Journal of Arid Environments (2001) 47: 403–424 doi:10.1006/jare.2000.0731, available online at http://www.idealibrary.com on

Smith dune field, Washington, U.S.A: relation to glacial outburst , the Mazama eruption, and paleoclimate

David R. Gaylord*, Franklin F. Foit Jr., Jeffrey K. Schatz & Angela J. Coleman

Department of Geology, Washington State University, Pullman, WA 99164-2812, U.S.A.

(Received 11 October 1999, accepted 20 September 2000)

Sedimentary deposits from the Smith Canyon dune field, south-central Col- umbia Basin, Washington, U.S.A. document climatically-influenced Late and Holocene aeolian and fluvial deposition in a region impacted by glacial outburst floods and tephra falls. The depositional history is sum- marized by five environmentally distinctive and climatically sensitive sedimen- tary units (temporal limits estimated): Unit 1 (c. 15)5}8 ka), pedogenically altered glacial outburst flood and minor aeolian silt and clay; Unit 2 (c. 8}6)9 ka), fluvial and minor aeolian sand; Unit 3 (c.6)9}6)8 ka), flood-induced fluvial sand with gravel-sized tephra clasts; Unit 4 (c.6)8}3)9 ka), aeolian dune sand; Unit 5 (c.3)9 ka to present), pedogenically altered, stabilized dune sand. Estimated age ranges are based on stratigraphic position, tephrochronology, and correlation with temporally constrained strata from elsewhere in the region.

! 2001 Academic Press

Keywords: sand dunes; Mazama; Holocene; palaeoclimate; glacial outburst flood

Introduction

Reconstruction of Pleistocene and Holocene from semi-arid and arid continental settings such as those that dominate the Columbia Basin of Washington are often hindered by the lack of temporal resolution and stratigraphic continuity in preserved sedimentary deposits. Perennial lake strata are a notable exception because of their potential to provide environmental and proxy climate records similar in complete- ness to those derived from ice cores and marine sediment. For example, pollen from Carp Lake in the south-western Columbia Basin has added significant insight into fluctuating ecologic and, inferentially, climatic conditions in the region (Barnosky, 1985; Whitlock & Bartlein, 1997). Unfortunately, perennial lakes are relatively rare in the Columbia Basin, necessitating reliance on proxy climate records from other sources. These sources consist primarily of Pleistocene and Holocene aeolian and fluvial sediment. Windblown aeolian silt (loess) blankets much of the eastern parts of the basin while relatively local sand dune deposits are distributed across the central

* Corresponding author. E-mail: [email protected]

0140}1963/01/040403#22 $35.00/0 ! 2001 Academic Press 404 D. R. GAYLORD ET AL. basin. Fluvial deposits tend to be best exposed in the northern, southern, and central portions of the basin. Loess deposits have yielded insight into former levels of plant activity and climate in the Columbia Basin during the Pleistocene and Holocene (O’Geen, 1998; Busacca, 1998). Similarly, the stratigraphic, sedimentary, and geomorphic record of sand dune deposits from the central Columbia Basin have contributed to understanding of the timing of aeolian activity, inactivity, and related climate changes (Gaylord & Stetler, 1994; Stetler & Gaylord, 1996). Temporal patterns and characteristics of fluvial sediment along the Columbia River and its tributaries have yielded insight into Holocene climate fluctuations in the northern Columbia Basin (Chatters & Hoover, 1992). Reconstruction of the depositional and erosional records of sedimentary deposits from arid and semi-arid regions relies primarily on the availability of well-preserved strata that can be constrained temporally. Given the relatively rapid rate of basin excavation by the Columbia River since the , as well as the minimal preservation of otherwise easily dated organic matter, this location would not at first seem appropriate. Fortunately, an unusual and recurring set of geologic events have combined to favor sedimentary aggradation throughout the latest Pleistocene and Holocene. In particular, enormous volumes of sediment were repeatedly delivered to the basin via Pleistocene glacial outburst floods (Bretz, 1969; Waitt, 1980, 1984, 1985; Atwater, 1984, 1987). Reworked sediments from this source are intercalated with geochemically distinct tephras whose ages are well established (Sarna-Wojcicki et al., 1983) and relatively inexpensive to obtain. In this paper, the depositional record of the Smith Canyon dune field as revealed primarily in exposures along the Eltopia Canal is examined to provide insight into the late Pleistocene to Holocene geologic, and to a limited extent, paleoecologic and paleoclimatic history of the south-eastern Columbia Basin. Relative ages of strata are estimated from evaluation of stratigraphic position, cross-cutting relations, and teph- rochronology. Previous sedimentary, stratigraphic, and geomorphic analyses from this area (Schatz, 1996; Gaylord et al., 1997) form the basis for environmental reconstruc- tions presented in the paper.

Background

Geologic setting

The Columbia River Basin is a regional-scale structural and topographic depression in central and southeastern Washington (Fig. 1). Basin geology is dominated by variably deformed Miocene continental flood basalt and Miocene(?)/Pliocene to Recent continental sedimentary deposits (Fig. 2). Bedrock units exposed in the Smith Canyon dune field include the Pomona, Elephant Mountain, and Ice Harbor Members of the Saddle Mountains Basalt, amongst the youngest flows from the Columbia River Basalt Group (Reidel & Fecht, 1994; Schuster, 1994). These tholeitic basalts are unconformably overlain by fluvial-lacustrine strata of the Miocene/Pliocene Ringold Formation (Lindsey & Gaylord, 1990; Reidel et al., 1994), quartzo-feldspathic and locally lithic-rich deposits with genetic ties to the ancestral Columbia and Snake Rivers. As the Columbia and Snake Rivers evolved, they and their tributary channels in- fluenced the paths of episodic Pleistocene glacial outburst floods (Bretz, 1969; Waitt, 1980, 1984, 1985; Atwater, 1984, 1987; McDonald & Busacca, 1988). Outburst floods created the Channelled Scabland, an irregular terrain characterized by broad, low channelways, or coulees, separated by elevated, loess-capped islands. As flood waters flowed towards the confluence of the Snake and Columbia Rivers they excavated RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 405

Figure 1. Generalized paleogeographic map of the U.S. Pacific Northwest highlighting: major glacial, volcanic, and sedimentary features, the margin of the Columbia Basin (as defined in this paper) and location of the Smith Canyon dune field (see also Fig. 2). CL"Carp Lake; ML"Mahoney Lake. Large arrows along coastline depict autumn-to-spring-dominant south- westerly to north-easterly winds, and summer-dominant, north-west to south-east winds. Cordil- leran ice sheet ( ), ( ), area inundated by outburst floods ( ), Volcano ( ), Columbia Basin ( ).

numerous 50}150-m deep, subparallel to parallel, NE}SW-oriented coulees, including Smith Canyon. During waning flood flows these coulees became partially or wholly filled with mixtures of gravel and/or finer-grained sand-silt-clay, primarily in so-called ‘slackwater’ areas. Sedimentary deposits from the most recent outburst flood events lie at or just beneath the surface throughout the Smith Canyon area (Gulick, 1994; Schuster, 1994). Earlier flood deposits, though recognized further north in the Colum- bia Basin (McDonald & Busacca, 1988), have not been identified from Smith Canyon or the surrounding area. Many of these now partially filled coulees have served as corridors for prevailing SW to NE winds that have promoted the transportation and deposition of late Pleistocene (?) and Holocene dune sands. Modern streams that flow to the south and southwest along the regional gradient, have been responsible only for modest and localized landscape modification since the last flood, c. 12)5 ka (Baker et al., 1991). Glacial outburst flood-derived sands have been identified as sources for dune sand elsewhere in the Columbia Basin, including the Hanford Site (Gaylord & Stetler, 1994; Stetler & Gaylord, 1996), near Moses Lake (Petrone, 1970), and the Juniper dune field north-east of Smith Canyon (Gaylord et al., 1997, 1998). Outburst floods and glacial meltwaters also were major sources for the silt and clay that comprise the thick loess deposits that blanket much of eastern Washington (McDonald & Busacca, 1988, 1992; Busacca & McDonald, 1994). Unlike the loess, dune sands in the Columbia Basin are much more localized and relatively close to their sediment sources. 406 D. R. GAYLORD ET AL.

Figure 2. Generalized geologic map of the Smith Canyon area delimiting major geologic units and position of Smith Canyon and Juniper dune fields. Eltopia Canal stratigraphic section site identified by circle with cross near center. Geology modified after Reidel and Fecht (1994), Schuster (1994), and Schatz (1996). SCDF"Smith Canyon dune field. Quaternary: sand dunes ( ), loess ( ), outburst flood sand, silt, and gravel ( ); Pliocene–Miocene: Ringold formation ( ); Miocene: Basalt ( ).

Climate

The climate of the Smith Canyon area largely reflects its location between the Cascade Range and the Rocky Mountains. The Cascade Range to the west modifies generally mild, wet, maritime flows from the Pacific Ocean, thus producing a rain shadow in the Columbia Basin. The Rocky Mountains shield the Columbia Basin from cold, polar air masses, further moderating the overall climate. The modern climate is marked by warm, dry summers and cool, damp winters (Liverman, 1975). Annual precipitation averages nearly 16 cm and the mean annual temperature averages about 123C. Sub-freezing temperatures usually occur only between November and February, and then for less than a total of 30 days. Winds tend to be from the north-west from May to August, and from the south-west and west-south-west during the remainder of the year; the highest winds usually come from the south-west and west-south-west (Gaylord & Stetler, 1994; Stetler & Gaylord, 1996) (Fig. 1). Paleoclimate simulations highlight the role that seasonal solar insolation, a shrinking Cordilleran ice sheet (Kutzbach & Guetter, 1986; Kutzbach & Webb, 1993) and expanding vegetation cover (Thompson et al., 1993; Whitlock & Bartlein, 1997) have had on the late Pleistocene and Holocene climate of the Pacific Northwest region (including the Columbia Basin). As the ice sheet retreated during the late Pleistocene, the jet stream shifted northward, strengthening the eastern Pacific sub-tropical high RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 407 located off the Washington and Oregon coastlines (Thompson et al., 1993; Whitlock & Bartlein, 1997). The relatively warm and humid post-glacial conditions of the late Pleistocene were supplanted by warmer and more arid conditions during the early and middle Holocene. By the late Holocene the climate had moderated to the relatively moist and cool conditions that have persisted until today (Whitlock & Bartlein, 1997).

Methods

The depositional history of the Smith Canyon sedimentary deposits was reconstructed using an integrated field and laboratory study. Field-based work included mapping of major depositional units on a 7)5-min USGS topographic base, and systematically sampling dune sands on a 1-km grid with hand-dug pits. Sample sites were located using a hand-held Global Positioning Satellite (GPS) device. Sediment samples also were collected from potential sources and detailed stratigraphic sections were measured where possible. The most informative stratigraphic sections were measured along Eltopia Canal, an irrigation canal where up to 14)5 m of outburst flood, loess, and dune sands are exposed along an approximately 100-m-long outcrop. Data from three stratigraphic sections were combined into a composite stratigraphic section central to this paper. Compositional and textural analyses were performed at Washington State University using standard thin section point count and analytical methods (Galehouse, 1971; van der Plas & Tobi, 1965) and sieving and settling tube analyses as outlined by Galehouse (1971). Grain size and sorting terminology follows that of Folk (1980). Ages of the environmentally distinctive units described in this paper are estimated and reflect stratigraphic placement of the units relative to outburst flood strata and the Mazama tephra. Preserved radiocarbon-datable material is virtually non-existent in the strata. Attempts prior to this study to acquire luminescence ages from basalt-rich sediment, compositionally comparable to that at Smith Canyon, yielded spurious re- sults, thus further limiting the dating options. Tephrochronologic methods and results are described in Table 1.

Smith Canyon dune field morphology

The Smith Canyon dune field blankets nearly 40 km2 of the lower Snake River drainage with up to 30 m of stabilized and active dune sand (Fig. 3). Situated immediately south-west of the Juniper dune field, the Smith Canyon dunes consist primarily of phytogenic parabolic and blowout dunes anchored by native and introduced grasses and shrubs that fall within Artemisia tridentata-Agropyron vegetation zone (Daubenmire, 1970). Widely spaced western juniper (Juniperus occidentalis) (Little, 1971) dot the landscape and are an important stabilizing influence. Actively migrating dunes generally have less than a 5% vegetative cover and account for less than 20% of the dune field. These active dunes tend to be concentrated in topographically elevated positions subject to higher wind speeds, zones of low subsurface moisture, or areas experiencing anthro- pogenic disturbance. Active and stabilized dune forms have morphologies that generally reflect south-west to north-east transport.

Dune field sedimentology and provenance

Sediments within the Smith Canyon}Juniper dune field consist of very fine to medium- grained, moderately to moderately well-sorted, quartzo-feldspathic to basalt-rich sand. Compositions vary directly with texture; finer-grained sands are more quartzo-feldspathic 408 D. R. GAYLORD ET AL.

Table 1. Electron microprobe analyses* of glass in Eltopia Canal tephras. Com- positions are compared to Mazama tephra standard from Kelly Site, Fort Rock Valley, Oregon, U.S.A.

Tephras from the Mazama climactic eruption of 6845 BP

Mazama Standard Kelly Site, Fort Eltopia Canal Eltopia Canal Rock Valley, OR Unit 4 Unit 3

SiO2 73)26 (0)19)** 73)02 (0)22) 73)01 (0)33) TiO2 0)42 (0)03) 0)43 (0)04) 0)43 (0)03) Al2O3 14)34 (0)11) 14)41 (0)10) 14)48 (0)11) Fe2O3 2)26 (0)10) 2)09 (0)08) 2)07 (0)14) MgO 0)43 (0)03) 0)45 (0)03) 0)46 (0)03) CaO 1)59 (0)05) 1)60 (0)08) 1)64 (0)05) Na2O4)80 (0)10) 5)05 (0)13) 4)94 (0)18) K2O2)74 (0)06) 2)77 (0)06) 2)76 (0)08) Cl 0)16 (0)04) 0)18 (0)03) 0)21 (0)03) Total 100 100 100 Number of shards analysed 46 19 12 Analytical total 98)77 (1)01) 97)76 (1)00) 96)39 (1)51) Similarity coefficient*** 0)97 0)97

* The tephra samples were dispersed in epoxy, smeared on a glass slide, polished and coated with carbon. The glass was analysed using a Cameca ‘Camebax’ electron microprobe under the following analytical conditions: acceleration voltage"15 Kv, beam diameter"8 m, beam current"11.7 nA, peak and back- ground counting times"10 sec. LiF, PET and TAP crystal spectrometers were used for the wavelength dispersive analyses. Standards used in the glass analyses were obsidian glass CCNM 211 for Na, K, Al, and Si; NBS glass K-411 for Mg and Ca; VG-A99 basaltic glass for Fe; titanite and KCl for Ti and Cl, respectively. ZAF corrections were applied to all analytical data. ** Weight percent oxides with standard deviations of the analyses given in parentheses. Analyses normalized to 100 weight % on a volatile free basis. *** Similarity coefficient calculations (Borchardt et al., 1972) based on unit weighting of the oxides of Si, Al, Fe, Ca, Na, K and 0)25 weighting of the oxides of Ti and Mg. Similarity coefficient reflects the similarity of glass composition in the unknown to that in the Mazama standard with a value of 1)0 indicating a perfect match.

and coarser-grained sands are more basalt-rich (Schatz, 1996; Gaylord et al., 1997). Mean grain sizes generally decrease from medium- to fine-grained sand eastward across the dune field (Schatz, 1996). Provenance was determined primarily by comparing dune sand compositions with those of potential sources including the Miocene(?)/Pliocene Ringold Formation, Quat- ernary alluvium from the Snake and Columbia Rivers, and Quaternary glacial outburst flood sediment (Schatz, 1996; Gaylord et al., 1997). Petrographic analyses of both dense and light minerals have demonstrated that sand-rich outburst flood sediments are compositionally most similar to the dune sands. Both dune sand and outburst flood sand are relatively enriched in clinopyroxene, green amphibole, and basalt lithics and depleted in orthopyroxene and garnet. An outburst flood provenance is further supported by the proximity of dune sands to sand-rich outburst flood deposits, the fining of sand downwind from these deposits, and paleowind orientations inferred from stabilized dune morphologies. RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 409

Figure 3. View towards the north along eastern margin of Smith Canyon dune field. Actively migrating parabolic dune in foreground is surrounded by stabilized dunes. Vegetation character- ized by Artemisia shrub steppe and sparse juniper trees visible in this photo.

Dune migration rates within the Smith Canyon dune field are not well established. However, measurement of a short-term migration rate of a partially stabilized dune from the nearby Juniper dunes (calibrated with tephra from the 1980 Mount Saint Helens eruption) revealed a migration rate of approximately 32 m in 17 years, or slightly less than 2 m year\1.

Dune field chronology and paleoclimatic reconstruction

Dune field depositional history is inferred primarily from evaluation of the stratigraphic record exposed along Eltopia Canal. Additional insight was gained from shallow pit stratigraphies and geomorphic evaluation of stabilized dunes. The 14)5-m-thick Eltopia Canal section reveals a late Pleistocene to Holocene record of glacial outburst flood, fluvial, and aeolian deposition that was punctuated by episodes of pedogenesis, tephra fall, and bioturbation. The composite stratigraphic section shown in Fig. 4 is sub- divided into five environmentally distinctive and climatically sensitive sedimentary units that are discussed below. Ages of the units have been inferred from stratigraphic positions relative to a temporally distinct tephra (from Mount Mazama) and tentative correlation with temporally constrained, published geologic and palynologic records within or near the Columbia Basin.

Unit 1

Unit 1 consists of a 1}2)5-m-thick, medium to light gray brown, massive to crudely horizontally stratified, well-sorted, massive silt, and ripple cross-laminated, very fine sand. Ripple cross-laminations in the upper 0)5 m of the section indicate north-north- west-directed paleocurrents. Sediment in the upper meter of the unit is clay-enriched, 410 D. R. GAYLORD ET AL.

Figure 4. Eltopia Canal composite stratigraphic section. Ranges of ages estimated as described in text. See Fig. 2 for location. Stratification ( ), cross-strata ( ), deformed strata ( ), rhizoliths ( ), Mazama tephra ( ), rip-up clasts ( ). calcite-cemented, more compacted than underlying strata, and contains both delicate filamentous and sturdy, calcite-encrusted rhizoliths (Fig. 5).

Environmental interpretation

The stratigraphic context, massive to crude horizontally stratified, well-sorted silt-rich character and ripple cross-laminations in this unit are consistent with deposition from decelerating outburst flood slackwaters. Late Pleistocene slackwater deposition in the Columbia Basin occurred at elevations up to 335 m above sea level (Myers et al., 1979) or up to 180 m above the elevation of the Eltopia Canal section. The proximity of the Eltopia Canal section to major outburst flood channelways (now occupied by the modern Snake and Columbia Rivers) suggests it is likely that Smith Canyon was inundated regularly during outburst flood events. The persistence of loess deposition in the Columbia River Basin throughout the Pleistocene and Holocene (McDonald & Busacca, 1988, 1992; Busacca & McDonald, RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 411

Figure 5. Variably compacted, clay-enriched, calcium-carbonate cemented, pedogenically alter- ed upper portions of Unit 1 unconformably overlain by low-angle planar cross-stratified fluvial sands of Unit 2. Rhizolith-bearing, pedogenically altered portions of Unit 1 delimited by double arrows. Rip-up clasts near base of Unit 2 were derived from Unit 1. Irregular dark areas in Unit 2 (in upper half of figure) are due to differential water saturation.

1994), suggests that windblown silt was regularly added to these slackwater deposits. Clay enrichment and associated compaction, calcium carbonate cementation, and phytoturbation in the uppermost strata of Unit 1 are attributed to pedogenesis and the influence of stabilizing vegetation. The delicate, filamentous rhizoliths are ascribed to grass roots while the sturdier rhizoliths are attributed to shrub roots.

Unit 2

Unit 2 lies in sharp contact with Unit 1, displaying up to 1)5 m of local relief (Fig. 5). This unit consists of 3)3}4)8 m of light gray brown, generally well-sorted, silt-enriched, horizontal to low-angle planar cross-stratified and locally convolutely deformed, fine to medium sand. Sand-rich strata in the lowermost 20 cm of this unit contain elongate, subangular to subrounded clasts of internally stratified and compacted, pedogenically- altered silt from Unit 1. The horizontal strata and planar cross-strata in this unit are thickly laminated to thinly bedded and cut by 1}3-cm-deep scours. A few 0)5}1)0-cm- thick, inversely graded laminations and planar cross-laminated sands are interspersed with massive to normally graded strata and cross-strata in the middle third of the unit, while a single 10}15-cm-thick bed of convolutely deformed and asymmetrically-folded strata occurs in the upper third of the unit (Fig. 6). Ripple cross-laminations and asymmetrically deformed flame structures indicate north-north-west-directed paleo- currents.

Environmental interpretation

The massive-to-normally grading, planar-stratification and cross-stratification, and in- corporation of silt rip-up clasts from Unit 1 suggests extensive fluvial reworking within this unit. Consistent with this interpretation are the convolutely deformed strata that most likely resulted from loading and/or shearing of saturated sediment. The inversely graded laminations are attributed to migration of aeolian ripples (see Hunter, 1977). 412 D. R. GAYLORD ET AL.

Figure 6. Convolutely deformed, very fine- to medium-grained and well-sorted fluvial sand from Unit 2. Low-angle, planar cross-stratified, mixed fluvial and aeolian sand near top of photo truncates these deformed strata.

The north-north-west-orientation of paleocurrents are opposite to modern Smith Can- yon drainage directions. These paleocurrents are attributed to the temporary blockage of Smith Canyon downstream from where the sections were measured. Plausible explana- tions include blockage and redirection of flow via outburst flood sediment or by an aeolian dune dam. Thick deposits of fine-grained and relatively impermeable slackwater sediment are common within the central and southern Columbia Basin. At the same time as demonstrated in the Nebraska Sand Hills (Loope et al., 1995; Mason et al., 1997) dune dams are also potentially major influences on geomorphic evolution. Whatever the case, the blockage has since been eliminated by fluvial incision and/or aeolian deflation.

Unit 3

Unit 3 consists of a 1)4}1)7-m-thick succession of light gray-brown, moderately well- sorted, low- to high-angle planar cross-stratified and horizontally stratified, fine to medium sand with gravel-sized tephra clasts in the lower third of the unit (Fig. 7). The tephra clasts are angular to subangular, internally massive to stratified, unweathered, and up to 8 cm wide by 12 cm long. Sparsely distributed, very fine- to medium-sized sand and pebble-sized tephra clasts also occur in the sand-rich, upper two-thirds of the unit. Sands occur in centimeters to decimeters thick tabular- and wedge-planar sets contain- ing laminated to thinly bedded and ripple cross-laminated sediment (Fig. 8). Locally, coarse-grained sand and granules are concentrated in a few, 1-cm-thick lenses and along bounding surfaces. Foreset orientations indicate paleocurrents were directed towards the north-north-west. Compositions of the volcanic glass from Unit 3 (and Unit 4) tephra clasts are presented in Table 1. Volcanic glass compositions from this tephra and tephra in the overlying unit are essentially identical (similarity coefficient"0)99; Borchardt et al., RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 413

Figure 7. Mixture of unweathered, stratified to massive Mazama tephra clasts in a medium to very coarse-grained, flood to flash flood-generated, fluvial deposit at the base of Unit 3.

1972). These glass compositions also closely match the Mazama climactic eruption standard (similarity coefficient"0)97). The climactic Mount Mazama eruption that resulted in present-day Crater Lake, Oregon (Fig. 1) occurred c. 6845 14C years B.P. (Bacon, 1983; Sarna-Wojcicki, 1991; Zdanowicz et al., 1999), when roughly

Figure 8. Arrow points to ripple cross-laminated fine- to medium-grained sand (paleoflow to the north-north-west) that incorporated detrital, sand-sized Mazama tephra clasts. These ripple cross-laminations occur in the upper part of Unit 3. 414 D. R. GAYLORD ET AL.

50 km3 of rhyodacitic magma was erupted (Bacon, 1983; Nelson et al., 1988; Suzuki-Kamata et al., 1993; Zdanowicz et al., 1999). Tephra from this eruption mantled an area of approximately 1)7;106 km2 (Zdanowicz et al., 1999) making it one of the most widely distributed Quaternary tephras in (Sarna- Wojcicki et al., 1983). Particulate fallout from this eruption was so voluminous that it continued for 2}3 years following the eruption (Mehringer et al., 1977; Zdanowicz et al., 1999) and sulfate aerosols remained in the atmosphere for up to 6 years (Zdanowicz et al., 1999).

Environmental interpretation

The scoured and incised features near the base, nature of the stratified and cross- stratified sand, and gravel-sized tephra clasts in Unit 3 strongly argue for fluvial deposition. Fluvial activity varied from flood or possibly flash flood conditions near the base of the unit to lower competence, normal stream flows near the top. The large unweathered and internally stratified tephra clasts suggest the tephra had resided on the ground for only a short time before being incorporated into the stream. The dominantly angular and subangular shapes of these soft, gravel-sized tephra clasts indicate travel distances were short. Zdanowicz et al. (1999), citing relatively elevated sulfate aerosol concentrations, suggested that the Mazama eruption could have triggered a 0)6}0)73C decrease in temperatures in the mid- and northern latitudes for 1}3 years following the eruption. It is possible that such a temperature decrease could have contributed to a heightened potential for flooding and/or flash-flooding; however, without additional corroborating evidence this scenario is conjectural.

Unit 4

Unit 4 consists of an approximately 8-m-thick succession of light gray brown, wedge- shaped sets of high-angle, planar cross-stratified, moderately to moderately well-sorted, fine- to coarse-grained sand. Individual cross-strata are from 0)5}3-cm-thick and consist of both massive and inversely-graded sediment that has been disturbed by burrowing. Inversely-graded cross-strata tend to pinch out down dip and are interstratified with cross-beds of massive, fine-grained sand. High-angle foresets dip at angles of 25}343 towards N50E to N120E. The high-angle cross-strata in the lower third of the unit are distinct, merge tangentially down dip and are inter-stratified with thinly laminated and inversely graded, low-angle cross-strata. In contrast, high-angle cross-strata in the upper two-thirds of the unit are less distinct. The cross-stratified sands of Unit 4 are highlighted by a single, 28}45-cm-thick, tephra-rich (Table 1) foreset bed that appears in the lower half of the section (Fig. 9). Mazama tephra within this distinctive foreset appears as large, intact granule to boulder-sized clasts and as irregularly mixed bands of sand and tephra. These intact tephra clasts are inversely graded, with large clasts concentrated near the top and toe of the bed. Large tephra clasts also are concentrated further down dip. The irregular bands of mixed sand and tephra are centimeters thick and either parallel the dips of foreset beds or are aligned horizontally (Fig. 10). Isolated, irregular, subangular to subrounded, gravel-sized concentrations of tephra appear within the uppermost cross-strata. The tephra-clast-bearing foreset bed as well as the sand-rich cross-strata 10}20 cm above and below it are cut by numerous cylindrical, backfilled, 0)75}2-cm-diameter burrows (Fig. 10). Backfill structures within the burrows have meniscus shapes that curve towards the burrow sides. Oblique transects through burrow cylinders can be traced for 5}10 cm. Larger, but less common, 4}6-cm-diameter burrows with massive infills were observed in cross-strata bordering the tephra-clast foreset bed. RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 415

Figure 9. Prominent, cataclastically emplaced tephra clasts within Unit 4 sand dune foreset bed that dips to the east-south-east. Note that the largest intact clasts occur at or near the top, as well as near the toe of the foreset bed, characteristic features in such grainflows. Foreset bed attitude and geometric relations indicate the dune was at least 5 m high.

Environmental interpretation

The textural character, grading, and geometry of the high-angle foresets indicate Unit 4 accumulated as dune sand under the influence of east, east-north-east, and occasionally east-south-east winds. The high-angle cross-strata that pinch out downdip are inter- preted as grainflow deposits while the intervening massive sands represent grainfall deposits. The thin, inversely graded laminations at the base of the section are interpreted as ripple cross-strata. Inverse grading of the large tephra clasts within the foreset as well as the concentration of large tephra clasts down slope are consistent with the transport and depositional behavior of particles in cataclastically deforming grainflows (Hunter, 1977; Vallance, 1994). The appearance of essentially unaltered Mazama clasts in Unit 4 as well as in the underlying deposits of Unit 3 presents a stratigraphic problem. How did the seemingly unreworked tephra clasts of Unit 3 reappear in Unit 4? One explanation favored here is that the Mazama tephra accumulated on a topographically irregular surface. Tephra from a lower level was more readily available for reworking and redeposition (as occurred in Unit 3). Tephra stored in elevated positions remained there until undercut by normal erosive processes. Subsequently, as was the case for the Unit 4 foreset bed, the tephra bed first collapsed and then flowed cataclastically onto the slipface of an adjacent, advancing sand dune. The unweathered, internally massive to stratified character of the clasts suggests that deposition of the tephra clasts for both units occurred shortly after the Mazama tephra first accumulated. Insect burrows in dune sands are a potentially valuable source of paleoecologic information, though they often are unrecognized or poorly understood (Ahlbrandt et al., 1978). The sizes, shapes, and infilling character of the 0)75}2-cm-diameter cylindrical burrows concentrated around the tephra-rich foreset bed are identical to those construc- ted by cicada nymphs in nearby exposures of the late Pleistocene Washtucna Soil 416 D. R. GAYLORD ET AL.

Figure 10. Close-up of tephra-rich Unit 4 foreset bed showing cicada burrows (note meniscus backfilling and cylindrical shapes), deformed and intact tephra clasts. Larger burrows attributed to rodents.

(O’Geen, 1998; O’Geen & Busacca, pers. comm. 1999). Cicada burrows are especially useful to paleoecological reconstructions because of the affinity cicada nymphs have for the roots of shrubs and particularly Artemisia. O’Geen (1998), who studied the burrowing and feeding habits of modern cicada nymphs, found an overwhelming preference of the nymphs (19:1) for sagebrush as compared to bunchgrass steppe. At Eltopia Canal, Artemisia roots likely reached the tephra-clast foreset bed because of the tephra’s tendency to retain moisture. The depth of the tephra clasts beneath the surface is inferred to have been on the order of one meter, the typical maximum depth to which cicada nymphs burrow in the northern Great Basin (Hugie & Passey, 1963). Sub- sequently, this partially vegetated surface was buried beneath advancing dune sands. The relatively large, 4}6-cm-diameter, and massively infilled burrows concentrated beneath the tephra-clast foreset reflect burrowing by rodents. The generally less distinct nature of dune cross-strata in the upper few meters of Unit 4 is attributed to the syn- or post-depositional influence of vegetation on the dune sands, a process that often leads to either partial or complete erasure of the original sedimentary structures.

Unit 5

Unit 5 consists of 1}1)5 m of dominantly massive to crudely stratified, fine to coarse sand, with occasional granules and pebbles, admixed organic matter, and relatively minor silt and clay (Fig. 11). The unit grades upwards from light gray brown at the base to a medium gray brown near its top. Modern rootlets and the traces of burrowing insects and rodents interrupt the stratified sediment. Plants stabilizing the surface lie within the Artemisia tridentata-Agropyron vegetation zone (Daubenmire, 1970).

Environmental interpretation

The poor sorting, phyto- and bioturbation, massive to crudely stratified character, relatively high concentration of organic matter, and subtle horizon development are RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 417

Figure 11. Composite view of upper three units of the Eltopia Canal section including the pedogenically altered dune strata of Unit 5, the tephra-bearing, high-angle cross-strata of Unit 4 (tephra-rich foreset depicted with lightly dashed lines), and planar-stratified fluvial strata of Unit 3. evidence for early soil formation. Soil forming processes on sandy substrate in the semi-arid to arid Columbia Basin are generally quite slow and, based on modern observations, may take thousands of years to develop recognizable pedogenic features (Busacca, pers. comm., 1999).

Discussion

Paleoclimatic implications and estimated temporal ranges for Eltopia Canal stratigraphic units

Reconstructing the depositional history and paleoclimates of the Smith Canyon dune field highlights the difficulties often encountered when working with deposits from persistently semi-arid or arid settings. Nevertheless, the available data provides useful clues about the scope of paleoclimatic changes that affected the southeastern Columbia Basin during the late Pleistocene and Holocene. Overall, the paleoclimatic record at Smith Canyon is one of moderating, late Pleistocene to early Holocene post-glacial conditions that were punctuated by variable episodes of decreased and increased moisture. Details of the sedimentary and climatic changes are summarized in Fig. 12, a chart that correlates the Eltopia Canal sedimentary record with better constrained (chronologically) paleoclimate reconstructions from fluvial deposits exposed at Mahoney Lake, British Columbia (Lowe et al., 1997) and the northern Columbia Basin (Chatters & Hoover, 1992).

Unit 1—late Pleistocene to early Holocene

Stratigraphic position, sedimentary character and paleocurrents suggest that the bulk of Unit 1 accumulated between c.15)5 and 12)5 ka, the approximate duration of the last glacial outburst floods (Baker et al., 1991). Regional vegetation histories (Barnosky et al., 418 D. R. GAYLORD ET AL.

Figure 12. Sedimentary-paleoclimatic-chronologic summary chart for the Smith Canyon area comparing the Eltopia Canal stratigraphic section with paleoclimatic reconstructions from Mahoney Lake (Lowe et al., 1997) and the northern Columbia Basin (Chatters & Hoover, 1992).

1987; Thompson et al., 1993) suggest that the climate in the Columbia Basin was favorable for pedogenesis from the end of the last full-glacial until the early Holocene. Palynologic evidence from Carp Lake, Washington indicates that relatively warm, dry conditions prevailed from c. 13)2}9)1 ka, as grasslands expanded within the Columbia Basin (Whitlock & Bartlein, 1997). The prevalence of grasses signalled a change from the Artemisia (sagebrush)-dominated steppe that had characterized the previous perig- lacial climate (Thompson et al., 1993). Grassland expansion apparently was interrupted by increasingly severe episodes of early and middle Holocene drought that promoted a mixed, grass-shrub steppe (Barnosky et al., 1987). Paleoclimate evidence from alluvial strata in the northern Columbia Basin, coupled with paleoclimate data from other sources, suggests that from c. 10}9 ka, cold winters, hot summers, and spring-dominant precipitation prevailed (Chatters & Hoover, 1992). These authors also suggested that warm winters, hot summers and winter-dominated precipitation (that included numer- ous rain on snow-induced flood events) were common from c.9}8 ka. Sedimentary data from a shallow a meromictic lake in southern British Columbia (situated at a similar elevation as Smith Canyon) (Fig. 1) reveal that episodes of low effective precipita- tion (low net annual precipitation) prevailed from c.10}9)2 ka and c.8)3}7)9 ka (Lowe et al., 1997). The calcite-encrusted, rhizoliths, calcite cementation, translocated clays in the upper- most Unit 1 strata, as well as data from other paleoclimate records suggest that a relatively warm and occasionally dry climate promoted grass growth and widespread RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 419 pedogenesis for thousands of years following the final outburst flood. The character of the rhizoliths as well as the pollen record from this time (Barnosky et al., 1987) indicate that a mixed shrub steppe that likely included Artemisia probably played a role in anchoring the outburst flood deposits. Overall, in the context of the Smith Canyon and Eltopia Canal geology and other regional reconstructions, it appears pedogenesis in- fluenced the Unit 1 deposits until c. 8 ka.

Unit 2—early Holocene

Placing limits on the timing of Unit 2 deposition is, like Unit 1, troublesome because of the absence of chronostratigraphic markers. Nevertheless, stratigraphic position and lithologic content indicate that Unit 2 deposition began after the Unit 1 soil was breached and apparently after c. 8 ka. The fluvial deposits of Unit 2 may have taken from years to centuries to accumulate but, in any case, their accumulation had ended by c.6)9 ka, when the Mazama tephra fell on this area. The mixed fluvial-aeolian character of Unit 2 deposition is consistent with paleocli- mate reconstructions from the northern Columbia Basin, where seasonally variable precipitation under relatively warm conditions prevailed after 9 ka and prior to the Mazama eruptive event (Chatters & Hoover, 1992). Notably, at Mahoney Lake, British Columbia, episodes of ‘high effective precipitation’ (i.e. relatively moist conditions when precipitation greatly exceeds evaporation) occurred between c.7)9 and 6)9 ka and episodes of ‘low effective precipitation’ (i.e. relatively arid, evaporitic conditions) prevailed between c.8)3 and 7)9 ka, and for (100 years preceding the Mazama eruption (Lowe et al., 1997). In sum, it appears these fluctuating climatic conditions influenced deposition and alteration of Unit 2 from c.8}6)9 ka.

Unit 3—early}middle Holocene

Regional paleoclimate records indicate that by c.6)9}6)8 ka portions of the Columbia Basin had started to experience more prolonged episodes of warmth and aridity. Channel incision and slow floodplain aggradation in the northern Columbia Basin led Chatters & Hoover (1992) to propose that the climate from c. 6.9 ka until nearly c.4)9 ka was characterized by hot summers, warm winters, and low levels of winter-dominant precipitation. Palynological evidence indicates that between c. 8 ka and 6)9kaArtemisia had become more prominent within the Columbia Basin (Barnosky et al., 1987). By c. 6)9 ka the climate at Mahoney Lake, British Columbia was about to enter into the first of a series of centuries-long episodes of low effective precipitation—after nearly 1000 years of high effective precipitation (Lowe et al., 1997). In light of the evidence for the beginnings of heightened regional drought at the time of the Mazama climactic eruption, and considering the long-term tendency towards aridity in the Smith Canyon area, it may seem at first unusual that the initial stratigraphic evidence of the eruption appeared in water-, as opposed to wind-, dominated sedimentary deposits. One possibility is that the flood/flash flood and the tephra fall were coincidental. Or, as suggested by Gaylord et al. (1997), perhaps the eruption and flood/flash flood were associated. And, if the eruption did decrease temperatures from 0)6to0)73Cinthemid and northern latitudes as Zdanowicz et al. (1999) contend, it seems reasonable that regional weather patterns between c.6)9and6)8kamighthavebeenperturbed.Additional evidence for such climatic perturbations associated with the Mazama eruption may exist elsewhere in the region’s geologic record, but have not yet been identified.

Unit 4—Middle Holocene The depositional history of Unit 4 extends from a time following the Mazama climactic eruption at c.6)8 ka until c.3)9 ka, or roughly within the interval of the middle Holocene 420 D. R. GAYLORD ET AL. maximum aridity that characterized many parts of the northwestern and western portions of the conterminous U.S. (Antevs, 1948; Deevey & Flint, 1957; Davis et al., 1986). Establishing when the interval of maximum Holocene aridity in the Columbia Basin ended is as elusive here as it is throughout much of North America, probably because of spatial and temporal variations in the climate of the time (Thompson et al., 1993; Mock & Bartlein, 1995). Relatively warm and dry conditions apparently prevailed at Carp Lake until c.3)9 ka and continued to support a mixed grass-shrub steppe (Whitlock & Bartlein, 1997); however, extended episodes of low to very low effec- tive precipitation at Mahoney Lake in British Columbia had apparently ended by c.4)9 ka (Lowe et al., 1997). Chatters & Hoover (1992) inferred that warm winters, hot summers, and winter dominant, but very low levels of precipitation dominated the northern Columbia Basin climate until c.4)4 ka. We suspect, based primarily on these other paleoclimate records, that a persistently dry climate promoted a relatively high level of dune activity until some time between c.4)9 and 3)9 ka.

Unit 5—late Holocene

Climatic conditions and vegetative cover in the Columbia Basin since the end of the Altithermal have tended to vary. Chatters & Hoover (1992) inferred that the climate in the northern Columbia Basin was generally cooler from c.4)4}2)4 ka, conditions that they also associated with increasingly dense steppe vegetation. They also surmised that as the climate warmed after c.2)4 ka, vegetation density decreased and rain on snow flooding increased. Lowe et al. (1997) suggested that the climate on the lowlands surrounding Mahoney Lake from c.4)9}3)3 ka was marked by moderate to high effective precipitation (i.e. relatively elevated precipitation to evaporation). These conditions were followed by a prolonged episode of relatively low effective precipi- tation that lasted until c. 1)5 ka, when the climate fluctuated between century-long episodes of high and low effective precipitation. Pollen from Carp Lake indicates that in general, from c.3)9 until present, the climate has promoted growth of the modern forest and mixed shrub steppe (Whitlock & Bartlein, 1997). Overall, the effect of having the basin vegetation stressed by episodes of drought during the past few thousands of years has prompted at least occasional sand dune reactivation. The widespread impact of humans on this area since the 1860s has added to the potential for destruction of stabilizing vegetation and dune reactivation.

Summary and Conclusions

The character of deposits preserved in the Smith Canyon dune field provides insight into aspects of aeolian and fluvial activity and climate in the southeastern Columbia Basin during the late Pleistocene and Holocene. Smith Canyon dune field development dates from the late Pleistocene or, at latest, early Holocene when episodes of drought and/or fluvial incision promoted disturbance and/or removal of vegetation and soil that had previously anchored outburst flood sediment. Palynological studies at Carp Lake in the south-western Columbia Basin strongly suggest that a relatively warm and moist post-glacial climate prompted forest expansion to lower elevations as well as expansion of increasingly dense shrub-steppe vegetation into basin areas (Barnosky, 1985; Barnosky et al., 1987; Whitlock & Bartlein, 1997). Between c.15)5 and 8 ka Unit 1 outburst flood deposits became stabilized and subject to pedogenesis. During relatively dry and windy episodes, saltating aeolian sand transported across poorly vegetated surfaces probably encouraged loess generation and buildup. Fluvial reworking of Unit 1 during accumulation of Unit 2 between c. 8 and 6)9ka coincided with a variably moist early Holocene climate elsewhere in the region (Lowe et al., 1997). Seasonal precipitation patterns (Chatters & Hoover, 1992) promoted RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 421

fluvial incision of surface sediments and short episodes of enhanced sand dune and loessial activity. Southerly-directed, pre-outburst flood drainage within Smith Canyon was blocked by outburst flood deposits and/or aeolian dune dam(s), causing streams within Smith Canyon to flow northward. Prompted by enhanced moisture and relatively moderate temperatures, surface vegetation expanded and increased in density in the Smith Canyon area. Unit 2 deposition ended prior to the c.6)9 ka climactic eruption of Mount Mazama. The eruption of Mount Mazama ejected into the atmosphere an enormous volume of tephra as well as a sufficient concentration of sulfate aerosols to influence the mid and northern latitude climate (Zdanowicz et al.,1999.Webelievesuchaclimaticperturbation could have produced flood, or possibly, flash flood events in Smith Canyon. Such rapid remobilization of the tephra clasts accounts for their unweathered, internally stratified and massive character within Unit 3 deposits (c.6)9}6)8ka).Thesurroundingvegetationat this time had already started to respond to increasingly warm and arid conditions that encouraged an expansion of shrub vegetation and a decreased density of grassland species in the Columbia Basin (Barnosky et al., 1987; Whitlock & Bartlein, 1997). The most dramatic expansion and mobilization of the Smith Canyon sand dunes is preserved in Unit 4 and occurred following the Mazama climactic eruption (c.6)8 ka). Migrating active dunes, evidenced by the high angle-foresets, were influenced, at least in part, by Artemisia shrub steppe vegetation whose root systems supported a cicada nymph population. Episodic transportation and deposition of dune sands led to con- tinued downwind (to the north-east and east) expansion of the dune field and likely enhanced loessial accumulation further downwind. The episodic nature of dune activity in the Columbia Basin may mirror the changes in effective precipitation docu- mented in southern British Columbia (Lowe et al., 1997). Ultimately, dune field activity was slowed by expansion of basin vegetation by c.3)9 ka, a stabilizing episode that was encouraged by generally cooler, moister conditions (Chatters & Hoover, 1992; Lowe et al., 1997; Whitlock & Bartlein, 1997). The final stages of the Smith Canyon dune field development are recorded in the modern soil of Unit 5. This soil probably developed during the past c.3)9 ka as climatic conditions in and near the Columbia Basin fluctuated between high and low levels of effective precipitation (Lowe et al., 1997), cold to warm winters, and cool to warm summers (Chatters & Hoover, 1992). The vegetation at this time also shifted towards a modern shrub steppe (Whitlock & Bartlein, 1997). Local reactivation of dunes by natural and anthropogenic causes has continued to promote aeolian activity in this area. However, understanding of aeolian-climatic thresholds that could prompt widespread sand dune reactivation in the Columbia Basin (Gaylord & Stetler, 1994; Stetler & Gaylord, 1996) remains elusive. As revealed from the generalized paleoclimatic reconstruction presented in this paper, there still is a compelling need for additional chronostratigraphic and paleoclimatic detail. Fortunately, the volume of preserved aeolian and fluvial sediment that remains in the Columbia Basin increases the probability that additional data will emerge to supple- ment and refine this reconstruction.

The authors wish to thank W. Neff, the Juniper Ranch, and K. Peterson for permitting access to their property. We also appreciated the helpful reviewer comments of M. Sweeney and useful discussions with A.J. Busacca and T. O’Geen. K.L. Petersen generously donated his time and archeologic expertise to help us gain essential permits to conduct parts of this study. Financial assistance was provided by a Meyers Grant and the Department of Geology, Washington State University.

References

Ahlbrandt, T.S., Andrews, S. & Gwynne, D.T. (1978). Bioturbation in aeolian deposits. Journal of Sedimentary Petrology, 48: 839}848. 422 D. R. GAYLORD ET AL.

Antevs, E.V. (1948). The Great Basin, with emphasis on glacial and postglacial times. Bulletin of the University of Utah, 38: 168}191. Atwater, B.F. (1984). Periodic floods from Glacial into the Sanpoil arm of Glacial Lake Columbia, northeastern Washington. Geology, 12: 464}467. Atwater, B.F. (1987). Status of glacial Lake Columbia during the last floods from glacial Lake Missoula. Quaternary Research, 27: 182}201. Bacon, C.R. (1983). Eruptive history of Mount Mazama and Crater Lake caldera, Cascade Range, USA. Journal of Volcanology and Geothermal Research, 18: 57}115. Baker, V.R., Bjornstad, B.N., Busacca, A.J., Fecht, K.R., Kiver, E.P., Moody, U.L., Rigby, J.G., Stradling, D.F. & Tallman, A.M. (1991). Quaternary geology of the Columbia Plateau, In: Morrison, R.B. (Ed.), Quaternary Nonglacial Geology, pp. 215}250. Boulder CO: Geological Society of America, The Geology of North America, 672 pp. Barnosky, C.W. (1985). Late Quaternary vegetation in the southwestern Columbia Basin, Washington. Quaternary Research, 23: 109}122. Barnosky, C.W., Anderson, P.M. & Bartlein, P.J. (1987). The northwestern U.S. during deglaci- ation: vegetational history and paleoclimatic implications. In: Ruddiman, W.F. & Wright, H.E., Jr., (Eds), North America and adjacent oceans during the last , pp. 289}321. Boulder, CO: Geological Society of America, 501 pp. Borchardt, G.A., Aruscavage, P.J., & Millard, Jr., H.T. (1972). Correlation of the Bishop Ash, a Pleistocene marker bed, using instrumental neutron activation analysis, Journal of Sedimentary Petrology, 42: 301}306. Busacca, A.J. (Ed.). (1998). Dust, aerosols, loess soils, and global change. Conference Proceedings, 11}14 October 1998, Seattle,. Seattle, WA, U.S.A: Washington State University, College of Agriculture and Home Economics Miscellaneous Publication No. MISC0190, Pullman, WA. 234 pp. Busacca, A.J. & McDonald, E.V. (1994). Regional sedimentation of late Quaternary loess on the Columbia Plateau: sediment source areas and loess distribution patterns. Washington Division of Geology and Earth Resources Bulletin, 80: 181}190. Bretz, J.H. (1969). The Lake Missoula floods and the Channeled Scabland. Journal of Geology, 77: 505}543. Chatters, J.C. & Hoover, K.A. (1992). Response of the Columbia River fluvial system to Holocene climatic change. Quaternary Research, 37: 42}59. Daubenmire, R. (1970). Steppe vegetation of Washington. College of Agriculture and Home Economics, Washington State University, Pullman, WA. Washington Agricultural Experiment Station Bulletin, 62:1}131. Davis, O.K., Sheppard, J.C. & Robertson, S. (1986). Contrasting climatic histories for the Snake River Plain, Idaho, resulting from multiple thermal maxima. Quaternary Research, 26: 321}339. Deevey, E.S. & Flint, R.F. (1957). Postglacial hypsithermal interval. Science, 125: 182}184. Folk, R.L. (1980). Petrology of Sedimentary Rocks. Austin: Hemphill Publishing, 182 pp. Galehouse, J.S. (1971). Point counting. In: Carver, R.E. (Ed.), Procedures in Sedimentary Petrol- ogy, pp. 385}408 New York: Wiley-Interscience, 653 pp. Gaylord, D.R. & Stetler, L.D. (1994). Aeolian-climatic thresholds and sand dunes at the Hanford Site, south-central Washington, U.S.A. Journal of Arid Environments, 28: 95}116. Gaylord, D.R., Schatz, J.K., Coleman, A.J., & Foit, F.F., Jr. (1997) Holocene to recent sedimen- tary, stratigraphic, and geomorphic development of the Juniper dune field, Abstracts with Programs), Geological Society of America Annual Meeting (Salt Lake City, Utah) 20}24 October 1997, 29:A}218. Gaylord, D.R., Foit, F.F., Schatz, J.K. & Coleman, A.J. (1998). Sand dune activity within sensitive agricultural lands, Columbia Plateau, USA. In: Busacca, A.J. (Ed.), Dust Aerosols, Loess Soils and Global Change, pp. 145}148. Washington State University College of Agriculture and Home Economics. Miscellaneous Publication No. MISC0190, Pullman, WA. 234 pp. Gulick, C.W. (1994). Geologic map of the Connell 1:1000,000 Quadrangle, Washington: Washington Division of Geology and Earth Resources Open File Report 94}14, 18 pp. 1 plate. Hugie, V.K. & Passey, H.B. (1963). Cicadas and their effect upon soil genesis in certain soils in southern Idaho, northern Utah, and northeastern Nevada. Soil Science Society of America, 27:78}82. Hunter, R.E. (1977). Basic types of stratification in small aeolian dunes. Sedimentology, 24: 361}387. Kutzbach, J.E. & Guetter, P.J. (1986). The influence of changing orbital patterns and surface boundary conditions on climate simulations for the past 18,000 years. Journal of Atmospheric Science, 43: 1726}1759. RELATION TO FLOODS, ERUPTION, AND PALEOCLIMATE 423

Kutzbach J.E. & Webb, III., T. (1993). Conceptual basis for understanding late-Quaternary climates. In: Wright, H.E. Jr., Kutzbach, J.E., Webb,T. III, Ruddiman, W.F., Street- Perrott, F.A. & Bartlein, P.J. (Eds), Global Climates since the last glacial maximum, pp. 5}11. Minneapolis: University of Minnesota Press. 569 pp. Lindsey, K.A. & Gaylord, D.R. (1990). Lithofacies and sedimentology of the Miocene- Pliocene Ringold Formation, Hanford Site, south-central Washington. Northwest Science, 64: 165}180. Little, E.L., Jr. (1971). Atlas of United States trees, conifers and important hardwoods. U.S. Depart- ment of Agriculture. Miscellaneous Publication Number 1146, 1:1}9 Liverman, J.L. (1975). Final environmental statement, waste management operations, Hanford Reservation, Richland, Washington. ERDA-1538, Vol. 1 and 2, Washington, D.C: Energy Research and Development Administration. Loope, D.B., Swinehart, J.B. & Mason, J.P. (1995). Dune-dammed paleovalleys of the Nebraska Sand Hills: intrinsic versus climatic controls on the accumulation of lake and marsh sediments. Geological Society of America Bulletin, 107: 396}406. Lowe, D.J., Green, J.D., Northcote, T.G. & Hall, K.J. (1997). Holocene fluctuations of a meromictic lake in southern British Columbia. Quaternary Research, 48: 100}113. McDonald, E.V. & Busacca, A.J. (1988). Record of pre-late Wisconsin giant floods in the Channeled Scabland interpreted from loess deposits. Geology, 16: 728}731. McDonald, E.V. & Busacca, A.J. (1992). Late Quaternary stratigraphy of loess in the Channeled Scabland and regions of Washington State. Quaternary Research, 38: 141}156. Mason, J.P., Swinehart, J.B. & Loope, D.B. (1997). Holocene history of lacustrine and marsh sediments in a dune-blocked drainage, southwestern Nebraska Sand Hills, U.S.A. Journal of Paleolimnology, 17: 67}83. Mehringer, P.J., Jr., Blinman, E. & Petersen, K.L. (1977). Pollen influx and volcanic ash. Science, 198: 257}261. Mock, C.J. & Bartlein, P.J. (1995). Spatial variability of Late-Quaternary paleoclimates in the western United States. Quaternary Research, 44: 425}433. Myers, D.W., Price, S.M., Caggiano, J.A., Cochran, M.P. Czimer, W.H., Davidson, N.J., Edwards, R.C., Fecht, K.R., Holmes, G.E., Jones, M.G., Kunk, J.R., Landon, R.D., Ledger- wood, R.K., Lillie, J.T., Long, P.E., Mitchell, T.H., Price, E.H., Reidel, S.P. & Tallman, A.M. (1979). Geologic studies of the Columbia Plateau: a status report. Rockwell Hanford Operations Report RHO-BWI-ST-4. Nelson, C.H., Carlson, P.R. & Bacon, C.R. (1988). The Mount Mazama climatic eruption (&6900 year B.P.) and resulting convulsive sedimentation on the Crater Lake caldera floor, continent, and ocean basin. Geological Society of America Special Paper, 229:37}57. O’Geen, T.A. (1998). Faunal paleocology of late Quaternary paleosols on the Columbia Plateau. M.Sc. thesis. Washington State University, Pullman, WA. 96 pp. Petrone, A. (1970). The Moses Lake sand dunes. M.Sc. thesis. Washington State University, Pullman, WA. 89 pp. Reidel, S.P. & Fecht, K.R. (1994). Geologic map of the Richland 1:100,000 Quadrangle, Washington: Washington Division of Geology and Earth Resources Open-File Report 94}8, 21 pp., 1 plate. Reidel, S.P., Campbell, N.P., Fecht, K.R. & Lindsey, K.A. (1994). Late Cenozoic structure and stratigraphy of south-central Washington. Washington Division of Geology and Earth Resources Bulletin, 80: 159}180. Sarna-Wojcicki, A.M., Champion, D.E. & Davis, J.O. (1983). Holocene volcanism in the conterminous United States and the role of silicic volcanic ash layers in correlation of latest- Pleistocene and Holocene deposits. In: Wright, H.E., Jr. (Ed.), Late Quaternary Environments of the United States, the Holocene, Vol. 2, pp. 52}77. Minneapolis: University of Minnesota Press. 295 pp. Sarna-Wojcicki, A.M., Lajoie, K.R., Meyer, C.E., Adam, D.P. & Rieck, H.J. (1991). Tephroch- ronologic correlation of upper Neogene sediments along the Pacific margin, conterminous United States. In: Morrison, R.B. (Ed.), Quaternary nonglacial geology, pp. 117}140. Con- terminous U.S.: Boulder CO: Geological Society of America, The Geology of North America, K}2. Schatz, J.K. (1996). Provenance and depositional histories of the Smith Canyon and Sand Hills Coulee dunes, south-central, Washington. M.Sc. thesis. Washington State University, Pullman, WA. 87 pp. 424 D. R. GAYLORD ET AL.

Schuster, J.E. (1994). Geologic map of the Walla Walla 1:100,000 Quadrangle. Washington: Washington Division of Geology and Earth Resources Open File Report 94}3. 18 pp. 1 plate. Stetler, L.D. & Gaylord, D.R. (1996). Evaluating aeolian-climatic interactions using a regional climate model from Hanford, Washington (USA). , 17: 99}113. Suzuki-Kamata, K. & Kamata, H. (1993). Evolution of the caldera-forming eruption at Crater Lake, Oregon, indicated by component analysis of lithic fragments. Journal of Geophysical Research, 98: 14059}14074. Thompson, R.E., Whitlock, C., Bartlein, P.J., Harrison, S. & Spaulding, W.G. (1993). Climatic changes in the western United States since 18,000 yr B.P. In: Wright, H.E. Jr., Kutzbach, J.E., Webb,T. III, Ruddiman, W.F., Street-Perrott, F.A. & Bartlein, P.J., Global Climates since the Last Glacial Maximum, p. 468}513. Minneapolis: University of Minnesota Press. 569 pp. Vallance, J.W. (1994). Experimental and field studies related to the behavior of granular mass flows and the characteristics of their deposits. Ph.D. thesis, Michigan Technological University, Houghton, Michigan. 197 pp. van der Plas, L. & Tobi, A.C. (1965). A chart for judging the reliability of point counting results. American Journal of Science, 263:87}90. Waitt, R.B. (1980). About 40 last-glacial Lake Missoula joK kulhlaups through southern Washing- ton. Journal of Geology, 88: 653}679. Waitt, R.B. (1984). Periodic joK kulhlaups from Pleistocene Glacial Lake Missoula}New evidence from varved sediment in northern Idaho and Washington. Quaternary Research, 22: 46}58. Waitt, R.B. (1985). Case for periodic colossal joK kulhlaups from Pleistocene glacial Lake Missoula. Geological Society of America Bulletin, 96: 1271}1286. Whitlock, C. & Bartlein, P.J. (1997). Vegetation and climate change in northwest America during the past 125 kyr. Nature, 388:57}60. Zdanowicz, C.M., Zielinski, G.A. & Germani, M.S. (1999). Mount Mazama eruption: Calendri- cal age verified and atmospheric impact assessed. Geology, 27: 621}624.