WARM STORMS ASSOCIATED WITH AVALANCHE HAZARD IN THE SIERRA NEVADA

Randall Osterhuber*Central Sierra Snow Laboratory, Soda Springs, California

Richard KattelmannSierra Nevada Aquatic Research Laboratory, Mammoth Lakes, California

ABSTRACT: Rain-on-snow events occasionally produce avalanches of varying magnitude depending onboth snowpack properties and storm characteristics. Under rain-on-snow conditions in the Sierra Nevada,avalanche release appears to be most likely if new snow falls a couple days previous to the rain. Incontrast, if the snowpack has already transmitted liquid water from the surface to the base, then even largeamounts of rainfall rarely produce significant avalanches in the Sierra Nevada. Winter storms in thismountain range typically have rain/snow levels between 1200 and 2000 m. Warm storms with higherrain/snow levels of up to 2500 m occur a couple times in most winters and have the potential to generaterain-on-snow floods and wet-snow avalanches. This paper describes the characteristics of warm stormsthat had the potential to generate wet-snow avalanches. It also examines the frequency of rainfallfollowing within three days of snowfall, which tends to be a hazardous combination.

A case study of a very warm storm at the beginning of 1997 describes snowpack response to rainfall athigh elevations where such warm storms have rarely been observed. At the snow research station at2930 m on Mammoth Mountain in the eastern Sierra Nevada, air temperatures exceeding 4° C and 220mm of rainfall was recorded. Although few avalanches were obseryed during this storm, flooding wassevere throughout much of the range.

KEYWORDS: avalanche, avalanche frequency, rain-on-snow, snow stability, wet-snow avalanche

1. INTRODUCTION

In a mountain range with a mostly maritimeclimate, the most important component forforecasting mid-winter avalanche activity is recentprecipitation. In the Sierra Nevada mountains ofCalifornia, widespread or destructive avalanchecycles not concurrent with precipitation events arerare. The largest storms to press upon the SierraNevada often precipitate rain at mid, andoccasionally high elevations. When rain falls ontoestablished snowpacks, avalanches can result.But not always. Preliminary evidence suggeststhat rain-induced avalanching is more apt if therehas been snowfall a few days prior to the rain. Thispaper discusses the anatomy of warm, mid-winterstorms and how they influence the snowpacks ofthe Sierra Nevada and avalanche activity.

2. WARMSTORMS

Approximately 90 percent of California'sprecipitation falls between early November and lateMay. At elevations above 2000 m, most of thatprecipitation falls as snow. However, because of

Corresponding author address: Randall Osterhuber,Central Sierra Snow Laboratory, PO Box 810, SodaSprings, CA 95728 USA; telephone: (530) 426-0318;fax: (530) 426-0319; email: randall@sierra.net

the Sierra Nevada's close proximity to the PacificOcean, the range tends to receive relatively warmmid-winter storms. Much of the snowfall at 2000 melevation precipitates within a few degrees of 0° C.Hence, it is not unusual for rain to fall at Sierranmid-elevations during any month of the winter.Average monthly air temperature at the CentralSierra Snow Laboratory (CSSL, 2098 m elevation)near Donner Summit during December, January,and February-historically the three wettestmonths of the snow season-is approximately _2°C for each month. During the month of March,which historically receives 15 percent of the annualprecipitation, the average monthly air temperatureis only _1° C (Osterhuber 1997a).

Typical mid-winter flow at the 500-mb level overthe Sierra Nevada incorporates air masses drawneastward by both the polar and subtropical jetstreams. Moist, unstable air masses can combinein any number of fashions, mixing cold air massesof the north with warmer southern air. With theSUbtropical jet drawn south, California receivespredominantly cold, moist air pulled by the polar jetfrom the Gulf of Alaska. These cold fronts from theNNW deposit mostly snow in the mid and upperelevations of the Sierra Nevada. Rain/snow levelsof 1600 mor lower are characteristic of thesestorms; precipitation totals are usually moderate,35 - 50 mm. With high pressures positioned overthe tip of Alaska's Aleutian Islands and northern

526

Mexico, the polar and subtropical jet streams areforced south and north, respectively, convergingin a strong west to east flow that hits the SierraNevada with maximum orographic uplift, resultingin high precipitation rates and totals. During thesestorms, precipitation intensities of more than 10mmlhour have been observed at the CSSL forseveral hours. Storm totals often exceed 225 mmof precipitation and 2 m of snowfall. Rain/snowlevels observed during these events are typically1700 - 2100 m. It should be noted that theatmospheric dynamics of these events varygreatly. For example, warm, frontal air splitting andoverlaying a colder retreating polar front hasresulted in dramatically different rain/snow levels atdiffering latitudes. Rain/snow levels of 1200 m inthe southern Cascades of northern California havebeen observed coincident with 2400 m rain/snowlevels around Lake Tahoe (Pechner 1998).

A large high pressure cell over the Aleutianscan effectively act as a blocking high, maintaining amore southerly flow of the polar jet stream. If thiscoincides with the subtropical jet stream oscillatingnorth into California, air flow from warm, moist, lowlatitudes result. This pattern not only causes highrain/snow levels in the Sierra Nevada, but candeliver the mountain range's wettest events. Ofthe 20 wettest storms on record at the CSSL, 11have not been associated with record snowfall. Alleleven of these events have occurred either lateautumn or mid-winter and fell as rain-on-snow(Osterhuber 1997b). The rain-on-snow storm ofDecember 1964 dropped 858 mm of precipitationover 5 days; the rain storm of February 1986precipitated more than 525 mm in the central SierraNevada over 8 days. The largest precipitationevents in the central Sierra Nevada are the rain-on­snow events. Storms with rain/snow levels ofapproximately 2500 m occur a couple times duringmost winters, but rain has been observed falling at3600 m in the southern Sierra Nevada during thespring on rare occasions (Kattelmann 1997).

During the past 47 years of record, the meanannual maximum rain-on-snow event at the CSSLis found to be 151 mm of precipitation-themajority of it rain-with a duration of about 4 days.This "average" rain-on-snow storm has arecurrence interval of approximately 2.6 years(Osterhuber 1997b). When the maximum annualrain-on-snow event precipitation totals are graphedas a time series, the slope of the plotted historicevents increases by about 2 mm/year.

The rain-on-snow storms of the Sierra Nevadaare of interest for several reasons. When warm,mid-winter rains fall onto the extensive snowpackof the range, high stream flows tend to result.Throughout history, the largest floods producedby the major river systems of California haveoccurred during rain-on-snow events (Kattelmann

527

et aJ 1991). Much of the direct runoff produced isfrom the storm rainfall, but the presence ofexpansive snow cover greatly increasesstreamflow potential. In a maritime climate, such asthat of the Cascade and Sierra Nevada mountainranges, rain-on-snow has been found to producegreater runoff than either rainfall or snowmelt alone(Harr 1981).

~. The largest rain storms (and therefore the<largest storms) tend to occur during mid-winter. Atthe CSSL, the mean date of commencement ofeach season's greatest rain-on-snow storm isJanuary 25, with a standard deviation of 49 days.The largest rain storms are more likely to fall ontosnow than not. Rarely does a rain-on-snow eventprecipitate only rain at mid-elevations in the SierraNevada. Rain/snow levels typically fluctuate withthe a9vent and passing of large frontal systems.During rain storms, the 2000 - 2500 m elevationsof the range commonly see a mix of rain and snow.

Because warm, maritime snowpacks tend torelease at least small amounts of liquid waterthroughout the winter, underlying soils are moreapt to be at or near saturation during winter andspring. When little soil moisture storage capacity isavailable, streamflow response to rainfall can berapid. In the two weeks previous to January 6,1997, the snowpack surrounding the CSSL(depth 173 cm, snow water equivalent 68 cm) hadabsorbed and outflowed more than 319 mm ofrainfall. The pre- and post-storm snowpack waterequivalents remained essentially equal.Underlying soils during this storm were consideredto be at or near their water-holding capacity.Lowland and upland flooding throughout theSierra Nevada was extensive during this event. Ifsnowpacks are already producing melt due to solarradiation during clear weather, the pack maycontain some appreciable amount of liquid (free)water that can be mobilized during a rain-on-snowevent. In addition, melt caused by convection­condensation during warm storms can besignificant, especially at the lower elevation,transient snow zont:l'of the range (Kattelmann andMcGurk 1989). All of these factors contribute togreat streamflow and potential flooding duringwarm mid-winter storms.

Mid-winter rain-on-snow storms also tend tomove soils. The ready recipe of high rates ofrainfall onto snowpacks and soils with little (if any)ability to store additional liquid water, proves idealfor instigating the movement of large masses ofsoil. Landslides reek insidious havoc onresidential and commercial structures andintermountain highways. Highway road cuts haveproven to be especially vulnerable to rain-on-snowinduced landslides: road cuts tend to be mostlyfree of soil-stabilizing vegetation, and often existnear the soils' maximum angle of repose.

Numerous landslides onto the highwayssurrounding the Tahoe Sierra have beenobserved during or immediately after mid-winterrain storms in 1982, 1983, 1984, 1986, 1995, and1997. These debris flows have ranged in size froma few cubic meters to several hundreds of cubicmeters. In the central and southern Sierra Nevada,25 of 33 (76%) documented landmass failuresduring winters 1982 and 1983 were attributed torain-on-snow (Bergman 1987). Researchers inwestern Oregon have associated 85 percent ofobserved landslides with rain-on-snow (Harr 1981).Wide-spread landslide activity during rain-on-snowis evidence that the snowpack is outflowing at leastmoderate-and often great-amounts of water.

Snow avalanche activity can also increaseduring rain-on-snow. Snowpack stability isinfluenced not only by storm characteristics, butalso by features within the snowpack.

3. WATER MOVEMENT THROUGH SNOW ANDSNOW STABILITY

The spatial and temporal distribution of liquidwater movement through snow is complex. Deepsnowpacks fashion a three-dimensional ice latticewith varying divisions of density, crystal and grainsize and type, temperature, hardness, permeabilityand porosity. Snowpacks accumulate during thewinter in a layer cake fashion. Each new layer ofsnow is unique to the atmospheric conditionsduring which it precipitated. Once on thesnowpack surface (or ground), metamorphism ofthe new snow is further determined by theimmediate atmosphere and surrounding snowlayers. In warm mountain ranges like the SierraNevada, daytime air temperatures greater than 00 Care typical during clear weather between storms.This varies widely with time of year, elevation,aspect, and latitude, ,but slight surface melting is

-"3. not unusual during clear days even in the dead ofwinter. Surface layers melting at day andrefreezing at night develop crusts. These also varyin thickness, hardness, and permeability. Surfacecrusts are subsequently buried under the nextsnowfall. Some of these crusts may impede thevertical movement of water (Berg 1982), insteadrouting' it laterally via capillary and gravitationalforces. The extent and ,rate of this lateralmovement is governed by the composition of thesurrounding snow layers and slope angle. Variousflow channels within the snowpack have beenobserved (e.g. Kattelmann 1985, McGurk andKattelmann 1988) that act as efficient conduits forrouting liquid water. Once formed, verticallyoriented flow ''fingers'' within the pack have beenknown to hasten the water movement through thepack (Kattelmann 1985). The snowpack need notbe either near isothermal (at 0 0 C) or at its water

holding capacity to outflow water. Under rain-on­snow conditions, the rate at which waterpermeates through the entire pack is dependentnot only on the snowpack makeup, but on theamount and rate of precipitation. At the CSSL,average rates of vertical water movement through alevel snowpack have ranged from 3 to 88 cm/hour(Berg et al 1991) at the onset of rain-on-snowevents.

Dry snow avalanches usually fail due to anincrease in shear stress; wet snow tends toavalanche because of a decrease in shear'strength. Unlike most mediums, snow existsnaturally very close to its melt point. When liquidwater is introduced into the ice lattice a generalincrease in temperature and decrease inmechanical strength results. At low amounts ofliquid water, bonding between individual grainscan increase slightly due to capillary forces. Aswater contents increase, these attractions weakenquickly. In the presence of water, snow has fewersmall grains, fewer contacts, and therefore fewerbonds (Kattelmann 1984). For an inclinedsnowfield, resistance to sudden avalanching ofone or more layers is a factor of those layers'combined strengths exceeding their combinedstresses. When combined strengths equalcombined stresses, avalanching occurs. Someobservations suggest as slope angle increases,wetted volume and water holding capacitydecrease (e.g. Kattelmann 1986). Rain falling ontoan inclined snowfield reduces snowpack strengthby reducing the number of bonds and thelubrication of layer interfaces. Increases in stressoccur through the addition of added mass andincreasing rate of creep. In the Sierra Nevada,avalanche cycles occur coincident with mid-winterrain storms often. Naturally occurring avalancheshave been observed during numerous rain stormsin the Sierra. The relatively recent events of 1986(Wilson 1986), 19891, and 19951 werewidespread and memorable. Not all rain storms,however, produce widespread or big avalanches.Avalanches have been observed at the initialonset of rain but decreased or ceased altogethereven though rainfall continued.

4. AVALANCHE ACTIVITY

An investigation of the frequency of avalancheactivity during rain storms was carried out at AlpineMeadows Ski Area in the late 1980s (Heywood1988). Alpine Meadows, one of the mostavalanche-active ski areas in the country, is a few

1Deep, wet slab avalanches occurred at several areasaround Donner Pass either during or immediately afterrainfall.

528

kilometers north of Lake Tahoe and receives rain­on-snow annually. 20 storms were analyzed thathad adequate data on precipitation-both rainfalland snowfall-and avalanche activity. Of eventswith snowfall within three days previous of the rain,Heywood found positive correlations betweenavalanche activity and total rainfall, and betweenavalanche activity and rainfall intensity. There wereonly three storms that had no avalanche activity butthat did have some snowfall within the threeprevious days. No unusual characteristics of thesethree events are evident. Heywood definedavalanche activity as "...widespread avalancheactivity on most exposures, with either or bothnatural or artificially released slab avalanches."Avalanche size was not considered. Positivecorrelations were found between total rainfall anddays since last new snow for both avalanche andnon-avalanche days. For events with no snowfallwithin 4 days of the rainfall, no "avalanche days"existed.

We expanded this investigation by looking atdata from an additional 55 (for a total of 75) mid­winter rain storms at Alpine Meadows. Rainstormswere identified from the records by notationsentered in the daily weather logs. 24 hourprecipitation and snowfall amounts were recorded;as were the number of days since last snowfall, andif any avalanche activity took place during thestorm. Data records were mostly complete.Precipitation and snowfall data originated near theski area's base at 2103 m. 24 hour precipitationvalues were generally totals from early morning toear1y morning. We recorded avalanche days as anyday with any observed avalanche activity at all,natural or induced. Size of events and avalanchetype were noted but not considered in thestatistics. Total precipitation of the events variedwidely, with a mean of 59 mm and standarddeviation of 96 mm. New snow fell during 55 of the75 events. Snowfall also varied widely: meansnowfall was 16 cm, standard deviation 33 cm.Data from the CSSL shows that the level snowpackthere (21 km NW of Alpine Meadows) had drainedsome amount of liquid water previous to most ofthe rain storms (CSSL data was unavailable for 10of the storms analyzed).

Of the 20 rain storms analyzed during which nonew snow was recorded, 4 (20%) eventsproduced avalanche activity. Of these events, 2storms had snowfall one day previous to the rain.The other 2 events saw snowfall six and eight days(respectively) previous to rainfall. Avalanchesobserved during these two latter events includedClass II explosive-induced slides and natural loose­snow avalanching. The mean rainfall amount forthe avalanche-producing "rain-only" storms was 17mm with a standard deviation of 16 mm. For thenon-avalanche "rain-only" storms, the mean rainfallwas 32 mm, standard deviation 62 mm. It shouldbe noted that for the "rain-only" storm data set, themean precipitation of the rainstorms producing noavalanches is greater than the storms that didproduce snow slides. The "rain-only" storms ofMarch 8, 1986 rained 166 mm, and December 11,1995 rained 227 mm-neither of them resulting inavalanche activity.

For the remaining 55 storms examined for thisstudy, some snowfall did accompany the rain. Wedivided these storms into two broad groups:storms with io < 0.28 and storms with io ~ 0.28,where io is a simple storm "density" index definedby

io =total storm precipitation (mm) I total storm snowfall (mm).

When plotted against percentage of days withavalanche activity, the group of storms (n =21) withio ~ 0.28 (low % snowfall) showed a weakcorrelation (statistically insignificant; r2 = .01)between increasing days since last new snow andincreased avalanche activity, independent ofprecipitation or snowfall amounts. For the io < 0.28(higher % snowfall) data, the trend was strongly theopposite. As days since last new snow increased,percentage of days with avalanche activitydecreased (r2 =.30). There were 34 storms withinthe low io data set. Results of the analysis arecompiled in Table 1.

Of the 75 storms examined, 39 events producedavalanches, 36 did not. A review of the avalanchedays reveals that 29 of the 39 (avalanche) days

mm mean days stddevdaysmean stddev mean stddev avaJanche % avalanche since last since last

n precipt precipt snowfall snowfall days days snow snow

vrain-only" storms 2) 3J 51 0 0 4 2) 7 5io< 0.28 storms 34 39 Zl 2a3 135 Z3 68 3 3io :! 0.28 storms 21 111 154 259 52B 14 01 4 4

all storms 75 53 96 100 ~ 39 s:! 5 5avalanche days 39 81 118 2ED 42) 39 100 3 4

non-avalanche days ::6 ::6 55 a:> 110 0 0 7 6

Table 1. Number of days with avalanche activity, precipitation totals, snowfall amounts, and days since lastsnowfall for 75 rain-on-snow storms, Alpine Meadows Ski Area, California.

529

Figure 1. 75 rain-on-snow storms at Alpine Meadows SkiArea, California.

Heywood's research suggests that avalancherelease is more likely when snow falls within threeto four days previous of rain. We found after threeor four days since snowfall, avalanche activityoccurred during less than or equal to 50 percent ofthe events. The mean number of days since lastsnowfall for the avalanche producing storms was 3days; for the non-avalanche storms, 7 days. Of the75 warm storms, 42 had snowfall within 3 days ofthe rain. Avalanche activity was observed during29 (69%) of these days. Of the remaining 33storms, 14 (42%) released avalanches.

80 Y = -3.76x + 61.44?:' •.:;:;:; • r2= 0.38()

0

<l> 60.c •()

c • •00> 40 •0

.c :2 15 days-~

·3(fJ 20 •>, •0

".-......0

~ 0 • • •

(74%) had snowfall mixed with the rain. In contrast,little more than half (54%) of the non-avalancheproducing storms had some snowfall. The mostsignificant difference between the two comes fromexamining their mean characteristics. On average,more than twice the amount of precipitation fellduring the events that produced avalanches thanthose that did not; snowfall was better than fivetimes as much. This strongly suggests that newsnowfall is still the overriding factor in producingavalanche activity, even during rain-on-snowevents. It should be pointed out that the 20 rainstorms that did not have any snowfall had thelowest mean storm precipitation: 30 rnm. Both thehigh and low io data sets reveal similar amounts ofmean snowfall (26 cm vs 20 cm, respectively), butthe standard deviation of the data sets variesconsiderably more (53 cm vs 14 cm). Figure 1represents all 75 rain-on-snow storms categorizedby number of days since the last snowfall. Thesedata are plotted against days of avalanche activity.

5. NEW YEAR'S STORM 1997, SOMEOBSERVATIONS

Although rain-on-snow events in the SierraNevada usually affect only starting zones below2500 m, rare storms with very warm temperaturescan deliver rain above timberline. On occasion,rain has been observed at great elevations in thehigher, southern part of the range.

At the snow research station (2930 m) onMammoth Mountain in the eastern Sierra Nevada,documented rain events include: January 1980when rain was recorded for five days (Davis andMarks 1980); and April 1982, when a storm thatdelivered rain up to 3000 rn-soon after a majorsnow storm-initiated a wet avalanche cycle in theSierra Nevada. Since then, mid-winter rainfall atthis site has been rare. A few millimeters of rainfallwere observed at this site during both the 1995and 1996 winters, and 70-80 mm of rain fell thereduring May 1996 (Kattelmann 1997). The NewYear's 1997 storm-a severe storm with intenseprecipitation and unusually warm temperatures­deposited more than 200 mm of rainfall atelevations above 3000 m but did not produce amajor avalanche cycle. The high rain/snow levelsprovided an opportunity to observe snowpackresponse to rainfall at high elevations where suchwarm mid-winter temperatures have rarely beenobserved.

At Mammoth Mountain, the season's snowpackbegan to accumulate with a storm on November21-22, 1996 that deposited more than 90 cm ofsnow at the study site. Three storms duringDecember each deposited more than a meter ofsnow at this site. Precipitation during December1996 was more than twice the average for themonth at almost all recording sites in the SierraNevada. Direct-action avalanches were commonbecause of the intense snowfall.

The series of storms that became progressivelywarmer began on December 26 with fluctuatingrain/snow levels averaging around 2000 m.Warmer air began entering the Sierra Nevada onDecember 29, and rain climbed to higher altitudes.Temperatures continued to rise, and precipitationintensified over the following four days. OnJanuary 1 and 2, 1997, air temperatures rangedfrom +2 to +4° C at the Mammoth Mountain studyplot. These high temperature measurementssuggest that rain was falling above 3000 m for atleast 36 hours and up to 3500 m for perhaps 12hours.

At the study site, steady rain began shortly aftermidnight on January 1. There may have beenminor amounts of rain mixed with thepredominantly solid precipitation as much as sixhours earlier. Rainfall intensities varied between 1and 6 mmlhour on January 1 and increased on

8 10 12 146

number of days since snowfall

2 4o

530

January 2 with a few periods of more than 10mm/hour. About 220 mm of rainfall was measuredduring the storm. Windspeeds averaged over 15­minute intervals varied from 4 to 13 m/s. Beforethe rain started, the snowpack was about 2.5 mdeep with an average density of about 300 kg/m3

and temperatures of -1 to _5° C. The first water thatpercolated through the snowpack took about 9hours to reach one of nine snowmelt Iysimeters atthe site (Kattelmann 1997). Snowpits wereexcavated at several sites on and near MammothMountain following the storm. The rainfallproduced a very complex layer structure in the top30 cm at all sites. In this near-surface region, wetzones alternated with ice lenses and dry zones.Thickness of the different zones varied across pitprofiles and between pits. Some of the wet zonesand ice lenses appeared to be only a millimeter ortwo in thickness. Other--ice lenses and complexesof ice lenses were up to 50 mm thick. On flatground, the intricate stratigraphy continuedthroughout the profiles to the soil surface.However, there were a few dry zones of up to 12cm thick. Snow temperatures were 0° C on levelground. Below the near-surface region, snowpitsexamined on sloping ground differed markedlywith their level counterparts. Evidence of waterflow below the top 30 cm was either lackingaltogether or present in bands of 5-10 emthickness at the presumed interfaces betweensnow layers deposited by different storms. A fewirregular ice blobs were found in the otherwise drylayers. Temperatures of these layers remainedbelow _3° C. Although these snowpits provideonly a very limited sample of snowpack responseto rain during this one event, they suggest thatwater bypassed much of the deeper snowpack onsloping terrain as it flowed down slope in the near­surface region. Most slopes in the region werecovered with surface expressions of a trellis-

":;;. pattern rill network after the storm (Kattelmann" 1997). These complex drainage networks may

have routed enough of the percolating rain waterthrough the snowpack, minimizing structuralinstability during this event.

6. CONCLUSIONS

Warm mid-winter rain storms strike the SierraNevada a couple times a year. Historically, manyseason's greatest storms have been warm, rain-on­snow events. Several of these storms havecaused landslides, extensive flooding, andwidespread avalanche activity throughout therange. Observations from 75 rain-on-snow stormsat a ski area in the central Sierra Nevada suggestthat avalanche activity is more imminent whenrainfall is accompanied with snowfall, especially ifnew snow falls within a couple days before the rain.

531

For 55 rain storms that had accompanying snowfall,avalanche activity seemed to be independent ofamount of new snow, even though the storms witha low percentage snowfall had a mean stormprecipitation value far exceeding those with moreconcurrent snowfall (111 mm vs 39 mm).Avalanche activity was more pronounced duringevents with mixed rain and snow than those withjust rainfall, but the "rain-only" storms had lessprecipitation overall.

Though some natural avalanche activity wasobserved during the 75 storms investigated, wedetermined avalanche release during orimmediately after rain-on-snow storms mainly bythe snowpack's positive response to explosivecharges. Other than naturally occurringavalanches, no evidence is available to determineif the snowpack would have avalanched otherwise.Ski areas have almost constant observers, andsince naturally occurring avalanches are relativelyrare (compared with the avalanche activity that'sexplosive-induced within a ski area), real-timeobservation of a snowpack's behavior with respectto stability remains somewhat limited to within­bounds.

Warm, wet snowpacks behave more like a fluidcompared to their colder, drier counterparts thatexhibit brittle tendencies. Consequently, oldernear-isothermal snowpacks respond poorly toexplosive-induced avalanche release (Heywood1988). Using explosives as a yardstick by which tomeasure the stability of snowpacks under rainfall istherefore limiting. Though somewhat rare, deep­slab instability in wet Sierran snowpacks doesoccur (e.g. 1989, 1995), most commonly duringspring melt and rain-on-snow. Explosives may bean ineffective gauge of stability for deep, wetpacks. Analysis of the 75 storms is thereforebiased toward the avalanching of storms with newsnow. Of course, deep, wet slab releases within aski area's boundary-explosive-induced or not­would most certainly have been noted. Since theforecasting of avalanches is of greatest concernwherever (and whenever) they cross paths withpeople or their property-namely ski areas,highways, and structures-using in-boundsavalanche control as a meter of snow stabilityremains valid-and in fact necessary. Morequalitative analysis is needed to determine howsusceptible any particular snowfield is to rain­induced avalanching-or not. Additional researchis needed to link snowpack pre-storm and post­storm stratigraphy with avalanche response to rain­on-snow, both for natural and mechanically altered(ski area) snowpacks.

Snowpit observations within level snowpackson and around Mammoth Mountain in the easternSierra Nevada following a rain-on-snow storm inearly January 1997 reveal evidence of rain water

movement through the entire snowpack. After thesame storm, inclined snowpacks exhibited wettingof primarily the near-surface snow layers. Withwater movement mostly confined to the nearsurface layers, free water-induced weaknessesmay have been restricted to the upper layers of thepack. Though widespread and/or destructiveavalanches were not observed during this storm,precipitation was intense, rain/snow levels werehigh, and flooding was severe and extensivethroughout the range.

The majority of avalanche activity in the SierraNevada is due to significant amounts of snowfall.Rain-on-snow storms that produce avalanches aremore prone to do so when snowfall is mixed withrainfall. But because some rain storms do notcause avalanching-even with new snow-onecould infer that for these events the rain is notintroducing enough of a decrease in shearstrength to cause instability. Since large amountsof water introduced into a parcel of snowdecreases its mechanical integrity, the deeperlayers of the snowpack during non-avalancheproducing rain storms must be directing liqUidwater away from shear planes. Conversely, rainstorms that do produce avalanches may be routingliquid water laterally over impenetrable layers actingas shear planes for overlaying snow slabs. Of the20 "rain-only" storms analyzed, the eventsproducing no avalanches had (on average) greaterprecipitation than those that did cause slides.Lesser amounts of percolating water unable topenetrate hard snowpack layers-while stillweakening interfaces-may be one explanation forthis. The presence of flow fingers or otherconduits to efficiently move water through thepack are an indicator that the pack has previouslytransferred some amount of free water. Thesesnowpacks would be less inclined to route waterover ice lenses, hard layers, crusts, or otherpotential shear planes. Establishing whether a

~ snowpack has developed subsurface flowchannels and/or has transmitted appreciableamounts of liquid water could be an importantcomponent when forecasting snow stability duringrain-on-snow.

ACKNOWLEDGMENTS

" Larry Heywood made available much of the rain­on-snow data. Mike Pechner supplied 500-mblevel circulation maps.

7. REFERENCES

Azuma, D., Berg, N. 1990. Precipitation Extremesat the Central Sierra Snow Laboratory,

California. Proceedings of the WesternSnow Conference 58: 141-144.

Berg, N. 1982. Layer and"crust development in acentral Sierra Nevada snowpack.Proceedings of the Western SnowConference 50: 180-183.

Berg, N., Osterhuber, R, Bergman, J. 1991.Rain-induced outflow from deepsnowpacks in the central Sierra Nevada,California. Journal of HydrologicalSciences 36: 611-629.

Bergman, J. 1987. Rain-on-Snow and Soil MassFailure in the Sierra Nevada of California.Earth Resources Monograph 12, USDAForest Service, San Francisco, CA. p.15­28.

Davis, R, Marks, D. 1980. Undisturbedmeasurement of the energy and massbalance of a deep alpine snowcover.Proceedings of the Western SnowConference 48: 62-67.

Harr, R. 1981. Some characteristics andconsequences of snowmelt during rainfallin western Oregon. Journal of Hydrology53: 277-304.

Heywood, L. 1988. Rain on Snow AvalancheEvents. Some Observations.Proceedings of the International SnowScience Workshop. p.125-136

Kattelmann, R 1984. Wet Slab Instability.Proceedings of the International SnowScience Workshop. p.102-108.

Kattelmann, R 1985. Macropores in snowpacksof Sierra Nevada. Annals of Glaciology 6:272-273.

Kattelmann, R 1986. Measurements of snowlayer water retention. Proceedings of theCold Regions Hydrology Symposium.p.377-386.

Kattelmann, R 1987. Methods of EstimatingliqUid Water Storage in Snow.Proceedings of the Western SnowConference 55: 158-161.

Kattelmann, R 1997. Very Warm Storms andSierra Nevada Snowpacks. Proceedingsof the Western Snow Conference 65:125·129.

Kattelmann, R, McGurk, B. 1989. SnowpackChanges and Water Release at a LowElevation Site in the Sierra Nevada.Proceedings of the Western SnowConference 57: 141-144.

Kattelmann, R, Berg, N., McGurk, B. 1991. AHistory of Rain-on-Snow Floods in theSierra Nevada. Proceedings of theWestern Snow Conference 59: 138-141.

McClung, D., Schaerer, P. 1993. The AvalancheHandbook. The Mountaineers, Seattle,WA. p. 56-73.

532

McGurk, B., Kattelmann, R. 1988. Evidence ofliquid water flow through snow from thicksection photography. Proceedings of theInternational Snow Science Workshop.p.137-139.

Osterhuber, R. 1997a. Climatic Summary ofDonner Summit, California 1878-1996.Presented at the Changing WaterRegimes in Dry Lands Conference, LakeTahoe, June 1997. On file at the CentralSierra Snow Laboratory.

Osterhuber, R. 1997b. Annual Rain-on-SnowEvents in the Central Sierra Nevada,California. Presented at the Fall Meetingof the American Geophysical Union, SanFrancisco, CA. On file at the Central SierraSnow Laboratory.

Pechner, M. 1998. Meteorologist, Golden WestMeteorology. Personal communication.

Wilson, N. 1986. A Widespread Cycle of UnusualAvalanche Events. Proceedings of theInternational Snow Science Workshop. p.153-154.

533