|Coordinates (Swiss map 1217 Scalettapass):||Release area approx.||778'850 / 173'350 / 2300|
|Toe of deposit approx.||779'500 / 173'800 / 1910|
||Saturday, February 21, 2004, approx. 10 AM|
||Tuesday, February 24, 2004, approx. 12–17 h|
||Hansueli Gubler (HG), Dieter Issler (DI), Bernardo Teufen (BT).|
Inneralp (1870–2000 m a.s.l.) is one of the summer settlements and pastures belonging to Monstein, the southernmost village of Davos. It is reached in a little less than an hour by ski from Monstein.
The release was not directly observed. Leading a party of tourists, BT passed Inneralp on Saturday, Feb. 21, 2004 around 9.30 AM and did not notice any traces of avalanches. Forced to return around 10.30 AM because of tempestuous southerly winds at high altitudes, he found that several avalanches had been released along the mountain slope west of Mittelalp and Mäschenboden (the middle and uppermost parts of Inneralp, respectively) in the meantime. The avalanche investigated in this report was the largest among these and caught BT's attention because it had a quite long runout on almost level terrain.
In the middle of February, the snowcover stability in the Davos area was generally high, except in a few steep north-facing gullies that were loaded with transported snow. On Saturday, February 20, a very strong föhn wind began to dramatically warm the snowpack and to transport large quantities of snow from south-facing to north or northeast-facing slopes. Sunday was exceptionally warm for this season, with temperatures above 0°C up to about 3000 m a.s.l. At lower altitudes, the danger of spontaneous wet-snow avalanches increased sharply. In fact, a series of such slides took place on the southwest-facing slopes just northeast of Inneralp.
During the night to February 23, the temperature dropped again when a north-westerly wind brought about a moderate snowfall (10–15 cm). The fresh snow bonded quite well to the warm and moist snowcover surface, and low temperatures during a clear night helped to stabilize the snowpack. On Tuesday morning, February 24, it was deemed safe to cross the south-facing slopes on the trail leading from Monstein to Inneralp. During the field work, the weather was constantly changing with regard to wind, sunshine and precipitation, but the temperature was fairly constant around −10°C.
Fig. 1. Map of the Inneralp avalanche, February 21, 2004. Equidistance is 10 m, the colors indicate classes of slope angles for open terrain (light yellow/pink/ocre/brown) and forest stands (light/dark green). The blue line indicates the conjectured release area, the red line corresponds to the deposit of the dense part whereas the green line shows where the fluidized part ran farther.
The release zone of the avalanche was not visible on February 20. By February 24, the very strong southerly winds had erased all traces of the crown. To judge from the map (Fig. 1), the most likely starting zone is between 2200 and 2300 m a.s.l. because a gentle ridge in the otherwise open slope probably caused deposition of considerable quantities of snow. The mean inclination of the assumed starting zone is approximately 33°, its projected area about 1 ha. It is planned to investigate the morphology of the starting zone in the summer of 2004.
Below 2200 m a.s.l., the avalanche track is channeled in a V-shaped gully (mark 1 in Fig. 2) about 50–70 m wide and 10–15 m deep. From 2200 to 2070 m a.s.l., the track inclination varies slightly about 31°. The track becomes somewhat steeper from 2070 to 1970 m a.s.l. (35–40°) and makes a noticeable turn at 2000 m a.s.l. On the ridges between this path and the adjacent paths, a forest (mainly spruce) has established itself up to 2060 m a.s.l. to the north, and to 2020 m a.s.l. to the south. Bushes and small larch trees grow on the flanks of the gully.
At 1950 m a.s.l., the transition from the gully to the alluvial fan takes place, the slope inclination decreasing steadily. At 1910 m a.s.l., the main branch of the avalanche reached a small stream flowing west of Mäschenboden, ascended slightly onto the moraine and stopped only a few meters short of the small road and about 20 m northwest of a stable. The runout angle is α = arctan(H/L) = 28.5° for the fluidized part and 30° for the dense part, where H and L are the drop height and the horizontally projected distance from crown to toe of the avalanche, respectively.
The avalanche was a dry-snow avalanche, as evidenced by the clearly visible front of the dense-flow deposits (mark 4 in Figs. 2 and 3) and an extended area of shallow deposits with snow blocks of various sizes distributed randomly on them in the distal area1 (mark 5). The avalanche may have had a powder-snow component, but no direct evidence could be found in the field because of the fresh snow that had fallen and the melting that had occurred since the event.
Some of the main features of this event are clearly visible on the photos (Figs. 2 and 3) and are also contained in the detail map (Fig. 1): A large fraction of the dense-flow deposits are contained in a “bulge” on the right side of the alluvial fan (as seen in the flow direction) that is about 10–20 m wide and extends from the beginning of the fan to about 30 m beyond the stream (mark 2 in Fig. 2). It is fairly straight, but shows several step-like changes in depth (mark 3), as if the avalanche had flowed in a number of surges.
Another clearly defined and rather narrow “bulge” is discernible on the left side of the alluvial fan (mark 2), but it did not reach as far as the first, and it bends towards it in a wide arc at the end. Beyond this arc, a field of shallow deposits is found (mark 5), as in the distal direction from the major deposit.
The major “bulge” is flanked on its right by shallow deposits (mark 5) that are, however, much more irregular than those mentioned previously.
In the area between the major and the secondary “bulge”, the snow surface is quite smooth. Two young larch trees (mark 6), about 3–4 m high and still fairly flexible, grow in this area and did not show signs of damage due to this avalanche. However, they are almost devoid of branches on the side exposed to the flow.
In the gully above the transition zone to the alluvial fan, the avalanche flowed clearly on the outside of the curve, ascending approximately 2–3 m on the flank. This was evidenced by a sharp boundary line, below which the avalanche had eroded the snowcover. On the inside of the curve, deposits not unlike levees in debris flows are found.
A small release occurred on the right flank just at the end of the gully. It was most likely induced by the main avalanche that passed underneath, eroding away the snowcover and destabilizing the steep area above it.
Due to the snowdrift in the release zone, safety considerations and time constraints, the investigations were limited to the runout zone, i.e. to the area up to 2000 m a.s.l. The focus was on understanding the flow mechanisms, the shape of the deposit and the erosion processes in the transition from track to deposition zone.
Fig. 4. Approximate location of snow pits AD and trenches E and F. See further description in the text.
To this end, we dug four snow pits and two trenches, the locations of which are indicated in Fig. 4. The pits A and B are in the area of shallow deposits downstream of the front of the dense-flow deposits. Pit C is located 1 m upstream of the two small larch trees between the two “bulges” in order to understand how this area was affected by the avalanche. Finally, pit D investigated the deposit structure in the transition from the gully to the alluvial fan. No pits were dug higher up in the main flow channel because the avalanche had eroded the snowcover to the ground, which was now only covered by the fresh snow.
The location of trench E was chosen at the outer edge of the zone of shallow deposits, and it was oriented perpendicular to the main local flow direction so as to highlight the lateral tapering off of the deposit. Trench F is a section across half of the secondary “bulge”.
||Fig. 5. Visualization of deposit texture in a snow pit: A mixture of writing ink and 2-propanol alcohol was sparingly sprayed onto one pit wall. HG warms the surface with a gas burner to enhance the migration of the ink into the pores. About 15 cm of new snow were removed to make the rough surface of the deposits of the fluidized part visible. Photo: D. Issler.|
Pits A, C and D served only to obtain qualitative profiles. The vertical density profile was recorded in trench F. In pit B and the two trenches, the texture of the deposits was made visible by first spraying the vertical wall with ink and then briefly heating it with a gas burner similar to those used for ski waxing (Fig. 5). The heat accelerates the migration of the ink through the interstitial pores. Where those pores are narrow, capillary effects suck the ink deeper into the snow and make it appear darker. The coloring of the pore spaces makes the texture and layering of the snow much better visible and thereby helps in distinguishing between the undisturbed snow cover, characterized by its fine striations, and avalanche deposits, which are more isotropic at small scales and may contain inhomogeneities (embedded particles) at larger scales.
This technique has already been successfully applied by HG in the 1990s. It turned out, however, that great care is necessary in choosing the material and especially the ink: At the low temperatures that persisted throughout the day, ink particles flocculated and fell out of the suspension, clogging the narrow tube leading to the nozzle. As a result, time was lost in cleaning the tube and nozzle several times, and the resulting spray was not as fine as desired. Yet even with these restrictions, this simple method proved very effective and useful.
Pit A: The location of pit A was chosen near the outer edge of the shallow deposits in the distal direction from the main “bulge” where a rounded block of snow with approximate dimensions 40×25×20 cm³ was visible on the snow surface. As was true for all other snow pits and the two trenches, the surface of the avalanche deposit underneath the new snow was quite rough at a scale of 1–5 cm because of a large number of small irregular-shaped but rounded snow clumps. The depth of the deposit was approximately 20 cm. Only a few clumps embedded in fine-grained snow could be found inside the deposit. Its lower boundary was marked by an abrupt change in hardness at the transition from the compacted deposit to the preexisting snowcover. At most pit locations, organic tracers such as larch needles were found occasionally or even in moderate numbers in the avalanche deposit, but not in the new snow above or the old snow below.
This location is quite exposed to winds blowing up or down the valley so that the depth of the pre-avalanche snow cannot be considered representative. It is therefore not possible to draw conclusions about whether or not some of the old snow was eroded by the avalanche sweeping over it.
||Fig. 6. Deposit structure at location B: About 30 cm of the original snow cover remain, the boundary between undisturbed snow cover and deposits being clearly visible. This deposit created by fluidized flow consists of a matrix of (unlayered) fine-grained snow with embedded, rounded snow clods of various sizes. Only clods with diameters below 10 cm were present in this pit, but very much larger blocks were found on the surface nearby. Photo: D. Issler.|
Pit B: This location was chosen about 15 m from pit A, but a little closer to the front of the dense-flow deposits. Only 2 m from the pit, two snow blocks of quite unusual dimensions were found. The larger of the two measured about 80×50×60 cm³ (Fig. V) and had an estimated weight of 50–60 kg., the other was smaller by a factor 2–3 in volume. The large block was disected and found to be homogeneous, consisting of fine-grained snow. In contrast to the smaller blocks found on the surface of the avalanche deposits, this one extended through the entire depth of the deposit.
The coloring of the snow profile in pit B (see Fig.6) confirmed the general findings from pit A: The deposit was homogeneous except for embedded rounded particles of up to 5 cm in length. It did not show the texture typical of a layered snowpack, which was apparent in the much softer snow below the deposit. The deposit was approximately 40–45 cm deep.
Pit C: The total snow depth at this location was about 1 m, of which 10 cm were fresh snow. Underneath this was a 15 cm deep layer of fine-grained snow with small snow particles embedded. The avalanche deposit appeared a little less hard than at the previous two locations, but it was still significantly harder than the old snow underneath. A person in ski boots could mostly, but not always walk on it without breaking through the deposit.
Pit D: The main difference from the previously described snow pits was that the avalanche deposit (about 25–30 cm deep, with a somewhat larger fraction of visible snow particles) did not rest on the old snow cover, but on about 60 cm of debris from an earlier wet-snow avalanche that had eroded the snowcover to the ground Accordingly, that old debris consisted almost exclusively of icy particles with typical diameters between 3 and 20 cm, which contained inclusions of organic material (leaves, needles, twigs, soil). The spaces between the irregular-shaped particles were not filled with fine-grained snow.
On Saturday, before the release of the more recent avalanche, this deposit was not visible under the snowcover. The latter must therefore have been eroded in its entirety. Since it is not known when the first avalanche occurred, the amount of snow eroded by the second avalanche cannot be determined.
||Fig. 7. Deposit texture in the very distal area of the deposits from the fluidized flow (trench E). The trench is oriented perpendicular to the flow direction and shows the tapering off of the deposit towards the right side. The application of dye made the approx. 10 cm thick layer visible that is located between the undisturbed snow at the bottom and the avalanche deposit (characterized by embedded snow clods). Photos: H. Gubler.|
Trench E: Over a distance of about 4 m, the decrease of deposit thickness from about 0.6 m to almost 0 was immediately apparent, but it proved to be surprisingly delicate to determine the precise location of the interface between old snow and avalanche deposit. Hand tests revealed an interface where the hardness of the snow changed abruptly, the softer snow being below it. When the dye was applied, however, another interface running in parallel about 10 cm above it was discovered. The snow in between exhibited some fine layering (Fig. 7). Combining both observations, one is led to the conclusion that this intermediate layer most likely does not represent part of the avalanche deposit. One may conjecture that its hardness is due to wind action, as that area is quite exposed to strong winds.
Trench F: The starting point of the trench was selected just upstream of the visible upper edge of the deposit, approximately 15 m from the two larch trees shown in Figs. 2 and 3. The excavation line followed the prolongation of the main flow direction and reached to the middle of the “bulge”; its length was approximately 6 m. At the upper end of the profile line, the transition from a very thin deposit similar to that found at pit location C to a much harder type of deposit rapidly growing in depth was quite abrupt. The ground was found to slope downward in the main flow direction. Near the downstream end of the trench, the avalanche deposit was 1.6 m deep and rested on 0.75 m of old snow.
A quite conspicuous property of the avalanche deposit was its hardness throughout its whole depth, corresponding to the highest degree (“knife”) on the qualitative scale. It proved to be time-consuming to excavate the trench manually; a steel-spade was required. Visually, the deposit was quite homogeneous except for embedded snow blocks, most of which were less than 10 cm in their largest dimension, and fir needles or blueberry leaves. Near the deepest part of the profile and close to the bottom of the deposit, a larger block of ice was found; it was probably ripped out of the deposit of the wet-snow avalanche discovered at snow pit D. The grain size in the deposit was about 0.1–0.3 mm; grains in the old snow underneath were angular with typical size 0.5 mm. The embedded snow particles consisted of somewhat larger grains. When the profile was dyed, these findings were confirmed: The layering in the old snow was quite apparent, but it was absent in the deposit.
snow cover [m]
|Snow water equi-
||Fig. 8. Density profile in trench F. The green part of the line represents the snow cover left intact by the avalanche (value extrapolated from measurement at 0.65 m).|
The density profile, Table 1 and Fig. 8, is rather constant, except for the much lower value at 1.15 m above the deposit bottom. Strangely, neither a corresponding variation in the apparent hardness nor any textural variation was found at this depth. The old snow underneath the avalanche deposit was compressed somewhat, but nevertheless was clearly less dense than the avalanche debris, as expected.
Early in the winter, a wet-snow avalanche descended the path but
stopped around the end of the gully. It eroded the existing snowcover
in its entirety and mixed in some organic material. At the same time it
deposited continuously between 0.4 and 0.8 m of very granular snow with
quite large voids. As Fig. 2 shows, the
distal end of the deposit cannot have been very deep.
No other information on this avalanche is available, but on the basis of many other observations in this project, we may conjecture that this first avalanche flowed as a dense granular flow. The humidity of the snow cover (at least in the lower part of the track) prevented the transition to fluidized flow.
The avalanche of February 21, 2004 was a dry-snow avalanche with a significant fluidized part that represents 5–10% of the total deposit mass. The runout distance of the fluidized part exceeded those of the dense part by up to 40 m while the runout angles differed by about 1.5°.
The fluidized part exhibited a surprising competence for rafting along large snow blocks weighing over 50 kg. The mechanism that enables this is unclear at present:
Open questions remain also in connection with the shape of the dense deposit. In particular the left branch is unusually short (about 15 m) and wide (more than 40 m). If the compression failure surfaces in the right branch deposit are indeed perpendicular to the flow direction, as we conjecture, a similar ratio obtains for the right branch as well.
A related enigma is the leaf-shaped zone between the two deposit branches that comprises the two small larch trees: It should have been overflowed by both the dense and fluidized parts of both branches, yet shows very little erosion and/or deposition and very moderate pressure that did not significantly damage the trees. One might speculate that the entire avalanche was essentially fluidized in the track, but most of it became dense again shortly before the mass stopped. As long as the mass was fluidized and its density low (approx. 30 kg/m³, say), the stagnation pressure should not have exceeded 10 kPa. However, we expect a stagnation pressure of 10 kPa would still damage the trees much more than was observed. Furthermore, the transition to the dense flow regime and the stopping would have had to take place over a very short distance. For example, with an initial speed of 25 m/s at the larch trees and a stopping distance of 25 m, a rather strong mean deceleration of −10 m/s² is required.
Oslo, March 3, 2004; revised and enhanced May 5, 2008