Regional Accounts Of Glacial History

The PEP 1 transect spans the cordillera of the western Americas and contains glaciers over most of its length (Fig. 1). However, in the central part of the transect, glaciers are commonly small and are usually widely scattered and restricted to high mountain areas, frequently well above the tree line. The primary research at most of these glaciers has been focused on the deposits and landforms of the last glacial and/or early Holocene. Dating control for many glacial events is poor and/or controversial (see, e.g., Davis, 1988), and, although LIA deposits are readily recognized by morphology, "freshness," and position in the landscape, few absolute age determinations are available. Even in the Sierra Nevada (ca. 38°N), where the LIA was originally defined by Matthes (1939), the most re-

cent account dates only the Matthes (LIA) moraines as younger than 720 14C years B.P. based on tephrochron-ology (Clark and Gillespie, 1997). Limited observations of former glacier extent and snow line levels are available from a variety of historical and archival sources in México and Ecuador that extend back 400500 years (e.g., Hastenrath, 1981, 1997; White, 1981; Vázquez-Selen and Phillips, 1998). Such records indicate, for example, significantly lower historical snow lines near Quito and little change in glacier equilibrium line altitude (ELA) from the 1500s through the early 1800s (Hastenrath, 1997). The most extensive glaciated area in the northern Andes is in Peru, but there are few dated records of LIA glacier fluctuations prior to the twentieth century (see Grove, 1988). Excellent paleo-environmental records have been recovered from the ice cores at Quelccaya, Huascarán, and Sajama in the high Andes (e.g., Thompson et al., 1984, 1998), but such records are not the focus of this chapter. There is widespread evidence of the rapid recession and disappear ance of tropical glaciers in the high Andes in the late twentieth century, plus the degradation of any potential ice core records they might contain (Thompson, 1995; White, 1981; Granados, 1997); the glaciers on Pico Bolivar near Merida (see Schubert, 1972, 1992), visited during the PEP 1 meeting, were simply vestigial ice patches. The few available studies of these tropical glaciers indicate that the mass balance-climate relationships are complex (Francou et al., 1995; Ribstein et al., 1995).

Therefore, in presenting regional summaries of the glacier history along the transect, we focus on records from the more heavily glaciated regions poleward of ca. 32°S and 47°N. The relative widths and lengths of the mountain belts in South and North America allow a clearer demonstration of latitudinal variation in the climatic controls of glacier fluctuations in the South American examples. Therefore, we begin with that record.

8.2.1. Southern South America (30°-550S)

The elevation of the Andes decreases southward from 6000-7000 m in the central Andes of Argentina and Chile to less than 2000 m in southern South America. Along this latitudinal gradient, climate varies from dry subtropical in the north to wet temperate in the south. Precipitation ranges from less than 1 m per year in the central Andes of Chile and Argentina to nearly 10 m on the western coast of southern Patagonia. There are also strong west-east precipitation gradients across the Andes. In the glaciated areas of the central and southern Andes, mountain peaks may be well over 1000 m above the regional snow line (Casassa, 1995).

More than 30 years ago, Mercer (1965) proposed a scheme of three major glacial advances in southern South America during the Holocene. A fourth neogla-cial advance during the Holocene has been recently proposed for southern South America (Aniya, 1995). The most recent of these events in southern Patagonia was dated between A.D. 1600 and 1850 (Mercer, 1965, 1968, 1976). Although there has been considerable subsequent interest in Holocene glacier fluctuations in this region (e.g., Clapperton and Sugden, 1988; Grove, 1997), almost all late Holocene glacier chronologies are based on radiocarbon dating of organics (usually wood) in subsurface deposits. The relatively large uncertainties associated with radiocarbon ages in this time frame (see earlier) and the associated problems of limiting dates suggest that most subsurface radiocarbon ages can be resolved only within ±50-100 years. Decadal-annual resolution is rarely attainable, and then only for a few studies that utilize tree-ring dating, lichens, or historical records during the last few cen turies. Synchroneity can be established only in the broadest terms for events predating the local LIA maximum.

An intensive search in the glaciological literature from South America indicates that high-resolution glaciological records for the past 1000 years are extremely rare. Most of these records were obtained during the twentieth century by using aerial photographs (since midcentury) or satellite images (e.g., Warren and Sugden, 1993; Aniya and Enomoto, 1986; Aniya et al., 1997) and are not directly relevant to this chapter. At most localities along the Andes, glaciers reached their LIA maxima prior to 1950. In the following sections, we briefly review the few high-resolution records of glacier fluctuations for the last 1000 years in the central Andes of Argentina and Chile, in northern and southern Patagonia, and in the subantarctic domain. For each of these regions, precisely dated glacial advances are compared with dendrochronological records to cross validate the glacial chronology and to evaluate the possible climate signal represented by these records.

8.2.2. Central Andes of Argentina and Chile

Videla (1997) used old scientific reports, maps, and photographs to document precisely changes in glacier extent on the eastern slope of the central Andes (32°-33°S). The glaciers of the Río del Plomo, Mendoza, reached their maximum LIA extents at the beginning of the twentieth century. According to Videla (1997), no Holocene moraines or other glacial deposits have been observed downstream from the 1910 limit. During the twentieth century, glaciers have generally retreated. Since ca. 1910, the Alto del Río del Plomo Glacier has retreated ca. 400 m in elevation and almost 5000 m in length.

On the western slopes of the Andes, at the same latitude as the Río del Plomo basin (32°S), LaMarche et al. (1979) developed the northernmost tree-ring chronology from Austrocedrus chilensis in central Chile. This chronology has been used largely as a proxy record of precipitation variations for central Chile (LaMarche, 1975; Boninsegna, 1988; Villalba, 1990a). Recently, the site was resampled, and the data from the new collection were merged with LaMarche's tree-ring series. The new chronology, which covers the interval A.D. 9561996, provides a basis to compare glaciological and dendrochronological records from the xeric Andes at 32°S. Replication exceeds 10 samples after A.D. 1200. Except for two short intervals centered in 1860 and 1890, the 1820-1906 interval is the longest period of high sustained growth in this record. Figure 2 shows that Austrocedrus ring widths are strongly correlated with precipitation during the period of the instrumental record. This result suggests that the 1820-1910 interval was the longest wet period in the central Andes during the past eight centuries.

Under the extreme dry conditions that prevail in the central Andes, glacier variations are strongly affected by precipitation variations (Leiva et al., 1986). As a consequence, tree-ring variations from the precipitation-sensitive chronology from central Chile support Videla's (1997) observations. At about 32°S, glaciers reached their maximum extent at the beginning of the twentieth century, coincident with the termination of the wettest period in the Andes during the past eight centuries (as inferred from tree-ring variations). In agreement with these results, decadal mean values of rainfall at two stations in central Chile, La Serena (30°S) and Santiago (33°S), show that marked wet years at the end of the nineteenth century were followed by a decreasing trend continuing to the present (Aceituno et al., 1993). At La Serena, decadal mean deviations in precipitation during the period 1880-1900 were greater than 1 SD above the long-term mean.

At 35°S, in a transitional zone to the wetter climatic conditions prevailing southward, glaciers reached their maximum extension during the last millennium sometime earlier than the glaciers located at 32°S. On the western Chilean slopes, radiocarbon ages indicate advances of the Los Cipreses Glacier (34°S) at the beginning of the fourteenth century and around the 1860s (Rothlisberger, 1987). Historical records show that the Ada Glacier extended down to 1800 m in 1858. The elevation of the toe had risen to 1930 m by 1882 and to 2500 m by 1980 (Cobos and Boninsegna, 1983). At about 35°S, in the upper Río Atuel basin in Argentina, the glaciers have undergone almost constant recession since they were observed by Hauthal (1895). At that time, the lower portion of the Humo Glacier, in the Río Atuel basin, consisted of large pieces of dead ice covered by debris (Cobos and Boninsegna, 1983). Cobos and Boninsegna also reconstructed streamflow of the upper Río Atuel from 1534-1968 using three Austroce-drus chronologies from central Chile. The reconstruction shows greatly increased runoff between 1820 and 1850, with a second runoff peak in the early 1900s. These high runoff periods could reflect higher precipitation, greater glacier melt and runoff, or a combination of these effects. However, the lack of dating for glacier fluctuations prior to 1895 prevents the exploration of possible linkages between these runoff peaks and glacier fluctuations. It seems probable that glaciers reached their maximum LIA extent in the mid-nineteenth century and underwent significant melting and stagnation by 1895. They have retreated consistently since the 1910s.

FIGURE 2 Precipitation and glacier fluctuations for central Chile and adjacent Argentina. The upper curves are instrumental precipitation records from Santiago and La Serena, Chile (25-year cubic spline filter). Glacier fluctuations are from Videla, 1997. The lower curve is an indexed tree-ring chronology for Austrocedrus chilen-sis in central Chile, A.D. 956-1996 (LaMarche et al., 1979, revised and updated). This is interpreted as a proxy precipitation record.

FIGURE 2 Precipitation and glacier fluctuations for central Chile and adjacent Argentina. The upper curves are instrumental precipitation records from Santiago and La Serena, Chile (25-year cubic spline filter). Glacier fluctuations are from Videla, 1997. The lower curve is an indexed tree-ring chronology for Austrocedrus chilen-sis in central Chile, A.D. 956-1996 (LaMarche et al., 1979, revised and updated). This is interpreted as a proxy precipitation record.

8.2.3. Northern Patagonia

For northern Patagonia, most information is available from the glaciers of Monte Tronador, located on the border between Chile and Argentina at 41°S. Radiocarbon ages of tree trunks overridden by the Río Manso Glacier indicate major glacier advances in ca. A.D. 1040, 1330, 1365, 1640, and during the period 1800-50 (Roth-lisberger, 1987). For the nearby Frías Glacier, two major glacial advances were identified for the intervals 1270-1380 and 1520-1670 (Villalba et al., 1990). However, only the latter advance has been adequately dated by using tree-ring records. Dating for the massive moraine, which represents the maximum extent of the Frías Glacier during the past 1000 years, is based on ring counts from the oldest tree found on the moraine, plus precise dates from trees damaged by the glacier along its margin during the latest neoglacial advance (Villalba et al., 1990).

Measurements of the terminal position of the Frías Glacier since 1976 indicate a strong relationship between summer climatic conditions (particularly tem perature variations) and fluctuations of the position of the Frías Glacier front (Villalba et al., 1990). On the basis of the sparse available data, Warren et al. (1995) also propose that glacier fluctuations on the more continental eastern slopes of the Andes are responding to temperature variations. Figure 3 compares these glacier records with a recently developed summer temperature reconstruction for the eastern side of the northern Patagonian Andes based on 17 multimillennial records for Fitzroya cupressoides in northern Patagonia (Villalba et al., 2000). The most striking feature in the 1000-year-long temperature reconstruction is the long, cold interval that extends from ca. 1500 to 1660. Severe, cold summers were particularly frequent from 1630-1650. The occurrence of this century-long cold interval is consistent with the glacial record, which indicates that the maximum extent of the Frías Glacier during the last millennium was between ca. 1660 and 1670 at the end of this cold period. Dating of several of the recessional moraines at the Frías Glacier also coincides with the timing of colder-to-warmer transitions of summer temperatures (see Fig. 3 and discussion in Villalba et al., 1990).

0 0

Post a comment