Introduction

6.1.1. Environmental Determinism in Pre-Columbian America

The concept of environmental determinism was debated in the anthropological and archaeological literature during the 1950s and 1960s. The fundamental tenet of the hypothesis posits that regional agricultural potential limits cultural development. Meggers (1954) concluded that the Maya lowlands of southern México and Petén, Guatemala (Fig. 1) have limited agricultural potential today. She suggested that the region would not have been conducive to the cultural florescence so evident in the archaeological record. A logical conclusion of the deterministic theory was that lowland Classic Maya culture (A.D. 250-850), with its monumental architecture, art, hieroglyphics, concept of zero, corbeled arches, calendrical system, and stela cult, did not develop in situ, but was imported from elsewhere and was destined to decline in its new geographic context. The simple theory apparently accounted for both the origin of the lowland Maya and their mysterious collapse in the ninth century A.D.

Meggers's (1954) claims were contested by those who argued that there was no archaeological basis for claiming that Classic Maya culture was imported from the highlands. Coe (1957) noted that several lowland Maya achievements, such as the corbeled arch, the Long Count calendar, and the stela cult, developed in the lowlands. Excavations in the lowlands yielded Early Preclassic (2000-1000 B.C.), Formative phase ceramics, supporting the claim that lowland Classic Maya cultural ontogeny had its roots in the low-elevation, tropical karst environment (Altschuler, 1958; Hammond et al., 1977, 1979). Today, it is generally acknowledged that the origins of lowland Maya civilization began in the Early Preclassic period, were complex, and involved interactions with peripheral coastal and high-

FIGURE 1 (a) Maya lowlands of the Yucatán Peninsula showing isohyets (mm) and locations of lake sediment records discussed in the text (PL = Punta Laguna, C = Chichancanab, PI = Petén-Itzá, S = Salpeten). Rainfall values are based on Wilson (1980). (b) The circum-Caribbean region showing the Yucatán Peninsula (box) and locations of paleoclimate studies referred to in the text, including Petén (Guatemala), Yucatán (México), Florida, Haiti, and Venezuela.

FIGURE 1 (a) Maya lowlands of the Yucatán Peninsula showing isohyets (mm) and locations of lake sediment records discussed in the text (PL = Punta Laguna, C = Chichancanab, PI = Petén-Itzá, S = Salpeten). Rainfall values are based on Wilson (1980). (b) The circum-Caribbean region showing the Yucatán Peninsula (box) and locations of paleoclimate studies referred to in the text, including Petén (Guatemala), Yucatán (México), Florida, Haiti, and Venezuela.

land regions (Sharer, 1994). Coe pointed out that there is no evidence for the gradual cultural decline referred to by Meggers. Instead, archaeological data point to abrupt, widespread collapse ca. A.D. 850, after 600 years of population growth and cultural expansion in the Early Classic (A.D. 250-550) and Late Classic (A.D. 550-850) periods.

Ferdon (1959) reassessed Peten's agricultural potential and challenged Meggers's application of environmental determinism to the lowland Classic Maya. Using modern temperature, rainfall, soil, and landform criteria, he reclassified the lowlands as a region favorable for agriculture. The analysis did not refute deterministic theory as applied to the Maya, but rather, freed the civilization from expectations dictated by an environment with limited agricultural potential. Ferdon, how ever, argued that there is no correlation between natural agricultural potential and cultural development and attributed the cultural decline to invasion of agricultural plots by grasses that interfered with Classic Maya crop cultivation.

Adams (1973) summarized theories that purport to explain the ninth century A.D. Maya collapse. Among the proposed causes that continue to be explored are overpopulation combined with soil erosion and exhaustion (Rice, 1978; Deevey et al., 1979; Paine and Freter, 1996), demographic models that show skewed sex ratios in favor of males (Cowgill and Hutchinson, 1963), health problems and disease (Saul, 1973), social upheaval, warfare, religion (i.e., the collapse was preordained), and insect infestation (Sabloff, 1973). Natural catastrophes such as earthquakes and volcanic eruptions (Ford and Rose, 1995) have also been invoked as factors contributing to the collapse. Others have suggested a correlation between climate and Maya cultural development and decline (Gore, 1992; Gill, 2000; Dahlin, 1983; Gunn and Adams, 1981; Folan et al., 1983; Messenger, 1990; Hodell et al., 1995; Curtis et al., 1996). Several theories for the collapse have been tested by archaeological and paleoenvironmental study, but some of the hypotheses are supported by few, if any, data.

Outside the Maya lowlands, other pre-Columbian cultures also arose, persisted for centuries, and ultimately collapsed in seemingly harsh environments. The Chiripa (1500-200 B.C.) and Tiwanaku (400 B.C. to A.D. 1100) cultures developed in the Lake Titicaca watershed of the Andean Altiplano (Fig. 2), where nitrogen-poor soils on steep slopes are prone to erosion and moisture deficit. Nighttime freezes often kill crops prior to harvest (Kolata and Ortloff, 1989). Flatlands (pampas) near Lake Titicaca are inundated periodically and subject to soil salinization. The prolonged rise and abrupt fall of a great civilization in this region of seemingly limited resources appears to defy environmental determinism.

Human ingenuity, combined with state-level agricultural organization, can overcome natural environmental limitations to food production. Intensive agricultural practices, such as raised-field construction and terracing, were widely used by pre-Columbian civilizations. Remnant raised fields have been found along waterways and in wetlands of the Maya lowlands (Adams, 1980; Adams et al., 1981; Siemens and Pule-ston, 1972; Matheny, 1976) and in the river drainages and pampas of the Titicaca basin (Kolata, 1991). The ancient agricultural strategy is documented archaeologi-cally throughout the tropics (Denevan, 1970; Denevan and Turner, 1974). In the Maya lowlands, constructed fields in wetlands served to drain land and raise crop

FIGURE 2 (a) Bathymetric map of the southern basin of Lake Titicaca (Lago Wiñaymarka) showing locations of the six sediment cores used to date late Holocene lake level changes. Cores A, B, C, D, G, and H were used to infer lake level shifts over the past ~3500 years. Also indicated are the Río Desaguadero outflow, the archaeological site of Tiwanaku, and the location of the Pampa Koani (see Fig. 8). (b) The Titicaca drainage basin showing the lake and major input rivers (redrawn from Boulangé and Aquize Jaen 1981). High mountain ranges of the Cordillera Oriental and Cordillera Occidental bind the watershed to the east and west, respectively. The Quelccaya ice cap ( • ) lies at the northwest margin of the drainage.

FIGURE 2 (a) Bathymetric map of the southern basin of Lake Titicaca (Lago Wiñaymarka) showing locations of the six sediment cores used to date late Holocene lake level changes. Cores A, B, C, D, G, and H were used to infer lake level shifts over the past ~3500 years. Also indicated are the Río Desaguadero outflow, the archaeological site of Tiwanaku, and the location of the Pampa Koani (see Fig. 8). (b) The Titicaca drainage basin showing the lake and major input rivers (redrawn from Boulangé and Aquize Jaen 1981). High mountain ranges of the Cordillera Oriental and Cordillera Occidental bind the watershed to the east and west, respectively. The Quelccaya ice cap ( • ) lies at the northwest margin of the drainage.

roots above the inundated soil zone (Pohl, 1990). Organic matter that accumulated in canals between raised fields was used to fertilize intensively farmed, nutrient-poor soils. Canals also provided an environment for cultivation of aquatic resources, including edible macrophytes, mollusks, fish, and turtles.

The multiple advantages of raised fields and their contribution to sustainable agricultual production have been demonstrated in the Andean Altiplano by both archaeological excavation (Kolata, 1991) and experimental study of reconstructed raised beds (Erick-son, 1988; Kolata et al., 1996). On the pampas, raised fields elevate crop roots above the phreatic zone. Canals between fields absorb solar radiation during the day and store heat, protecting crops against nighttime freezes (Kolata and Ortloff, 1989). Cyanobacteria in canals fix atmospheric nitrogen (Biesboer et al., 1999), and the canals retain nutrients and provide fertilizer for N-limited soils (Binford et al., 1996; Carney et al., 1993). Canals between raised fields receive fresh stream water or groundwater from the base of surrounding hills, and low ion concentrations in these waters prevent soil salinization (Sanchez de Lozada, 1996).

Our purpose in revisiting the concept of environmental determinism is not to debate its validity as applied to the Maya and Tiwanaku cultures. Instead, we note that in the Maya and Tiwanaku cases an implicit assumption used to assess natural agricultural potential was indefensible, rendering the discussion moot. Although it is never explicitly stated, evaluations of natural agricultural potential ignored technological innovations and assumed that climate conditions were constant and similar to those of the present. Recent archaeological investigations indicate that many pre-Columbian societies utilized intensive agricultural practices (Denevan, 1970; Pohl, 1990). Paleoclimate studies in the circum-Caribbean (Hodell et al., 1991, 2000; Curtis and Hodell, 1993) and South American Altiplano (Abbott et al., 1997a; Thompson et al., 1985, 1995) provide evidence for late Holocene climate fluctuations. We contend that modern agricultural potential alone cannot establish whether past environments stimulated or constrained cultural development. Instead, we argue that although agricultural potential is, in part, dependent on landscape characteristics such as landforms, soils, and vegetation, crop production varies over time as a function of human social organization, land use practices, and climate change. About 3000 years ago, climate changes in the Maya lowlands and Andean Altiplano generated environmental circumstances that were conducive to human agricultural development. Two millennia later, climatic change leading to droughts exceeded environmental thresholds for Maya and Tiwanaku agricultural sustainabili-ty and led to the collapse of both civilizations.

In this chapter, we review paleoclimatological and archaeological results and discuss the correlation between Holocene climate and both Maya and Tiwanaku cultural development and collapse. The two cultural areas were chosen for several reasons. First, they have been subject to intensive archaeological study, so their demographic and agricultural prehistories are well known. Second, both civilizations arose, prospered, and collapsed in low-latitude areas adjacent to lakes that contain paleoclimate archives in their sediments. Third, although the Maya lowlands lie at low elevations (0-300 m above sea level [asl]) and the Tiwanaku

FIGURE 3 Map of the Americas indicating the Maya lowlands and Andean Altiplano where paleoclimate studies were done.

FIGURE 3 Map of the Americas indicating the Maya lowlands and Andean Altiplano where paleoclimate studies were done.

region is located at high altitudes (—3600 to >4000 m asl), both regions are in climatically marginal areas for agriculture, where small changes in available moisture can have profound impacts on crop yields. Finally, the two study areas lie along the Pole-Equator-Pole: Americas (PEP 1) transect (Fig. 3). Combined investigation of the Maya and Tiwanaku cultural areas allows temporal correlation of tropical climate changes north and south of the equator, establishment of interhemispheric climate linkages, and differentiation of forcing factors driving long-term and short-term climate change in the two regions.

6.1.2. The Maya Lowlands

Here, we consider the Maya area of the Yucatán Peninsula encompassed by what is today Belize; the northern department of Petén, Guatemala; and the Mexican states of Campeche, Yucatán, and Quintana Roo (Fig. 1). The region is characterized principally by karst topography that varies in elevation from sea level to —300 m asl. Annual rainfall across the peninsula is heterogeneous, ranging from <500 mm/ year in the extreme northwest to —2000 mm/ year (Fig. 1) in central Petén (Wilson, 1980). The precipitation gradient is reflected in soil development and vegetation (Flores and

Carvajal, 1994). Soils in the flat, northern part of the peninsula are thin, with limestone bedrock exposed at the ground surface. Farther south, in the karst hills of Petén, soils are deeper and more productive. Dry-adapted vegetation in the semiarid northwest is of low stature and diversity. The Petén forest, sometimes referred to as a "quasi-rain forest" (Lundell, 1937), is taller and more diverse.

Rainfall on the Yucatán Peninsula is highly seasonal, with dry conditions from January through April (Dee-vey et al., 1980). Rainfall is high in May, June, September, and October. A drier interval, referred to as the canicula, usually occurs in July or August (Magaña et al., 1999). Interannual variation in the amount and timing of rainfall can be pronounced (Wilson, 1980). Under the influence of the Northeast Trades, summer rains on the peninsula coincide with the northward migration of the Intertropical Convergence Zone (ITCZ) (Hastenrath, 1976, 1984). The Azores-Bermuda high-pressure system also moves northward in summer. Sea surface temperatures (SSTs) are warm in the tropical / subtropical North Atlantic and Caribbean during the Northern Hemisphere summer, providing ample moisture for precipitation. Summertime rainfall is frequently delivered during violent, convective thunderstorms, and total rainfall in any given year can be influenced by tropical storms and hurricanes that contribute substantial precipitation within a short time period.

Dry conditions prevail on the Yucatán Peninsula in the Northern Hemisphere winter, when the ITCZ and Azores-Bermuda high-pressure system move southward (Hastenrath, 1976, 1984). Low winter precipitation is a consequence of relatively low SSTs in the tropical North Atlantic, a steep pressure gradient on the equatorward side of the Azores-Bermuda High, and a strong temperature inversion associated with enhanced trade winds. Differences in precipitation, rather than temperature, define interannual climate variability on the Yucatán Peninsula.

Both pre-Columbian and modern agriculturists on the Yucatán Peninsula have had to contend with strong seasonality and unpredictability of annual precipitation. Although the intraannual pattern of rainfall is known, delayed onset of summer rains or other disruptions of seasonal precipitation can have disastrous consequences for farmers, many of whom still depend on slash-and-burn (swidden) techniques. In rural Yucatán, the contemporary Maya still perform the ancient Cha Cha'ac ceremony to invoke rainfall. Ubiquitous portrayal of the rain god, Chac, on monumental architecture in the northern Maya lowlands (Fig. 4 [see color insert]) demonstrates that the quantity and timing of rainfall were also of great importance to the ancient Maya.

High-resolution Holocene paleoclimate inferences for the Yucatán Peninsula are based on stable isotopic (S18O) study of multiple ostracod or gastropod shells in samples taken at 1-cm intervals from lake sediment cores. Here, we summarize core results from Lakes Chichancanab (Hodell et al., 1995) and Punta Laguna (Curtis et al., 1996) on the northern part of the peninsula and from Lake Petén-Itzá (Curtis et al., 1998) in Petén, Guatemala (Fig. 1). We are currently studying cores from Lake Salpeten, in the Petén lake district (Fig. 1), to examine the details of climate change in the southern Maya lowlands. The S18O of shell carbonate from these lake sediment cores was governed principally by three factors: (1) the S18O of the lake water from which the shells were precipitated; (2) taxon-specific fraction-ation, i.e., vital effects; and (3) water temperature. In these stratigraphic studies, we used shells of monospe-cific, adult snails and ostracods in each core to minimize vital effects. Temperature changes during the Holocene were probably not responsible for the dramatic shifts (>3%o) in shell S18O, because a 1%o shift in S18O requires a ~4°C change in temperature (Craig, 1965). It is unlikely that long-term temperature fluctuations were solely or largely responsible for the strati-graphic shifts in shell S18O, as this would require a change in mean water temperature of > 12°C. Likewise, intraannual temperature fluctuations probably did not contribute significantly to stratigraphic changes in shell 518O. Seasonal differences in diurnal water temperatures from these tropical lakes are typically on the order of only 3°-4°C (Covich and Stuiver, 1974; Deevey et al., 1980), and shells collected from 1-cm sample intervals in the cores integrate, on average, material deposited over ~5- to 50-year periods. Although temperature changes may contribute minimally to stratigraphic S18O variation, the primary determinant of large shifts in the S18O of shell carbonate during the Holocene has been the S18O of the lake water from which the carbonate was derived.

In tropical lakes that lack overland outflows, the two major factors that control the S18O of lake water over time are the isotopic signature of input waters (rain, runoff, and groundwater) and the relation between evaporation and precipitation (E/P) (Fontes and Gonfi-antini, 1967). Rozanski et al. (1993) reported weighted mean isotopic values for several low-elevation, circum-Caribbean sites that were part of the International Atomic Energy Agency (IAEA)/World Meteorological Organization (WMO) precipitation monitoring network. Samples were collected between 1961 and 1987, and in each case, weighted means integrated data obtained over a period of at least 14 years. The isotopic signatures of rainfall from stations at Veracruz, México (—4.13%o, n = 169); Howard Air Force Base, Panama (—5.65%o, n =

165); Barranquilla, Colombia (—5.09%o, n = 85); and Maracay, Venezuela (—4.01%o, n = 64) are similar.

Our mean values for S18O in precipitation samples from Petén, Guatemala (—2.86%o) (Rosenmeier et al., 1998) and near Punta Laguna, México (-3.91%o) (Curtis et al., 1996) reflect only a few rainfall events, but are only slightly higher than weighted mean values for the circum-Caribbean reported by Rozanski et al. (1993). Isotopic values for groundwaters in Petén (—3.38%o) and near Punta Laguna (—3.92%o) are similar to their respective rainfall values, suggesting that input waters to Yucatán lakes, whether they are delivered via direct rainfall, runoff, or subterranean infiltration, possess the same isotopic signature.

Lake waters on the Yucatán Peninsula yield isotopic values that are more positive relative to input waters, indicating that 18O is evaporatively concentrated in the water bodies (Covich and Stuiver, 1974; Hodell et al., 1995; Curtis et al., 1996; Rosenmeier et al., 1998). Differences between S18O values for precipitation and lake water show enrichment ranging from ~3.6%o at Lake Petén-Itzá (Curtis et al., 1998) to >7%o at Lakes Chichancanab (Hodell et al., 1995) and Salpeten (Rosenmeier et al., 1998). Although we cannot demonstrate definitively that the source water 518O signature has not changed over the course of the Holocene, we believe that the major process affecting lake water S18O on the Yucatán Peninsula for the last —10,000 years has been shifting E/P. Thus, stratigraphic study of S18O in mollusk and crustacean shells in Yucatán sediment cores can be used to reconstruct past changes in moisture availability. Ostracods and snails often occupy different habitats within lakes and display different growth characteristics. Whereas ostracods molt as they pass through instar stages, mollusks simply add shell as they grow. The two taxonomic groups are ecologically and developmentally different, but yield similar stratigraphic S18O signals, indicating that the sedi-mented shell remains faithfully reflect past changes in the oxygen isotope ratio of the lake water.

Shallow lakes (<30 m) and shallow areas of deep lakes on the Yucatán Peninsula were first covered by water and began depositing lacustrine sediments between ca. 8030 and 5840 B.C. (9000-7000 14C B.P.) (Leyden et al., 1994, 1998). Lakes and sinkholes (cenotes) filled in the early Holocene, after a long, late Pleistocene arid period (Leyden et al., 1994). On the northern part of the peninsula, early Holocene basin filling was a consequence of increasing moisture availability as well as of sea level rise, which raised the local water table. Farther south in the Petén, where the local aquifer lies deep below the land surface (Gill, 2000), increased rainfall at the Pleis-tocene-Holocene transition filled lakes and supplied moisture for tropical forest synthesis (Leyden, 1984).

FIGURE 5 High-resolution paleoclimate record from Lake Chichancanab, in the northern Maya lowlands. Percent CaCO3, percent sulfur (gypsum), S18O of gastropod shells (Pyrgophorus coronatus), and S18O in shells of ostracods Physocypria sp. (+) and Cyprinotus cf. salinus (O) at 1-cm intervals, plotted against 14C years B.P. Oxygen isotope results are 3-point running means and are expressed relative to the PDB (PeeDee Belemnite) standard. Arrows between about 7800 and 7300 14C years B.P. within the gastropod S18O plot indicate depths at which the foraminifer Ammonia beccarii was found. The arrow at the right margin indicates the driest episode of the late Holocene and is dated at 1140 ± 35 14C years B.P. (A.D. 922), coinciding with the Classic Maya collapse. (From Hodell, D. A. et al. (1995), with permission.)

FIGURE 5 High-resolution paleoclimate record from Lake Chichancanab, in the northern Maya lowlands. Percent CaCO3, percent sulfur (gypsum), S18O of gastropod shells (Pyrgophorus coronatus), and S18O in shells of ostracods Physocypria sp. (+) and Cyprinotus cf. salinus (O) at 1-cm intervals, plotted against 14C years B.P. Oxygen isotope results are 3-point running means and are expressed relative to the PDB (PeeDee Belemnite) standard. Arrows between about 7800 and 7300 14C years B.P. within the gastropod S18O plot indicate depths at which the foraminifer Ammonia beccarii was found. The arrow at the right margin indicates the driest episode of the late Holocene and is dated at 1140 ± 35 14C years B.P. (A.D. 922), coinciding with the Classic Maya collapse. (From Hodell, D. A. et al. (1995), with permission.)

The stable isotope (S18O), geochemical, and microfossil stratigraphies (Fig. 5) of a core from Lake Chi-chancanab (19°50' N, 88°45' W) provide a high-resolution record of changes in E/P over the last —8200 14C years (Hodell et al., 1995). Radiocarbon dating of gastropod and ostracod shells from Yucatán lakes is confounded by the effects of hard water lake error (Deevey and Stui-ver, 1964), which makes dates on lacustrine carbonates in Lake Chichancanab ca. 1200 years too old (Hodell et al., 1995). Therefore, the age/depth relation for the core (Fig. 5) was developed by regression using only accelerator mass spectrometer (AMS) 14C dates on terrestrial carbon (Hodell et al., 1995).

Initial filling of Lake Chichancanab ca. 8200 14C B.P. (7250 B.C.) is marked by gypsum precipitation, relatively high S18O values for gastropod and ostracod shells, and the presence of the benthic foraminifer Ammonia beccarii (Fig. 5). A. beccarii can tolerate a wide range of temperatures (10°-35°C) and salinities (7-67 g/L), but is capable of reproducing only at salinities between 13 and 40 g/L (Bradshaw, 1957). Dry conditions in the earliest Holocene are inferred from the biological and geochemical indicators that point to low lake stage, saline waters, and high E/P. After -720014C B.P. (6000 B.C.), the lake filled rapidly. Gypsum precipitation was replaced by carbonate deposition, stable isotope values were lower, and A. beccarii disappeared from the record, indicating wetter conditions. Relatively moist conditions persisted from ca. 7200 to 3000 14C B.P. (6000-1250 B.C.) (Fig. 5).

Palynological evidence suggests that swidden activity may have begun in northern Guatemala as early as -5600 14C B.P. (4410 B.C.) (Islebe et al., 1996), but this gradual decline of forest in Peten may be attributable to the onset of a regional drying trend. The archaeological record indicates humans did not establish sedentary settlements in the Maya lowlands during the relatively moist early to middle Holocene. Archaeologists date the earliest sedentary populations in the region to the Middle Preclassic period, ca. 1000-300 B.C. (Rice and Rice, 1990; Turner, 1990). Initial settlement in the Maya lowlands corresponds generally to a period of increased regional drying at about 3000 14C B.P. (1250 B.C.). This pronounced drying trend is reflected in the

FIGURE 6 Oxygen isotope (518O) records based on ostracods from Lakes Punta Laguna and Chichancanab (México), spanning the last 3500 calendar years. Both 518O records display increasing or relatively high values, i.e., high E/P, within or just following the interval from A.D. 800-1000, corresponding to the period of the Maya collapse.

FIGURE 6 Oxygen isotope (518O) records based on ostracods from Lakes Punta Laguna and Chichancanab (México), spanning the last 3500 calendar years. Both 518O records display increasing or relatively high values, i.e., high E/P, within or just following the interval from A.D. 800-1000, corresponding to the period of the Maya collapse.

Chichancanab sediment core by increased sulfur (gypsum) concentrations and higher 518O values for gastropod and ostracod shells (Fig. 5). Concurrent late Holocene drying has also been documented in Lake Miragoane, Haiti (Hodell et al., 1991), and Lake Valencia in northern Venezuela (Bradbury et al., 1981; Ley-den, 1985).

The paleoclimate record from Lake Chichancanab sediments indicates that early Maya agriculturists faced several challenges. Despite the drying trend that began ca. 1250 B.C., general conditions were apparently suitable for shifting agriculture. Nevertheless, in addition to interannual variations in the timing and amount of rainfall, relative E/P on the Yucatán Peninsula between ca. 1250 B.C. and ca. A.D. 1000 varied substantially over decadal to centennial scales. This variability is seen most clearly in the high-resolution, 3500 14C-year paleoclimate record from Punta Laguna (Fig. 6), which lies about 150 km northeast of Chichancanab at 20°38' N, 87°37' W (Fig. 1). Hard water lake dating error at Punta Laguna is also on the order of 1200-1300 years, and the age/depth relation for the core was developed by linear interpolation between five AMS 14C-

dated terrestrial wood samples, assuming linear sedimentation between dated horizons (Curtis et al., 1996). Dates were calibrated by using CALIB 3.0 with a 100-year moving average of the tree-ring data set (Stuiver and Becker, 1993; Stuiver and Reimer, 1993).

In the Chichancanab core, the highest gypsum concentrations and S18O values of the middle to late Holocene period were found in samples at ~65 cm depth (Fig. 5). An AMS 14C date on a terrestrial seed at 65 cm shows that the protracted dry episode culminated at 1140 14C B.P. (A.D. 922), coinciding closely with the Classic Maya collapse (Hodell et al., 1995). In the Punta Laguna section, mean 518O values for the period 1750-940 14C B.P. (A.D. 300-1100) are higher than mean values for the preceding and following periods (Fig. 6). The 518O peak at A.D. 862 (1210 14C B.P.) in the Punta Laguna section may correspond to the Late Classic dry event recorded in the Chichancanab core.

Paleoclimate records from Lakes Chichancanab and Punta Laguna, in combination with the archaeological record, demonstrate a temporal correlation between drought and cultural demise. These water bodies lie in the dry, northern Maya lowlands. This region, howev er, was apparently least affected by the "collapse" (Lowe, 1985). Demographic consequences of the ninth-century decline were felt most acutely in the wetter, southern lowlands. This pattern may be a consequence of the fact that groundwater is less accessible in the southern lowlands than it is in the north because the depth to the water table increases southward. Therefore, the southern lowlands were more reliant upon surficial water supplies (Gill, 2000). Archaeological survey and test excavation of house mounds in several Peten watersheds yielded estimates of Late Classic Maya population densities in excess of 200 persons per square kilometer (Rice and Rice, 1990). By the Terminal Classic period (A.D. ~900), population densities in nearly all the studied drainage basins had declined to <100 persons per square kilometer, and by the Late Postclassic period (A.D. 1500), many watersheds were abandoned. Drought is implicated in the Classic Maya collapse. Evidence for this climatic change should therefore be found in Peten lake cores.

Recent attention has turned to paleoclimate reconstruction in the Peten lake district (Fig. 1). Lake Peten-Itza (16°55' N, 89°52' W), the largest Peten water body (area, A = 99.6 km2), yielded a ~9000 14C-year paleo-environmental record based on sediment geochemistry, magnetic susceptibility, pollen, and stable isotopes (Curtis et al., 1998). Chronology for the core was based on AMS 14C dating of terrestrial wood and charcoal samples (Curtis et al., 1998). Ages between dated horizons were interpolated by assuming constant linear sedimentation. Three 518O records from the basin, two based on snail taxa and another on ostracods, display little variability over the past 5000 14C years (Fig. 7). Rather than indicating climatic constancy during the past five millennia, however, they may simply reflect the fact that lake water 518O in large-volume lakes is insensitive to short-term changes in E/P. In large, deep lakes with long residence times, even protracted droughts may not reduce lake volume sufficiently to alter the lake water isotopic signature.

Lake Salpeten (16°58' N, 89°40' W) lies near Lake Peten-Itza (Fig. 1), but is smaller (A = 2.6 km2) and more saline (TDS = 4.76 g/L). Lake Salpeten is being studied because its sediments possess abundant ostra-cod and gastropod shells, it is at saturation with respect to gypsum, and its lakewater 518O is demonstrably enriched by >7%o relative to rainfall 518O. The relatively greater 18O enrichment in Lake Salpeten suggests it is more effectively "closed" than Lake Peten-Itza. Seismic reflection studies in Lake Salpeten were completed in summer 1999. Imaging of the sediment stratigraphy indicates periods of low lake stage during the Holocene, and consequent sedimentation hiatuses at shallow-water locations. Efforts to reconstruct late Holocene cli

Lake Péten-Itzá, Guatemala

Cochliopina sp. Pyrgophorus sp. Ostracods

0 0

Post a comment