Paleofloods on the Hells Canyon Reach of the Snake River, Idaho/Oregon
by
Rhodes, Gwendolyn
Department of Geology, University of Maryland, College Park 20742

The Snake River is a significant water resource in the northwestern United States. The basin is approximately 282,000 km2 in area and the river is the largest tributary to the Columbia River annually discharging an average of 4.54 x 109 m3 of water into the Columbia River. It is therefore, important to understand the frequency of flooding in the region. Several ancient flood terraces line the walls of the Hells Canyon reach of the Snake River and preserve a long-term record of extreme floods on the river. The terraces have been examined for evidence of pre-historical human occupation and higher terraces were examined for catastrophic flooding. However, no prior study has examined lower flood terraces in the canyon to chronicle the occurrence of extreme floods or evaluate the coincidence of flooding and pre-historic human occupation. This study analyzes late-Holocene slackwater deposits in flood terraces at two sites on the Snake River in Hells Canyon Idaho/Oregon US to chronicle the occurrence of extreme floods. The paleoflood chronology was then compared to the timing of human occupation.

Study area For purposes of this study, the Hells Canyon reach of the Snake River begins at Hells Canyon Dam and ends at the mouth of the Grande Ronde River (Figures 1). The Tin Shed and China Rapids sites were analyzed to determine paleoflood frequency. The Tin Shed site forms a 270-m long 4-to 5-m high flood terrace on the west bank of the Snake River approximately 51 km downstream from Hells Canyon Dam. An archeological study notes three periods of human occupation there (Reid, 1991). The China Rapids site is 49 km from the dam and forms a 5-m high gently sloping 30-m long flood terrace on the east bank of the river. No archeological study has taken place at this site.

Flood terraces at the two sites were used in this study because they meet the paleoflood analysis criteria described by Baker (1987). The river flows through a bedrock channel in an arid to semi-arid climate where bioturbation of flood terrace sediments is minimal and sand-sized sediment supply is sufficient enough to contribute to the development and preservation of slackwater stratigraphy (Figure 2).

Hells Canyon Dam regulates runoff into the canyon, however the timing of pre- and post dam seasonal discharge patterns have not changed (Muckleston, 1993). Climate in the basin is arid to semi-arid. Most moisture in the region is brought from the Pacific Ocean, and rainfall is controlled by the orographic effect because of mountain ranges between the western coast of the United States and the Continental Divide. Annual precipitation is highest in December, January and February and amounts range from 25 cm per year in valleys to 102 cm per year at higher elevations (Jackson, 1993). Flood-producing meteorological events in the Hells Canyon reach of the Snake River have included rain on snow, rapid snowmelt, ice jams, rainfall persistence and rainfall repetitiveness (Paulson et al., 1991, U.S. National Weather Service, 1988 and 1997, Table 1).

Slackwater Deposition and the Archeological Record Chronicling pre-historical floods was done primarily by absolute dating based on the ages of organic-rich sediment, charcoal and shells that underwent radiocarbon (14C) laboratory analysis conducted at Beta Analytic, Miami, Florida US (Table 2). Radiocarbon ages were also used to correlate stratigraphic units from both sites. Analytic unit descriptions from Root et al. (1998) were consulted to assist in identifying cultural horizons in the stratigraphic sections and determining the coincidence of human occupation and flooding at the Tin Shed site. Paleoflood unit characteristics were described in the field based on lithology, soil cohesion, color, grain size, boundary distinctness, boundary topography and depth below surface. Upper units in sections at the Tin Shed site were also traced in the field and correlated based on unit characteristics (Figure 3). The timing of historical slackwater deposition at the Tin Shed and China Rapids sites was obtained from a list of the dates of 10-year recurrence interval (R.I.) or greater discharges on lower and central basin streams by Rhodes (2001). Other information about extreme discharge and precipitation was obtained from Paulson et al. (1991), the U.S. National Weather Service (1997) and Perry (2001) (Table 1).

Twelve extreme flood events are reported in the stream gauge records for the central and lower Snake River basin (Table 1). The most likely events to contribute to Tin Shed and China Rapids flood terraces were the June 1894, December of 1955 and December 1964 events because of the areal extent and magnitude of floods in the Snake River basin. Evidence of the 1894 flood is probably recorded by the inundated burn area containing the sample with a calibrated age of 192-81 cal yr BP and the sand and silt units above it could account for the 1955 and 1964 floods. Since construction of Hells Canyon Dam record discharge in the canyon occurred between January 1 and 3, 1997 but no above-terrace evidence of flooding was found when this study was conducted a year later. Schmidt et al. (1995) indicate erosion of terraces in the canyon is occurring since construction of Hells Canyon Dam. Flood events that could have contributed to erosion of the Tin Shed site flood terrace occurred in January 1997 spring 1984, January 1974 and June 1974.

As many as 27 extreme floods over more than 5,000 years are recorded in the stratigraphy at the Tin Shed and China Rapids sites (Table 2, Figures 3 and 4). From approximately 600 years ago to the present extreme floods left 8 deposits and after a 700-year temporal gap 19 other deposits were left beginning 1,300 years ago. A charcoal sample from a flood-inundated burn area was retrieved at approximately the same depth (28.5 cm) below the surface and had a calibrated age between 192-81 cal yr BP. In the uppermost portion of Sections 1 and 2 archeological and paleoflood ages coincide. Root et al. (1998) noted the presence of a cultural horizon between 30 cm and 50 cm below the surface that was used for domestic purposes in historical times.

In the next two sections there is general agreement between paleoflood and archeological unit ages with the exception of a unit containing a midden. Also, the midden suggests a cause for the temporal gap. The radiocarbon calibrated ages of two charcoal samples from paleoflood analysis range from 543 to 339 cal years B.P (Table 2, Figures 3 and 4). Uncertainties associated with the radiocarbon dating process e.g., Stuiver and Polach (1977) do not allow for distinction of slackwater depositional episodes when samples have radiocarbon ages in the 500 to 50 cal yr BP range. As a result, units 3 and 8 above are considered the same age but the stratigraphy there are 4 thick units 3 and 8. Further, unit 8, Section 2 is at least 700 years younger than the underlying unit 9. The archeologists report a cultural horizon between 50 and 90 cm below the surface that was probably the edge of a village with a midden where animals were buried in a pit between 1000 and 500 years BP. The depth range would put the midden in unit 8, however, the archeological analytic unit description better matches the description of paleoflood study unit 9, Section 2, which is >100 cm below the surface. Root et al. (1998) reports an older cultural horizon with fewer artefacts at >100 cm. Since archeological test pits were dug on the surface and the paleoflood analysis was conducted along the embankment, the discrepancy could be attributed to the difference in analysis locations. If the cultural activity associated with the base of paleoflood unit 8 (radiocarbon age range 543-339 cal yr BP) was a burial pit, then the record of slackwater deposits may have been disturbed.

Four calibrated ages were obtained from three shells and organic sediment for the paleoflood study (Table 2, Figures 3 and 4). Initially, the radiocarbon ages of shells suggested a 3,000-year temporal gap as well as an age reversal. So, the organic sediment sample was analyzed to confirm the three shell ages. The calibrated age of the sediment was 1322-1098 cal yr BP while the shell ages ranged from 4790 to 4403 cal yr BP (Table 2, Figures 3 and 4). Mussel shells could yield anomalously old ages from the intake of old carbon (Trumbore, 2000) but the younger sediment sample age still indicated a temporal gap in the stratigraphic record at this site. The units at this depth are considered to be slackwater deposits with the midden discussed above at the upper boundary. The third cultural horizon identified by the archeologists was between 100 and 140 cm below surface and age and human activity was unknown.

The cause of a radiocarbon sample age reversal lower in Section 2 is uncertain. Sediment from a lens of coarse sand had an age range between 8715-8505 cal yr BP and from the same unit near the lower boundary the age was between 4546-4492 cal yr BP. Perhaps at this level sediment rich in charcoal preserved at another site was delivered to this location in the flood. The age of the lowest sample was derived from organic sediment and had an age of 5994-5909 cal yr BP. No archeological test pits were dug to this depth.

Fourteen floods are represented in the stratigraphy at the China Rapids site. No archeological test pits were hollowed out at this site and there are no age reversals. Four flood layers in the stratigraphy are in the younger portion of the section and 10 slackwater deposits are in the older portion of the section (Table 2, Figures 3). The radiocarbon ages and depth below surface of samples retrieved from here are consistent with ages and depths for samples from the upper stratigraphy at the Tin Shed site.

All stratigraphic sections in this study show a cluster of flood layers hosting samples with calibrated radiocarbon ages between 543 cal yr BP and present and older flood layers with sample ages between 5994 cal yr BP and 1098 cal yr BP. At most 2 units separate older and younger layers thus implying a temporal gap. The sections do not record the same number of flood deposits probably because of original underlying topography; anthropogenic influences; clarity of boundaries and/or preservation of the record. Still, the temporal gap exists at approximately the same depth below surface in all but Tin Shed site Section 3 (Figure 3).

Discussion and Conclusion The reason for a temporal gap in the record at the study sites is unclear. The existence of a cultural horizon that is thought to have been the site of a burial pit lends credence to the idea that the temporal gap exists due to soil disturbance from digging by humans living in the area but, the gap exists in all sections including China Rapids, which is approximately 2 km away from the Tin Shed site and on the opposite side of the Snake River. Other possibilities for the consistent temporal gap in all sections, are as follows: 1) a high magnitude flood eroded the record between 1,000 and 300 years ago, 2) there was a change in local channel hydraulics or, 3) after the earlier period of flooding there was a hiatus in slackwater deposition. Evidence of landslides is abundant in Hells Canyon (Vallier, 1998); therefore, the possibility exists that a dam outburst flood could have occurred and removed a portion of the record at the Tin Shed and China Rapids sites. But no study details the timing of these events for comparison to paleoflood chronology. Channel hydraulics was not studied in any detail and therefore cannot be addressed. The fact that all four sites have the same temporal gap hints at a hiatus in floods capable of leaving deposits above the unit hosting the 1098 cal yr BP- aged sample but no comparison has been made to climate proxies for the area. As the archeological findings are based on test pit excavations instead of a complete archeological excavation no definite conclusion can be drawn without further examination of more flood terraces along the Hells Canyon reach of the Snake River.

Baker, V.R., 1987, Paleoflood hydrology and extraordinary flood events: Journal of Hydrology, v. 96, no. 1-4, p. 79-99.

Jackson, P.L., 1993, Climate, in Jackson, P.L., and Kimerling, A.J., ed., Atlas of the Pacific Northwest: Corvallis, Oregon State University Press, 152 p.

McCormac, F.G., Reimer, P.J., Hogg, A.G., Higham, T.F.G., Baillie, M.G.L., Palmer, J., Stuiver, M., 2002, Radiocarbon, v. 44:, p. 641-651.

Muckelston, K.W., 1993, Water resources, in Jackson, P.L., and Kimerling, A.J., ed., Atlas of the Pacific Northwest: Corvallis, Oregon State University Press, p. 152.

Paulson R.W., Chase E.B., Roberts, R.S., Moody D.W., eds., 1991, National water summary 1988-89 - hydrological events and floods and droughts: US Geological Survey Water-Supply Paper 2375, p. 99-104.

Reid, K.C., 1991, Prehistory and Paleoenvironments at Pittsburg Landing: data recovery and test excavations at six sites in Hells Canyon National Recreation Area, west central Idaho: Center for Northwest Anthropology, Washington State University, 15 p.

Reid, K.C., 2000, Account of circumstances leading to erosion of the Tin Shed site: e-mail correspondence dated June 27, 2000, from K.C. Reid to G.B. Rhodes regarding the event that led to embankment erosion at the Tin Shed site.

Rhodes, G.B., 2001, A small look at the big picture: paleofloods on the Snake River as indicators of Holocene climate in the Northwest United States: Masters thesis, Central Washington University, Ellensburg, unpublished.

Root, M.J., Chatters, J.C., Ferguson, D.E., and Reid, K.C., 1998, Test excavations at the Tin Shed site, Hells Canyon National Recreation Area: Rainshadow Research Report, 42 p.

Schmidt, J.C., Grams, P.E., and Webb, R.H., 1995, Comparison of the magnitude of erosion along two large regulated rivers: Water Resources Bulletin, v. 31, no. 4, p. 617-631.

Stuiver, M., and Polach, H., 1977, Discussion: Reporting of 14C Data: Radiocarbon, v. 19, no. 3, p. 355-363.

Stuiver, M., and Reimer, P. J., 1993, Extended 14C database and revised CALIB radiocarbon calibration program, Radiocarbon v. 35, p. 215-230.

Trumbore, S.E., 2000, Radiocarbon geochronology, in Quaternary Geochronology: Methods and Applications, edited by J.S. Noller, J. M. Sowers, and W.R. Lettis, p 41-60, American Geophysical Union, New York.

U.S. Army Corp of Engineers, 1941, Snake River, Idaho, Washington, and Oregon: Letter from the Secretary of War: U.S. House of Representatives and 77th Congress, Document Number 452.

U.S. National Weather Service, 1988, Probable maximum precipitation estimates: United States between the Continental Divide and the 103rd meridian, hydrometeorological report no. 55A /prepared by E.M. Hansen . . . [et al.]: Silver Spring, U.S. Department of Commerce, 242 p.

U.S. National Weather Service, 1994, Probable maximum precipitation, Pacific Northwest states: Columbia River (including portions of Canada), Snake River and Pacific coastal drainages, hydrometeorological report no. 57/ prepared by E.M. Hansen. . .[et al.]: Silver Spring, U.S. Department of Commerce, 338 p.

U.S. National Weather Service, 1997, Disastrous floods from the severe winter storms in California, Nevada, Washington, Oregon, and Idaho: December 1996 – January 1997, Natural disaster survey report 98-0116-P / prepared by Western Region Headquarters: Silver Spring, U.S. Department of Commerce, National Oceanic and Atmospheric Administration, 51 p.

Vallier, T.L., 1998, Islands and Rapids: A geologic story of Hells Canyon: Lewiston, Idaho, Confluence Press, 151 p.

Date received: January 27, 2004

Copyright © 2004 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # camu-30.