NATURAL HAZARDS IN MOUNTAIN COLORADO

JACK D. IVES, ARTHUR I. MEARS, PAUL E. CARRARA, AND MICHAEL J. BOVIS*

ABSTRACT. Interdisciplinary field studies and remote sensing techniques were used to delineate mountain areas in Colorado subject to such natural hazards as snow avalanches, mudflows, rockfalls, and landslides. The old mining townsite of Ophir in the northwestern San Juan Mountains was used as a case study. Its serious snow avalanche hazard has been made even more critical with prospects of new housing developments. Techniques in remote sensing and geoecology have been applied to the solution of practical land management problems at the county and township levels of local government. The rapidly increasing hazard to human life and property results directly from accelerated growth of the winter recreation industry and construction of mountain homes. Many of the world's temperate zone high mountains urgently need development and application of new land management policies. KEY WORDS: Avalanches, Geoecology, Hazards, Land management, Mountains, Remote sensing.

THE mountain section of Colorado has experienced accelerating pressures from rapid development of the recreation industry, principally winter sports expansion and the spread of second homes. The population explosion along the Front Range urban corridor over the past ten years has induced the completion of the Eisenhower Tunnel bypassing Love land Pass and bringing large sections of Summit, Eagle, and Pitkin counties within two to three hours' driving time of Denver; the twinning of I-70 (partially complete); and the creation of a new type of boom town, the ski resort, as exemplified by Vail.

The inflow of population has placed large numbers of people with little or no mountain experience in high mountain terrain. Land values exceed $70,000 per acre in some of the more attractive sites, and land speculation is rife. Only a limited amount of land in the Rocky Mountains is suited for home and condominium construction. The inevitable result- a combination of speculation, ignorance, and the very speed of the development itself-has been land sales and actual construction in areas subject to a variety of natural hazards: avalanche, landslide, mudflow, rockfall, and mountain flood.

Over the past three years the Institute of Arctic and Alpine Research (INSTAAR) has been seeking to develop methodologies, including a combination of remote sensing techniques and interdisciplinary field studies, to assist governmental agencies at the township, county, and state levels to alleviate this serious land management problem. The initial studies were conducted near Vail, with smaller scale studies in Telluride, Crested Butte, Silverton, and Ophir. The special situations and problems of Ophir, San Miguel County, provide an excellent case study to demonstrate the methodologies used. No new development has taken place near Ophir, and the use of these methodologies to prepare hazard maps can give local planning authorities a better opportunity to control future growth patterns. The possibilities for the success of such an approach were greatly augmented in 1974 with the passage of Colorado State House Bill 1041 which, in part, requires each county to prepare maps of land subject to a variety of natural hazards. The legislative step has been reinforced by the development of hazard criteria and definitions by the Colorado Geological Survey.¹

OPHIR AND SPRING GULCH

Ophir is one of many relics of the early Colorado mining boom. During most of the present century it has remained a small, almost forgotten, group of houses with a total new migrant population of fewer than thirty persons. The town is in the northwestern San Juan Mountains at an elevation of 2,973 meters, 9.6 kilometers south of Telluride, the county seat of San Miguel County and the site of a recent ski development. Mountain ridges exceeding 3,962 meters separate the two settlements (Fig. 1). Ophir occupies part of the floor of a spectacular glaciated valley which is drained by Howard Fork, a tributary of the San Miguel River. The townsite is north of the stream on the western sector of a large alluvial fan emanating from Spring Gulch.

FIG. 1. Location of the Ophir-Telluride
area, San Juan Mountains,
southwestern Colorado.

The main source of avalanche hazard is Spring Gulch (Fig. 2 and PLATE I). The total vertical range of its catchment basin is 1,100 meters from the summit of Silver Mountain (4,100 meters) to the vicinity of Ophir, making it one of the largest in Colorado. The snow accumulation zone above 3,300 meters is almost entirely above treeline. Less than fifteen percent of the total area is too steep to accumulate a deep snowpack. Most of the accumulation basin consists of smooth slopes with average gradients of 30 degrees to 40 degrees. Much of this basin could probably release simultaneously, given appropriate snow and weather conditions. These steep, smooth slopes also have many active mudflow channels and extensive areas of soil creep, indicating instability that would provide serious difficulties for any future attempt to construct supporting structures to anchor the snowpack.

Below about 3,300 meters the mass of moving snow released from the accumulation basin (starting zone) becomes concentrated into the deeply entrenched channel of Spring Gulch, which serves as the avalanche track. All avalanches, regardless of type or size, utilize this channel, which has an average gradient of 26 degrees (45 percent) between 3,400 and 3,150 meters. Cross sections of previous avalanches have been surveyed (Fig. 3). The cross section of the April, 1973, wet snow avalanche indicates that the major powder avalanches of the past were much larger, partly because of the turbulent, high-velocity powder cloud which is assumed to have accompanied them. Measurement of broken trees along the margins indicates that the depth of the destructive moving fronts of past major events exceeded sixty meters.

The lower part of the Spring Gulch catchment basin (run-out zone) is a gently undulating alluvial fan. The undulations, with low ridges approximately perpendicular to the contours, are the result of numerous mudflows and/or debris flows. A local relief on the order of two meters is of considerable importance for wet snow avalanches, but has much less effect on dry snow events. A small stream channel extends from the apex of the fan down its western edge, and a steep-sided gully cuts into the surface east of the center line. Occasional conifers grow near the town, south of the county road, and in the upper part of the stream channel, which also contains patches of aspen and willow. Otherwise, the alluvial fan is treeless, although the eastern forest border (PLATE I) is abruptly uneven and indicates that timber probably has been cut in the past. The color infrared air photograph gives an excellent overview of the townsite and the immediate hazards that threaten it. Coniferous forest (dark red on the photograph) can be distinguished from the aspen forest cover, which should always be viewed as an indicator of potential instability. The linear patterns in the vegetation, perpendicular to the contours, are diagnostic as a preliminary sign of avalanche hazard. This type of photograph has been a vital tool in all phases of the natural hazard delineation.

The present residents have come to Ophir over the last three years. They have reincorporated the town and have formed a small but very active group of modern "mountain men" who obtain their livelihood largely by working in Telluride. In addition, the land owners, deriving their land from early mining claims, are moving to place many housing lots on the market; ski resort speculation is apparent, and thus the ingredients for serious problems in local planning are already assembled. This study was requested both by the people of Ophir and by the San Miguel County Planning Office.

FIG. 2.  Generalized avalanche hazard map of the Howard Fork Valley between Ophir and Ophir Loop. 
The arrows indicate location of minor paths, principally wet, spring avalanches.  Delineation of secondary
hazard to the access road incorporates detail from work by Arthur I. Mears for the Colorado Geological Survey.

THE PROBLEM

A reconnaissance of Ophir and Howard Fork Valley in September, 1974, indicated that the major hazards threatening the existing houses, and especially the undeveloped area of the platted townsite to the east, were periodic wet and dry snow avalanches from Spring Gulch. The inhabitants were also in danger from avalanches crossing the access road between Ophir and Telluride. Secondary hazards include the Waterfall Avalanche path, which ran and temporarily knocked out the town's water supply in January, 1975; a series of small avalanche paths north of the town and west of Spring Gulch; and a variety of mudflow, debris flow, rockfall, and associated problems (Fig. 2 ). A growing tendency for cross-country skiers to use Ophir as a car park and ski up the valley toward Ophir Pass constitutes an additional hazard not considered in the present study.

The difficulties of assessing avalanche magnitude and frequency (recurrence interval) in areas such as the European Alps, where hundreds of years of historical data are available, are formidable.² In Colorado historical data frequently are entirely lacking, especially for Ophir, where none of the present residents has lived in the area for more than three years, although we have some information dating back to the early years of the century. Since the physical properties of snow vary rapidly in time and space, the difficulty of predicting avalanche size is basically a problem of inadequate snow mechanics theory. Two main forms of torrential snow mass movement must be considered: dry powder avalanches, sometimes accompanied by an airborne powder cloud, that may travel up to 120 m/sec (250 mph); and wet snow avalanches that travel much more slowly (up to 22 m/sec, or 50 mph), but also produce formidable pressures in the run-out zone.³ Assessment of hazard must consider the maximum possible run-out zones both of wet and of dry snow avalanches, recurrence intervals, and probable pressures in the run-out zone. Two extreme cases would be one in which an avalanche discharges at least once each winter and one in which infrequent occurrence-perhaps less than once in 100 years-even allows reafforestation of the track and run-out zone. The first should be so self evident that it is usually avoided automatically, but the second type may escape recognition. Serious loss of life and property may result in areas such as Colorado, which have rapid population growth and few historical data. On the other hand, the indirect methods of prediction, if indicating a recurrence interval of more than 100 years, may limit otherwise usable land and will probably be more difficult to maintain in a legal action, given the obvious margin of error in interpretation of the field data. The concept of the 100-year avalanche (best described as a one percent chance of an avalanche in any one year) has not yet remotely attained the legal and planning respectability of the 100-year flood.

FIG. 3. Cross sections of three Spring Gulch avalanches.


This study used indirect and direct field methods, applied available, albeit imperfect, flow laws, and used any historic data that could be collected from interviews with local residents. To the problems of determining the magnitude and frequency of natural catastrophic events must be added the challenge of translating the research results into meaningful recommendations so that the responsible decision-makers can improve mountain land management within the limits set by the democratic process of local government. We recognized that snow avalanches were the major source of hazard facing Ophir. Although other natural hazards, including mudflow, debris flow, rockfall, and mountain flood, are present, major emphasis had to be placed on the determination of avalanche magnitude and frequency.⁴



PLATE I. Color infrared photograph taken from 70,000 feet as part of a NASA
underflight mission in support of LANDSAT I. The townsite of Ophir, Spring
Gulch, and the alluvial fan are conspicuous. Aspen and coniferous forest and
vegetation trimlines emphasize the avalanche paths. A recent mudflow, which
originated right of center, has swept down the gulch east of Spring Gulch and
run along the bed of Howard Fork. (Enough reproductions of this photograph
for insertion in each copy of this issue of the Annals were provided under the
auspices of NASA Grant NGL-06-003-200 without cost to the Association of
American Geographers.)

TYPES OF AVALANCHES

The types of avalanches in Spring Gulch differ greatly in extent, velocity, flow characteristics, and mechanics of impact, and they must be considered separately if defense structures and new habitations are to be planned.

Wet Snow Avalanches
Wet snow avalanches have a density of 300 to 400 kg/m³, although they may attain maximum velocities of 22 m/sec in the main gully of Spring Gulch. Because of their relatively low velocities, they tend to follow irregularities in the terrain fairly closely and are more easily controlled in the run-out zone than are dry snow avalanches. Nevertheless, the paths of wet snow avalanches are less predictable because channel blockage by the debris itself can cause lobes to break out into entirely new courses. Wet snow avalanches can also produce high impact pressures and could conceivably reach any section of the Spring Gulch alluvial fan. Three houses in Ophir have been moved by such events.

Dry Snow Avalanches
An avalanche of mixed dry flowing and powder snow is the most dangerous and destructive type emerging from Spring Gulch. It is also the most difficult to control. It occurs as large releases of cold, dry snow, generally in midwinter, and consists of two parts. A lower part, with a density of 60 to 90 kg/m₃, tends to follow terrain irregularities and probably attains velocities of up to 90 m/sec in the avalanche track. These velocities will drop, fairly rapidly in the run-out zone because of the great reduction in gradient. The widespread open ground encourages the flowing snow mass to extend laterally and become more shallow. High velocity in the gully, however, creates a low density, high velocity suspension of snow and ice particles which is called the powder cloud. Its density probably ranges between 2 and 10 kg/m³. Damage to tree limbs on the sides of Spring Gulch apparently was caused by this portion of past avalanche events, indicating a flow depth of at least sixty meters. Although the powder cloud will also tend to widen and decelerate on the alluvial fan, it can overtake the denser body of flowing snow, completely cross the fan, and damage mature coniferous trees on the south side of Howard Fork, a full 800 meters from the mouth of Spring Gulch (Fig. 4).

FIG.4. Mature conifers at the extreme limit
of the Spring Gulch run-out zone have been
trimmed by the impact of the powder cloud of
fast-moving snow and air blasts (photo by
Jack D. lves).

DENDROCHRONOLOGY AND DEBRIS

The avalanche paths themselves are rendered conspicuous on the air photographs and in the field by major vegetation differences resulting from the magnitude and frequency of avalanche occurrence (PLATE I). An idealized cross section of the middle reaches of an avalanche path has an inner zone of alpine plants, or aspen and willow, where avalanches are frequent and relatively small; an intermediate zone of destroyed mature trees with seedlings or saplings of either conifers and/or aspen where avalanches are less frequent and larger; and an outer undamaged zone of mature conifers (Fig. 5). The outer edge of the undamaged mature stand is usually trimmed by the rare major avalanche. The height of snapped limbs can be used to calculate the cross section of the major event, the marginal pressures generated, and the maximum horizontal spread if the edge of the run-out zone has mature stands.

FIG.5. Idealized mid-track cross
section of an avalanche path showing
vegetation trimlines.

This idealized description is frequently developed in the field to a sufficient degree to facilitate the application of standard dendrochronological methods.⁵ Scars, discernable in cross section or increment core, are produced by physical damage to the tree, including breakage of limbs. In addition, the occasional pressure against trees at the edges of the avalanche path may bend rather than break limbs and stem. A bent coniferous tree forms reaction wood (compression wood) on the down slope side and frequently has compressed tree rings on the upslope side. The reaction wood in conifers is reddish yellow and shows thick walled cells under the microscope (Fig. 6). Ring compression was not observed in aspen, but reaction wood is common and has a dark red-brown color.

Several natural limitations in the Ophir area restricted the collection of data through application of these principles. The primary limitation is the age of the tree itself. Coring of Engelmann spruce (Picea engelmannii) in the avalanche-damaged forest area southeast of Ophir revealed that most dated from the turn of the century. A few were more than a hundred years old, although there were insufficient numbers of these to provide data with a high level of statistical significance. Trees were also cored in reforested areas at the bottom of Spring Gulch which have several aspen (Populus tremuloides) trimlines but, again, age of the trees limited the historical record. The other major limitation was the general absence of trees on the main part of the Spring Gulch fan. Thus, allowance must be made for the possible occurrence of quite large avalanches (which could cover much of the townsite) that left no record in the forest stand at the extreme edge of the run-out zone.

FIG. 6. Diagram of cross section of conifer showing avalanche
impact damage and reaction wood. The inner zone of reaction
wood, produced while the tree was very young indicates bending
in several directions. The gap on the right side faces the avalanche
track and was caused by bark abrasion. The overgrowth is slowly
healing the wound which occurred 19 years before the section was
cut. The main area of reaction wood to wards the top of the section
resulted from avalanche impact, the point of the impact being on
the opposite (bottom) side. This shows one-direction bending since
the tree had become strong enough to resist bending in all directions
except the main one.

These applications indicated that the forested area southeast of Ophir was struck by a large avalanche in the late 1950s. The damage indicates a dry powder avalanche, which suggests the January event of 1958, rather than the wet slide of April, 1959, known from reports of local residents. An avalanche in the early 1950s is evident in several trees northeast of the town, but the recorded avalanche of January, 1951, although large, apparently did not cross Howard Fork.

Trees cored in a control forested area show no recent avalanche damage, yet indicate disturbance in the middle to late 1880s and possibly in the early 1860s. Few trees cored possess a tree ring record that extends back this far, but it appears that avalanches from Spring Gulch crossed Howard Fork at least once and possibly twice in the latter half of the nineteenth century.

A histogram showing the number of disturbances (compressed rings and reaction wood) noted in the tree ring analysis has been weighted to account for the fact that many trees did not have an early tree ring record. No tree ring evidence indicated the avalanches of 1918 and 1959 which ran close to Ophir. Evidently these avalanches did not run across Howard Fork to be recorded in the tree ring record. Other avalanches also may have gone unrecorded because of the lack of forest.

Finally, tree and rock debris scattered across Spring Gulch fan (Fig. 7) were mapped systematically, since their distribution provides good evidence for the minimum extent of avalanche activity. Such debris, however, is probably the result of multiple events: debris may not necessarily be carried all the way to the extreme end of the run-out; and tree debris may be absent from some areas because of disturbance by man. Nevertheless, useful supplementary data were obtained and used in the compilation of the hazard maps (Figs. 2 and 10).

 
FIG. 7. Avalanche debris on the Spring Gulch fan.
This photograph gives a graphic impression of the
proximity of the existing settlement of Ophir to
avalanche activity (photo by Jack D. Ives).

HISTORICAL DATA

Old photographs, newspaper files, and the recollections of long-term residents add confidence to the indirect evidence, but this type of data also must be used with caution, since human recollection of events can give indications larger than reality; remembrance of actual dates can be particularly faulty. Convergence of different types of evidence becomes a valuable test of reliability, and in Ophir such convergence indicates a high degree of accuracy in reconstruction. Mr. and Mrs. Randolphe Belisle, long-term residents of the area who currently live at Ophir Loop, say that avalanches from Spring Gulch have reached the vicinity of Ophir four times in the last fifty-six years. Large wet slides reached the town during May, 1918, and April, 1959, and dry snow avalanches approached the town in midwinter 1951, and in January, 1958. The 1958 event crossed the creek at the extreme edge of the alluvial fan and hit mature trees, causing damage to limbs. Snow accumulated in mid-fan to the height of the telephone poles (nine meters). Mr. Fred Eanes, a present Ophir resident, reported that a moderately large wet snow avalanche from Spring Gulch in April or May, 1973, split into three lobes; one ran to within 100 to 200 meters of the existing houses. Since this avalanche is the best known to the present residents, it provides a useful base for comparison with larger events of the past. The relative size of the track cross sections indicate that the destructive front of the 1973 event, as it passed through the lower gully of Spring Gulch, was small in comparison with past events, but wet snow deposits on the upper and middle part of the fan were up to ten meters deep (Fig. 3).⁶

FIG. 8. Photograph taken about 1950 showing
how major avalanches cross the access road between
Ophir and Ophir Loop (photograph by Mrs.
Randolphe Belisle).

Mrs. Belisle was also able to provide information on the avalanche paths that threaten the access road. They may be expected to cut the road every three to four years (Fig. 8). The present residents had their first experience with this phenomenon in January, 1975, when the road was buried in at least four places by the Howard Fork, Magnolia, St. Louis, and Badger avalanche paths (Fig. 2); the Colorado avalanche reached the edge of the road, as did the Needles avalanche, while the Butterfly and Terrible ran out onto the highway west of Ophir Loop. This type of hazard is significantly more severe today with daily movement between Ophir and Telluride than it was fifty years ago, when the residents were more or less closed in for the winter. The historical record is impressive enough, but other large avalanches may have gone unnoticed if their debris was covered by new snow during midwinter snowstorms.

FREQUENCY OF LARGE AVALANCHES

Historical data indicate that four avalanches have either reached or closely approached Ophir during the last fifty-six years. Tree ring analysis substantiates and reinforces this recollection of local residents. From a combination of the two lines of enquiry, the broad picture of avalanche activity has been put together (Fig. 9) as the basis for subsequent recommendations on land management.

FIG. 9. Frequency of avalanche occurrence as
interpreted from the tree-ring record.

Powder avalanche impact on trees in the run-out zone, south of Howard Fork, occurred in the late 1950s and probably in the middle to late 1880s. Two cores suggest avalanche impact south of Howard Fork in the early 1860s. Young, uniformly aged aspen stands on the lateral track boundaries between 3,200 and 3,300 meters elevation are fifteen to twenty years old. They correspond to the lateral flow boundaries of the avalanches of the late 1950s, which were considered large since they reached the town limits. Avalanche damage and trimlines extending farther up the sides of the gully indicate that Spring Gulch has run much larger in the past.

When the historical and tree ring records are combined, there is substantial evidence for six major avalanches, all capable of reaching Ophir (1860?, 1885, 1918, 1951, 1958, 1959), in the last 114 years. We conclude an average recurrence interval of approximately twenty years, indicating a five percent probability of occurrence in any one year. The total number of events and the length of the record weaken any statistical approach, but, as a first approximation, we argue that the conclusion is highly relevant to land use decision-making. In addition, the Spring Gulch fan has no forest cover, so that the six avalanches identified represent a minimum number of occurrences. The 1918 and 1959 avalanches are not revealed in the tree ring record. Evidently they did not run out across Howard Fork into the forested area. The very absence of trees on the fan it self is an indicator of a geomorphologically active environment, although some timber may have been cut, especially along its eastern margin.

EXTENT AND IMPACT PRESSURE OF RUN-OUTS

Mapping of debris and damage to living trees, historical data, and dendrochronology give good indications of the frequency of avalanche occurrence. They also assist in delineation of the extent of the run-out zone and in calculation of impact pressures. As a further cross check, the extent was calculated mathematically by using Voellmy's equations of avalanche flow, which are applicable to dense, flowing avalanches, both wet and dry.⁷ They do not consider lateral spreading in diffuse powder avalanches of great height, so we used independent methods to calculate the forces associated with the high velocity powder head which accompanies dry powder avalanches in Spring Gulch.⁸ Additional modifications were made to Voellmy's approach following the work of Schaerer.⁹ The basis for these computations, however, is an expression derived by Voellmy equating avalanche kinetic energy with frictional work, viscous energy dissipation, turbulent energy dissipation, and potential energy, solved for calculating run-out distance. To check the applicability of the Swiss work to Ophir, the run-out distance was measured in the field to coincide with the outer limit of timber destruction on the south side of Howard Fork. This agreed very well with the computed figures.

The next step was to calculate impact pressures across the run-out zone. It was necessary to estimate the deceleration of the flow as it crossed the fan. For the powder avalanches, the velocity at the top of the fan was calculated using Voellmy's equations, and the velocity at Howards Fork was calculated from observed impact effects on mature trees.^l0 We assumed that velocity decreased between these two points proportionately. The velocity remaining at the bottom of the run-out was calculated by assuming that the flow was nine meters (+ one meter) deep as it hit the trees. This figure was obtained by measuring impact trimming of limbs. The velocity was assumed to have a logarithmic velocity profile, as is common in turbulent shear flow. Diameters of broken trees compared with adjacent surviving trees provided data for derivation of impact pressures, again using Voellmy's methods.¹¹ We took the conservative approach of assuming that trees failed by "static" rather than by "dynamic" loading.

An alternate method calculated the velocity through simple conservation of energy. The kinetic energy per unit of Dowing mass is 1/2 V². This is transformed into potential energy gained, gh, friction work, (g cos0),ud, flow work, and drag on surrounding air, where h is the height climbed, 0 the average slope angle, u the coefficient of friction, d the slope distance, and g the acceleration caused by gravity. If flow work and drag are assumed small as the avalanches climb the slope south of the Howard Fork then

The distance, d, was measured as 75 meters in the field, h is 13.5 meters, 0 is 13 degrees, and u is assumed to have been 0.5. The velocity calculated in this manner is 31 m/sec.

The two methods give velocities at the Howard Fork of approximately 30 to 50 m/sec if dynamic loading is assumed, and 30 to 65 m/sec if static loading is assumed. If an average velocity of 45 m/sec is taken and a velocity of 100 m/sec is calculated at the top of the fan, then a velocity decay between these points can be obtained. The velocities calculated in this way were converted to impact pressures, P, through the relationship

where y is the density, in order to subdivide run-out hazard maps into two zones of impact pressure.

These calculations and a plot of the debris distribution were combined for construction of preliminary hazard maps. Subsequent discussion with local residents, examination of winter field conditions, and collection of more detailed information on the location of wet snow avalanche lobes led to modifications and the production of the final maps (Figs. 2 and 10).

The final avalanche hazard maps follow the traditional Swiss and Austrian approach and show three zones of intensity.¹² In Zone I, avalanches will occur every twenty years or less and produce impact pressures greater than 3 t/m².^l3 Zone II will have avalanches with a recurrence interval greater than twenty years and with impact pressures below 3 t/m². Zone III is considered free of avalanche hazard. Any method of avalanche prediction has built-in uncertainties and limitations, but combining them in hazard assessment maps provides a reasonable first approach. This approach should be supplemented by a coordinated program to observe and survey avalanche events.

The avalanche run-out zones that cross the access road from Ophir to Ophir Loop have not received the detailed attention given to Spring Gulch. The run-out zones as plotted present a conservative viewpoint, and the recurrence interval of three to four years, based upon Mrs. Belisle's recollection, is short enough to emphasize that a considerable hazard exists, but it is an entirely different hazard from that facing houses. A house needs to be hit only once with its owners inside for danger to life and property to be high. Avalanches may cross the access road many times with little chance of hitting a vehicle, and inconvenience is the more probable result. Nevertheless, the hazard will grow in proportion to any increase in population, so that development of effective land management policies is vital.^l4

PLANNING RECOMMENDATIONS

We recommend that no construction be permitted within Zone I on the avalanche hazard maps. Any new buildings in Zone II, south of Howard Fork, where damage is primarily the result of powder avalanches, should be de signed to withstand 3 t/m² impact loading. The uplift force of the aerodynamic loading must also be considered. Wind blast from powder avalanches may also occur close to the indicated run-out limits, and even pressures less than 0.5 t/m² are potentially destructive for normal buildings. Windows, for instance, should not face the apex of the Spring Gulch alluvial fan. The other areas of Zone II indicate a recurrence interval greater than twenty years and diminished impact pressures, although the same building restrictions should apply.

FIG. 10. Detailed avalanche hazard map, Ophir.


Dry and wet snow avalanches originate on the aspen-covered slope just west of Spring Gulch and on the hillside southeast of the town. We recommend that consideration be given to the feasibility of evacuating the threatened section of the town of Ophir at times of extreme danger from large wet snow avalanches from Spring Gulch unless defense measures are undertaken. A successful evacuation policy will depend upon improvement in current forecasting. Although much progress has been made in predicting the timing of wet snow avalanches, based upon recent work in the Red Mountain Pass-Silverton area on the far side of Ophir Pass, much more is required before a practical evacuation scheme can be developed.^l5 Finally, there is some undeterminable possibility that an even larger avalanche in the future will sweep through most of the existing built-up area. In the absence of historical evidence for an event of this magnitude, we are dealing with an extremely long recurrence interval that cannot be incorporated into any realistic land use policy.

For reduction of existing hazards that threaten Ophir, six standard mitigation approaches should be considered.

Warning and Evacuation: Local residents might be evacuated before a major avalanche if competent local observers are available, but a successful evacuation program depends upon the credibility of the scheme to the local residents. Prediction is extremely difficult and, with a recurrence interval of twenty years for major events, the Austrian and Swiss experience would indicate that a high degree of success is unlikely.

Explosives: Control of avalanching snow by explosives is widely practiced at ski resorts and along highways. The run-out zones are evacuated before release. This system is not used for built-up areas, because permanent buildings cannot be moved from run-out zones, and controlled releases are sometimes much larger than anticipated. These methods would pose complex legal problems in the event of property damage or personal injury.

Structures in the starting zone: Such structures have been used in the Alps with some success, although there is virtually no experience in the United States with large-scale structural control in the starting zone, and in addition, costs would probably exceed $200,000 per acre of defense structure.^l6

Structures in the run-out zone: Dense, low level avalanches, both wet and dry, may be controlled by placing obstacles in the run-out zone to dissipate avalanche energy or to deflect the flow. These structures are largely ineffective against high velocity dry snow avalanches, especially when accompanied by an airborne cloud, but wet snow avalanches are the greatest hazard to the existing houses. The most promising structure would be a large earthen dam designed to split the flow 150 to 300 meters northeast of Ophir. This dam might be combined with an array of earthen mounds to dissipate the flow energy and with an afforestation program (Fig. 11). An alternative approach, which could also be used in conjunction, would be to barricade the mouth of the small stream channel running down the western margin of the fan. Such a barricade could deflect wet snow avalanches down the fan's center line.^l7

Protection structures for individual buildings: Special building design has proved effective in the Alps when individual buildings required protection. Such structures are designed to withstand high impact pressures or to split the flow of snow, but diverted snow may damage adjacent, closely spaced buildings in a town. Nevertheless, development of new individual buildings in Ophir may produce candidates for such an approach.

Afforestation: Extensive afforestation of the Spring Gulch alluvial fan northeast of the existing buildings could be beneficial. Such a scheme should be used only in conjunction with earthen deflecting structures, and would render them more acceptable esthetically. Afforestation is used primarily in the avalanche starting zones. Large avalanches may sweep away a forest in the run-out zone which does not have adequate earthen structures, and the ram effect of the tree trunks carried down with the slide may increase the damage.^l8

An additional and obvious alternative is to do nothing, let avalanches occur, and accept the risk. This risk may be approximated statistically through the concept of "encounter probability."¹⁹ For instance, if an avalanche has a recurrence interval of twenty years and a building in its path has an estimated life of forty years, there is an eighty-six percent chance that the building wil1 be hit by an avalanche once during its life. If it is occupied by one family for ten years, that family has a thirty-nine percent chance of being hit. The probability of impact carries the possibility of death or personal injury. Also, it is one thing to adopt a "do nothing" policy for buildings which have stood for many years, but quite another to permit erection of new buildings. Future construction should be vigorously controlled by the county planning authorities.

CONCLUSIONS

The Alpine countries are experiencing a rapid acceleration in the rate of avalanche and other hazard-induced death, injury, and property damage.²⁰ This accelerating loss, and the concomitant increase in expenditures for protection, is a result of a rapid growth in population based primarily upon modern two-season tourism which has become characteristic of high mountains in temperate latitudes.²¹ The phenomenon is acute in the Alps; it is becoming acute in Colorado and other parts of the North American mountain west. Natural hazard mapping, now in its infancy, still awaits development of prototype thematic maps at different scales-the general scale of 1: 24,000 or 1: 50,000, and the site scale down to 1:1,000. There is also the opportunity for application of remote sensing techniques, particularly NASA-LANDSAT underflight imagery interpretation. Satellite imagery should be useful for rapid reconnaissance mapping at scales of 1:100,000 to 1:500,000 for the state as a whole. Such highly generalized maps, although of little direct value for site survey and design, would delineate critical areas and provide a powerful tool for assault on another associated and complex problem: public awareness.

FlG. 11. Projected outline of avalanche defense possibilities for Ophir.
Any detailed planning would require an in-depth engineering site survey.

Another major problem is establishment of criteria for designation of the 100-year avalanche run-out zone-the analogue of the 100 year floodplain-for planning and legal purposes. The solution of this problem would be facilitated by systematic collection of data relating to avalanche events. A start could be made through the training of local volunteers for recording size, type, and date of avalanche events; additional mountain weather observation stations would also be useful. Finally, detailed mapping and derivation of hazard maps for individual communities such as Ophir would assist in the identification of alternate building sites.²²

ACKNOWLEDGMENTS

We are indebted to Joseph Vitale of the NASA Office of University Affairs, monitor of Grant No. NGL-06-003-200 to the senior author. His enthusiasm and encouragement have been invaluable. Mark Frauhiger, County Planner, San Miguel County, was largely responsible for the initiation of the project. The residents of Ophir and vicinity have provided hospitality and historical data, and have offered to become Colorado's first volunteer avalanche observer team; this study is dedicated to them and their children. Paula V. Krebs, INSTAAR research ecologist, established the dendrochronological approach, and Hans Frutiger andEdward LaChapelle have provided advice and encouragement in the development of IN STAAR's applied mountain geoecology program.

Throughout the development of INSTAAR's work under the NASA-PY grant-Application of Space Technology to the Solution of Land Management Problems in Montane Colorado-numerous graduate students and staff members have assisted and thereby have strengthened the research base for the present paper. These include Jim Clark, D. M. Glenn, D. P. Groenveld, R. F. Madole, Janet Nichol, Betsy Palmer, Marith Reheis, D. R. Sharpe, and L. D. Williams. The cartography was undertaken by Marilyn Joel.

The work was inspired by the UNESCO MAB-Project 6. The senior author, in particular, has benefitted from the infectious enthusiasm of Francesco di Castri and Gisbert Glaser of the UNESCO MAB Secretariate, Paris, and of Donald King, U. S. Department of State, Chairman of the U. S. National Committee for MAB.

*Dr.Ives is Director of the Institute of Arctic and AIpine Research (INSTAAR) at the University of Colorado in Boulder, CO 80302; Mr. Mears is a natural hazards consultant in Boulder; Mr. Carrara is a geologist with the United States Geological Survey in Denver; and Dr. Bovis is a Research Associate at INSTAAR.

1 W. P. Rogers et al., Guidelines and Criteria for Identification and Land-use Controls of Geologic Hazard and Mineral Resource Areas, Special Publication No. 6 (Denver: Colorado Geological Survey, 1974).

2 H. Aulitzky, "Endangered Alpine Regions and Disaster Prevention Measures," Nature and Environment, No. 6 (Strasbourg: Council of Europe, 1974); H. Frutiger, The Avalanche Zoning Plan, Translation No. 11 (Alta, Utah: U. S. Forest Service Alta Avalanche Study Center, 1970); and O. Voellmy, On the Destructive Force of Avalanches, Translation No. 2 (Alta, Utah: U. S. Forest Service Alta Avalanche Study Center, 1964).

3 M. Mellor, Avalanches, Monograph A-IIId (Hanover, New Hampshire: U. S. Army Cold Regions Research and Engineering Laboratory, 1968).

4NASA EROS underflight imagery, false color, flown at high altitude (20,000 meters) in support of

5 N. Potter, Jr., Tree-ring Dating of Snow Avalanche Tracks and the Geomorphic Activity of Avalanching, Absaroka Mountains, Wyoming, Special Paper NO.123 (Boulder, Colorado: Geological society of America, 1969).

6 Fred Eanes, personal communication, February, 1975.

7 Voellmy, op. cit., footnote 2. In practice, modifications were made to Voellmy's approach. Any one who would like a detailed explanation should write to the senior author.

8 Voellmy, op. cit., footnote 2.

9 P. A. Schaerer, personal communication, 1975.

10 Voellmy, op. cit., footnote 2.

11 Voellmy, op. cit., footnote 2.

12 Aulitzky, op. cit., Frutiger, op. cit., and Voellmy, op. cit., footnote 2.

13 The Swiss Federal Government prohibits construction at pressure above 3.0 metric tons per square meter (t/m²) = 615 psf.

14 E. R. LaChapelle, Encounfer Probabilities for Avalanche Damage, Miscellaneous Report 10 (Alta, Utah: U. S. Forest Service Alta Avalanche Study Center, 1966).

15 R. L. Armstrong, E. R. LaChapelle, M. J. Bovis, and J. D. Ives, Development of Methodology for Evaluation and Prediction of Avalanche Hazard in the San Juan Mountain Area of Southwestern Colorado, Occasional Paper 13 (Boulder, Colorado: Institute of Arctic and Alpine Research, University of Colorado, 1974)

16 H. Frutiger, personal communication, 1975.

17 Fred Eanes, personal communication, 1974.

18  Frutiger, op. cit., footnote 2.

19  LaChapelle, op. cit., footnote 14.

20 Aulitzky, op. cit.. footnote 2.

21 This phenomenon has been identified as a major study area under the UNESCO Man and the Biosphere Programme (MAB). MAB Report 14: Programme on Man and the Biosphere, Working Group on Project 6: Impact of Human Activities on Mountain and Tundra Ecosystems (Lillehammer, November 20-23, 1973), Final Report (Paris: UNESCO, March 20, 1974).

22 The work described here has been identified as part of the United State's contribution to the UNESCO Man and the Biosphere Program (MAB), Project 6: Impact of Human Activities on Mountain and Tundra Ecosystems.