M a n a g e m e n t

a n d

A n a l y s i s

o f

S n o w,

A v a l a n c h e

a n d

C l i m a t e

D a t a

Aspect P_{AS P}

The weight factor for aspectchangesfrom situation to situ-
ation (and bulletin resoectively) according to what aspects
are considered to be dangerous.In this example aspects
fromthe weights are distributed as follows

P_ASP

P_ASP

P_ASP

P_ASP

= 5:NW yo R (315 - 90º)

= 3:E to ESE (90 - 112.5º) and
WNW to NW (292.5 - 315º)

= 2:ESE to SE (112.5 - 315º)
and W to WNW (270- 292.5º)

= 1:SE to W (135 -270º)

superimposed to the map of the predicted hazard.Fig. 6
shows an examplefrom a very extreme situation. On march
23,1996mostof the possible avalanches departed. All of
them started in the dark aerea: PHG_max= 6 *5 * 5 *5 =
750.For reliablevalidationexperimentalavalanchesshould
be performed (Bolognesi 1996). Spontaneous or accidental
avalancheshave the disadvantagethat they are not always
reconstructable.Veryoftentheywerenotobservedand
manyoftheconnectedparametersremainunknown.
Moreoveritisdifficulttolocalizethereleaseareasof
potentialavalanches.Theymanifestthemselfesonlyby
externalinfluencessuchasskitourists,animalsor
explosives. Since suchavalanchesarerare (fortunatly) and
not always predictable, it is difficult to gather abig enough
sampleforstatisticaltesting-afactwhichcomplicates
the work of all avalanche experts.

CONCLUSIONS AND OUTLOOK

ThevisualizationoftheSwissavalanchebulletinusing
GIS is an attempt to hand more detailed information to the
public: mountainprofessionals as well aspeople withlittle
avalanchetrainingandexpert knowledge. All these people
expectandeventuallydependondetailedinformation
concerning the snowcoverand avalanchesituation.
As shown,it ispossible to"pin down" the avalanche
hazardand toreproduce itina scale inwhich theindi-
vidualskitouristmovestoandfroaswell.Settingup a
model like LAWIPROGand producingsuch maps is possi-
ble as soon as the underlaying data like a high resolution
digitalterrainmodel are available.It wouldevenbe possible
to model optimal skiroutes (between departure andarrival
point) on top of predicted hazard maps by using so called
"crosscountrymovement"algorithms(XuandLathrop,
1995). However, thecritical question must be asked, if such
maps should be produced at all. To most forecast systems
agreatmanysources ofuncertaintiesareinherent.Very
often simple models are used, which can only be justified
iftheyare applied byexperts.Inthe wrong hands such
maps may be disastrous.On the otherhand with modern
computing facilities it will not take long until such maps
willbeproducedand spread bythemassmedia.Most
obviouslynotmuchexpertknowledgewillthenbein-
volved.Thereforeitisplanedtofurtherdevelop
LAWIPROG at the SFISAR.LAWIPROG will be tested in-

Fig. 4: Calculation of weight-factor for ASPECT

Again flexible rules forexpressionslike"all aspects",
"facing north" or"shady slopes" exist.
Inour(fictive)example,a38degreesteep,northeast
facing slope at an altitude of 3000 mwould be estimated
with PHG_max= 8 *5 * 5 *5 = 1000 (see Fig. 5).

RESULTS AND VERIFICATIONOF LAWIPROG

The output of the LAWIPROG-model is a raster data set of
continous "hazard values" foreach cell. These values can
now be transformed into a colorramp for cartographic pur-
poseshowingalltransitionsfromnon-dangerous(non-
hazardous) to dangerous (hazardous)areasfor a givenbul-
letin.AccordingtotheEuropeanhazardstandarda
colorramp from green to red over yellow was choosen.In
the present paper dark grey/black indicates high danger,
light gray low danger (see Fig. 5).
With LAWIPROG's interactiveuserinterface thePHG
can now be analysed. For eachcellsall values can be inter-
rogated,areas of equalvaluescanbeselected,altitude,slope
and aspect of a selected cell will appear on the screen etc.
Moreover input parameters can be changedand newmaps
be calculated immediately. Thanksto modern data storage
technologiescomputationaloperationsperfomedbythe
GRID module require only little computing time and new
PHGsmaybeobtainedwithinmoments.Thisisagreat
advantagefor theexperts whowantto evaluatetheaccuracy
of theirforecast before its release.
Themultiplicativemodellisbutaninitialapproach.
LAWIPROG wasdevelopedduring winter1995/96.The
weight-factors were definedby avalanche experts, but not
yetempiricallytestedunderfieldconditions.Thefirst
method to do sois to compare the predicted hazard with
observed events.Released avalanches - spontaneous and
man-madeones-canbemapped,digitizedand finally

4