FLOOD04278 - Laserfiche WebLink

Home Browse Search

2 M. Grimm etal. /GeomorphoJogy 00 (1995) {)()()....()()( a few minutes to a few hours, followed by a rapid decline to preflond stages (Follansbee and Sawyer, 1948; McCain and Ebling, 1979). Overland and stream flow velocities are swift, enhancing erosion on hillslo- pes and in channels (Shroba et aI., 1979; Jarrett and Costa, 1988) and creating boulder bars. Rain-on-snow floods seldom occur and are assumed to have a negli- gible influence on coarse sediment distribution at the study sites. Snowmelt-runoff floods have very broad hydrograph peaks and smaller magnitudes than rainfall floods. We assume that snowmelt flood peaks are less competent than rainfall floods to transport the coarse portion of available sediment at the study sites and that snowmelt floods are less likely to create boulder bars. This assumption is supported by observations of newly created boulder bars following rainfall floods (Costa, 1978, 1983; Shroba et aI., 1979; Jarrett, 1987, 1990; Jarrett and Costa, 1988; Jarrett and Way thomas, 1995) and by the lack of such deposits following snowmelt floods. Using U.S. Geological Survey gaged discharge records for the Rocky Mountains, Jarrett (1993) iden- tified variable elevation limits to rainfall-produced flash flooding, The limit varies with latitude and dis- tance from moisture sources in the Gulf of Mexico. It ranges from about 2350 m in New Mexico to 1650 m in Montana, and is 2300 m in Colorado (Jarrett and Costa, 1988; Jarrett, 1987, 1990). Rash flooding rarely occurs above this elevation limit. but the intense rain- falls below the limit produce frequent floods, The disjunct spatial (downstream) distribution of rainfall. induced flash floods along channels in the Rocky Mountains facilitates a study of the geomorphic role of these floods by providing contiguous channel reaches affected and unaffected by rainfall floods. In this paper we characterize downstream trends in coarse sediment deposition along Bear Creek as an illustration of how flash floods affect coarse-sediment deposits in a mountain channel. The grain-size characteristics and spatial distribution of coarse sediments above and below the flash-flood elevation limit are distinctly dif- ferent. We begin by estimating maximum flood discharge at four sites along Bear Creek using paleostage indi- cators. Understanding of flood magnitude-frequency relations can be enhanced by using paleohydrologic data to reconstruct the characteristics of floods not included in systematic data. Paleoflood data can indi- ----- cate probable upper limits for the largest floods that have occurred in a basin (Costa, 1983). Both the occur- rence and absence of paleohydrologic indicators within a given channel reach indicate the spatial distribution of flooding in a basin. In channels of the Colorado Rocky Mountains, paleoflood indicators include flood boulder bars, alluvial fans, impact scars on trees, and slackwater deposits (Jarrett, 1990), These indicators are found between elevations of approximately 2300 m and 1670 m (the base of the mountain range). We used peak flood discharges estimated from paleostage indicators at Bear Creek to explain coarse- sediment distribution along the creek. We hypothesize that values of D,o and D_ are inversely proportional to elevation within the drainage basin as a result of decreasing peak unit discharge with increasing eleva- tion. The relation between unit discharge and elevation may be explained in terms of an upper elevation limit for rainfall-floods. Based on the abundance of cobble-and boulder-sized clasts at all study sites, we assume that the availability of coarse sediment to the channels is not a limiting factor in the Bear Creek basin. Under that assumption, changes in sediment-size distribution between simi- larly sized tributary drainages can qualitatively indicate changes in flood peak magnitude and in stream power per unit area as a result of runoff regime (Grimm, 1993). For example, if the proposed 2300 m elevation limit for rainfall-dominated floods is correct, tributaries downstream from the limit should show substantially coarser particle size distributions than similarly sized tributaries above 2300 m. In addition, particle-size dis- tributions upstream and downstream from tributary junctions below 2300 m should be substantially differ- ent. Because these small tributary basins are entirely below 2300 m, they are often the source areas of large floods (Diebold, 1939; Follansbee and Sawyer, 1948). The larger peak unit discharge for these tributaries implies greater sediment-transport competence than most flows in the main channel. When a flood peak from a tributary enters the main channel, much of the coarse sediment is deposited as a bar, analogous to the debris fans described for the Colorado River in the Grand Canyon (Webb et a!., 1989). In contrast, we hypothesize that a tributary at or upstream from 2300 m with approximately the same drainage area, but that is snowmelt dominated, should show similar particle- Journal: GEOMOR Article: 368