Terminus fluctuations of five glaciers and the correspondence of these fluctuations to temperature and precipitation patterns were assessed at Oregon's Mount Hood over the period 1901–2001. Historical photographs, descriptions, and climate data, combined with contemporary GPS measurements and GIS analysis, revealed that each glacier experienced overall retreat, ranging from −62 m at the Newton Clark Glacier to −1102 m at the Ladd Glacier. Within this overall trend, Mount Hood's glaciers experienced two periods each of retreat and advance. Glaciers retreated between 1901 and 1946 in response to rising temperatures and declining precipitation. A mid-century cool, wet period led to glacier advances. Glaciers retreated from the late 1970s to the mid-1990s as a result of rising temperatures and generally declining precipitation. High precipitation in the late 1990s caused slight advances in 2000 and 2001. The general correspondence of Mount Hood's glacier terminus fluctuations with glaciers in Washington and Oregon suggests that regional, decadal-scale weather and climate events, driven by the Pacific Decadal Oscillation, play a key role in shaping atmosphere-cryosphere interactions in Pacific Northwest mountains. Deviations from the general glacier fluctuation pattern may arise from local differences in glacier aspect, altitude, size, and steepness as well as volcanic and geothermal activity, topography, and debris cover.
Glaciers thicken and advance when accumulation (e.g., snowfall and avalanches) exceeds ablation (e.g., snowmelt and calving); conversely, glaciers thin and retreat under negative mass balance conditions. The termini of small alpine glaciers are especially sensitive indicators of climate change (Menzies, 2002). Ablation season temperature changes of as little as 0.5°C or accumulation season precipitation changes of 10% may be sufficient to alter glacier mass balance (Tangborn, 1980) and ultimately shift termini. Measurable alpine glacier size changes may occur in as few as 1–5 years (Burbank, 1982) to a more decadal scale (Johannesson, 1989; Pelto and Hedlund, 2001; Kovanen, 2003) following climate forcing (Burbank, 1982). Glacier termini measurements have long been used to assess glacier response to climate (Nesje and Dahl, 2000). While lacking the accuracy of mass balance determinations, termini measurements provide a spatially and temporally comparable and economical way to assess glacier-climate relationships (Harper, 1993). Alpine glacier termini may thus be important tools for determining the direction, frequency, and magnitude of past climate change.
North America's Pacific Northwest glaciers have generally receded during the past century, a period characterized by a region-wide temperature rise of ∼0.8°C and precipitation increase of approximately 7 cm (JISAO/SMA Climate Impacts Group, 1999). Most such glacier change has been attributed to climate fluctuations; however, glacier aspect, altitude, size, and steepness as well as volcanic and geothermal activity, accumulation area, topography, and debris cover can complicate a direct response to climate forcing.
Glaciers on Oregon's Mount Hood (Fig. 1) have been studied since the late 19th century when Hague (1871) recorded the character and extent of the White River, Sandy, and Reid glaciers. Reid (1905, 1906) described the Coe, Eliot, Ladd, Newton Clark, and White River glacier termini in 1901. His termini photographs are a remarkable baseline for 20th century glacier studies at Mount Hood. Gannett (1903) and Sylvester (1908a) further described Mount Hood's early 20th century glaciers. The Mazamas, a Portland, Oregon-based mountaineering organization, began measurements from fixed points to the termini of the Eliot, Ladd, and Coe glaciers in the 1920s (Mazamas Research Committee, 1925, 1927, 1928), and the White River and Newton Clark glaciers in the 1930s (Phillips, 1935). Termini measurements continued until 1946 (Phillips, 1946). The Mazamas Research Committee (MRC) also contracted for oblique aerial photographs of Mount Hood's glaciers in 1935, 1938, 1946 and 1956 (Mazamas Research Committee, 1935, 1938, 1946, 1956). Ice volume mapping (Driedger and Kennard, 1986), cross-section and mass balance measurements (Mason, 1954, 1955, 1957; Handewith, 1959; Dodge, 1964, 1971, 1987), superglacial debris analysis (Lundstrom, 1992, 1993), and moraine mapping and dating (Lawrence, 1948; Lillquist, 1988) have shed further light on Mount Hood's post-Pleistocene glaciers.
This paper addresses two questions: (1) how did Coe, Eliot, Ladd, Newton Clark, and White River glacier termini fluctuate between 1901 and 2001; and (2) how did these fluctuations correspond to Mount Hood's temperature and precipitation patterns?
Mount Hood, at 3426 m, is the highest summit in Oregon and the fourth highest mountain in the Cascade Range. Quaternary andesite and dacite lavas and interbedded pyroclastic debris compose this stratovolcano (Wise, 1969; White, 1980). Collapsing lava domes formed pyroclastic flows on Mount Hood's south flanks, ash plumes that drifted downwind, and lahars that extended far down valleys radiating from the mountain (Crandell, 1980; Cameron and Pringle, 1986, 1987). Crater Rock (Fig. 1) has been the focus of historic geothermal and minor eruptive activity (Sylvester, 1908a; Phillips and Collins, 1935).
Mount Hood's mid-latitude location atop the north-trending Cascade Range, approximately 160 km inland of the Pacific Ocean and the Coast Range, plays a strong role in its climate patterns. Temperatures are generally lower and precipitation is typically higher at the mountain than in adjacent lowlands. Eastward-moving mid-latitude cyclones bring more precipitation to west-facing slopes than east-facing slopes. Pacific Northwest (PNW) summers are generally dry because of the influence of the North Pacific High, while winters are generally wet because of the influence of the Aleutian Low (Walters and Meier, 1989). Deviations from “normal” climatic conditions result from changing intensities in the pressure systems (Miller et al., 2004) associated with the coupled atmosphere/ocean system (e.g., Arctic Oscillation—Thompson and Wallace, 1998; El Nino/Southern Oscillation (ENSO)—Walters and Meier, 1989; Pacific Decadal Oscillation (PDO)—Mantua et al., 1997; and Pacific North America pattern—Wallace and Gutzler, 1981). During 1971–2000, Government Camp (1213 m elevation) (Fig. 1) averaged 6°C temperature and 225 cm precipitation annually with January averages of −1°C and 33 cm and August averages of 14°C and 4 cm. Annual snowfall over this same period averaged 643 cm (Oregon Climate Service, 2004).
Eleven glaciers cloak Mount Hood (Fig. 1; Table 1). Late Pleistocene glaciers terminated at approximately 800 m on the West and Middle Forks Hood River (Glisan, Ladd, and Coe), 1200 m on East Fork Hood River (Eliot and Newton Clark), 1050 m on Polallie Creek, 1150 m on White River (White River), 1100 m on Salmon River, and 730 m on Sandy River (Zig Zag, Reid, and Sandy) (William Scott, written communication, 11 August 2004). During the early 1980s, Eliot Glacier (Fig. 1) was the thickest and most voluminous but Newton Clark Glacier covered the greatest area (Driedger and Kennard, 1986) (Table 1). With a total glacier surface area of 13.5 km2 and a volume of 0.4 km3, Mount Hood ranked fifth in glacier cover in the Cascade Range (Driedger and Kennard, 1986; Kennard and Driedger, 1987; Carolyn Driedger, written communication, 16 December 2003).
We assessed historical changes in Mount Hood's glacier termini with repeat ground and aerial photographs, historical written descriptions, historical field surveys, and global positioning system (GPS) surveys. These data were then compiled in a geographic information system (GIS). Ground photographs taken by Reid at Coe, Eliot, and Newton Clark glaciers in 1901 (photographs on file at the Mazamas office), Langille at White River Glacier in 1902 (Langille et al., 1903), and Richards at Ladd Glacier in 1932 (Richards, 1932) provided benchmarks for each of the five glaciers. Photographs taken from essentially the same photo points in summer 2001 allowed us to visually compare glacier change during the past century.
United States Forest Service (USFS) 1:12,000–1:20,000-scale vertical airphotos from 1946, 1959, 1967, 1972/1973, 1979, 1982, 1989, and 1995, and 1-m-resolution 1994 and 2000 USFS orthophotos were used with select GPS measurements (Coe and Eliot glaciers) to quantify 1946–2001 glacier terminus changes. Airphotos were scanned and georectified to the 1994 digital orthophotos using ERDAS® software. Each terminus was digitized on the georectified airphotos and orthophotos using ArcGIS® software following the Mazamas' criteria for identifying glacier termini—i.e., “the point at which a glacial stream emerges” (Phillips, 1935). Active glacier termini were differentiated from stagnant, debris-covered ice by: (1) an identifiable glacier front; (2) a lobate planimetric form; and (3) a visible connection of the lobate form to a debris-free glacier upslope. Lateral moraine-parallel distances of terminus advance or retreat were subsequently determined using ArcGIS® by plotting the locations of the heads of main outflow streams at successive time periods. We estimate our measurement and mapping error to be approximately 12 m based on the U.S. Geological Survey lowland vertical mapping standard of one-half the contour interval of the map (U.S. Geological Survey, n.d.) and the likelihood that such errors are greater in mountainous areas.
In connecting dating from the post-1946 airphotos, orthophotos, and GPS measurements to the pre-1947 Mazamas surveys, photographs, and historical descriptions, we assumed that the glacier termini shown on the 9 September 1946 vertical airphotos were in essentially the same positions as those measured by the Mazamas in late August and late September 1946. Mazamas survey measurements and estimates (Phillips, 1935, 1938, 1946) were used to back-calculate from 1946 to the 1930s (all glaciers), 1920s (Coe, Eliot, and Ladd), and 1902/1901 (Eliot, Newton Clark, and White River). We estimated the 1901 locations of Coe and Ladd termini based on Reid's descriptions and photograph. Unfortunately, the small scales and large contour intervals of early topographic maps (Langille et al., 1903; U.S. Geological Survey, 1913) precluded their use in this study.
The climatic foundation for assessing glacier-climate relationships came from the 103-Year High-Resolution Climate Data Set for the Conterminous United States (see ftp://ftp.ncdc.noaa.gov/pub/data/prism100) gridded, GIS PRISM dataset created by the Oregon Spatial Climate Analysis Service (OSCAS). Four additional years of OSCAS data extended the data set from 1895 to 2001. One 2.5′ latitude × 2.5′ longitude (approximately 4 km × 4 km) data cell centered on Mount Hood's summit was used to represent the climate at each of the five glaciers. Data were divided into a 1 October–30 April accumulation season and 1 May–30 September ablation season following Tangborn (1980) in a water year (WY) format. Five-year running average (RA) temperature and precipitation were calculated for each season to identify trends and to compare with glacier termini variation (Burbank, 1982). Historical climate and glacier change literature was used to assess regional patterns of climate and glacier fluctuations and the causes of the temperature and precipitation patterns.
A variety of statistical techniques were used to assess the linkage between Mount Hood's glaciers and climate. We used Spearman's rank correlation to identify significant correlations among Mount Hood's cumulative glacier terminus fluctuation records. Spearman's rank correlations, multiple regression, and cross correlation were used to assess the relationship between glacier terminus fluctuations and the five-year RA climate data (i.e., accumulation and ablation season temperature and precipitation). In the glacier-climate analysis, we attempted to correlate the actual (e.g., 1946 and 1959) glacier measurements with climate data averaged over the same period (e.g., 1947–1959). All statistical analyses were accomplished using Quattro® and Statistix® software.
HISTORICAL GLACIER TERMINUS FLUCTUATIONS
The termini of Coe, Eliot, Ladd, Newton Clark, and White River glaciers all receded between 1901 and 2001; however, magnitude, timing, and rate of glacier terminus change varied considerably among these glaciers.
Coe Glacier originates in a cirque immediately north of Mount Hood's summit and extends ∼3.3 km to the north (Figs. 1 and 2A; Table 1). Reid's photograph from the right lateral moraine (Fig. 2B) shows that the debris-covered terminus extended nearly to the top of the prominent Little Ice Age (LIA) lateral moraines (Lillquist, 1988) and downvalley to a point where the lateral moraine crests dip steeply to the north. Boulders painted near the terminus during the initial MRC survey (Mazamas Research Committee, 1928) are located 82 m upvalley of the 1901 terminus. Mazamas Research Committee measurements reveal 37 m of retreat between 1928 and 1938 (Phillips, 1938) and 58 m of recession between 1938 and 1946 (Phillips, 1946) (Table 2). Advance was noted in only one year, 1935, of the MRC measurements. Coe Glacier rapidly receded 295 m between 1946 and 1959 to its most upvalley position in the past century. A thickening of the upper profile of the glacier (Handewith, 1959) (i.e., kinematic wave) resulted in a terminus advance of 104 m by 1967. This advance continued through 1972. Although the terminal position over the past 30 years has oscillated, it showed a net retreat of 94 m. As of summer 2001, Coe Glacier's ice front was still 64 m downvalley of the 1959 terminus. The net 1901–2001 change in the terminus was −408 m, a 12% loss in glacier length, as dramatically illustrated by the 1901 and 2001 photographs (Figs. 2B and 2C).
Eliot Glacier originates in a cirque immediately northeast of Mount Hood's summit and extends about 3.5 km to the northeast (Fig. 1 and 3A; Table 1). In summer 1901, the glacier terminated in the deep canyon of Eliot Branch (Reid, 1905) (Fig. 3B) and extended nearly to the top of the prominent, bounding LIA moraines (Lawrence, 1948; Lundstrom, 1992). The MRC measured 88 m of terminus recession between 1901 and 1925, 42 m of recession between 1925 and 1938, (Phillips, 1938), and 51 m of recession between 1938 and 1946 (Phillips, 1946) (Fig. 3; Table 2). During the 1925–1946 period, advance occurred only in 1932 and 1933 (Phillips, 1938). Significant recession continued through at least 1959 despite a thickening in an upper profile noted by Handewith (1959) and Dodge (1964). The thickening resulted in 79 m of terminus advance between 1967 and 1972. Eliot Glacier retreated 609 m between 1979 and 2000 with 378 m of that retreat occurring between 1995 and 2000 when the lower portion of Eliot Glacier downwasted sufficiently that its terminus was only identifiable as an ice-cored moraine. Eliot Glacier's summer 2000 terminus was at its farthest upvalley position in the past century; however, the summer 2001 terminus had advanced 18 m downvalley. Eliot Glacier had a net 1901–2001 change of −775 m, representing a 22% loss in total length, as illustrated by the recession from the lower trough during the past century (Figs. 3B and 3C).
Ladd Glacier originates with Coe Glacier in a cirque northwest of the summit and extends 1.8 km to the northwest (Figs. 1 and 4A; Table 1). Snow cover prevented Reid (1905) from photographing the summer 1901 terminus; however, his description allowed us to estimate that the 1901 glacier terminated ∼70 m downvalley of the 1927 snout (Fig. 4A; Table 2). Only 3 m of recession were recorded by MRC surveys between 1927 and 1936 (Phillips, 1938). Mazamas Research Committee measurements show that Ladd Glacier receded 40 m from 1936–1946 (Phillips, 1946). The terminus receded an additional 111 m between 1946 and 1967. Snow cover prevented the identification of the 1959 terminus. However, a 1956 oblique airphoto (Mazamas Research Committee, 1956) revealed little recession between 1946 and 1956, thus it is likely that the majority of the 111 m of recession occurred between 1956 and 1967. A brief, minor advance between 1967 and 1973 preceded the 1973–1989 retreat of 992 m. Unlike the Eliot Glacier's extensive retreat, the majority of this major recession occurred between 1984 and 1989, leaving the former terminus as ice-cored ground and lateral moraine, and the active terminus at the top of an icefall. The 1989 terminus was at its farthest upvalley point of the past century; however, Ladd Glacier advanced 100 m from 1989–2000. Despite the recent advance, the net 1901–2000 change in Ladd Glacier terminus was −1102 m, a 61% loss in length. These values are the highest of any of the five glaciers analyzed (Figs. 4B and 4C).
Newton Clark Glacier
Newton Clark Glacier extends approximately 1.2 km east of Mount Hood's summit (Figs. 1 and 5A; Table 1). It differs from the other glaciers analyzed in that it comprises a broad icefield constrained by Cooper Spur to the north and Steel Cliff to the west (Fig. 1), and it had termini at the heads of Cold Spring, Newton, and Clark creeks. We measured the Newton Creek terminus because it was the focus of previous glacier observations and measurements. Reid's 1901 photograph (Fig. 5B) (Phillips, 1935) reveals the glacier draped over the edge of a prominent cliff. Mazama Research Committee members, using Gilardi's 1937 photographs (Phillips, 1937), estimated that the glacier receded about 150 m and thinned by about 30 m between 1901 and 1937 (Phillips, 1938) (Table 2). We measured 111 m of retreat from the estimated 1937 terminus to the 1946 ice front shown on the vertical airphoto. The 1946 terminus was at its most upvalley point of the past century. Post-1946 termini were characterized by alternating advances and retreats (+217 m between 1946 and 1967, −127 m between 1967 and 1972, +223 m between 1972 and 1984, and −271 m between 1984 and 1995). The 2000 position of the Newton Clark Glacier terminus extended 157 m downvalley of the 1995 position. Ground photographs from 2001 reveal a dramatically thinner ice front than that of 1901 (Figs. 5B and 5C). Despite the differences seen in the ground photos, the net 1901–2000 change in the Newton Clark terminus was only −62 m, a 5% decline in length. This represents the least terminus retreat of any of the glaciers analyzed.
White River Glacier
White River Glacier originates south of Mount Hood's summit in a cirque bounded by Steel Cliff, Crater Rock, and an unnamed ridge to the west (Figs. 1 and 6A). It extends approximately 2.1 km south to a point well above timberline (Table 1). In summer 1870, White River Glacier was significantly longer, extending approximately 150 m below timberline (Hague, 1871). Reid (1905, p. 196) described the 1901 White River Glacier terminus as 2100 m elevation “at the head of a very deep canyon”. Langille's (1903) photograph (Fig. 6B) reveals that White River Glacier may have been connected with Coalman Glacier above Crater Rock. However, J. N. LeConte's photograph in Williams (1912) shows that White River Glacier was separated from Coalman Glacier. A Mazamas survey in 1936, coupled with Reid's estimate of his photopoint being “500 yards downvalley of the terminus” indicates that White River Glacier retreated 561 m between 1901 and 1936 (Phillips, 1938) (Table 2). Since 1936, White River Glacier has experienced alternating periods of advance and retreat—i.e., +299 m between 1936 and 1959, −258 m between 1959 and 1967, +163 m between 1967 and 1979, −141 m between 1979 and 1984, +22 m from 1984 to 1989, and −141 m from 1989 to 2000. Thickening of the upper glacier noted in 1958 and 1959 by Handewith (1959) likely impacted the terminus advance between 1967–1972. Despite the alternating nature of advances and retreats over at least the past 60 years, the net 1901–2000 change in White River Glacier terminus was −617 m, a 30% decline in length (Figs. 6B and 6C).
HISTORICAL CLIMATE PATTERNS
Five-year RA temperatures increased during the 1896–2001 WY period—∼1.5°C in the accumulation season and ∼4°C in the ablation season (Fig. 7A). Accumulation and ablation season temperatures were at their lowest levels in ∼1900, 0.7°C and 3.1°C, respectively, cooler than the 1896–2001 averages. By 1940, temperatures had climbed to reach their highest levels of the century—i.e., five-year RA accumulation season temperatures were approximately 1.4°C higher and ablation season temperatures were 1.5°C higher than the 1896–2001 averages. Temperatures began to decline soon after 1940 and generally continued to drop until ∼1975. Five-year RA accumulation season temperatures fell 0.6°C, and five-year RA ablation season temperatures dropped 0.1°C below the 1896–2001 averages during this second coldest period of the past century. Since ∼1975, five-year RA accumulation and ablation season temperatures have generally risen 1.0°C and 1.2°C, respectively, above the 1896–2001 averages. The second warmest five year period during the past century occurred from 1996–2000.
Overall, five-year RA accumulation and ablation season precipitation increased during the 1896–2001 WY period (Fig. 7B). Within this pattern, precipitation generally declined between 1896–1948 and increased from 1948–present. Five-year RA accumulation season precipitation has been above the 1896–2001 mean by as much as 70 cm (+20%) during seven different periods. Conversely, five-year RA accumulation season precipitation has been below the long-term mean six times by as much as 105 cm (−37%). The wettest five-year RA accumulation seasons occurred in 1995–1999 when each averaged 336 cm of precipitation. The driest five-year RA accumulation seasons occurred in 1927–1931 with an average of 238 cm.
GLACIERS AND CLIMATE
The pronounced period of retreat displayed by all of Mount Hood's glaciers from 1901 through the mid-1930s, and all but White River Glacier through 1946 (Fig. 7C), coincided with generally rising temperatures (especially following 1924) and falling precipitation (1925–1930 and 1938–1944). Falling temperatures and rising precipitation beginning in the early 1940s appears to have initiated advances in all glaciers; however, the timing of these advances varied. Newton Clark and White River glaciers were advancing by 1959, Coe and Eliot by 1967, and Ladd by 1973. Retreat was the overall trend of glaciers from the mid-1970s to the mid-1990s, corresponding with generally rising temperatures since the mid-1970s and periods of declining precipitation in 1975–1980 and 1986–1993. Increasing precipitation appears to have dampened the impacts of rising temperatures on glacier termini, especially during the past decade, when the Coe, Eliot, Ladd, and Newton Clark glaciers each advanced.
Spearman rank correlation analysis of cumulative glacier terminus fluctuation data reveals that Mount Hood's northern glaciers fluctuated similarly over the past century (Table 3). Coe, Eliot, and Ladd glacier terminus fluctuations were strongly correlated, while Newton Clark and White River were not (p < 0.05). These results suggest that the Coe, Eliot, and Ladd glacier terminus fluctuations responded to common factors or forcing mechanisms. Unfortunately, none of the statistical tests showed significant, strong to moderate correlations among glacier terminus fluctuations and the various climate variables. Thus, statistical analysis did little to support the qualitative observations of glacier-climate relations mentioned above.
Each of Mount Hood's glaciers retreated overall during 1901–2001. Statistical analysis shows that Coe, Eliot, and Ladd glaciers displayed generally synchronous patterns of advance and retreat within this overall period of retreat. The similar pattern of Mount Hood's northern glacier terminus fluctuations, combined with qualitative analysis of glacier terminus fluctuation and climate data, suggests that temperature and precipitation played a significant role in Mount Hood's glacier terminus fluctuations during the past century.
Supporting evidence for glacier-climate linkages comes from other mountainous areas around the PNW. The general patterns of glacier terminus change seen at Mount Hood—i.e., retreat from 1901 through the mid-1940s, advance from the mid-1940s through the mid-1970s, and retreat from the mid-1970s through the mid-1990s—have been observed in Washington's North Cascades (Hubley, 1956; Harper, 1993; Pelto and Reidel, 2001), South Cascades (Heliker et al., 1983; Driedger, 1993), and Olympics (Spicer, 1989), and in Oregon's Central Cascades (Hopson, 1960; O'Connor et al., 2001). The historical glacier terminus fluctuations may have been set in motion by rising, mid- to late 19th century temperatures that caused significant ablation, making glaciers more susceptible to 20th century climate changes (Burbank, 1982). The general correspondence of glacier terminus fluctuations across ∼6° latitude in mountainous terrain suggests that regional scale, interannual- to interdecadal-scale weather and climate events helped shape Mount Hood's glacier fluctuations over the past century.
PNW weather and climate changes are forced by atmosphere and ocean interactions. One such coupled atmosphere-ocean circulation forcing mechanism, the Pacific Decadal Oscillation (PDO) (Mantua et al., 1997; JISAO/SMA Climate Impacts Group, 1999), affects glacier climate on decadal and interdecadal scales throughout the PNW (e.g., McCabe and Fountain, 1995; Bitz and Battisti, 1999; Kovanen, 2003). The PDO (Mantua et al., 1997) is characterized by 20- to 30-year-long phases (JISAO/SMA Climate Impacts Group, 1999) in the north Pacific Ocean. Warm phase October–March air temperatures are higher and precipitation is lower than normal, with below-normal PNW spring snowpack (Mantua, 2002). Cool phases have the opposite characteristics. Mantua et al (1997) and Mantua (2002) identified four PDO phases during the past century—cool from 1890 to 1924, warm from 1925 to 1946, cool from 1947 to 1976, and warm from 1977 to the mid-1990s. These phases correspond well with Mount Hood's historical temperature and precipitation patterns, and glacier terminus fluctuations (Fig. 7), as they do elsewhere in the PNW (Kovanen, 2003).
OTHER FACTORS AFFECTING GLACIER TERMINUS FLUCTUATIONS
Much circumstantial evidence supports the linkage between Mount Hood's glaciers and ocean-atmosphere forced climate changes; however, the lack of strong, statistically significant correlations between glaciers and climate variables indicate that these glaciers were not well synchronized with climate changes. Variations in the terminus responses of Mount Hood's glaciers over time further support this point—i.e., Coe, Eliot, and Ladd glacier termini fluctuated very differently than did the termini of Newton Clark and White River glaciers. Further, even the termini of the Coe, Eliot, and Ladd glaciers occupied their most upvalley points at different times during the past century. Such differences in overall glacier response and response time may occur because of a myriad of local factors (Nesje and Dahl, 2000; Pelto and Hedlund, 2001; Klok and Oerlemans, 2004).
Some of the differences between the terminus fluctuations of Mount Hood's glaciers may be attributed to the physical characteristics of each of the glaciers (Table 1). The three corresponding glaciers are similar in type, aspect, elevation of source area, and thickness. However, the relatively small size of these glaciers may reduce their climate sensitivity (Klok and Oerlemans, 2004). Further, the climate sensitivity of the Newton Clark Glacier may also be hampered by its wide terminus (Klok and Oerlemans, 2004).
Given Mount Hood's setting, it is also likely that volcanic and geothermal activity, subglacial topography, and debris cover have also played an important local role in Mount Hood's historical glacier termini fluctuations. Heat associated with volcanic and geothermal activity can accelerate melting of glacial ice, firn, and snowpack, thus decreasing glacier mass balance and ultimately reducing glacier length. Conversely, such activity may enhance basal melting and basal sliding, thus causing glacier advance (Frank and Krimmel, 1978; Zimbelman et al., 2000; Sturm et al., 1991). A series of minor eruptions, including glowing lava at Crater Rock (Fig. 1), occurred between 1853 and 1869 (Harris, 1988), and in 1907 (Sylvester, 1908a, 1908b). Fumaroles and warm ground near Crater Rock produced caves beneath snowfields (Phillips and Collins, 1935) and bisected White River Glacier (Sylvester, 1908a, 1908b; Cameron 1988), thus reducing the accumulation area feeding the terminus. Separation of White River Glacier from its upper accumulation zone (now the Coalman Glacier) occurred between 1894 (see M. W. Gorman's photograph in Russell, 1904, plate 12) and 1912 (see J. N. LeConte's photograph in Williams, 1912, p. 75). Minor eruption and geothermal effects likely enhanced a climatically reduced glacial mass, thus collectively resulting in rapid recession and associated thinning of White River Glacier between 1901 and the mid-1930s. Volcanic and geothermal activity does not appear to have changed since 1907.
Subglacial topography may have influenced the sensitivity of Mount Hood's glacier terminus fluctuations to climate change. Icefalls form as glacial ice passes over abrupt, convex-upward topography. Tensile stresses at icefalls (Benn and Evans, 1998) thin glaciers, thus reducing the amount of ice reaching the terminus and ultimately affecting glacier front positions (Venteris et al., 1997). Newton Clark Glacier terminates atop an escarpment beyond which it has not extended since before 1901. Post-LIA warming, combined with icefall-induced thinning of the downwasted glacial ice, likely severed the upper glacier and its accumulation zone from a lower portion of the glacier, ultimately resulting in recession to the top of the escarpment. Similarly, the terminus of the Ladd Glacier thinned and receded 922 m between 1984 and 1989 to the top of a bedrock escarpment that once underlaid a prominent icefall. Since 1989, the portion of the glacier below the icefall has become an ice-cored ground moraine complete with thermokarst terrain. The tensile stresses imposed on glaciers at icefalls, combined with the generally thin nature of Mount Hood's glaciers—i.e., all were less than ∼120 m thick in the mid-1980s (Driedger and Kennard, 1986)—will likely keep the Newton Clark and Ladd glaciers from advancing beyond the escarpments unless very positive mass balance conditions again develop.
Differences in superglacial debris cover may have affected the climate sensitivity of Mount Hood's glaciers. Debris cover decreases glacier ablation rates (e.g., Ogilvie, 1904; Sharp, 1949; Nakawo and Young, 1982; Kirkbride and Warren, 1999; Nakawo et al., 1999), thus reducing glacier sensitivity to mass balance changes (Benn and Evans, 1998). Twentieth century warming has amplified mountain mass wasting processes, thus adding debris to alpine glacier surfaces (Diolaiuti et al., 2003). Coe, Eliot, and Ladd glacier termini have been debris covered since before late ablation season airphotos were taken in September, 1935. Lundstrom (1992) determined that 60% of Eliot Glacier's ablation area was covered with >1 cm of debris which has decreased thermal conduction and ultimately ablation rates. Further, increased debris cover downglacier resulted in a net ablation decrease thus a less negative net mass balance. Despite the extensive debris cover, Eliot Glacier experienced significant terminus retreat (this study) and thinning (Lundstrom, 1993) in recent years. Extensive ablation zone debris cover has also likely reduced ablation rates on Coe and Ladd glaciers rendering them less sensitive to changing temperatures during the past century.
Ken Cameron, the late Donald Lawrence, the late Kenneth Phillips, Larry Price, and Pat Pringle provided valuable suggestions during the early stages of this research. Mazamas and Central Washington University–College of the Sciences research grants funded the project. Mazamas archivist Sharon Howe and librarian Bob Lockerby provided access to historic reports and photographs. Mount Hood National Forest employees Jaime Bradbury, Thomas DeRoo, and Diana Nead helped with acquisition of airphotos and digital orthophoto quadrangles. Chris Daly, Wayne Gibson, Amanda Matske, and George Taylor of the Oregon Climate Service provided climate data, and Marc Fairbanks converted these data to a more user-friendly format. Anthony Gabriel provided valuable statistical assistance. Paul Blanton, Ellen Cameron, Ken Cameron, Andrew Fountain, Martin Kaatz, Nancy Lillquist, William Scott, and Luke Swan critiqued earlier versions of this manuscript. We also thank an anonymous reviewer for comments that significantly improved our manuscript. Ultimately, the responsibility for this paper's form and content is ours alone.
Selected physical characteristics of Mount Hood's glaciers
Measured changes in Mount Hood's glacier termini at select intervals over the period 1901–2001
Spearman rank correlations and associated significance (P) values of cumulative terminus changes, Mount Hood's glaciers, 1901–2000