A Rapid Return

There is a small campsite in Roger Mills County, in the headwaters area of Brokenleg Creek, which we recorded in 1984 as 34RM451, or Thurmond Ranch #30. It is a surface exposure (or was back when vegetation did not totally obscure the surface) of burned rock, lithic debitage (mostly biface thinning and retouch flakes of Ogallala quartzite, with some Alibates and locally available cherts), and flake tools. The latter which we have collected include a drill, a graver, and five other flakes with one or two edges (most straight, two strongly concave and of the right diameter for working arrow shafts) retouched or damaged through use in cutting durable materials. We have also collected three arrow points from the site (Figure 1, Table 1): two corner-notched points with expanding stems, classifiable as type Scallorn (after Suhm and Jelks 1962), and one basally notched point classifiable as type Deadman’s (after Willey and Hughes 1978). It is worth noting that a campsite which has also produced type Deadman’s points (not known for their ubiquity in Oklahoma), 34RM449, is only 600 m to the north.

Specimen	Weight	Thickness	Material
A	.7 gm	3 mm	Alibates
B	.7 gm	3 mm	Local chert
C	1.0 gm	4 mm	Ogallala quartzite

34RM451 covers about 6000 m², or roughly 1.5 ac. It wraps around the head of what was a dry, shallow, grassy upland swale in 1984, in a U-shaped arc, opening to the east, some 40 m wide and 150 m in length from one U-crest to the other. The swale is about 25 m wide adjacent to the site, tapering to a rounded end inside the “U”. It is gentle in cross-section and about two meters lower at its floor than the higher, sandy ground of 34RM451. Again, in 1984, this entire area, including the floor of the swale, was upland sandy-land pasture, covered in little and big bluestem, Indian grass, and sand dropseed, with scattered yucca and sand plum. A lone cottonwood tree stood about 120 m east of the site. While recording 34RM451 in 1984, we commented on the survey form that the site might have marked the upper limit of the water table during the wet climatic conditions of the first millennium AD.

The headwaters springs of Brokenleg Creek in 1984 were 400 m to the east and 11 m lower in elevation in the winter, and 750 m east and 15 m lower in elevation in the summer (that’s half a mile away and 50’ lower in elevation). Counterintuitive as it may seem, spring discharge along the eastern base of the Ogallala Formation declines during the summer, when rainfall tends to be higher, and increases during the winter, when precipitation is generally much lower. Presumably, spring discharge picks up in the winter when groundwater uptake by the heavy parkland vegetation of the upper Brokenleg Creek and Sergeant Major Creek basins shuts down. There is probably also some time lag in the local recharge, but we suspect the evapotranspiration effect is dominant.

What a difference thirteen years have made. Our average annual rainfall in Roger Mills County has been gradually increasing for decades, but has skyrocketed the past three years (Figures 2 and 3). Average annual rainfall increased from around 500 mm (20") in the 1930s to 660 mm (26") in the 1970s and 1980s. Average rainfall during the period 1995-1997 has averaged nearly twice that, at 1245 mm (49"), and has varied within a narrow range of 1170-1320 mm (46-52"). Spring discharge along the eastern edge of the Ogallala aquifer has mushroomed, and many small springs and seeps have opened up both along the outcrop edge of the Ogallala Fm. and in the underlying Permian bedrock units, presumably as a result of hydraulic head forcing through bedding planes and joints. There were nine springs on the ranch in 1984. We mapped 229 significant seeps and springs on the 9275 ac ranch in January of 1998 (an average of one spring per 40 ac).

The headwaters springs of Brokenleg Creek have risen the full 15 m in elevation from their summertime low in the 1980s, and have migrated 750 m west to the foot of 34RM451. The site now rings a bog, which has drowned out the former grasses and grown up in sedges and black willow. An open pond, ringed with cattails, willows and cottonwoods, fills the swale from the east edge of the site, some 50 m to the east. A running stream, lined with young willows and cottonwoods, leads away from this pond and down to Brokenleg Canyon. There is no longer any seasonal difference in the elevation of spring discharge, although the volume of flow still picks up in the wintertime.

The projectile points from 34RM451 indicate that our northeast Asian-descended predecessors camped here during the first millennium AD (Boyd 1997, Briscoe 1987, Willey and Hughes 1978), during the Woodland mesic climatic episode, when the Copan paleosol was forming (Hall 1990). Deep, organic matter-enriched soils relatable to the Copan paleosol we have dated on and around the ranch at the Big Kitty Canyon, Brokenleg Bend #3, Finch Canyon, Herring Ranch, Higgins Creek, and Plum Creek localities indicate that this wet climatic episode began locally around 50 BC, and lasted until AD 1000.

Given the shape of 34RM451 and its spatial relation to the head of the swale, one would presume that the spring head was inside the ”U” of 34RM451 then, as now. This would indicate that the water table has now returned to the elevation at which it stood when the site was occupied. So what does this mean? We have overwhelming evidence from paleoclimatic proxies that climate can make drastic changes in a matter of a few years (c.f. Allen 1993; Allen and Anderson 1993; Alley et al. 1997; Curtis, Hodell, and Brenner 1996; Dowdeswell, et al. 1997; Phillips and Grantz 1997; Zielinski & Mershon 1997).

Previous research has documented at least two Late Holocene mesic climatic episodes on the Southern Plains. An apparently long one spanned the first millennium AD (producing the Copan or Caddo County paleosol; Ferring 1982, Hall 1990, Lintz and Hall 1983) which, again, we have dated locally from 50 BC to AD 1000. A much briefer one has been documented around AD 1450, producing the Delaware Canyon paleosol (Ferring 1982). However, our recent radiocarbon dating of Late Holocene buried soils in Roger Mills County suggests that the period 50 BC to AD 1650 may be best regarded locally as a series of rapid state shifts from mesic to more xeric conditions every 200 years or so, with three particularly benign mesic episodes (known to climatologists as pluvials) during the first millennium AD. Our observations at 34RM451 over the past two decades indicate that local effective precipitation is as high today as it was over a thousand years ago, during the wettest parts of the Woodland period.

1993 Late Quaternary Lacustrine Record of Paleoclimate from Estancia Basin, New Mexico. Ph.D. dissertation, Department of Geology, University of New Mexico, Albuquerque.

1993 Evidence from Western North America for Rapid Shifts in Climate during the Last Glacial Maximum. Science 260: 1920-1923.

1997 Holocene Climatic Instability: A Prominent, Widespread Event 8200 Years Ago. Geology 25(6): 483-486.

1997 Caprock Canyons Archeology: A Synthesis of the Late Prehistory and History of Lake Alan Henry and the Texas Panhandle-Plains. Prewitt & Associates, Inc., Reports of Investigations 110.

1987 Analysis of Cultural Materials from the Swift Horse Site, an Early Plains Woodland Site on the Black Kettle National Grasslands, Roger Mills County, Oklahoma. Briscoe Consulting Services, Butler, OK.

1996 Climate Variability on the Yucatan Peninsula (Mexico) during the Past 3500 Years, and Implications for Maya Cultural Evolution. Quaternary Research 46(1):37-47.

1997 The Mass Balance of Circum-Arctic Glaciers and Recent Climate Change. Quaternary Research 48(1): 1-14.

1982 The Late Holocene Prehistory of Delaware Canyon, Oklahoma. Institute of Applied Sciences, North Texas State University, Denton.

1990 Channel Trenching and Climatic Change in the Southern Plains. Geology 18: 342-345.

1983 The Geomorphology and Archaeology of Carnegie Canyon, Caddo County, Oklahoma. Oklahoma Conservation Commission Archaeological Research Report 10.

1997 Quaternary History of Sea Ice and Paleoclimate in the Amerasia Basin, Arctic Ocean, As Recorded in the Cyclical Strata of Northwind Ridge. Geological Society of America Bulletin 109(9): 1101-1115.

1962 Handbook of Texas Archeology: Type Descriptions. Texas Archeological Society Special Publication 1.

1978 The Deadman’s Shelter Site. In “Archeology at Mackenzie Reservoir”, Texas Historical Commission Archeological Survey Report 24, pp. 148-232.

1997 Paleoenvironmental Implications of the Insoluble Microparticle Record in the GISP2 (Greenland) Ice Core during the Rapidly Changing Climate of the Pleistocene-Holocene Transition. Geological Society of America Bulletin 109(5): 547-559.

Specimen

Weight

Thickness

Material