skip navigation

Previous | View all abstracts | Next


MUHS, DANIEL R.  U.S. Geological Survey.
Budahn, James R.  U.S. Geological Survey.

Loess (eolian silt) is geographically the most extensive surficial deposit in Alaska (Fig. 1). In central Alaska, near Fairbanks, loess is several tens of meters thick and the periods of deposition may span the entire Quaternary and extend into the Pliocene. Nevertheless, there have been few studies of loess origins in Alaska and it was not even until the mid-1950's that an eolian origin for these deposits was finally accepted, based on Pw's (1955) pioneering study.

The source areas of loess in central Alaska (near Fairbanks and Nenana) are not immediately apparent. The loess is composed of quartz, plagioclase, K-feldspar and mica, as well as a suite of accessory minerals (e.g., magnetite, zircon, etc.). The high mica content suggests that local, schist-bedrock-derived particles, carried by first-order streams, is a possible source. However, Pw (1955) showed that there are many accessory minerals in the loess that do not occur in the local bedrock. Immediately to the south of the main loess bodies near Fairbanks and Nenana, the Tanana River drains the glaciated Alaska Range. Although silt is abundant in this river even now (and likely was greater during glacial periods), other evidence suggests that dominant paleowinds were not from the south during the last glacial period. Orientations of sand dunes across much of Alaska indicate that last-glacial, or at least late-glacial, paleowinds were dominantly from the northeast (Lea and Waythomas, 1990). To the north of Fairbanks and Nenana, the Yukon River (which drains parts of the glaciated Brooks Range) provides an alternative source. Silt derived from the Yukon River would be consistent with the northeasterly dune-derived paleowind pattern. However, this river is 140-170 km from Fairbanks and Nenana. Fairbanks/Nenana-area loess has 50-70% coarse (53-20 microns) silt, which seems far too coarse to have traveled distances of over 100 km from a source, based on analogs with areas of active loess deposition in southern Alaska and past loess deposition in the North American midcontinent.

In this study, we investigated the origin of central Alaskan loess using trace element geochemistry as a provenance tool. We collected loess from three stratigraphic sections in the Fairbanks-Nenana area, Birch Hill, Gold Hill and Halfway House, from east to west, over a distance of ~ 45 km. Loess samples were taken from those portions of the section that are thought to date from the last glacial period, based on the stratigraphy and chronology reported by Muhs et al. (2003). We also sampled three potential loess sources: (1) local schist bedrock, (2) Tanana River silts and (3) Yukon River silts. Schist was analyzed as bulk rock after pulverization, but Tanana and Yukon River sediments were first fractionated into loess-sized (53-2 microns) material. Trace element abundances were determined by instrumental neutron activation analysis (INAA). We chose immobile elements Sc, Cr, Hf, Ta, and Th and a suite of rare earth elements (REE), including La, Ce, Sm, Eu, Tb, Yb and Lu to characterize the source rocks and sediments.

Results indicate that the three possible loess sources have very different compositions. Local schist bedrock has a highly variable composition, typical of schists in many other regions. Because the loess has a much more uniform composition, schist is unlikely to be the sole source of the loess, which supports Pw's (1955) earlier mineralogical work. Tanana River and Yukon River silts have distinctly different compositions from each other. For example, Tanana River silts have higher La/Yb, Th/Ta and Th/U values, whereas Yukon River silts have higher Cr/Sc and Zr/Hf values as well as Eu/Eu*, a measure of the negative Eu anomaly in the REE suite. Simple bivariate plots indicate that central Alaskan loess has compositions that overlap closely the range of Tanana River silts for some element ratios (Th/Ta, Zr/Hf, and Eu/Eu*). However, a measure of the full suite of REEs (La/Yb) and Cr/Sc ratios indicate that loess has compositions intermediate between Tanana and Yukon River sediments.

In order to provide a more quantitative analysis of the relative contributions of the three possible source materials, we also used a geochemical modeling approach that has been utilized in igneous petrology (Budahn and Schmitt, 1985). The approach uses a multiple regression method for the mean abundances of a suite of immobile elements for each source material compared to each loess sample. The calculations yield the estimated contribution of each source material (in percent) and chi-square analysis is used to assess the degree of explanation of the resultant regression equation. Results of the geochemical modeling approach indicate that central Alaskan loess is probably derived mostly from a mixture of Tanana River sediment and Yukon River sediment, with minor contributions from local schist bedrock.

Our interpretation of the results is that loess-generating winds in central Alaska were complex during the last glacial period. The orientations of dunes across the state suggest a dominance of northeasterly winds and the geochemical evidence for a Yukon River contribution to central Alaskan loess is consistent with this pattern. However, contributions from the Tanana River, also indicated by the geochemical data, require southerly winds. An explanation for this apparent dilemma may be the importance of katabatic winds during the last glacial period. Expanded, north-flowing glaciers from the Alaska Range (to the south of Fairbanks) during the last glacial period may have generated local, but strong southerly winds, a hypothesis proposed by Thorson and Bender (1985) and Muhs et al. (2003). These winds would have allowed transport of silt out of the Tanana River and into the Fairbanks-Nenana area. Alternation of regional-scale northeasterly winds and local, southerly, katabatic winds during the last glacial period can explain the mixture of sediment sources apparent in the geochemistry of central Alaskan loess. Multiple sources and paleowinds that could have as much as 180 degrees of directional variability demonstrate that loess deposition is a complex process. The results suggest that reconstruction of paleowinds or interpretation of loess-paleosol sequences in other regions first require an understanding of loess source sediments.

Budahn, J.R., and Schmitt, R.A., 1985, Petrogenetic modeling of Hawaiian tholeiitic basalts: A geochemical approach: Geochimica et Cosmochimica Acta, v. 49, p. 67-87.

Lea, P.D., and Waythomas, C.F., 1990, Late-Pleistocene eolian sand sheets in Alaska: Quaternary Research, v. 34, p. 269-281.

Muhs, D.R., Ager, T.A., Bettis, E.A., III, McGeehin, J., Been, J.M., Begt, J.E., Pavich, M.J., Stafford, T.W., Jr., and Stevens, D.S.P., 2003, Stratigraphy and paleoclimatic significance of late Quaternary loess-paleosol sequences of the last interglacial-glacial cycle in central Alaska: Quaternary Science Reviews, v. 22, p. 1947-1986.

Pw, T.L., 1955, Origin of the upland silt near Fairbanks, Alaska: Geological Society of America Bulletin, v. 66, p. 699-724.

Pw, T.L., 1975, Quaternary geology of Alaska: U.S. Geological Survey Professional Paper, 835, 145 p.

Thorson, R.M., and Bender, G., 1985, Eolian deflation by ancient katabatic winds: A late Quaternary example from the north Alaska Range: Geological Society of America Bulletin, v. 96, p. 702-709.

Figure 1. Map showing the distribution of loess in Alaska (shaded areas) and location of the Fairbanks area. Loess distribution from Pw (1975) and additional sources cited in Muhs et al. (2003).

Previous | View all abstracts | Next