Introduction

Observation of the different terminus behaviors of North Cascade glaciers, Washington, U.S.A., in response to climate changes during the last century (Reference Tangborn, Fountain and SikoniaTangborn and others, 1990) has prompted the evaluation of the terminus behavior of 38 of these glaciers for the 1890s–1990s period. The objective is to determine the characteristics that lead to the differential terminus behavior.

To complete this task requires examination of the response time of glaciers to climate change. For any glacier there is a lag time (T _s), or reaction time, between a significant climate change and the initial observed terminus response (Reference PatersonPaterson, 1994). It should be noted that T _s cannot be considered a physical property of a glacier and is expected to depend on the mass-balance history and physical characteristics of the glacier. In this paper, T _s is simply defined as the time lag from an observed climate shift to the initial observed change from an advancing to a retreating, or from a retreating to an advancing, terminus.

In addition, for each glacier there is a response time, the time taken to approach a new steady state for a given climate-driven mass-balance change (T _m), referred to as length of memory by Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others (1989). Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others (1989) defined T _m as the time-scale for exponential asymptotic approach to a final steady state (approximately 63% of a full response), resulting from a sudden change in climate to a new constant climate. The magnitudes of T _s and T _m are crucial for interpreting past and current glacier fluctuations and predicting future changes (Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others, 1989; Reference PatersonPaterson, 1994).

In nature a step change in climate causing an evolution from an initial to a final steady state never occurs (Reference Schwitter and RaymondSchwitter and Raymond, 1993). The variability of climate forcing and the continuous changes in the glacier superimpose new disturbances on the response of the glaciers to previous climate changes. This leads to the conclusion that T _m cannot be defined directly from observations in nature. However, T _m cannot be accurately defined from models alone either.

In this study we recognize that the terminus response is influenced by ongoing changes in forcing, limiting the accuracy of determination of T _m from terminus observations. Rather than attempt direct calculation of T _m from terminus observations, we use observations to establish limits on the range of T _m. Our estimates are based on the understanding that, during a period of relatively constant climate, the glacier should advance/retreat to about two-thirds of the final adjustment in its terminus positions, and the rate of advance/retreat of the terminus should be reduced to approximately one-third over a time period of length T _m. The difficulty of examining terminus response to a specific step change is minimized with respect to the post-Little Ice Age (post-LIA) warming, because the climate change and associated terminus response is large compared to climate changes and resulting terminus changes that have occurred in the last half-century (Reference BurbankBurbank, 1981; Reference PorterPorter, 1986; Reference Schwitter and RaymondSchwitter and Raymond, 1993).

In the North Cascades the disparate terminus behavior of glaciers has been noted by Reference HubleyHubley (1956), Reference Post, Richardson, Tangborn and RosselotPost and others (1971), Reference Tangborn, Fountain and SikoniaTangborn and others (1990) and Reference PeltoPelto (1993). The primary goal of this paper is to examine the observed terminus behavior and glacier geographic characteristics of 38 North Cascade glaciers to determine why the terminus behavior history varies. The secondary goal is to identify the ranges of T _s and T _m for the response of North Cascade glaciers to climate changes.

Data Used

The North Cascade Glacier Climate Project (NCGCP) has been observing terminus behavior on 47 North Cascade glaciers and measuring annual balance on 9 glaciers since 1984 (Reference PeltoPelto, 1993, Reference Pelto1996). The primary objective of this research has been to identify the magnitude and timing of glacier response to climate change. This NCGCP dataset is extensive in its breadth, but not in its length of record. Extensive U.S. Geological Survey (USGS) aerial photographic collections and ongoing research by the USGS on South Cascade Glacier have been indispensable to this research, providing long-term records.

Glacier characteristics

On 17 of the aforementioned 38 glaciers more detailed observations are used to calculate theoretical estimates of T _m for comparison with observations of terminus behavior. Each of these 17 glaciers had at least five terminus observations during the period 1850–1998 (Table 1) (Reference LongLong, 1955; Reference HubleyHubley, 1956; Reference Meier and PostMeier and Post, 1962; Reference Post, Richardson, Tangborn and RosselotPost and others, 1971; Reference PeltoPelto, 1993). In order to calculate T _m, terminus velocity u _t, mass balance near the terminus b _t, ice thickness h and glacier length l must be determined. Annual balance measurements have been performed on 9 of the 17 glaciers over a span of > 10 years (Table 1) (Reference KrimmelKrimmel, 1996; Reference PeltoPelto, 1996). Reference KrimmelKrimmel (1999) has observed b _t and u _t on South Cascade Glacier. On the other 16 glaciers b _t and u _t have been directly observed for at least two hydrologic years using stakes drilled into the glacier terminus area by NCGCP (Reference PeltoPelto, 1996; Reference KrimmelKrimmel, 1999). Measurement of u _t relied on at least three points, in the lower 25% of the glacier’s length, which is within 200 m of the terminus in each case. The longer-term annual balance measured on nine of these glaciers indicates that the period of record for terminus mass-balance measurement (1994–96) was close to the 1984–97 mean.

Table 1. Number of terminus observations used in this study for each of the 17 glaciers where T _m has been both derived from field observations and calculated from theoretical equations, and the interval of mass-balance and velocity observations at the terminus of each glacier (b _t and u _t, respectively)

Reference KrimmelKrimmel (1999) observed the length of South Cascade Glacier in 1998. The length of the other 16 glaciers has been taken from the most recent USGS maps compiled from aerial photographs between 1982 and 1984, and adjusted for the observed retreat up to 1998.

Glacier thickness h has been measured on three North Cascade glaciers, namely, South Cascade Glacier (Reference KrimmelKrimmel, 1970), Easton Glacier (Reference HarperHarper, 1993) and Lewis Glacier. The first two are comparatively large glaciers, each with a maximum mean profile depth of 60–80 m. These measurements and those of Reference Driedger and KennardDriedger and Kennard (1986) on Cascade volcanoes, which yielded a mean thickness of 50 m on Mount Rainier and 35 m on Mount Hood, show that these glaciers are comparatively thin. On Lewis Glacier the thickness was measured in a moulin that reached the base of the glacier in 1985, 1986 and 1987. The thickness of each of the other 14 glaciers, where direct measurements were not made, is assumed to be 50 m on thin concave or slope glaciers and 100 m on thicker convex or valley glaciers. In both cases the resulting h is a maximum value, so the T _m also would be a maximum value (Reference Post, Richardson, Tangborn and RosselotPost and others, 1971).

Climate data

A good single climate indicator for North Cascade glacier behavior is the Pacific Northwest Index (PNI), developed by Reference Ebbesmeyer, Cayan, McLain, Nichols, Peterson, Redmond, Betancourt and TharpEbbesmeyer and others (1991). Figure 1 plots the PNI during the period 1896–1997. The index is based on 15 March snow-pack depth at Paradise on Mount Rainier, mean annual temperature at Olga and total annual precipitation at Cedar Lake (all sites in Washington). Positive values reflect negative glacier mass balances.

Fig. 1. Five-year running mean of PNI, 1985–1997 (Ebbesmeyer and others, 1990). Note the increasingly high values from 1895 to 1944, low values from 1945 to 1976 and high values from 1977 to 1997.

Terminus Observations

Since the maximum advance of the LIA there have been three climate changes in the North Cascades sufficient to substantially alter glacier terminus behavior. During the LIA, mean annual temperatures were 1.0–1.5°C cooler than at present (Reference BurbankBurbank, 1981: Reference PorterPorter, 1986). The lower temperatures in the North Cascades led to a snowline lowering of 100–150 m during the LIA (Reference BurbankBurbank, 1981; Reference PorterPorter, 1986). Depending on the glacier, the maximum advance occurred in the 16th, 18th or 19th century (Reference LongLong, 1956; Reference MillerMiller, 1969). North Cascade glaciers maintained advanced terminal positions from 1650 to 1890, emplacing one or several LIA terminal moraines.

Retreat from the LIAM was modest prior to a stillstand in the 1880s (Reference LongLong, 1956; Reference BurbankBurbank, 1981). Reference MillerMiller (1969) mapped the age of terminal moraines in front of Chickamin and South Cascade Glaciers and found in each case that a late-19th-century moraine was emplaced within 100 m of the LIAM.

Just south of the study area, mapping of terminus changes by Reference BurbankBurbank (1981) indicates that rapid and continuous retreat of Mount Rainier glaciers from their LIAM began after the 1880–85 stillstand. Reference LongLong (1953) noted that retreat on Lyman and Easton Glaciers became substantial only after 1890. Reference LongLong (1955) noted that Lyman Glacier has been retreating steadily since the 1890s. Climate warming and retreat began about 1850, but because of the modest retreat and subsequent stillstand or advance of North Cascade glaciers around 1880, 1890 is taken as the approximate time for the climate change that initiated a continuous substantial retreat from the LIA advanced positions in the North Cascades.

This first substantial climate change was a progressive temperature rise from the 1880s to the 1940s, which led to ubiquitous rapid retreat of North Cascade Range alpine glaciers from 1890 to 1944 (Reference RuskRusk, 1924; Reference LongLong, 1955; Reference HubleyHubley, 1956; Reference BurbankBurbank, 1981). Each North Cascade glacier retreated significantly from its LIAM. The retreat did not occur exclusively in the 1890–1944 time period; observations do not exist on most glaciers to distinguish the exact timing of the initial post-LIAM retreat, though retreat was minor before 1890 on glaciers where observations exist. The average retreat of glaciers on Mount Baker was 1440 m from the LIAM to 1950 (Reference PeltoPelto, 1993). The average retreat of the 38 North Cascade glaciers in this study was 1215 m.

The 1895–1923 PNI is rising, but has a comparatively low mean (−0.34) that is still capable of generating retreat on North Cascade glaciers that remain in advanced positions from the LIA. The 1924–44 PNI average was high at 0.44. This warm, dry period was one of rapid glacier retreat worldwide (Reference LongLong, 1955; Reference HubleyHubley, 1956; Reference BurbankBurbank, 1981). The second substantial change in climate began in 1944, when conditions became cooler and precipitation increased (Reference HubleyHubley, 1956; Reference TangbornTangborn, 1980). The climate change in 1944 is evident in the PNI record (Fig. 1). The 1945–76 mean PNI was −0.37. This drop of 0.76 in the PNI average compared to the previous time period marks the climate change that initiated the advance of some North Cascade glaciers and the more positive mass balance for that period (Reference HubleyHubley, 1956; Reference TangbornTangborn, 1980).

Reference HubleyHubley (1956) and Reference LongLong (1956) noted that North Cascade glaciers began to advance in the early 1950s, after 30 years of rapid retreat. This change was reflected in mass-balance values. A runoff–precipitation model constructed for South Cascade Glacier (Reference TangbornTangborn, 1980) yielded a mean annual balance of −1.15 m a⁻¹ from 1924 to 1944, compared to −0.15 m a⁻¹ from 1945 to 1976.

Approximately half the North Cascade glaciers advanced during the period 1950–79 (Reference HubleyHubley, 1956; Reference Meier and PostMeier and Post, 1962). Advances of Mount Baker glaciers ranged from 120 to 750 m, an average of 480 m, and ended in 1978 (Reference HeikkinenHeikkinen, 1984; Reference HarperHarper, 1993; Reference PeltoPelto, 1993). All 11 Glacier Peak glaciers that advanced during the period 1950–79 emplaced an identifiable maximum advance terminal moraine; the mean advance was 295 m. Of the 47 glaciers observed by NCGCP during the period 1984–98, 15 advanced during the period 1950–78.

The final climate change was a step change in 1977 to a drier, warmer climate in the Pacific Northwest (Reference Ebbesmeyer, Cayan, McLain, Nichols, Peterson, Redmond, Betancourt and TharpEbbesmeyer and others, 1991). The mean PNI for 1977–98 was 0.53, even higher for 1924–44, indicating a warmer, drier period that re-established the ubiquitous retreat of North Cascade glaciers (Reference PeltoPelto, 1993). This change impacted glacier mass balance, alpine stream-flow and alpine snowpack (Reference Ebbesmeyer, Cayan, McLain, Nichols, Peterson, Redmond, Betancourt and TharpEbbesmeyer and others, 1991). The impact on North Cascade glacier mass balance is evident from the USGS long-term record of South Cascade Glacier (1958–98), where mean annual balance was −0.15 m a⁻¹ for the period 1958–76, in contrast to −1.00 m a⁻¹ for the period 1977–98 (Reference KrimmelKrimmel, 1999).

The retreat and negative mass balances of the 1977–98 period have been noted by Reference HarperHarper (1993), Reference PeltoPelto (1993, Reference Pelto1996) and Reference KrimmelKrimmel (1994, Reference Krimmel1999). By 1984, all the Mount Baker glaciers, which were advancing in 1975, were again retreating (Reference PeltoPelto, 1993). In 1997–98, NCGCP measured the retreat of eight Mount Baker glaciers from their recent maximum position (late 1970s-early 1980s). The average retreat had been −197 m. Photographs from J. B. Richardson and observations by A. Post (personal communication, 2000) show that all of the Mount Baker termini are still in advance of their 1940 position. However, the glacier region between the current terminus and 1940 terminus is nearly stagnant on each of the glaciers. Between 1979 and 1984, 35 of the 47 North Cascade glaciers observed annually by NCGCP had begun retreating. By 1992 all 47 glacier termini observed by NCGCP were retreating (Reference PeltoPelto, 1993) and two, Lewis and Milk Lake Glaciers, had disappeared.

Initial Terminus Response

The time between the onset of a mass-balance change and the onset of a significant change in terminus behavior is called the initial terminus response time or reaction time (T _s). As indicated previously, T _s is a descriptive quantity that quantifies the time lag between climate forcing and terminus response for a particular climate event, rather than a physical property of a glacier. T _s in this study is based solely on the first observed terminus change from retreat to advance after 1944, and from advance to retreat after 1976. T _s has been identified from the response of North Cascade glaciers to the cooler, wetter weather beginning in 1944 (Reference LongLong 1955, Reference Long1956; Reference HubleyHubley, 1956; Reference TangbornTangborn, 1980), and to the subsequent warmer, drier conditions beginning in 1977 (Reference PeltoPelto, 1988, Reference Pelto1993; Reference Ebbesmeyer, Cayan, McLain, Nichols, Peterson, Redmond, Betancourt and TharpEbbesmeyer and others, 1991; Reference HarperHarper, 1993; Reference KrimmelKrimmel, 1994).

Focusing on 21 North Cascade glaciers that responded to these two climate shifts, all having an area of < 10 km², the initial terminus response invariably is <16 years (Reference HubleyHubley, 1956; Reference HarperHarper, 1993; Reference PeltoPelto, 1993). Table 2 lists 21 North Cascade glaciers where T _s has been noted for both advance and retreat in the post-1944 period by Reference LongLong (1955, Reference Long1956) or Reference HubleyHubley (1956). In each case the glaciers were observed to be in retreat during the 1940s, and subsequently advanced within 16 years of the climate change. Table 2 also notes the response of the same glaciers from a period of advance by each glacier in the early 1970s to one of retreat by 1988, 12 years after the climate change (Reference PeltoPelto, 1988, Reference Pelto1993; Reference HarperHarper, 1993). The observed T _s is not significantly different on individual glaciers for initiation of advance vs initiation of retreat.

Table 2. Date of first observed advance following the 1944 climate change, and of first observed retreat following the climate change in 1976/77

Many North Cascade glaciers did not respond to these two climatic changes. This may be due to a longer T _s, or more likely is due to either an ongoing retreat caused by continuing negative mass balances, or the fact that climate change was insufficient to substantially alter mass balance on these glaciers.

Analysis and Classification of Glacier Response

The 38 North Cascade glaciers where the terminus history has been determined for the period 1890–1998 exhibit three distinct patterns (Table 3): (1) Retreat from 1890 to 1950, then a period of advance from 1950 to 1976, followed by retreat since 1976. (2) Rapid retreat from 1890 to approximately 1950, slow retreat or equilibrium from 1950 to 1976 and moderate to rapid retreat since 1976. (3) Continuous retreat from 1890 to the present. Assignment of a glacier’s type is based solely on its terminus behavior in this analysis.

Table 3. Retreat of North Cascade glaciers for three intervals since the LIAM. The mean slope of the glacier surface, the mean altitude of the glacier and the surface area of the glacier are also listed

Characteristics of Glacier Types

Each of the three glacier types was established based solely on terminus behavior history, but it is apparent that each type has specific characteristics (Figs 2–5). Figure 6 illustrates the different terminus behavior histories of North Cascade glaciers. The slower initial response to climate change of type 3 glaciers is evident. The long-term result of the slow start is a more persistent retreat. Table 3 lists the mean slope, mean altitude and area of each of the 38 glaciers. Mass-balance measurements in the accumulation zone exist for 12 of the glaciers in this study. Type 1 glaciers have the highest mean elevations (2200 m a.s.l), largest mean slope (0.42 ± 0.07), highest measured mean accumulation (Reference PeltoPelto, 1988, Reference Pelto1996), most extensive crevassing and highest measured terminus-region velocity (20 ± 3 m a⁻¹). Type 3 glaciers have the lowest slopes (0.23 ± 0.06), least crevassing and lowest mean terminus velocity (5 ± 3 m a⁻¹) of any of the glacier types. Type 2 glaciers have on average a lower slope (0.35 ± 0.08), a lower terminus-region velocity (7 ±4 m a⁻¹), less crevassing and a lower mean accumulation rate than type 1 glaciers (Fig. 3).

Fig. 2. (a) Glacier Peak in 1972. From right to left, Chocolate to North Guardian and Dusty Glaciers are typical type 1 glaciers, steep, crevassed and with high accumulation rates. The glaciers are just ending a period of advance at the time of this photograph (A. Post, USGS). (b) Squak Glacier on Mount Baker in 1990. This longitudinal view of the glacier indicates the consistent degree of crevassing and rapid flow of this type 1 glacier.

Fig. 3. (a) Lynch Glacier in 1972. A type 2 glacier, it is steep, but with fewer crevasses, and has not stopped retreating during the past century. Foss Glacier on the right is a type 3 glacier with a lower slope and less crevassing (A. Post, USGS) (b) Columbia Glacier in 1986. This type 2 glacier receives particularly large avalanche accumulation.

Fig. 4. (a) Honeycomb Glacier in 1967, a type 3 glacier, which has no crevassing or motion in the terminus region. This is still a large glacier, though it did retreat 480 m during the period 1979–97 (W Long). (b) Foss Glacier in 1995. The low slope and lack of crevassing of this type 3 glacier is evident.

Fig. 5. North and South Klawatti Glaciers from the north. Note the difference in surface slope and crevassing of South Klawatti Glacier (left) and North Klawatti Glacier (right) (A. Post, USGS).

Fig. 6. Cumulative terminus-position change on seven North Cascade glaciers since approximately 1850. Easton, Mazama, Boulder and Rainbow Glaciers are type 1 glaciers and all show a period of advance. Columbia Glacier, a type 2 glacier, has a moderate but continuous retreat. Lyman and South Cascade Glaciers retreated slowly at first, but have now accelerated.

There is no significant relation between aspect and glacier type. A larger mean slope, higher accumulation rates, more extensive crevassing and higher velocity either reflect or lead to increased glacier velocities and longitudinal strain rates. The greater the longitudinal strain rate the faster the glacier can adjust to changing climate conditions (Reference PatersonPaterson, 1994). The terminus-region velocities on large alpine glaciers may be quite independent of glacier velocity as a whole. On the smaller North Cascade glaciers, the terminus-region velocity is, on the other hand, a good indicator of mean glacier velocity.

The three glacier types show that persistent differences in glacier behavior can be explained in terms of the basic characteristics of the glacier which in turn determine response time, and are not specific to individual glaciers. This is reinforced by the exceptionally high degree of correlation in annual balance between North Cascade glaciers (Reference PeltoPelto, 1997). The lowest cross-correlation value for annual balance, between any pair of the 14 glaciers observed by the North Cascades National Park Service, USGS and NCGCP is 0.79 (Pelto and Riedel, in press).

The different behavior of adjacent glacier termini based on differing topographic characteristics in the North Cascades was observed on South and North Klawatti Glaciers (Reference Tangborn, Fountain and SikoniaTang-born and others, 1990). From 1947 to 1961, North Klawatti Glacier lost a volume equivalent to a mean thickness of 8.3 m, continuing its ongoing retreat, while South Klawatti Glacier gained a thickness equivalent to a mean thickening of 5.8 m (Reference Tangborn, Fountain and SikoniaTangborn and others, 1990). North Klawatti Glacier is a type 3 glacier, and South Klawatti a type 1 glacier. The difference in the degree of crevassing alone indicates a significant difference in flow. The adjacent glaciers’ differing responses fit the overall pattern for glaciers of their respective types in the North Cascades, and are not an anomalous case. We identified no important microclimatic effects that created differing mass-balance conditions on glaciers across the North Cascades (Reference PeltoPelto, 1996, Reference Pelto1997).

An even more striking example is that of Neve and Ladder Creek Glaciers, which share the same accumulation zone and have termini that both end at approximately 1680 m. The shared accumulation zone between 2000 and 2400 m flows into a pass at 2000 m, where the glacier turns both east and west. Ladder Creek Glacier flows northwest and is a type 1 glacier with a steep, shorter terminus region, 1200 m from pass to terminus, comparatively rapid velocity and was noted to advance by Reference HubleyHubley (1956). Neve Glacier is a type 2 glacier, slightly larger than Ladder Creek Glacier and with a longer terminus region, 1920 m from pass to terminus. This results in a gentler slope and consequently slower velocity. This glacier did not advance, though crevassing did increase within 500 m of the terminus, slowing the retreat to a standstill.

Observation of different responses of glacier types to climate change is not unique to the North Cascades. In Switzerland a sample of 38 glaciers with 150 year long terminus records was divided into four classes, with different types of terminus response for each class (Reference Herren, Hoelzle and MaischHerren and others, 1999):

1. (1) Large valley glaciers such as Gornergletscher which has retreated rapidly and continuously.

2. (2) Small mountain glaciers such as the Saleina, which advanced twice during the 20th century, though retreat was more pronounced.

3. (3) Large mountain glaciers, such as the Tschierva, which follow the same pattern as Saleina except with more significant retreat.

4. (4) Small cirque glaciers, such as Gran Plan Névé, which retreated slowly throughout the 20th century.

Theoretical Estimates of Response Time

A major difficulty in the identification of T _m is that after a climate change climate conditions do not reach a new steady state for periods comparable to the T _m of glaciers. Each glacier is then adjusting to the continually changing climate conditions and never achieves a steady state, due to the non-steady climate (Reference Schwitter and RaymondSchwitter and Raymond, 1993). Reference Schwitter and RaymondSchwitter and Raymond (1993) noted that these difficulties are minimized with regard to changes from the LIAM to the present, because the basic climate change since the late 19th century has been from a LIA climate favorable to glaciers and a post-LIA climate unfavorable for glaciers. Changes in North Cascade terminus behavior and glacier thickness from the LIAM to the present are large compared to changes in response to more recent climate changes (Reference Schwitter and RaymondSchwittter and Raymond, 1993).

The observed terminus record of North Cascade glaciers indicates a T _m range of 20–30 years on type 1 glaciers, approximately 40–60 years on type 2 glaciers and a minimum of 60–100 years on type 3 glaciers. How do these values compare to those calculated from the equations of Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others (1989)?

Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others (1989) compared two means of calculating T _m:

(1)

and

(2)

T _m in these equations is potentially dependent on four variables: the glacier length, L, velocity of the glacier at the terminus, u _t, the thickness of the glacier, h, and the net annual balance at the terminus, b _t. Equation (1), which was proposed by Reference NyeNye (1960), produces longer response times of 100–1000 years, and Equation (2) response times of 10–100 years (Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others, 1989). The variable f is a shape factor that is the ratio between changes in thickness at the terminus and changes in thickness at the glacier head (Reference Schwitter and RaymondSchwitter and Raymond, 1993). Similar changes in ice thickness will yield a value of f = 1; f – 0.5 corresponds to a linear decrease of thickness change from a maximum at the terminus to zero at the head. The mean value of f has been determined as 0.3 (Reference Schwitter and RaymondSchwitter and Raymond, 1993), and this value of f is applied. This equation is quite sensitive to terminus velocity, which is often spatially inconsistent.

Table 4 displays the variables used in determining T _m for 17 North Cascade glaciers, the calculated T _m from Equation (1) and T _m from Equation (2). Each variable, except h, has been observed on each glacier by the USGS (South Cascade Glacier) or NCGCP. Equation (2) yields values that are lower than the terminus observation-based estimates of T _m for North Cascade glaciers of types 2 and 3, but the difference is smaller for type 1 glaciers. Equation (1) overestimates T _m and, because of the wide spatial variability of u _t, is not expected to yield a consistently accurate result on alpine glaciers.

Table 4. Variables used in Equations (1) and (2) and the calculated and observed T _m

South Cascade Glacier, like all type 3 glaciers, is still adjusting to post-LIA warming and the discontinuous but progressive warming of the past century. This is not unique to the North Cascades: many alpine glaciers have not yet fully adjusted to post-LIA warming. Ptarmigan and Lemon Creek Glaciers, Alaska; Gornergletscher and Rhonegletscher, Switzerland; Athabasca Glacier, Canada; and several glaciers in the Darwin Cordillera, Chile have retreated continuously during the past century (Reference Holmlund and FuenzalidaHolmund and Fuenzalida, 1995; Reference Marcus, Chambers, Miller and LangMarcus and others, 1995; Reference Herren, Hoelzle and MaischHerren and others, 1999).

Conclusions

North Cascade glaciers had a varied terminal response to the 1944 climate change, with only type 1 glaciers advancing. Based on this study, the failure of type 2 and 3 glaciers to advance is a function of their incomplete adjustment to the post-LIA progressive warming. Thus, they were still significantly out of equilibrium in 1944, after approximately a half-century of retreat (Reference BurbankBurbank, 1981), and the modest positive mass balances did not trigger a glacier advance. Reference PeltoPelto (1996, Reference Pelto1997) noted the high cross-correlation in observed annual balance on North Cascade glaciers (Fig. 7). This similarity is true regardless of glacier type, and shows that microclimates are not the key to differences in behavior. Instead it is the physical characteristics slope, terminus velocity, thickness and accumulation rate of the glacier that determine its recent terminus behavior and response time. Response is thus the key to understanding the differences in terminus behavior of North Cascade glaciers during the study interval.

Fig. 7. Annual balance of nine North Cascade Glaciers, 1984–98, in m w.e. (Reference PeltoPelto, 1996; Reference KrimmelKrimmel, 1999).

An example is the adjacent North Klawatti and South Klawatti Glaciers. South Klawatti advanced during the period 1950–75 and North Klawatti continued to retreat. The glaciers have different area–altitude distributions, to which Reference Tangborn, Fountain and SikoniaTangborn and others (1990) attributed the differential terminus response. However, the different area–altitude distribution is both a result of slower response and a reflection of the differing topographic setting.

Reference PorterPorter (1986) noted that many alpine glaciers experienced nearly synchronous reversals in terminus behavior around 1950. This change in glacier terminus behavior prompted Reference Jóhannesson, Raymond and WaddingtonJóhannesson and others (1989) to suggest that alpine glacier behavior is dominated by short-term climate effects. In the North Cascades, this synchronous reversal to advance and later to retreat of only type 1 glaciers indicates that in the North Cascades only glaciers close to equilibrium underwent a reversal in terminus behavior due to shortterm climate effects. Glacier termini such as those of Honey-comb, Lyman, Columbia, Milk Lake and South Cascade Glaciers were only modestly affected by recent short-term climate changes in the North Cascades.

North Cascade glaciers occupy an exceptionally temperate maritime climate for glaciers. T _s on North Cascade glaciers is short (Reference HubleyHubley, 1956), ranging from 4 to 16 years in response to both positive and negative mass-balance changes.

T _m varies considerably even between similarly sized glaciers in this region. The key variables that decrease T _m are factors that increase mean glacier velocity, accumulation rates and glacier slopes in particular. The response times of 30–100 years for most of these small glaciers indicate that with a substantial climate change the initial response may be rapid, but full adjustment is not rapid in the North Cascade Range.