North

Eolian sand

Soil there was widespread lunette formation during full glacial time (Holliday, 1997a), and extensive eolian sand sheet deposition is dated to ca. 11,000-10,000 14C B.P. (Holliday 1997b, 2000). Extensive eolian sheet sands in Texas and adjacent parts of New Mexico are confidently correlated to the late glacial period because of their association with well-dated artifacts (Fig. 10). Forman et al. (1995) stated that eolian activity in the mid-continent ceased between ca. 12,000 and 9000 14C B.P., based on examination of a few sections in a limited area of northeastern Colorado. However, the studies from Nebraska (Loope et al., 1995; Mason et al., 1997), Texas, and New Mexico (Holliday, 2000) indicate that the late glacial period was an important time of eolian sand deposition in the midcontinent.

Because few dunes (as opposed to sand sheets) can be confidently linked to the last glacial period in the Great Plains, it is difficult to reconstruct paleowinds. However, distribution of last glacial eolian sands relative to probable source sediments indicates that winds were probably northwesterly on the central Great Plains and westerly on the southern Great Plains during last glacial time (Loope et al., 1995; Mason et al., 1997; Muhs et al., 1996; Holliday, 1995a, 1997a), similar to the present.

12.6.3. Mid-Holocene: Loess and Sand

Mid-Holocene eolian sediments have been reported in some parts of midcontinental North America. A thin (1-4 m) eolian silt called the Bignell Loess is found in a patchy distribution west of the Missouri River in the Great Plains region of Nebraska, Kansas, and Colorado (Johnson and Willey, 2000; Mason and Kuzila, 2000). Radiocarbon ages indicate that it was deposited between ca. 11,000 and 9000 B.P. at some localities and possibly during the mid-Holocene, between ca. 10,000 and 1500 B.P., at others (Fig. 8). Direct dating of Bignell Loess using TL methods gives ages ranging from —9000 to —3000 cal. years B.P. (Pye et al., 1995; Maat and Johnson, 1996). A recent study by Olson et al. (1997) indicates that a post-mid-Holocene loess covers extensive parts of southwestern Kansas and may be younger than the Bignell Loess found farther north. Holocene loess has also been described in Saskatchewan (David, 1970; Vreeken, 1996) and, as with Holocene loess farther south, has a patchy distribution.

Although there is little direct geomorphic evidence for it, some stratigraphic evidence suggests mid-Holocene eolian sand activity in midcontinental North

FIGURE 11 Stratigraphy and ages of selected eolian sand sections in Colorado, Minnesota, and Texas that show evidence of mid-Holocene eolian activity. Colorado data are from Forman and Maat (1990) and Forman et al. (1995), Minnesota data are from Grigal et al. (1976), Texas data are from Holliday (1995b). Locations are shown in Fig. 7.

FIGURE 11 Stratigraphy and ages of selected eolian sand sections in Colorado, Minnesota, and Texas that show evidence of mid-Holocene eolian activity. Colorado data are from Forman and Maat (1990) and Forman et al. (1995), Minnesota data are from Grigal et al. (1976), Texas data are from Holliday (1995b). Locations are shown in Fig. 7.

America (Fig. 11). Radiocarbon ages and stratigraphic data from the western Nebraska Sand Hills indicate that mid-Holocene eolian activity probably created a large dune dam that blocked a major drainage in the area, as was the case during the late glacial period (Loope et al., 1995; Mason et al., 1997; Muhs et al., 2000). In eastern Colorado, there are a few closely spaced localities where evidence of mid-Holocene eolian sand movement has been documented (Forman and Maat, 1990; Forman et al., 1995). Mid-Holocene dune activity took place in parts of the intermountain basins of Wyoming to the west of the Great Plains (Gaylord, 1990). Holliday (1995b, 1997a,b) reported radiocarbon and stratigraphic evidence of mid-Holocene lunette formation in the southern Great Plains, and dry valleys or draws in this region contain thick eolian sands that are of mid-Holocene age. However, the upland dunes in this region, mostly parabolic forms, do not date from the mid-Holocene and are largely of late Holocene age (Holliday, 1995c). Preservation of mid-Holocene eolian sand in draws and not on uplands may be a function of the more protected sedimentary setting in the draws, where late Holocene reworking of sediment would be less likely. Elsewhere, the record for mid-Holocene eo-lian sand deposition is scanty and is based on indirect lines of evidence such as dunes with a degree of soil development that is greater than that on late Holocene eo-lian sand, but less than that on late Pleistocene sand (Madole, 1995; Muhs et al., 1996). In some parts of the Great Plains, dunes are depleted in carbonate minerals, thought to be the result of physical reduction of sand-sized grains from extended periods of eolian activity in the mid-Holocene period (Muhs et al., 1997b; Arbogast and Muhs, 2000). Although it seems likely, from other proxy paleoclimate evidence (discussed later), that conditions were optimal for eolian sand activity over the semiarid Great Plains during the mid-Holocene extensive late Holocene, eolian activity may have removed much of the record.

Interestingly, subhumid areas around the margins of the Great Plains may have a better record of mid-Holocene eolian activity, such as that documented for northern Minnesota (Fig. 11) by Grigal et al. (1976). It can be hypothesized that these currently subhumid to humid areas were affected by a mid-Holocene dry period, but, unlike the drier Great Plains to the west, were unaffected by reworking during the late Holocene. Stabilized eolian sand is extensive in parts of Minnesota, Wisconsin, Illinois, and Indiana (Fig. 7) and in smaller areas in Kansas and Michigan. It would be worthwhile to conduct detailed stratigraphic and geochronologic studies of these areas to see if the last period of eolian activity dates to the mid-Holocene. Some of the best records of mid-Holocene eolian activity may be found in lake sediments and in areas marginal to the Great Plains (Dean et al., 1996).

12.6.4. Late Holocene

In the past, a number of investigators assumed that dunes in the Great Plains were last active during full glacial time, at ca. 22,000-16,000 14C B.P. (Watts and Wright, 1966; Wright, 1970; Warren, 1976; Sarnthein, 1978; Wells, 1983; Kutzbach and Wright, 1985), or at least no later than during what is perceived to be a warm and dry mid-Holocene, at ca. 8000-5000 B.P. Recent studies have confirmed (as discussed earlier) that the last glacial period was indeed a time of eolian sand activity over much of the region. However, a number of recent studies also show that much eolian sand in the Great Plains region is of late Holocene age. Indeed, in the Great Plains of both Canada and the United States, the areal extent of late Holocene eolian sand is much greater than that of last glacial sand and is a more important part of the record than previously thought. For example, dunes of the Nebraska Sand Hills cover —50,000 km2, yet all have only simple A/ AC/ C soil profiles (Entisols), similar to the young dune fields of the Pampas in Argentina discussed earlier. This observation, along with late Holocene radiocarbon ages at numerous localities, suggests that most or all of the Nebraska Sand Hills region has been active at some time in the late Holocene, though not all parts of the dune field necessarily were active simultaneously (Muhs et al., 1997a).

An indirect piece of evidence supporting late Holocene ages for eolian sand comprises new data that show that sand and loess in the Great Plains are geo-chemically and mineralogically distinct, indicating different source sediments. This is a departure from the long-held concept, articulated by Lugn (1935, 1939, 1962, 1968), that eolian sand and loess in the Great Plains are essentially the same deposit, with eolian sand a coarse-grained, proximal facies and loess a fine grained, distal facies. In both eastern Colorado and Nebraska, loess is highly calcareous, whereas eolian sand is not. This mineralogical difference is reflected in the higher Ca concentrations in loess compared to eolian sand; Ca/Sr values are also significantly different between the two sediment types (Fig. 12). Eolian sand has concentrations of both Ti and Zr that are lower than in loess that occurs to the south and east of the dune fields. Furthermore, ratios of Ti to Zr in eolian sand are lower than these values in loess in both areas (Fig. 12). These compositional differences, which reflect both carbonate mineral and heavy mineral assemblages, cannot be explained by differences in transport processes alone, and they suggest that eolian sand and loess have different sources. Muhs et al. (1996) presented evidence to show that eolian sand in Colorado is derived mainly from South Platte River sediments. However, loess in Colorado is derived partly from South Platte River sediments, but also from Cretaceous and Tertiary bedrock sources (Aleinikoff et al., 1998, 1999). Thus, because different source sediments supply dune fields and loess, eolian sand and loess entrainment need not be synchronous in time.

Stratigraphic, soil geomorphic, radiocarbon, and luminescence methods demonstrate that eolian sand over a wide range of midcontinental North America has been active in the past 3000 years (Ahlbrandt et al., 1983; Muhs, 1985; Swinehart and Diffendal, 1990; Madole, 1994, 1995; Holliday, 1995a,c, 1997a,b; Forman et al., 1995; Loope et al., 1995; Arbogast, 1996; Muhs and Holliday, 1995; Muhs et al., 1996, 1997a,b; Wolfe et al., 1995, 2000; Stokes and Swinehart, 1997). In addition, most of these studies have stratigraphic data indicating multiple periods of eolian activity in the late Holocene (Fig. 13). The number of radiocarbon ages and their analytical uncertainties do not yet make it possible to test the hypothesis of regional synchroneity of activity. However, these observations indicate that, contrary to earlier beliefs, eolian sand in this region can be active under an essentially modern climatic regime. In fact, observations by explorers in the nineteenth century indicate that many parts of the Great Plains had active dune sand where it is now stable (Muhs and Holliday, 1995; Muhs and Wolfe, 1999). The evidence for multiple periods of activity and stability in the past few thousand years indicates that Great Plains eolian sand is quite sensitive to small changes in the overall moisture balance and degree of vegetation cover. Most of the late Holocene dunes in the midcontinent of North America are parabolic forms, which are excellent paleowind indicators. Orientations of late Holocene dunes indicate northwesterly winds in the central Great Plains and southwesterly winds in the southern Great Plains, similar to modern wind regimes (Fig. 14).

0 0

Post a comment