|
<br />12th Symposium on Global Change and Climate Variations, 81st AMS Annual Meeting,
<br />Albuquerque, NM, January 14-19,2000, American Meteorological Society, Boston.
<br />
<br />t
<br />
<br />J2.14 PALEOHYDROLOGIC ESTIMATES OF CONVECTIVE RAINFALL IN THE ROCKY MOUNTAINS
<br />
<br />Robert D. Jarrett .
<br />U.S, Geological Survey, Lakewood, Colorado
<br />
<br />1. INTRODUCTION
<br />
<br />Increased emphasis on extreme rainstorm and flood
<br />documentation will maximize benefits from emerging
<br />technologies, will help improve hydrologic modeling, and
<br />improve flash-flood forecasting, Convective rainfall is
<br />characteristically localized and can have large
<br />gradients in both rain rates and rainfall amounts over
<br />very short distances, often a few kilometers or less. In
<br />remote areas, systematic precipitation networks may be
<br />sparse or nonexistent. Limited resources often
<br />preciude extensive, instrumented data collection
<br />efforts, and there also is a likelihood rainstorms and
<br />floods will not occur in instrumented basins.
<br />Two paleohydrologic (geomorphic and hydrologic)
<br />techniques to estimate rainfall amounts of convective
<br />rainstorms are being developed and evaluated for
<br />mountainous areas. Paleohydrology includes the study
<br />of flood-transported sediments and botanic information
<br />from past floods preserved in river basins (Jarrett,
<br />1991; Jarrett and Tomlinson, 2000), Paleohydrology,
<br />which can be viewed as forensic hydrology, uses this
<br />physical evidence to make inferences of
<br />hydrometeorologic information. Elements of this
<br />presentation include discussions of the: (1) approach,
<br />results, and benefit$ for recent, historic, and prehistoric
<br />rainstorm reconstructions in the Rocky Mountain region;
<br />(2) limitations and uncertainties of rainfall estimates,
<br />and; (3) transfer to other hydroclimatic regions,
<br />In this paper, the approach is discussed for the
<br />community of Buffalo Creek, located about 50 km
<br />southwest of Denver, Colorado (fig. 1 a). On May 18,
<br />1996, an intense wildfire (Buffalo Creek fire) bumed
<br />about 50 km2 of forest, produced hydrophobic (water-
<br />repellent) soils, and making the are more susceptible to
<br />flooding. Subsequently, two people were killed and
<br />several million dollars in public and private property
<br />damage were caused by a flood on the evening of July
<br />12, 1996 (Colorado Water Conservation Board, 1997).
<br />Maximum water depths as much as 4 m occurred within
<br />about 30 to 45 minutes of the storm's onset in Buffalo
<br />Creek, Spring Creek, and the North Fork South Platte
<br />and South Platte Rivers. The paleohydrologic rainfall
<br />estimate for the July 12th storm was at least 110 mm in
<br />about an about an hour, and the storm footprint (within
<br />the 25 mm isohyet) was about 120 km2,
<br />For comparison, Henz (199B) estimated a maximum
<br />rainfall of about 130 mm, and Fulton (1999) estimated a
<br />
<br />'Correspondlng author address: Robert 0, Jarrett,
<br />U,S. Geological Survey, P,O, Box 25046, MS 412,
<br />Lakewood, CO 80225; e-mail: rjarrett@usgs.gov,
<br />
<br />maximum of 72 mm; however, their storm footprints were
<br />iocated slightly different and were much larger, The
<br />paleohydrologic results were obtained by July 16, 1996
<br />(two days of field and office work) and subsequently
<br />were used for emergency flood response, The
<br />paleohydrologic methodology is a flexible, "storm-
<br />chasing" approach that provides independent, cost-
<br />effective rainfall estimates, and can be used to
<br />compiement conventional instrumented monitoring,
<br />
<br />2. STUDY AREA
<br />
<br />The community of Buffalo Creek is located in the
<br />foothills of the Colorado Rocky Mountains (fig. la). The
<br />community, at an eievation of about 2,012 m, consists
<br />of several hundred homes within a montane forest
<br />(predominantly ponderosa pine, lOdgepole pine, Douglas
<br />fir, and aspen). Accumulation of organic litter (duff) in
<br />forested areas primary pine needles has an average
<br />depth of about 75 mm. Topography is rugged (slopes
<br />range from 5 to 60 percent) and sails are shallow (- 1 m
<br />to bedrock with numerous outcrops), moderately well
<br />drained, and composed of coarse sandy gravel (Sphinx-
<br />Legault-Rock granite complex). The climate is semiarid
<br />and mean annual precipitation is about 400 mm. The
<br />100-yr, 1-hr rainfall is about 55 mOl for the Buffafo Creek
<br />area (Miller et aI., 1973). Most streams in the study area
<br />are ephemeral. These streams flaw Into Buffalo Creek
<br />and the North Fork South Platte and South Platte
<br />Rivers, which primarily are fed by melting snowpack and
<br />trans-basin flow diversions, Stream gradients typical
<br />range from about 0.005 to 0.06 m1m. Flood flows in the
<br />Colorado foothills can result from generalized
<br />rainstorms, spring snowmeit, but primarily result from
<br />intense, localized thunderstorms (Jarrett, 1990).
<br />
<br />3. METHODS
<br />
<br />In the first or geomorphic methOd, rainfall amounts
<br />can be inferred from the amount of hillslope erosion,
<br />maximum size of sediments transported, and deposition
<br />characteristics, preferably an sparsely vegetated
<br />hillslopes. The hillslopes used should have as similar
<br />characteristics as possible. The dimensions of fresh
<br />rills, gullies, and headcuts as well as maximum size of
<br />sediments transported and their deposition
<br />characteristics are obtained and located on topographic
<br />maps. Local residents can often provide valuable
<br />information about the rainstorm including rainfall "bucket
<br />data," storm duration, and hail (which also can be
<br />
|