Geological Survey Professional Paper 1180 Grinnell and Sperry Glaciers, Glacier National Park, Montana: A Record of Vanishing Ice

LOCATION AND ACCESSIBILITY

Grinnell Glacier, in Glacier National Park, is on the east slope of the Continental Divide, 17 miles south of the international boundary between the United States and Canada, at lat 48deg 45' N., and long 113° 44' W.; it occupies a cirque formed by a sharp, narrow, serrated ridge connecting Mt. Gould (elev. 9,553 ft) and Mt. Grinnell (elev. 8,851 ft). The ridge is part of a spectacular feature known as The Garden Wall, which forms the Continental Divide in this area (fig. 1, pl. 1). The glacier is 6 miles southwest of Many Glacier Hotel.

DISCOVERY AND EARLY DESCRIPTIONS

George Bird Grinnell (fig. 3), after whom Grinnell Glacier is named, first visited the general area in 1885, while he was editor of Forest and Stream, a well-known outdoor magazine of the period. He was so impressed by his visit that he returned annually for several years and was among the early explorers to advocate the establishment of Glacier National Park (Grinnell, 1901). Available records indicate that the actual glacier was first visited in 1887, by Grinnell, James W. Schultz, and Lt. J.H. Beacon, U.S. Army. Schultz and Beacon each claimed to have named the glacier in honor of Grinnell, but Robinson (1957) maintained that Beacon should have the credit. Glacier National Park was ultimately established in 1910.

FIGURE 3.—George Bird Grinnell (1849-1938), editor, author, and explorer, was the discoverer of Grinnell Glacier. The exact date of the portrait is not known; it was made sometime before 1936, when it was given to the Glacier National Park museum by the Grinnell family. Photograph 7667 from the archives of Glacier National Park, courtesy of the National Park Service.

In 1926, 16 years after establishment of the park, Grinnell revisited the glacier, accompanied by M.J. Elrod, Park Naturalist. Grinnell then described to Elrod the glacier in 1887, as he remembered it. Together with the 1887 photographs (now in the Park's historical files), these reminiscenses enable us to mentally reconstruct the Grinnell Glacier as it existed at time of discovery.

Francois Matthes topographically mapped the Chief Mountain 30-minute quadrangle, which encompasses Grinnell Glacier. He described (1904, p. 262) the glacier as he viewed it in 1900 from Mt. Grinnell:

It covers a series of broad flat shelves with an aggregate surface of about one mile square, the ice descending in irregular cascades from the higher to the lower shelves. Around its upper end stretches a serrated comb-ridge, known as "the Garden Wall," connecting Mt. Grinnell on the north with Mt. Gould *** on the south, and forming for several miles the crest of the Continental Divide. A short distance in front of the glacier its shelving site abruptly terminates in the precipitous, curving wall of a deep amphitheatre. A pretty green lake—Grinnell Lake—lies in its middle, surrounded by dense woods.

The lower, main body of the glacier is separated by a ledge-topped cliff from the glacier's upper, smaller part, called The Salamander (pl. 1). Visible in the 1887 and 1952 photographs (fig. 4) and labeled on the recent maps (Many Glacier 7-1/2—minute quadrangle, 1950 and 1960 maps of Grinnell Glacier), The Salamander has scarcely changed since the glacier's discovery. In 1900 (when the Chief Mountain quadrangle was mapped) the terminus was slightly lower than 6,200 feet elevation, and the top of the glacier was at 7,500 feet elevation.

The next studies were in 1911 by Alden (1914, p. 19), who described the glacier as follows:

This glacier has a width from the northwest to the southeast of about 1-1/2 miles and a length from southwest to northeast of about 1 mile. Its area is a little over 1 square mile. It consists of a neve covered upper part, lying on an upper bench in the western part of the cirque, and the main glacier, whose lowest part is not far from the crest of the cliff which rises abruptly nearly 1,000 feet from the valley floor above Grinnell Lake. Through most of its lateral extent the upper mass of ice [The Salamander] ends at the crest of the bare rock ledge below the upper bench. South of this, however, the ice cascades over the ledge with a much crevassed surface to the main glacier below. From the encircling cliffs the ice flow converges toward the lowest point in the lip of the cirque. A large part of the surface is crevassed, showing that the ice is moving down over an uneven bed, and nearly the whole surface is banded with the soiled zones which mark the outcropping of the ice strata. A morainal embankment, consisting of narrow sharp-crested ridges of drift 30 to 100 feet in height, closely borders the ice margin on the east and north. *** This is in part [a] lateral and in part a terminal moraine.

PICTORIAL RECORD

Grinnell Glacier, like other glaciers in Glacier National Park and most glaciers in western North America, has significantly diminished in area and volume since it was first observed. Changes are well illustrated by selected photographs.

The earliest available photographs are those taken by the Grinnell party in 1887 (figs. 4A and 5A). Later photographs taken from the same locations (figs. 4B and 5B) and other comparative pairs of photographs illustrate the changes in the glacier's extent (figs. 5-8).

FIGURE 4.—Comparative views of Grinnell Glacier from the upper end of lake Josephine photographed by Lt. J.H. Beacon, U.S. Army, with the Grinnell party in 1887 (A) and by Gerald Baden in 1952 (B). The terminus in 1887 was near the top of the cliff encircling the upper end of Grinnell Lake and near the crest of the waterfall shown in the 1952 view. The dark area in the lower part of the glacier in the 1887 view appears to be a protruding rock but is probably scree, broken rock debris that originated from rockfalls at the head of the glacier. In 1887 the north edge (at right in photograph) of the glacier nearly reached the top of the narrow, sharp-crested moraine, which is conspicuous in later photographs. The upper and lower parts of the glacier (A) became separated into two distinct parts in 1926 or 1927. The lower, larger part is barely visible in the 1952 view (B). A, Photograph 4383 from the museum of Glacier National Park, courtesy of the National Park Service.

FIGURE 5.—Front of Grinnell Glacier as viewed (A) by the Grinnell party in 1887 and (B) by Arthur Johnson on August 28, 1969, from near the end of the present horse trail. Moraine, clearly shown in 1969 view, was partly covered by the glacier in 1887. Tall tree in 1969 view is probably the larger of the two trees shown in left side of 1887 view; in 1961 this tree had 260 rings (Gerald Baden, oral commun., 1969). Smaller trees in 1969 view have sprung up since 1887.

AREA

Grinnell Glacier, as shown on the Chief Mountain 30-minute quadrangle map, occupied 530 acres in 1900, a measurement probably accurate to within 5 percent. Periodic remapping has documented its diminishing area, as shown in the accompanying table. Of the two sections of the glacier that resulted from the disappearance of the previously connecting icefall in 1926 or 1927, Dyson (1948) considered only the lower part to be Grinnell Glacier.

The upper part of the glacier, known as The Salamander because of its shape, is hemmed in by steep rock walls and so its area has scarcely changed since it was first mapped in 1900.

The lower, main part of the glacier, however, has steadily decreased in extent. The average annual decrease in area during successive periods was: 1900-37, 4.1 acres; 1937-46, 5.3 acres; 1946-50, 2 acres; 1950-60, 1.3 acres; and 1960-66, 2.5 acres. At the present lower edge of the ice the bedrock surface slopes rather gently and uniformly toward the glacier and so a measurable change in the surface elevation shifts the position of the ice front less than it would have in earlier years when the ice overrode the crest of the rock ledge. Gibson and Dyson (1939, p. 684) observed that the rock strata dipped 15° upvalley. The back, or top, of the glacier lies against the steep, almost vertical, cliffs of The Garden Wall.

A pond appeared at the north end of the glacier, probably in the early 1920's. The park naturalist (file report) reported in 1927 that "A lake of small size has formed on the north side of the glacier, between the ice front and the moraine. The face of the ice is a steep hill down which streams rush continuously." As the ice has retreated, this lake (shown on 1960 map as Grinnell Lake and on 1968 map of Many Glacier 7-1/2—minute quadrangle as Upper Grinnell Lake) has spread from 3.4 acres in 1937 to 31.3 acres in 1968 (see table):

The 1950-60 net increase of only 1.6 acres reflects mainly a sudden 8-foot drop of the lake level in August 1957, when some undetermined change in configuration of the glacier's underside shifted the direction of major melt-water drainage. After the sudden drop in lake level, there was no longer any direct drainage from the lake; major runoff of melt water bypassed the lake and followed the Grinnell Creek channel (pl. 1). The 17.4-acre increase in lake size between 1937 and 1950 represented about one-third of the glacier's concurrent areal shrinkage.

The apparent 1887 outline of the glacier as inferred from photographs was sketched on the 1950 USGS map; the size of the area was estimated to have been 565 acres at that time. In 1900 the glacier's area was only 530 acres. During roughly the same time, on Mount Rainier, Wash., both Nisqually Glacier (Johnson, 1960) and Carbon Glacier (Russell, 1898, p. 390) also receded.

RECESSION

Measurements of terminus changes generally are of limited value to short-term glacier studies because positions of the ice terminus are affected by both ice movement and ablation. Furthermore, as at Grinnell Glacier, local topography at the terminus may profoundly affect rates of recession or advance. Measurements to only one point on the terminus are even less meaningful because changes along the ice front are not uniform. Terminal recession does not necessarily indicate a shrinkage or lowering of the ice surface of the entire glacier. The surface elevation of the upper reaches of a glacier may be increasing even when the terminus is receding. This was demonstrated by investigations on Nisqually Glacier on Mount Rainier, Wash. (Johnson, 1960; Veatch, 1969).

The first known recorded observations of recession of Grinnell Glacier were made by M.J. Elrod, nature guide and ranger-naturalist, during the 1925, 1926, and 1927 summer seasons. He measured the location of the ice front by pacing to a large, conspicuous boulder near the present foot trail, shown as "Elrods Rock" on plate 1. Elrod's 1927 report stated "I am sure that the ice was around this rock in 1922 ***" His measurements were as follows:

The widely varying rate of recession—from 35 feet during the summer of 1926 to only 8 feet for the year between the last two observations—reflects differing weather conditions. According to Elrod, the park season, which is weather-controlled, opened very early in 1926 and was very warm, and the warm weather continued until late in the year. By early July the previous winter's light snowpack had disappeared from the main body of the glacier, and ablation of glacial ice was rapid for the remainder of the season. In contrast, the 1927 park season began late and was cool, cloudy, and rainy. The preceding winter's heavy snowfall did not melt from the glacier until late August.

Elrod's measurements related to a single point on the front of the glacier. This section of the front, when mapped by Dyson in 1937, was 440 feet from "Elrods Rock," or 315 feet beyond the last position recorded in 1927, an average annual recession of 31.5 feet for that 10 years.

From 1932 until 1937 recession was measured annually by the park naturalist, George C. Ruhle. The following measurements were reported to the Committee on Glaciers of the American Geophysical Union (Matthes, 1937, p. 298; 1938, p. 319):

Measurements since 1937 have been based on mapping and are more accurate.

A comparison of the 1937 and 1946 maps by Dyson (1948, p. 101) shows a total average recession of 318 feet along the glacier front during the 9 years, an average annual recession of 35 feet.

In 1945 park naturalist M.E. Beatty and I selected and permanently marked several points that have served as convenient instrument stations for mapping the ice front. These points are shown on the 1960 map (pl. 1) as planetable bench marks 6459, 6425, 6413, and 6454.

Changes measured along a 2,000-foot section of the front of the glacier extending southeastward from Upper Grinnell Lake (pl. 1) are tabulated below. The ice front along the lake was not measured, inasmuch as changes there caused by ice breaking into the lake would not necessarily indicate a true recession.

The positions of the ice front in the years indicated were determined as follows: 1937, from map by Dyson; 1945 and 1968, from field surveys; 1950 and 1960 from USGS maps.

The changes along the 2,000-foot section were not uniform, mainly because of the irregular steplike topography of the bedrock at the glacier's edge. The increase in annual precipitation and the decrease in annual temperature since the mid-1940's (fig. 2) would account for the decreasing rate of recession.

These recession measurements record change in the position of the glacier's edge rather than the recession of the true terminus. At a much earlier stage in the history of the glacier this edge was, no doubt, essentially the terminus. Now, however, it seems logical to consider the actual terminus as the south shore of Upper Grinnell Lake, which is consistent with the direction of movement shown on plate 1. Accordingly, the recession figures were adjusted to the changing shoreline of the lake:

Recession, from 1946 to 1969, measured normal to a 1,125-foot base line, totaled 769 feet, an average of 33 feet annually during the 23 years.

FIGURE 6.—Comparative views of Grinnell Glacier from locations on or near the trail along the south side of Mt. Grinnell: A, August 27, 1900 (Francois Matthes photograph 33-A); B, August 1911 (T.W. Stanton, photograph 704); C, 1935 (Grant); D September 1, 1956 (H.B. Robinson). Gem Glacier, the small ice mass visible in the upper left of all four views, has appeared much the same over the years; it occupies about 7 acres, and fills a shallow rock-rimmed basin. A, Earliest available general view of Grinnell Glacier and surroundings; though viewpoint not identical to the 1887 views (figs. 4A and 5A), the photograph suggests that no appreciable recession or shrinkage occurred between 1887 and 1900. B, 1911 view indicates some recession and shrinkage since 1900. The upper part of the glacier, particularly the part extending up toward the col (saddle) in The Garden Wall, was noticeably lower than in 1900 and the front seemingly has receded somewhat. C, The glacier has separated into upper and lower parts, and has receded and shrunk considerably during the 24 years since 1911. D 1956 view shows further recession of the lower part since 1935. The upper part appears virtually identical in the 1935 and 1956 views and in views as late as 1969. (click on image for a PDF version)

FIGURE 7. — Grinnell Glacier as it appeared in 1911 and 1966 from viewpoint north of the lateral moraine (foreground). A, photograph 711 by T.W. Stanton; B, photograph by Arthur Johnson. A, By August 1911 the glacier had already retreated from its extent in 1887 (fig. 5A). B, By August 1966 the glacier had retreated even farther and separated into an upper part (The Salamander) and a lower part. Upper Grinnell Lake (not visible in this view) had formed between the moraine and the glacier.

FIGURE 8.—Comparative 1911 and 1956 views eastward along crest of the lateral moraine north of Grinnell Glacier. A, August 1911 (T.W. Stanton, photograph 710); B, August 31, 1956 (H.B. Robinson). In 1911 edge of glacier extended well up on side of moraine (photograph A and line on B). In the 1956 photograph from the same position, none of the glacier is visible; its nearest edge is out of view to the right, more than 1,000 feet from the camera station.

MOVEMENT

The first recorded measurements of the glacier's surface movement were made by Alden (1923) for the period August 26-30, 1920. Four markers set on the glacier near its frontal edge moved 1 to 4-3/4 inches during the period. These measurements, characterized by Alden as "crude," indicated merely that the rate of movement during the brief period was rather slow.

Gibson and Dyson (1939, p. 689) estimated the rate of movement by considering the distance between stratification lines, which are the surface expression of boundary planes between successive annual accumulations of ice. Careful study of photographs identified 60 bands from the south cirque wall to a point 1,800 feet distant on the east front, which Gibson and Dyson interpreted to indicate an average movement of 30 feet. Probably because their measurements were at an angle to the direction of movement, this finding is less than from the subsequent 1947-69 measurements.

Movement since 1947 was determined by periodically plotting the location of prominent boulders on the ice surface. The boulders have been identified by year and sequence of marking; for example, 50-2 indicates the second boulder marked in 1950.

The 1947-69 data are summarized in table 3 and are plotted on plate 1. Movement has been generally northeastward. The movement was undoubtedly more to the east when the glacier was both larger and higher, for example, when it was discovered in 1887. As both area and surface elevation decreased, the direction of movement gradually changed from eastward to northeastward, as indicated by the 1947-69 observations.

TABLE 3.—Movement of marked rocks on Grinnell Glacier, 1947-69

[Vectors shown on plate 1]

The annual rate of movement observed since 1947 ranged from 32 to 52 feet, more than half the observed values being in the range of 35-45 feet. Measurement points with annual movement exceeding 45 feet were in a small area southwest from reference point bench mark 6454 feet, and that movement was toward the head of the outlet stream from the glacier. The glacier surface in the area of most rapid movement was steeper than in other areas where marked rocks were located.

Since 1947 the rate and direction of movement have continued virtually unchanged, even though the surface elevation of the glacier decreased 25-30 feet during the period. In general, decreasing surface elevation of a glacier would be expected to be accompanied by a slower rate of movement.

FLUCTUATIONS IN SURFACE ELEVATION

PROFILES

Changes in the surface elevation of Grinnell Glacier have been measured periodically, usually annually, since 1950. Periodic profiles, measured by planetable or transit and stadia along three lines of profile, are shown on plate 1; profiles A—A' and B—B' originate from planetable bench mark 6425 (1960 map) and profile C—C', established in 1957, originates from planetable bench mark 6454. Also shown on plate 1 are profiles based on Dyson's 1937 and 1946 maps. Dyson's measurements, particularly for the higher parts of the glacier, should be considered as only approximate, owing to uncertainty of datum correlation between Dyson's maps and subsequent field surveys; Dyson's contouring on higher areas of glacier does not match later maps, profiles, and observations.

Profile A—A' is alined along the crest of the waterfall below The Salamander. The August 26, 1969, line shows a reversal in slope approximately at midpoint and a 25-foot drop in elevation at a point about 200 feet from the cliff, near the base of Salamander Falls. Mr. Grinnell, during his 1926 visit to the glacier, recalled that in 1887 a "well" or conical depression existed at the base of the falls. Dyson's 1937 map shows a depression, but with the lowering of the ice surface this depression disappeared. The general ice surface in 1968 sloped gently toward the lake, 800 feet from the base of the waterfall. The water from the waterfall does not appear as surface runoff, instead disappearing through or underneath the ice.

A comparison of the 1937 profile developed from Dyson's map with the 1969 profile shows a difference in surface elevation of about 100 feet to more than 120 feet. Even allowing for an error of as much as 20 feet (one contour interval on Dyson's map), the average decrease in the surface elevation was about 100 feet during the 32-year period. The 1946 elevations based on Dyson's map appear fairly reasonable except for the central part of profile A3A' where the 1946 and 1952 surfaces are shown as coincident. In view of later observations, the 1946 surface apparently should have been higher in that part of the profile.

Profile B3B' slopes rather uniformly to within a few hundred feet of the headwall, and then steepens abruptly. A comparison of the 1937 profile developed from Dyson's map with the 1969 profile shows a drop of more than 100 feet near the terminus and lesser differences higher up the glacier. A comparison of the 1946 profile (also from Dyson's map) with the 1969 profile shows a decrease in surface elevation of 50-60 feet.

Profile C3C' is much steeper than the other two profiles, particularly beyond 1,500 feet from the reference point, where the slope increases rapidly. The 1937-to-1969 decrease in surface elevation near the lower edge of the glacier was about 80 feet; the difference at midglacier was about 40-50 feet. The profile developed from Dyson's 1937 map indicates that beyond 1,800 feet from the reference point the ice surface was coincident with or below the 1969 surface, a condition considered unlikely. The developed 1946 profile conflicts with later observations and is therefore not included on plate 1.

As evident from profiles B3B' and C3C', the glacier surface has dropped more near the northeast margin than near the headwall.

Changes in surface elevations from year to year (or, for earlier measurements, over a span of several years) for segments of each profile are listed in tables 4-6. The dates of measurement varied slightly from year to year, a fact which must be considered when comparing results, but nearly all measurement dates fell between mid-August and mid-September. Composite changes during the 1-month period for 1957, 1958, and 1959 along profiles A3A' and B3B', shown in the accompanying table, illustrate the direct relation between glacier melting and prevailing temperatures.

TABLE 4.—Mean elevations, in feet, of segments of profile A3A', Grinnell Glacier, on specific dates

(Elevations are above assumed datum for glacier surveys (pl. 1). Distances measured from reference point, planetable bench mark 6425.---, no data)

TABLE 5.—Mean elevations, in feet, of segments of profile B3B', Grinnell Glacier, on specific dates

[Elevations are above assumed datum for glacier surveys (pl. 1). Distances measured from reference point, planetable bench mark 6425. ---, no data]

TABLE 6.—Mean elevations, in feet, of segments of profile C3C', Grinnell Glacier, on specific dates

(Elevations are above assumed datum for glacier surveys (pl. 1). Distances measured from reference point, planetable bench mark 6454.---, no data)

A graph of the mean annual elevations of selected segments of the three profiles (fig. 9) shows that from year to year elevations both increased and decreased; but the general trend was a decrease in elevation for these segments and for the entire glacier.

FIGURE 9. —Mean annual elevations of selected segments of profiles A3A', B3B', and C3C'. 1950-69, Grinnell Glacier. Segments measured from respective reference points shown on plate 1.

Rainfall does not significantly affect the glacier's melting rate. Theoretically, 1 inch of rain at 50°F (10°C) produces only one-eighth inch of melt water. From August 14 to September 12, 1959, for example, rainfall at the Grinnell Creek gaging station was 13.1 inches. If we assume that precipitation was uniform over the entire glacier and was at a temperature of 50°F, the depth of resulting melt water runoff would have been only 1.6 inches. The equivalent loss of ice thickness—less than 2 inches— would be only a small fraction of the actual 2.4 feet of loss (profile B3B') during that period.

The fluctuations in surface elevations of representative segments of the profiles are shown graphically in figure 9. The general trend has been a decrease in elevation, indicating a shrinkage of volume. From 1950 to 1969 the representative elevations on profiles A3A' and B3B' decreased 35 feet and 30 feet, respectively, approximately 1.7 feet per year. From 1957 to 1969 the elevation on profile C—C' decreased 18 feet, approximately 1.5 feet per year. The other segments of the profiles showed corresponding decreases.

ABLATION

To supplement information obtained from profile measurements, ablation (the amount of snow and ice removed from the glacier's surface by melting, evaporation, and wind erosion) was measured. Ablation was determined by drilling 3/4-inch-diameter holes in the ice, with a specially designed ice auger to depths as great as 20 feet, actual depths being governed by conditions of the ice. A wooden stake, 3/4 inch by 6 feet, was placed in each hole, marked with an identifying location number and date of emplacement. For holes deeper than 6 feet, several stakes were emplaced, end to end, and numbered from bottom to top. The following example illustrates the procedure. On August 29, 1964, a hole was drilled to a depth of 19.7 feet. Three stakes were placed in the hole, the top of the uppermost (No. 3) stake being 1.7 feet below the ice surface. On September 6, 1965, 3.3 feet of stake 3 was exposed, indicating an ablation of 5.0 feet (1.7 + 3.3) since August 1964. On August 16, 1966, stake 3 had fallen away and 0.2 foot of stake 2 was exposed, indicating an ablation of 2.9 feet (2.7 + 0.2) since the 1965 observation. By August 27, 1968, the ice had melted so much that the hole was no longer evident and the remaining two stakes were lying on the ice, indicating an ablation of at least 11.8 feet since the 1966 observation. The results of ablation movements are recorded in tables 7 and 8.

TABLE 7.—Seasonal measurements of ablation at Grinnell Glacier, 1960-65

[Location of stakes shown on plate 1.---, no data]

TABLE 8—Cumulative annual ablation measurements of Grinnell Glacier

For 1960, 1962, and 1963 (table 7) during periods of 75-80 days from about mid-July to early or mid-October, ablation averaged about 11.5 feet. According to observations on intermediate dates, ablation was most rapid in July and August.

Usually some snow from the previous winter was on the glacier at the time the stakes were placed. At stakes 4 and 7, from 1961 to 1965, the snow from the previous winter either was entirely melted away or was less than 1 foot deep. At those two locations the observed ablation was essentially the actual loss from the main ice body.

PRECIPITATION AND RUNOFF

Studies of the relation between precipitation, runoff, and glacier variations at Grinnell Glacier have included measurement of precipitation at two storage gages near the glacier and measurement of runoff at two gaging stations on Grinnell Creek.

In 1949 the U.S. Weather Bureau, in cooperation with the National Park Service, installed a storage precipitation gage (precipitation gage No. 1 on 1960 plan of Grinnell Glacier, plate 1) near the end of the horse trail. The gage, within half a mile of the glacier, is at an elevation of 6,227 feet, 200 feet lower than the terminus. Precipitation gage No. 2 was installed in 1955 by the Weather Bureau and the National Park Service 2,100 feet south-southeast of the first installation, at an elevation of 6,113 feet. The gages are designated as Grinnell 1 and Grinnell 2 in Weather Bureau records.

The gages are maintained by the Weather Bureau and the National Park Service and are read and serviced yearly in July or August. The periods between observations have ranged from 327 to 383 days. Precipitation in this area during July and August is usually light; consequently, the exact dates of observations generally are not critical to measurements of annual totals. The precipitation measurements that have been obtained and the Grinnell Creek runoff measurements for the corresponding periods are recorded in table 9. The catch at Grinnell 2 has always been greater than at Grinnell 1, the ratio between the two ranging from 1.34 to 1.81. The mean ratio for the 13 years of concurrent records, 1955-66 and 1967-69, was 1.55. The difference in the observed values at the two gages illustrates the pronounced effect of the wind patterns and air currents in this rugged mountain terrain, which causes the average measured catch to be greater at the lower gage.

TABLE 9.—Precipitation and runoff in vicinity of Grinnell Glacier, 1949-69

[Gage locations shown on plate 1. Runoff measured at gaging station on Grinnell creek below outlet of Upper Grinnell Lake. Runoff in inches: the depth to which the drainage area would be covered if all the runoff during the period were uniformly distributed on it]

In 1949 the USGS installed a gaging station on Grinnell Creek just below the outlet of Grinnell Lake (Many Glacier 7-1/2—min quadrangle). In Geological Survey publications its location is designated as Grinnell Creek near Many Glacier, Mont. The elevation at the gage is about 4,900 feet, and the drainage area is 3.47 square miles. Streamflow at this station includes the runoff from Grinnell Glacier, which covers about 14 percent of the drainage area. The mean annual runoff of 99.8 inches for 1950-69 closely reflects average precipitation in this area, although runoff is affected somewhat by evapotranspiration and by net changes in glacier volume.

Monthly and annual runoff at Grinnell Creek gaging station for the 20-year period October 1, 1949-September 30, 1969, is shown in table 10; 85 percent of the annual runoff occurred during the 5 months May-September (the main snowmelt period), and only 15 percent during the 7 months October-April (the main snow accumulation period). Monthly runoff was greatest in June.

TABLE 10.—Monthly and annual runoff at Grinnell Creek near Many Glacier, Mont.

[Gage installed in July 1949. Lat 48°46.2'N., long 113°41.9'W.; elevation 4,925 feet; drainage area 2,221 acres (3.47 mi²). Water year begins previous Oct. 1 and ends Sept. 30. Acre-foot: the quantity of water required to cover an acre to the depth of 1 foot, equivalent to 4-3,560 cubic feet; runoff in inches: the depth to which the drainage area would be covered if all the runoff during the period were uniformly distributed on it. Source of data (U.S. Geological Survey, 1959, 1964, 1971, 1976), by water year: 1950, Water-Supply Paper 1308, p. 34; 1951-60, Water-Supply Paper 1728, p. 17; 1961-1965, Water-Supply Paper 191-3, p. 35-38; 1966-69. Water-supply Paper 211-3, p. 25-27]

(click on image for a PDF version)

During the 20-year period of record on Grinnell Creek (table 10), the estimated loss in volume of Grinnell Glacier, determined from profile measurements, totaled 7,720 acre feet or an average annual loss of 386 acre feet. This annual loss is equivalent to 2.1 inches runoff from the area above the Grinnell Creek gaging station, or about 2 percent of the average annual runoff. The runoff contributed by melting of the glacier in recent years has, therefore, to some extent compensated for the loss due to evapotranspiration, and the recorded runoff has probably been only slightly less than the average precipitation over the Grinnell Creek basin. Snow accumulation occurs during October-April and snowmelt occurs during May-September.

In 1959 the USGS installed a second gaging station on Grinnell Creek about 1,000 feet from the glacier. The location is shown on the Grinnell Glacier 1960 map (pl. 1). This station is only operable during the summer and early fall. Trail conditions preclude access until late June or early July. The recording gage has continued to operate through October of most years, and in some years, through November. The runoff for the months of complete record is listed in table 11, along with the measurements for the corresponding months at the station on Grinnell Creek near Many Glacier, and a comparison of the percent of runoff of the two stations.

TABLE 11.—Runoff measured at two gaging stations on Grinnell Creek

[A: installed 1959, "gaging station" on plate 1, drainage area 704 acres. B: installed 1949, location shown on map of Many Glacier 7-1/2—min quadrangle, drainage area 2.221 acres]

The unusually high runoff-percentage relation in some months, such as 96 percent in August 1961 and 97 percent in September 1966, may reflect late-summer glacial melting or localized heavy rainshowers over the glacier. In some years almost all the previous winter's snow over the basin has melted by September or possibly even by August, and nearly the entire late-summer runoff represents melting of the glacier. Occasionally, rainstorms occur in the upper elevations of the basin while Grinnell Lake and Lake Josephine below are in sunshine. From the lower valley heavy clouds which envelop the glacier are frequently seen above an elevation of about 6,000 feet.

A monthly record of precipitation was obtained at the gaging station on Grinnell Creek below Grinnell Lake for May-September of 1956-65 (table 12). The five mean monthly values for the period totaled 22.2 inches. Concurrent records at Summit showed May-September precipitation during these years was 30 percent of the annual total there. Assuming the two locations show a constant relationship between monthly values, we may infer that the mean annual precipitation in the Grinnell gaging station vicinity during 1956-65 was about 74 inches.

TABLE 12.—Monthly precipitation, May-September, 1956-65, recorded at Grinnell Creek gaging station near Many Glacier, Mont.

[Precipitation in inches. Tr., trace]

The mean annual runoff at Grinnell Creek gaging station for 1956-65 was 96 inches. To account for the observed runoff, precipitation in the upper part of the basin must have been appreciably greater than in the area of the gaging station. The precipitation near the glacier, upstream from the near-glacier gaging station, was estimated by extrapolating the July-November runoff values recorded at the low gaging station, as follows:

Runoff (inches) at Grinnell Creek gaging station at Grinnell Glacier, 1950-69

The estimated mean annual runoff, 147-171 inches, includes about 6 inches per year from the reduction in volume of the glacier. Thus the annual precipitation over the area above the gaging station near the glacier must have been in the range 141-165 inches. A reasonable estimate for runoff would be about 150 inches per year.

Additional information on precipitation and runoff is provided by a record of snow surveys in Montana (Farnes and Shafer, 1975) and runoff at a gaging station on Swiftcurrent Creek at Many Glacier. Grinnell Creek is a tributary of Swiftcurrent Creek. The snow survey course on Allen Mountain, 1.5 miles southeast of Grinnell Glacier, and 800 feet lower than the glacier, has been measured annually on or about May 1 since 1922. A snow course at Marias Pass (near Summit, fig. 1) has been measured on or about April 1 and May 1 since 1936. The results obtained at these two courses are summarized in table 13.

TABLE 13.—Mean snow depth, water content, and density at Allen Mountain and Marias Pass snow courses

[Allen Mountain on USGS maps; Mt. Allen in Weather Bureau records. From Farnes and Shafer (1975)]

At Marias Pass the April 1 water content for the two periods was 47 percent of the average precipitation, which was 38.2 and 40.9 inches. The April 1 water content of the snowpack therefore represents roughly half the annual precipitation in the Marias Pass area.

Snow at the Marias Pass snow course shows an appreciable decrease in average water content from April 1 to May 1, although May 1 values were equal or greater in 11 years of the 34-year record, 9 of these occurring during the 20 years 1950-69.

For the Allen Mountain snow course the less pronounced decrease in water content from April 1 to May 1 may reflect its higher elevation and probably lower temperature. The Allen Mountain record should reasonably indicate conditions at Grinnell Glacier at the end of the snow accumulation period and the beginning of the melting or runoff period. This should be particularly true for the 1950-69 period in which, at Marias Pass, nine May 1 values equaled or exceeded April 1 values.

Runoff at a gaging station on Swiftcurrent Creek at Many Glacier for May to September has been recorded annually since 1912 (fig. 10). Grinnell Creek is a tributary of Swiftcurrent Creek through Swiftcurrent Lake. In general, the May-September runoff comprises the melt from the previous winter's snow accumulation plus the May-September precipitation. The average May-September runoff for the 58 years 1912-69 was 48.9 inches, about 75 percent of the annual total of about 65 inches. The May-September average for the period 1950-69 was slightly higher, 51 inches.

FIGURE 10.—Cumulative departures from mean water content at Allen Mountain snow course and from mean May-September runoff at Swiftcurrent Creek at Many Glacier. (click on image for a PDF version)

Cumulative annual departures from mean water content at the snow course on Allen Mountain for 1922-69 are shown in figure 10, with the mean May-September runoff for Swiftcurrent Creek at Many Glacier for 1912-69. These curves exhibit the same general trends shown by departure-from-mean precipitation curves in figure 2.

A study of the departure from mean annual flow of the Kootenai River at Libby, Mont., for 1911-69 showed the same general characteristics indicated by the above curves and for the departure-from-mean precipitation curves shown in figure 2: a downward trend in streamflow from 1920 to the mid-1940's and an upward trend thereafter.

VEGETATIVE SUCCESSION

The following account of the area's vegetation is excerpted from Park Naturalist M.J. Elrod's report for 1927 (National Park Service files).

A fine problem is presented to someone who has the time and inclination, to study the ascent of vegetation as the glacier retreats. A few observations were made this season which may stimulate someone to further study.

Just below the present ice, and not extending down as far as the end of the horse trail when it is completed, twenty-eight species of trees, shrubs, herbaceous plants, sedges, grasses, and mosses were observed, in larger or smaller numbers. Within a hundred yards or so from the ice, alpine beardtongue, milfoil, yellow paintbrush, mountain timothy, white phacelia, Carex, cottonwood, dock, a red-top grass, and low alpine willow had made a start. Along the numerous small streams, among the boulders, and in protected places other species, both alpine species and those with wide altitudinal distribution, were found. The grasses, sedges, alpine beardtongue, yellow monkey flower, and cottonwood make the closest approach [to the ice], and are the apparent forerunners of a rather complex vegetation, whose roots soon will hold the soil made by the disintegrating rocks. Less than a good quarter of a mile from the ice a willow, head high, has a good start. Wild currant has a footing at several places. Small spruce trees, a foot high, are in small numbers showing conspicuously here and there. One was determined as 15 years old.

The other species were moss, Cystopteris and shield ferns, milkvetch, carpet pink, a short-headed sedge, a chickweed, a northern anemone, a mustard, goldenrod, a vetch, nodding onion, purple vetch, and valerian.

When Dr. George Bird Grinnell first saw the glacier in 1887 it is quite certain [that] ice covered all the space in which these twenty-eight species (perhaps others) are growing, This gives an idea of the change that has taken place in forty years. In another half century visitors to the ice, which will be still higher on the mountainside, will pass through an open forest, ten to twenty feet high, with a floor carpeted with flowers.

The "problem" suggested by Elrod was taken up by Gerald Baden, seasonal Ranger-Naturalist, during 1952-61. His attention was directed primarily toward obtaining the age of trees from increment borings. Many of the trees sampled by Baden and me during the 1961 field season were located with a planetable. The tree locations are shown on plate 1. The identifying numbers are described in table 14. Several additional trees mentioned in the table are described with reference to known locations although they were not actually located in the field in 1961.

TABLE 14.—Annual-growth-ring dating of trees near Grinnell Glacier, as determined from coring in 1959-61

[Ring count was at boring height, several feet above ground level. Each tree, at time of count, was at least 5-10 years older than the age indicated by the stated number of rings]

Additional information was provided by Mr. Baden (written commun., 1961): Borings were obtained in 1952 from some of the oldest-appearing trees in the stand on the bench west of the high moraine bordering the melt-water pond. A 9.5-foot-tall fir had 93 rings, a 15-foot white pine 52 rings, and a 10-foot spruce 38 rings. Also in 1952 a spruce with cones, located near the front of the upper part of the glacier (The Salamander) and near the edge of the cliff, was found to have a ring count of 46.