Radiocarbon Years Bp

FIGURE 8.1 A reconstruction of Holocene treeline fluctuations in southwestern Keewatin, Northwest Territories, Canada.Treeline position is based on radiocarbon-dated tree macrofossils in situ north of the present treeline, and on dates on buried forest and tundra soils north and south of the modern treeline (Sorenson and Knox, 1974).

more extensive boreal forest zone show an additional warming of 1 °C (in summer) and 4 °C in spring. This was the result of higher net radiation due to lower albedo under forest cover (compared to tundra), which was especially critical in the snow-covered spring season.

If air mass boundaries are a determinant of the forest border, then mapping pa-leoforest limits may provide an important insight into the dynamic climatology of the past (Ritchie and Hare, 1971). Unfortunately, a number of factors make such interpretations difficult. Northward treeline migration during a climatic amelioration is more rapid than southward treeline migration in response to a climatic deterioration. Once established, trees may survive periods of adverse climate, and the treeline will only slowly "recede" as the trees that die are not replaced (see alpine treelines; LaMarche and Mooney, 1967). This process may differ from one location to another. For example, in the forest-tundra zone of northeastern Quebec (east of Hudson Bay), charcoal from formerly extensive coniferous forests indicates that the treeline (i.e., the limit of continuous forest) does not appear to have migrated north-south en masse. Rather, the evidence suggests that the modern tundra-forest ecotone in this area is the product of a formerly more extensive (mid-Holocene) forest that experienced lower temperatures and periodic destruction by fire in the late Holocene, leading to the predominantly tree-less landscape, with the isolated stands

Westerlies Arctic View Bryson

FIGURE 8.2 Modern treeline In relation to modal, mean, or median position of Arctic Front in recent years (for definitions of positions, see Krebs and Barry, 1970; Bryson, 1966). Proposed location of front at 8000 yr B.P. is shown, based on macrofossil and palynological evidence.The 8000 yr B.P position implies a higher amplitude upper level westerly flow pattern at that time (Ritchie and Hare, 1971).

FIGURE 8.2 Modern treeline In relation to modal, mean, or median position of Arctic Front in recent years (for definitions of positions, see Krebs and Barry, 1970; Bryson, 1966). Proposed location of front at 8000 yr B.P. is shown, based on macrofossil and palynological evidence.The 8000 yr B.P position implies a higher amplitude upper level westerly flow pattern at that time (Ritchie and Hare, 1971).

of trees seen today (Payette and Gagnon, 1985). Trees were unable to reproduce in the cooler conditions after 3000 yr B.P. (and especially from -650-450 yr B.P.) so the forest-tundra boundary seen today reflects the combined influence of climate and fire history (Payette and Morneau, 1993). Furthermore, detailed studies show the importance of changing growth form in response to climatic change, particularly changes in snow depth. Black spruce (Picea mariana) at the northern treeline in Quebec can grow in both upright and prostrate (krummholz) forms, as is typical of many species at their arctic and alpine range limits. In the warmer interval from -A.D. 1435-1570, spruce was growing mainly as erect trees, but after 1570 cold conditions led to stem-dieback so only the snow-protected krummholz forms survived (Payette et al., 1989). Milder temperatures (and possibly heavier snowfall) in the eighteenth century allowed more shoots to form, but very severe winters from -1801-1880 again killed many exposed stems (Lavoie and Payette, 1992). Thus, trees at the northern forest boundary may adapt to cooler climatic conditions by adopting a more prostrate growth form, awaiting more favorable conditions to assume erect growth. In such areas, the idea of a north-south migration of treeline is clearly too simplistic. No doubt further studies of 14C-dated macrofossils, pa-leosols, growth form analysis, and tree ring variations will produce a more coherent picture of paleoclimate at this important ecotone.

8.2.2 Alpine Treeline Fluctuations

In a survey of upper treelines in different climatic zones, C. Troll remarked: "It is absolutely clear that upper timberlines in different parts of the world cannot be climatically equivalent, not even in a relatively small mountain system such as the Alps or Tatra mountains" (Troll, 1973, p. A6). In the paleoclimatic interpretation of macrofossils from above the modern treeline, we must therefore recognize that treeline variations in one area may result from different climatic factors than in another area, and that controls on tree growth may be complex.

In the arid subtropics, treeline is influenced by both temperature and the availability of moisture. In the humid tropics, where seasonal temperature differences are extremely small, the transition to a treeless zone is generally quite abrupt, apparently related to a critical temperature threshold. In mid to high latitudes (particularly in the Northern Hemisphere) treelines are more diffuse, often reflecting strong topoclimatic controls. Climatic studies generally point to summer or July temperature as the major controlling factor in these latitudes (Wardle, 1974), although mean temperatures are only a convenient proxy for the actual controls, which probably involve the frequency and timing of extreme events (Tranquillini, 1993; Holtmeier, 1994). Nevertheless, evidence of higher treeline in the past is generally interpreted as indicating warmer summer temperatures, a lapse rate of 0.6-0.7 °C per 100 m commonly being used to assess the magnitude of temperature change. For example, Dahl and Nesje (1996) estimated that summer temperatures in northern Sweden have varied by -1.5 °C over the Holocene, based on treeline variations of <220 m (Fig. 8.3). However, Karlen (1976) pointed out numerous problems in interpreting such data, among them the following:

(a) The incomplete nature of the macrofossil record; the highest trees may not have been found, or indeed may not have been preserved. Furthermore, in some areas, mountain summits may not extend far enough beyond the modern treeline to give a maximum estimate on former treeline extent (LaMarche, 1973). Thus, paleotreeline evidence should be considered as providing a minimum paleotemperature estimate only.

(b) The present altitude of the treeline is often not precisely determined and may not be in equilibrium with modern climate (Ives, 1978); a recent history of fire, overgrazing, avalanches, gales, or insect infestation may have resulted in a treeline well below the potential maximum for modern climatic conditions (Griggs, 1938).

(c) Trees take many years to become established during periods of favorable climate and may not have reached their highest position until after the temperature maximum; again this factor would tend to make any paleotemperature estimates based on treeline fluctuations minimum only.

(d) Treeline in some areas may have been affected by regional isostatic uplift following deglaciation of the region; such an effect must be taken into account when assessing the treeline record, particularly from the early Holocene.

It is also worth noting that the time of treeline establishment may be of more significance than the time of tree death (which may not have been climatically related). Denton and Karlen (1977), for example, have dated wood from trees growing above the modern treeline that were killed by explosive volcanic eruptions around 1800 and 1150 yr B.P. Ideally then, one should attempt to date the innermost wood fraction either by radiocarbon analysis, or dendrochronology, or both (LaMarche and Mooney, 1967, 1972). Often, the inner wood has weathered away, or decayed, making this impossible and providing only a minimum estimate on the timing of alpine treeline advances. Karlen (1976) considers that higher treelines in the past probably indicate mean temperatures above contemporary values for periods of 50-100 yr, in order for young seedlings to become established and for the treeline as a whole to "advance." Once established, trees can survive periods of adverse climate that are perhaps as long as favorable climate periods; thus treeline variations are a low frequency record of past climate and are biased towards recording warmer intervals of >50 years in duration.

Bearing all these factors in mind, it is perhaps not surprising to find that the alpine treeline record, considered on a worldwide basis, is extremely complex. Nevertheless, there is evidence that treelines were higher during the early to mid-Holocene in many areas, perhaps reflecting a globally extensive warmer interval. The most comprehensive studies have been carried out in Scandinavia where hundreds of birch, alder, and pine macrofossils from above the modern treeline have been 14C-dated. These show unequivocally that the altitudinal limit of trees was well above modern treeline elevation from soon after deglaciation of the mountains (-9000 yr B.P.) to the mid-Holocene. In northern Fennoscandia the maximum extent of pine was -5 ka B.P., becoming lower especially after -3.5 ka B.P. (Eronen and Huttunen, 1987; Karlen, 1993). In central Sweden (Fig. 8.3), pine grew at least 220 m above its mid-twentieth century limit from 9-7 ka B.P., but was replaced by birch as the highest elevation species around 6 ka. At that time pine reached its maximum Holocene abundance, and birch forest extended 200 m above modern limits (Kullman, 1989, 1993). This suggests July temperatures were -1-2 °C above mid-twentieth century levels (using a lapse rate of -0.65 °C/100 m). Cooling set in after -3500 B.P., at which time many glaciers that had completely disappeared in the early Holocene reformed (Kvamme, 1993; Karlen, 1993; Matthews, 1993). This marked the first of several neoglaciation episodes, culminating in the most recent cold episode from the sixteenth-nineteenth century A.D. (the "Little Ice Age").

Detailed studies of dead larch trees above the modern treeline in the northern Urals showed that prior to the Little Ice Age trees grew 60-80 m above the modern limit (from the ninth to the thirteenth century A.D.) but nothing was established after that date until the mid-twentieth century (Shiyatov, 1993). Trees from this medieval warm period did survive into the subsequent colder period, but many died in the late thirteenth and early fourteenth century, in the early 1500s, and in the early and late nineteenth century (Kullman, 1987). This pattern echoes the views of many

Radiocarbon years x 103BP

Calibrated years x 10 3 BP

- 220 200 ■ 180 160 - 140 120 ■ 100 80 60 40 20 0 -20 ■ -40 -60 -80 -100 -120

Calibrated years x 103 BP

FIGURE 8.3 Elevation of l4C dated subfossil pine wood samples (Pinus sylvestris L.) in the Scandes mountains, central Sweden (black bars) relative to modern pine limit in the region (after adjustment for isostatic rebound during the Holocene). Upper limit of pine growth is indicated by the dashed line, but this view may change somewhat as new samples are recovered. Changes in temperature are estimated by assuming a lapse rate of 0.6 °C 100 m '.Wood samples were l4C dated close to the core to obtain an age near the time of germination (Dahl and Nesje, 1996, based on samples collected by L. Kullman and G. and J. Lundqvist).

paleoclimatologists that there was an earlier onset to the Little Ice Age (in the fourteenth century) and that the subsequent 500 yr experienced a succession of both mild and sharply colder episodes, culminating in the coldest period in the early nineteenth century (Jones and Bradley, 1992; Bradley and Jones, 1992, 1993; Mann etal., 1998).

Alpine treeline studies elsewhere mirror to a large extent the picture emerging from Scandinavia (see e.g., the parallel changes found in the Carpathian mountains by Rybnickova and Rybnicek, 1993; and in the Swiss Alps by Tinner et al., 1996). Similarly, in western North America, several studies indicate treeline was above modern limits in the early Holocene. In the San Juan Mountains, Carrara et al. (1991) dated >50 macrofossils of conifer wood from above treeline and found evidence for trees up to 140 m above modern limits from 9600-5400 yr B.P. indicating July temperatures were up to 0.9 °C warmer than present. From 5400-3500 the treeline was near modern limits, then a climatic deterioration led to a decline in treeline after 3500 B.P. This is similar to evidence from the Canadian Rockies (Luckman and Kearney, 1986; Clague and Mathewes, 1989) and from other parts of the western U.S. (Rochefort et al., 1994). In some areas the absence of radiocarbon-dated wood at certain intervals suggests brief colder episodes within the early Holocene but more extensive sampling may revise that picture (Kullman, 1988). Macrofossil evidence is a rather blunt instrument and apparent "data gaps" are difficult to interpret. Karlen (1993) supports his interpretation of cooler episodes with lake sediment evidence, and this points to the need for integration of many different proxies to fully comprehend the complexity of Holocene climates in mountain areas.

8.2.3 Lower Treeline Fluctuations

Throughout the arid and semiarid regions of the southwestern United States, there is an altitudinal zonation of vegetation with xerophytic desert scrub (commonly sagebrush, evergreen creosote bush, and evergreen blackbrush) at lower elevations, grading into mesophytic woodland (juniper, pinon pine, and live oak) at successively higher elevations (Fig. 8.4). The precise elevation of the woodland/desert scrub boundary varies more or less with latitude, being lowest in the Chihuahuan Desert of Mexico and highest in the interior Great Basin of Nevada. This eleva-tional gradient, decreasing to the south, is related to distance from the source of summer moisture (the Gulf of Mexico and the tropical Pacific); the interior Great Basin is both farthest from these source regions of tropical maritime air and isolated from temperate Pacific moisture sources by the mountain ranges to the west (Wells, 1979). The lower treeline elevational gradient thus strongly reflects the importance of moisture for tree growth and for this reason it is sometimes referred to as the "dryness treeline."

Fluctuations of the lower treeline have been greatly facilitated by the analysis of fossil middens of the packrat (genus Neotoma) from caves throughout the southwestern United States (Wells and Jorgensen, 1964; Wells and Berger, 1967). Packrats forage incessantly within a very limited range (~1 ha) of their dens, which are constructed of plant material from the surrounding site. Because of their propensity for

Upper Treeline LimitVegetation Elevations Tatra Mountains
FIGURE 8.4 Transect through mountains of the southwestern U.S. to show vegetation change with elevation. During glacial times, vegetation boundaries were generally lower due to increased effective moisture (resulting from lower temperatures and/or higher precipitation).

collecting items at random, and not simply food stocks, the dens or middens effectively provide a remarkably complete inventory of the local flora (Wells, 1976; Spaulding et al., 1990; Vaughan, 1990). Middens are cemented together into hard, fibrous masses by a dark brown, varnish-like coating of dried Neotoma urine (known as amberat). The amberat cements the deposit to rocky crevices in caves and prevents its destruction by fungi and bacteria. Because the cave sites are so dry, Neotoma middens may remain preserved for tens of thousands of years; in fact over a thousand macrofossils from Neotoma middens in the western U.S. have so far been dated, ranging in age from the late Holocene to >40,000 yr B.P. (Webb and Betancourt, 1990). Studies of rat middens in other arid and semiarid regions of the world are just beginning (see Part IV of Betancourt et al., 1990; Pearson and Dodson, 1993).

Middens are constructed over relatively short intervals, until the rock crevice is filled, so continuous stratigraphic records are not available; rather, they represent samples of vegetation near the site, from discrete time intervals in the past. Macro-fossils recovered include branches, twigs, leaves, bark, seeds, fruits, grasses, invertebrates such as snails and beetles, and even the bones of vertebrate animals. Such prolific inventories of macrofossils have enabled quite detailed pictures of the local vegetation around the midden site to be reconstructed, although the material collected may not represent a random sample of plants in the area. Because packrats may collect certain types of material preferentially, middens do not necessarily give a complete picture of the relative abundance of plants in the collection area. Furthermore, different species of packrat have different collection preferences, so sequential occupancy of the same site by different species might give the erroneous impression of a change in local vegetation (Dial and Czaplewski, 1990). In spite of these potential problems, regional comparisons of midden composition for different time intervals have enabled the broad-scale patterns of vegetation change since the last glaciation to be established. Most significantly, the results demonstrate a dramatic increase in the area of pinon-juniper woodland throughout the South during the late Wisconsin period of maximum glaciation (Van Devender and Spaulding, 1979; Van Devender, 1990a). In the Great Basin, Mojave, Sonoran, and Chi-huahuan deserts of today (Fig. 8.5), such woodlands are restricted to higher elevations, often on isolated peaks surrounded by vast areas of desert scrubland. However, over periods ranging from >40,000 to -12,000 yr B.P., Neotoma middens from currently hyperarid sites document the presence of juniper or pinon-juniper woodlands over a vastly enlarged area. In the Great Basin, forest vegetation (now

Desert Regions The United States
FIGURE 8.S Desert areas of the western United States. Changing altitudinal vegetation zonation within desert areas is recorded in fossil packrat middens.

restricted to isolated mountain peaks) extended well below its present range; for example, subalpine conifers grew as much as 1000 m below modern limits (Thompson, 1990). In the Mojave Desert, pinon-juniper woodland occupied terrain down to -900 m, in areas that today support only thermophilous desert scrub. Similarly, in the Sonoran and Chihuahuan Desert, pinon-juniper-oak woodland extended down to 500-600 m, into areas which are extensive deserts today (Van Devender, 1990a, 1990b). Interestingly, in these southern desert regions the vegetation communities that predominated have no modern analog in the region today that has both forest vegetation and frost-sensitive desert succulents. This suggests a lower frequency of winter cold air outbreaks leading to killing frosts; possibly the forest cover also helped by reducing night-time radiational cooling.

The patterns of glacial period vegetation change in the region all indicate lower summer temperatures, by 6-10 °C, reduced evaporation, and enhanced winter rainfall (by up to 50%) (Table 8.1). This is related to the displacement of winter storms southward into the region, and a reduction in summer monsoon airflow from the Gulf of Mexico. Such conditions appear to have prevailed, with little change, from at least 22,000 yr B.P. to 12,000 yr B.P. Rapid changes in climate and vegetation took place in the following few thousand years, so that by 8000-9000 yr B.P., vegetation patterns had begun to look more like those seen today; the winter rainfall regime had ended, temperatures had risen to within a few degrees of modern levels, and rainfall was largely delivered by summer monsoons (Fig. 8.6). Maximum aridity appears to have occurred in the mid-Holocene in some areas (e.g., the Grand Canyon and Mojave Deserts; Cole, 1990; Spaulding, 1990, 1991) but elsewhere there is evidence for a stronger mid-Holocene monsoon rainfall regime, with drier conditions and the most extensive desert vegetation developing later in the Holocene (Van Devender, 1990a; Van Devender et al., 1994). In the Mojave Desert, evidence for increased effective moisture in the late Holocene (3.8-1.5 ka B.P.) suggests a "neopluvial" of cooler and/or wetter conditions, perhaps correlative with neoglacial episodes in the mountains to the north and east (Spaulding, 1990).

TABLE 8.1 Estimates of Climatic Conditions in the Western and Southwestern United States During the Last Glacial Maximum, Relative to Today, Based on Fossils Found in Packrat Middens


AT (°C)




Colorado Plateau

>6.3 (Su)

Higher (Wi.)

Drier summers

Betancourt, 1990

Grand Canyon

-6.7 (Ann)


Wi. rainfall max

Cole, 1990

Great Basin

10 (Ann)


Thompson, 1990

Sonoran Desert

8 (July)


Wi. rainfall max

Van Devender, 1990b

Chihuahuan Desert

Strong Su.


Winter rainfall


regime; few freezes

Van Devender, 1990a

Mojave Desert

6 (Ann)


Wi. rainfall max

Spaulding, 1990

Inferred Paleovegetation

Current Vegetation

0 0

Post a comment