Central America And The Circumcaribbean Region

A relatively limited number of paleolimnological records exist for México, Central America, and the cir-cum-Caribbean region, between 30° and 8°N latitude. Records for Central America south of México are predominantly palynological studies (e.g., Horn, 1993; Islebe et al., 1996; Islebe and Hooghiemstra, 1997) and will not be considered in this chapter (see Grimm et al., 2000). The locations of sites reviewed in this section are shown in Fig. 5.

The highest density of Holocene lacustrine records comes from the basins of the trans-Mexican volcanic belt, which crosses México from east to west at about 19°N. In spite of the large number of published records, the Holocene dating control for many records is surprisingly poor. In addition, many records show significant human disturbance over the last 3500 years or longer, and, in some cases, deliberate drainage of basins in the post-Hispanic period has resulted in deflation and loss of late Holocene sediments.

The best-dated Holocene sequence comes from the Zacapu basin in Michoacán (Petrequin et al., 1994; Metcalfe, 1995). The magnitude of fluctuations over the Holocene is small relative to those in the late Pleistocene; however, diatom sequences indicate that conditions may have been slightly wetter between ca. 7000 and 6000 B.P., followed by a dry period until ca. 5000 B.P. Water levels then recovered, and there may have been a brief episode of lake-level rise at ca. 2900 B.P. The largest fluctuation in lake level, however, was a distinct drying at ca. 1100 B.P.

Although the basin of México has been the focus of much research (Bradbury, 1989; Lozano-Garcia et al., 1993; Caballero-Miranda and Ortega-Guerrero, 1998), the dating control for Holocene sequences is poor, and catchment disturbance and probable deflation of later sediments compromise good Holocene records in many parts of the basin. Water levels increased during the terminal Pleistocene (14,000-10,000 B.P.), but the early Holocene was marked by low lake levels throughout, followed by generally wetter, although possibly quite variable, conditions. One of the sequences from Chalco, in the southern part of the basin (Caballero-Miranda et al., 1995), records a dry period at ca. 5500 B.P.

Other Holocene sequences are available for Lake Chichonahuapan in the upper Lerma basin, Lake Pátzcuaro in Michoacán, and La Piscina de Yuriria in southern Guanajuato. The upper Lerma diatom record (Metcalfe et al., 1991) indicates a lake transgression after 8000 B.P. which ended by 6000 B.P. (Metcalfe, unpublished data). A strong drying episode occurred prior to 4600 B.P., and a late Holocene transgression peaked at ca. 1600 B.P., which was followed by a marked regression and desiccation. The interpretation of Holo-cene changes in the Pátzcuaro basin (Watts and Bradbury, 1982; O'Hara et al., 1994) is again constrained by the scarcity of dates. A diatom record (Bradbury, in press) indicates that Holocene lake levels were generally lower than those in the Pleistocene. A possible shift to more arid conditions at ca. 5000 B.P. was reported by Watts and Bradbury (1982). Ostracode data (Bridgwater et al., 1998) indicate three wet phases over the last 4000

Circum Caribbean
FIGURE 5 Map of Central America and the circum-Caribbean region showing locations of sites mentioned in the text.

years separated by longer dry periods. The most complete Holocene record for La Piscina de Yuriria (Davies, 1995) is based on ostracode isotopic and trace element analyses. The early Holocene was marked by fluctuating conditions, followed by generally wetter conditions between 6700 and 5500 B.P. The diatom record for La Piscina de Yuriria (Metcalfe et al., 1994) covers only the last 4000 years and indicates conditions drier than those now at the base of the sequence and between ca. 3700 and 3300, 2950 and 2700, and 2300 and 800 B.P.

For northern México, the only published lacustrine records come from the Alta Babícora basin in Chihuahua (Ortega Ramirez, 1995; Metcalfe et al., 1997). Early Holocene conditions were clearly very much wetter than those now, whereas the mid-Holocene was dry, possibly at its driest at ca. 6000 B.P. Slightly wetter conditions were then reestablished until ca. 3000 B.P., after which there was basin desiccation, although the record is poorly preserved.

There are a number of records from the Mexican part of the Yucatán Peninsula, which should be considered together with the sequences from adjacent lowland Guatemala and Belize (see Sections 15.3.2 and 15.3.3). In this karstic geology, hard-water errors and inputs of colluvium of unknown isotopic composition have resulted in the rejection of many radiocarbon dates

(Vaughan et al., 1985). Many lake levels are directly, or indirectly, affected by sea level. Records covering most of the Holocene include Lakes Chichancanab, Coba, and San José Chulchaca. The original work on Chichancanab (Covich and Stuiver, 1974) extended back to the Pleistocene, whereas a more recent paper by Hodell et al. (1995) provides a full Holocene record. Interpretations based on the oxygen isotopic composition of bio-genic carbonate indicate that lake levels rose slowly after 8200 B.P. and rapidly from ca. 7200 B.P., and wet conditions prevailed between 7100 and 3000 B.P. (Hodell et al., 1995). The climate then became drier, with the driest conditions of the last 8000 years between 1300 and 1100 B.P. A core from Lake Coba (Whitmore et al., 1996) has a date near the base of 7410 ± 80 B.P., but only a patchy diatom record. The lake did not fill until ca. 7600 B.P. and was then dry and saline. Lake levels subsequently rose to form a deep freshwater lake prior to 2600 B.P., followed by a series of poorly dated fluctuations. The sequence for San José Chulchaca (Leyden et al., 1996) covers most of the Holocene, with pollen and some isotopic data from a long core and diatom data from a short core. The early Holocene was dry, but moist conditions, possibly wetter than those now, were established by 6800 B.P. There was a dry period between 6000 and 5000 B.P., followed by the reestablish-

ment of wetter conditions until 3500 B.P. After this, there was drying until ca. 1500 B.P. The diatom record indicates lower salinities between 1860 and 1010 B.P., followed by increased salinity, which suggests drier, shallower conditions. The record for Sayaucil (Whit-more et al., 1996) suggests that the modern lake is deeper and fresher than at any time in the past.

High-resolution studies of the last 2000 years are derived from the 818O record of ostracode carbonate in two Yucatán lakes (Curtis and Hodell, 1996) and suggest significant decadal to centennial climate variability. The data indicate that the interval from A.D. 3101050 was frequently dry, with periods of extreme aridity that persisted for multiple decades. The collapse of Mayan culture is dated at between A.D. 800 and 900 and is correlated with an interval of severe drought. Moisture generally increased from A.D. 1050 onward, but within this time are several intervals of moderate aridity lasting one to two decades.

15.3.2. Belize

A record covering the mid-Holocene has been obtained from Cobweb Swamp in Belize (Jacob and Hallmark, 1996). This record provides a detailed sedimen-tological history of the Holocene, but the site has been influenced by sea-level changes and long-term anthropogenic impacts.

15.3.3. Guatemala

Pioneering tropical paleolimnological work in Guatemala began in the 1960s (Cowgill et al., 1966; Tsukada and Deevey, 1967). Studies in the Petén lake district of northern Guatemala have been frustrated, however, by problems with obtaining a reliable radiocarbon chronology (see Leyden et al., 1994). With the absence of a radiometric chronology, pollen records were divided into zones, which were then correlated with archaeological events of known ages (Vaughan et al., 1985). Conditions appear to have been wetter than those now during the early part of the Holocene, with drier conditions becoming established from ca. 5600 B.P. Later records are significantly affected by human disturbance.

15.3.4. Panama

Paleolimnological data are confined to Lake La Yeguada in central Panama (Piperno et al., 1990). Between 10,000 and 6000 B.P., the diatom record is poor (Bush et al., 1992). Phytoliths are reduced greatly in number between 7000 and 5000 B.P., and sedimentation rates slow significantly (Piperno et al., 1990). Piperno et al. (1990) suggest that streamflow to the site was greatly reduced between 7000 and 3800 B.P., lowering the water level and indicating drier conditions. The absence of Podostemataceae (river weed) phytoliths from ca. 7650 B.P. points to reduced streamflow into the lake, and lower lake levels are inferred at this time, as deposition of lacustrine sediments ceased on a raised beach deposit at the lake margin. However, the mechanisms that caused loss of the diatom record remain unclear (Bush et al., 1992). The reappearance of diatoms after 5000 B.P., albeit in small quantities, may suggest a slight increase in effective moisture. The vegetation record for this site shows evidence of human disturbance since the beginning of the Holocene.

15.3.5. The Caribbean

Published paleolimnological records are available for only two sites in the Caribbean: Wallywash Great Pond in Jamaica (Street-Perrott et al., 1993) and Lake Miragoane in Haiti (Hodell et al., 1991). A lack of freshwater lakes and potential dating problems in karstic environments account for the scarcity of paleolimnologi-cal information for the region. Of the two records, only the record for Lake Miragoane is of sufficient resolution to examine Holocene climatic trends in any detail. Analysis of stable isotopes in the core from Wallywash Great Pond indicates three cycles of wet and dry conditions during the Holocene. However, accurate radiocarbon dating of these events has not been possible because of evidence of significant contamination of sediments with dead carbon from springs surrounding the lake (Street-Perrott et al., 1993).

Oxygen isotope data for ostracodes in the Lake Mi-ragoane core show major changes in precipitation and evaporation through the Holocene (Hodell et al., 1991). For the early Holocene, declining 818O values indicate a shift from the more arid conditions of the Pleistocene to a wetter climate and higher lake levels. The lowest 818O values in the Miragoane record are between 7000 and 5300 B.P. Following a slight increase in evaporation relative to precipitation at ca. 5200 B.P., lake levels remained relatively high until approximately 3200 B.P. (Hodell et al., 1991). The short, dry episode at ca. 5200 B.P. may correspond to other abrupt dry phases evident elsewhere in the region at ca. 5000 B.P. The pollen record indicates that mesic forests developed in the vicinity during the mid-Holocene when lake levels were relatively high. After 3200 B.P., a marked increase in 818O points to drier conditions. By 2400 B.P., evaporation rates reached their highest levels, marking the beginning of a late Holocene dry phase.

15.3.6. Venezuela

The only study for northern South America is for Lake Valencia, where diatom, mineralogical, and sedi-mentological data (Bradbury et al., 1981) show a transition at ca. 8500 B.P. from a shallow, saline, closed-basin lake to an overflowing, low-salinity system, driven largely by rising sea levels. Diatom assemblages and sedimentology suggest an interval of lower lake levels between ca. 7000 and 6000 B.P., followed by generally high levels until at least 3000 B.P. Effective moisture has decreased over the last 2200 years, marked by falling lake levels and less extensive forests (Leyden, 1985).

15.3.7. Regional Climate Patterns

Central America and the circum-Caribbean region today are within a summer precipitation regime, with moisture derived mainly from the Atlantic through the Northeast Trades (McGregor and Nieuwolt, 1998). Locally, Pacific moisture sources may be significant, particularly from tropical cyclones. A monsoon-type circulation occurs over México and into the southwest and central United States in the Northern Hemisphere summer, with moisture from both the Gulf of México / Caribbean and the Pacific Ocean (Douglas et al., 1993).

Figures 6 and 7 summarize available data for the period 7000-4500 B.P. for Central America and the cir-cum-Caribbean region. A combination of poor dating control and variable preservation of certain types of microfossils means that the number of data points for different time periods is variable. It has been clear from review of the literature that there is a need for much better dating control on Holocene sequences. This need is particularly evident for periods when it appears that the climate may have fluctuated quite rapidly (e.g., 6000-5000 B.P.).

Lake levels were generally low during the late glacial period because of southerly displacement of the Intertropical Convergence Zone (ITCZ) and/or because of shifts in the position and strength of the Northeast Trades (see Bradbury et al., 2000). The only exception to this trend is in northern México, which probably came under the influence of the westerlies which were displaced equatorward (see earlier). The picture for 7000 B.P. (Fig. 6) is rather mixed. It is clear that a number of low-lying sites were still affected by sea levels well below those now. Modern summer rainfall systems do not appear to have become fully established. The data for northern México (Babicora), taken with records for other data sources and adjacent areas north of the United States—México border (see earlier), seem to indicate conditions still wetter than those now (persisting from the late Pleistocene), possibly due to enhanced winter precipitation. It would be valuable to have data for sites to the west of the Sierra Madre Occidental and around the Gulf of California for this time period. How far south any enhanced winter precipitation was effective remains unclear.

The picture for 6000 B.P. (Fig. 6) suggests that wetter conditions had become established over much of the region, although Panama may be an exception. Lake levels in the lowlands of the Yucatán had risen rapidly by this time, whereas conditions in northern México were becoming drier. This pattern may reflect a strong summer precipitation regime, together with warmer temperatures, possibly associated with shifts in the position and strength of the Bermuda High (Hodell et al., 1991; Metcalfe et al., 1997) and changes in the strength of the summer Mexican monsoon. By 5000 B.P. (Fig. 7), a number of sites in central México record increased aridity, although this may have been quite short lived. Whether this was due to a further increase in temperature or a decrease in precipitation cannot be determined. Conditions in the Chihuahuan desert were moving toward those now. Data for 4500 B.P. (Fig. 7) show an intensification of the drying in central México, suggesting a decrease in the westerly penetration of summer storms. The records for many sites are hard to interpret over this interval.

The data for 3000 B.P. show considerable variation between lakes in the same areas. This variation seems to be indicative of another period of fluctuating climate, with short episodes of higher or lower lake levels. An increasing number of lacustrine records are affected by human impact, in some cases apparently completely obscuring the paleoclimatic signal. Indeed, it has been argued that the climate of Central America did not change significantly during the late Holocene (Rue, 1987). This assumption was based entirely on pa-lynological records available at the time, and it is now evident that significant and abrupt climatic changes did occur throughout the Holocene.

After ca. 3000 B.P., some lakes in central México showed rising lake levels, and high lake stands initially persisted in the Yucatán, whereas the Caribbean became dry. Between 2000 and 1000 B.P., however, the Yucatán and central México showed some of the lowest lake levels of the Holocene, whereas moisture appears to have increased in the Caribbean (Hodell et al., 1991, 1995).

Across the region as a whole, it is possible to identify areas that often display coherent patterns of change. Northern México seems to be quite different from the rest; the records from the Mexican Yucatán and Haiti often show the same trends. The lakes in central Méx

7000 14C yr BP

7000 14C yr BP

A More available moisture than at present

V Less available moisture than at present o As present

? No clear signal

A More available moisture than at present

V Less available moisture than at present o As present

? No clear signal

6000 14C yr BP

6000 14C yr BP

FIGURE 6 Maps of lake-level status of lakes in Central America and the circum-Caribbean at 7000 and 6000 B.P.

ico are very varied—in spite of being within the same summer rainfall regime as the Yucatán and Caribbean today, they show the same signal as these regions only at 6000 B.P. It should be noted perhaps that a number of central Mexican records are affected by volcanic activity over the period 5200-4300 B.P., which complicates paleoclimatic interpretation. The site in Panama is different from most of the others, indicating dry con

5000 14C yr BP

5000 14C yr BP

A More available moisture than at present

V Less available moisture than at present o As present

? No clear signal

A More available moisture than at present

V Less available moisture than at present o As present

? No clear signal

4500 14C yr BP

4500 14C yr BP

FIGURE 7 Maps of lake-level status of lakes in Central America and the circum-Caribbean at 5000 and 4500 B.P.

ditions over most of the Holocene. Some countries have no published Holocene data (e.g., Cuba), and there may be significant paleoclimatic gradients that cannot be identified from the present coverage of sites.

This may be an area where implementation of a regional-scale climate model would yield particularly interesting results and offer up many testable hypotheses.

Was this article helpful?

0 0

Post a comment