousses'J.At this site we have 8 seasons of data.

4. CLIMATOLOGY

The subject of this study is to describethe different parameters relevant for thesnowdrift phenomenon. Thus 5 kinds of graphare shown:.:. Compass card The wind velocity

thresholds are not standards, but werechosen according to our knowledge onsnowdrift.

•:. Air temperature histogram and graphs :We have chosen to visualize the monthlyevolution of the air temperature over theseason. The chosen thresholds are -10°c; -

5°C and O°C. We also presented the hourlyevolution with the aim of showing theperiods when air temperature overtook O°C.

•:. Mean wind velocity : In order to bettervisualize strong wind periods, we drawcurves of temporal evolution of the wind forwhich the colour of the line corresponds to aclass of direction (only when the windvelocity is greater than 4 mls - windthreshold for the start of snowdrift).

.:. Snow depth and precipitation : It wasinteresting to superpose on the same graphthe snow depth and the water equivalent ofprecipitation (most of the time snow at thisaltitude for considered period): Exceptduring snowfalls, the snow-depthfluctuations are due essentially to thepacking down of the snow cover and to theblowing snow periods. This type of graphallows to give prominence to suchphenomena.

•:. Snow depth differences and wind gusts :We compared the hourly evolution of snowdepth measurements (difference betweenhourly maxi and mini values - if it is greaterthan 2 cm) and the wind gust coefficient(ratio between hourly maximum velocity andhourly mean velocity - only when the windspeed is greater than 4 m/s).

5. ANALYSIS OF PARAMETERS

5.2 Wind parameters

At the main experimental location ("du Lac Blanc") the wind direction are stronginfluenced by the topography. Due to thischaracteristic the northerly winds representabout 50% of cases and the southerly winds10%. Because of this feature we have zoomed inthe compass card on the first 5% of frequency(figure 5). The figure 6 exhibits the detailedfrequency for each class of direction andvelocity. We can observe the lack of wind fromwest and about 1% of wind velocity lower than 1m/s.

Compass Card "Col du Lac Blanc" (1988 to 2000)

360

180

Figure 5 : Zoom on the first 5% (frequency) ofthe compass card for the 12 seasons at the "Coldu Lac Blancn site.

Figure 6 : Distribution of the data by classdirection and velocity.

For every seasons, we have drawnthe same graph (figure 7) the snow depththe snow water equivalent accumulated overwhole period. By comparison _with 0

parameters (wind, snow surface conditions,..·

Equivalent5.3 Snow depth

10 11 12 1 2 3 4

Month

298

Air Temperature Evolution

5.1 Air temperature

70 +============1leo+-------------1~ 50 +-._-----------1CD6- 40 t-.--..,-- -._---==--_._-.,CD

~ 30 t-.--If---­~:20c~ 10

o

Figure 4 : Frequency of the 4 classes of airtemperature for each month of the 12 winterseasons.

For this parameter we have 12 seasonsat the "Col du Lac Blanc" site from 1988 to 2000.Below (figure 4), as an example, we show thegraph for all the seasons. The highest frequencyis between 0 °c and -5°C, the positivetemperatures are more marginal.

6.2 Calendar of snowdrift periods

(the sensor precision is +1- 1 cm) during a periodof at least 5 hours.

..

-..- >=2cm~ >=3cm

-.- >=4cm~ .. >=5cm

Number of consecutive hours

' .... ......:.......

600 umber of snowdrift periods........

100 ••

1l. JOO ••..1zoo ..i

Figure 8 : Snowdrift periods in connection withthe hourly difference of the maxi and mini valuesof snow depth sensor.

uivalent 1993-1994

this graph will be able .to b~ used for theidentification of the snowdrrft perrods.

It is very difficult at this altitude to takeinto account th~ water equivalent . ofprecipitation, essentially because of the windconditions. Nevertheless, if we compare thesnow depth with. th~. accumulati~n ofprecipitation, we can Identified some perrods ofsnowdrift.

Figure 7 : Example of a graph for the 1993-1994winter season.

6. STUDY ON THE SNOWDRIFT PERIODS

6. 1 Criteria ofselection

For every periods recorded parametersallow to describe in detail the weather and snowconditions. It will also be possible to completethese parameters with the results of snow packsimulations (Safran-Crocus [ Brun and others,1992], [Durand and others, 1993]). An extract ofthis calendar is shown below (figure 9).

This experimental site is particularlyconvenient for the observation of thephenomena of snow transport by the wind.

How to detect these snowdrift periods?For two years, we have been able to use severalacoustic sensors [Michaux and others, 2000],but for the previous seasons we only have gotthe data recorded by the automatic station andin-situ observations. Therefore, we selected thesnowdrift periods by using the number ofconsecutive hours for which difference betweenhourly maxi and mini values of the snow-depthsensor exceeded a threshold (figure 8). Weassume that this difference is big when the waveof ultrasonic sounds is intercepted by snowflakes above the snow surface. It happens whenit snows or when the snow is moved by the wind.These periods were only chosen if the hourlymean value of the wind velocity is always greaterthan 2 m/s. It is possible to discriminateafterwards between events with precipitation ornot by using the data from the rain gauge.. By using this method, we have selected,In a first approach, 119 periods for which thesnow depth difference is greater or equal to 4 cm

Figure 9 : All periods (according to the previouscriteria) are described.

6.3 statistical study

Some calculations have been made onthe whole sample of snowdrift periods, then onthe data without snowdrift periods (figure 10).

299

- -- s..L-.... 1·~ - W_ ..........ThnIion ll.2 6.0 4 S.O 6.0 8.0 11.0

Me.. Velocity 10.6 10.4 2.8 8.3 9.8 11.1 12.7

Guso ........... L5 L4 0.3 1.3 L4 L5 L7-- 22.2 18.0 18.1 10.0 13.0 21.0 .n.ODepth

3.6 3.4 L4L7 L7 L4

Figure 10 : Statistical calculation over thesnowdrift or no-snowdrift periods.

It brings us to some conclusions:•:. wind speed is on hourly average much

stronger when snowdrift is present (10.6 m/sagainst 4.4 m/s for lack of the phenomenon).We can observe besides an importantdistance between the average and the

median of the wind velocity of no-snowdriftperiods because of the distribution of sJ>eeds(40 % of wind velocities are in that caselower than 2.8 m/s).

.:. The coefficient of wind gusts is weak incases of snowdrift. It is surprising, we COuldthink, on the contrary, that wind gusts allowto extract more easily the snow from thesnow surface. Moreover recent stUdies[Michaux a.nd others, 2000) have shown thatby calculating average and the maxi over ashorter period results are appreciablydifferent.

6.5 Examples of snowdrift episodes

We have selected below (figure 11) agraph which represents snowdrift periods.Different types of snowdrift periods are shown .with or without precipitation, with erosion or withaccumulation.

Col dU,Lac Blanc January 199625 ...------------~---------=---~-----_,

~205'~ 15

~~ 10

'3~ 5

140

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Figure 11 : On this graph several different periods of snowdrift can be obseNed.

300

4,5 :Ea

4,0 !!lCD

.&I3,5 ~

3,0 ~

9­2,5 ~n @2,0 .g.

;:;

15 ~, 0='

1,0 '32.

Thus, each selected period can bedetailed in terms of weather and snow conditionsand we are able to discriminate between periodswith or without precipitations. We can also refinethe wind velocity thresholds in connection withthe snow type at the snow cover surface thanksto the acoustic sensor. These types of snowparticles are determined either by groundobservations or through the results of snow-packsimulations (Safran-Crocus).

It is also interesting to examine periodswith strong wind (over 12 mls on hourly average)without snowdrift observed (as shown on figure12). In that case, the quality of the snow grainsat the snow surface is the determining elementto explain the absence of the phenomenon.

the "Pole Grenoblois" for the study of majorrisks.

9. REFERENCES

Brun E., P. David, M. Sudul and G. Brunot.1992. A numerical model to simulatesnow-cover stratigraphy for operationalavalanche forecasting. J. Glacial.,38(128), 13-22.

Durand, Y., E Brun., L Merindol., Guyomarc'hG., B Lesaftre. and E Martin. 1993. Ameteorological estimation of relevantparameters for snow models. An. Glaciol.,18,65·71.

"..",,1121

380~---..,--------,,------r 25Col du Lac Blanc 1997·1998 Season

Guyomarc'h, G. and T Castelle. 1992. A study ofwind drift snow phenomena on an alpinesite. Proceedings, Intemational SnowScience Workshop, October 4-8 1992,Breckenridge Colorado, 57-67

Durand Y., Guyomarc'h G. and Merindol L.2000. Numerical Experiments of WindTransport over a MountainousInstrumented Site. (Part. 1: Regionalscale) in press An. Glaciol.

. - Hourly MoanHoul1y Moxi

. _. Houl1y Ioini 20

J!!15.5-~...

10.2

~~5 ~

~

3'0

370

300.1....--__4- -+- --'

Figure 12 : Strong wind periods without snowdriftobserved.

7. DATA SUPPORT

Guyomarc'h G. and L. Merindol. 1998. Validationof an application for forecasting blowingsnow., An. Glaciol., 26,138-143

We have chosen to make a model ofdata server for a later integration in a web site inorder to facilitate the data access. In a first timethe data are available on a CD-rom.

At the end of the study, the details of thedata could be put at the scientific communitydisposal in the framework of future scientificcollaborations.

8. ACKNOWLEDGEMENTS

The first work of this study was realizedby Mathieu Joly for his "end of study" report(Meteo-France National School).

We are grateful in Philippe Pugliese andJean-Michel Panel (electronic team of the CEN)for the precious help for the fitting andmaintenance of the equipment. The SATA(safety service of Alpe d'Huez resort), as usual,helped us on the ground.

This study was partially funded by theGeneral Council of Isere department thanks to

Guyomarc'h G, Merindol L. and Olafsson H.1998. A Method for the Forecasting ofWind in Mountainous Regions.Proceedings, Intemational Snow ScienceWorkshop, October 1998, SunriverOregon, 171~177

Michaux J-L, Naaim-Bouvet F., Naaim M.,Guyomarc'h G. 2000, The drifting snowacoustic sensor : interests, calibration andresults. Proceedings, Intemational SnowScience Workshop, October 2000, BigSky Montana.

Michaux JL, Naaim-Bouvet F. and Naaim. M.2000. Numerical Experiments of WindTransport over a MountainousInstrumented Site. (Part. 2: Avalanchepath scale) in press An. Glacial.

301