consistent overall pattern: (1) 1000- to 2000-yr aggradational episodes coinciding with phases of lower flood frequency, and (2) 300- to 500-yr phases of higher flood frequency involving river entrenchment or avulsion. The overall period of 2000 to 2500 yr coincides with other detailed proxy records of Holocene climate such as Denton and Karlen's (1973) analysis of mountain glaciation and the oxygen isotope variations in the Camp Century, Greenland, ice core (Dansgaard et al., 1986).
Paleoflood hydrological investigations are just beginning to provide a perspective on extreme river flows over the past 2000 yr (Figure 6.3). The most interesting patterns are appearing in the southwestern United States, where approximate 1000- and 500-yr periods are observed (Figure 6.4). A recent compilation of SWD-PSI studies (Ely et al., in press) has established the regional coherence of these patterns of flooding.
Baker (1983) suggested that the effects of Quaternary climatic change on South American rivers might be used as a surrogate to explore potential effects of man-induced environmental change in the region. During the last full glacial, most of South America north of the Tropic of Capricorn was dominated by climates drier than present (Tricart, 1985). The Amazon Basin rivers have a remarkable diversity of patterns, reflecting the influence of Andean source region and the relative abilities of lowland rivers to rework relict alluvium deposited during the drier full-glacial periods (Baker, 1978). Although little fluvial paleohydrological work has been done in the region, the global importance of Amazon and Orinoco Basin hydrology would seem to warrant further attention.
Central-western South America is strongly influenced by the El Niño-Southern Oscillation (ENSO) phenomenon. This coupled oceanic-atmospheric oscillation is associated with anomalous periods of flood and drought in diverse parts of the Earth (Yarnal, 1985). The ENSO is of considerable interest for understanding the clusterings of floods and droughts observed in paleohydrological records. Preliminary work from northern Australia suggests that patterns of tropical storm activity and related floods may reflect ENSO variations (Wohl, 1988). Western South America would appear to be very important for establishing the pulse of ENSO variation from studies of Holocene flood sediments (Wells, 1987).
Fluvial responses in much of equatorial Africa, northern Australia, and southern Asia are dominated by the annual monsoonal weather cycle (Figure 6.5). Monsoons
are extremely complex meteorological phenomena characterized by dramatic variability from year to year, as documented by the historical record (Mooley and Parthasarathy, 1984). Both droughts and wet anomalies occur in runs, suggesting clumping or persistence of wet and dry anomaly values (Kutzbach, 1986).
Paleohydrological data from tropical Africa indicate relatively dry conditions, reduced annual discharge in the Nile, Niger, and Senegal rivers, but with high flood peaks between 17 and 25 ka (Williams, 1985). The latest Pleistocene and early Holocene is a markedly wetter period (COHMAP, 1988). Exceptionally high Nile floods occurred around 12 to 11 ka, and fluctuations in flow levels occurred later into the Holocene, but at reduced levels from the late-glacial extremes.