When streams unravel: the tale of Plum Creek, CO

Streams that have recently undergone disturbance provide a natural laboratory for understanding how fluvial landscapes evolve. Plum Creek, CO, which experienced catastrophic flooding in the 1960s, has provided scientists with an opportunity to learn more about a type of stream morphology that is relatively rare in North America.

The perfect storm

On the morning of June 16, 1965, Plum Creek was a typical sand-and-gravel meandering stream draining dissected Tertiary sandstone and siltstone between Denver and Castle Rock, Colorado, USA (Fig. 1). Just above its confluence with the South Platte River, it occupied an alluvial valley approximately 1 km wide and exhibited a meandering morphology, bordered by a riparian community of willows (Salix spp.) on the banks and cottonwoods (Populus deltoids monilifera) atop rarely inundated quaternary alluvial terraces.

As the summer sun heated the waters of Plum Creek that day, it also heated a low, moist air mass from the south that was in the process of converging with a higher, cooler mass of air from the west. Intense convective thunderstorms developed. Over a period of four hours, the storms released 360 mm of precipitation—more than 80% of the annual total! The resulting flood washed out bridges, caused major structural damage in Denver, and eliminated 53% of the woody vegetation on Plum Creek's floodplain. Peak discharge at Louviers, CO was estimated to be 4360 m³s^-1, or 15 times the 50-year flood—the highest such ratio recorded at any US gaging station. The recurrence interval of the 1965 flood has been estimated at 900-1600 years.

Before the end of the day, Plum Creek had "unraveled" down to its confluence with the South Platte, a term referring to catastrophic bank erosion and the cutting of new channels into its floodplain. For 2.5 hours, the stream overtopped its second set of terraces and scoured the supply of fine sediment from the floodplain, leaving behind three-meter-high deposits of coarse sand. Then, like water from a hose in a sandbox, the stream began to cut into those deposits in multiple braided rivulets, incising a headcut of about two meters within two days of the flood.

Due to the lack of riparian vegetation, Plum Creek persisted in a braided configuration, continuing to widen during subsequent minor high flow events until 1973 (Fig. 2). Between 1973 and 1984, a drier climate persisted, and vegetation began to take hold and stabilize sediment deposits that had accumulated around snags and large objects from a refuse pile washed out in the flood. This period was characterized by the formation of fluvial islands and anabranches (i.e., multiple, persistent channels separated by semi-permanent vegetated bars or islands).

Channel morphology and the attainment of graded equilibrium

In most locations, anabranching is an inherently unstable morphology. Bed shear stress in rivers increases proportionally to channel depth, so when, by chance, one anabranch becomes slightly deeper than another, it erodes preferentially, becoming even deeper in a positive feedback cycle. As it incises, it captures discharge and induces deposition in the other anabranch(es), which can lead to anabranch infilling and the coalescence of islands with the floodplain (Fig. 3). Indeed, this process has led to gradual channel narrowing (Fig. 4) and a slight increase in sinuosity in Plum Creek over the past three decades (Fig. 2). It appears that through island formation and anabranch infilling, Plum Creek is undergoing a gradual transition back to a single-threaded, meandering morphology.

However, Plum Creek's presumed return to a "graded equilibrium"—in which the stream's transport capacity equals its sediment supply and the bed neither aggrades nor degrades—will likely take decades to centuries longer. Today's Plum Creek exhibits reaches in various transitional stages, from single-thread meandering, to anabranching (with both persistent anabranching reaches and reaches with infilling anabranches), to braided. Overall sinuosity remains much closer to its post-flood than pre-flood value, and the channel gradient remains approximately 20% higher (Fig. 2). As such, dynamic Plum Creek is an ideal location to test theories in fluvial geomorphology about channel pattern and its relationship to sediment transport dynamics and hydraulic geometry.

Does anabranching lead to channel stability?

One persistent puzzle in geomorphology has been why anabranching is a stable to quasi-stable configuration in some locations such as dryland rivers in central Australia. Empirically, these stable anabranching morphologies occur in locations with large sediment supply relative to the energy available to transport sediment, and, compared to upstream or downstream single-thread reaches, anabranching channels often have a greater transport capacity. Using simple hydraulic geometry-based relationships for stream power, Nanson and Huang (1999) show that stream velocity (and hence transport capacity) reaches a maximum value at a channel width-to-depth ratio that is smaller than that of most nonequilibrium streams that adopt an anabranching morphology. They argue that the aggregate width-to-depth ratio decreases with a transition to an anabranching channel pattern and suggest that anabranching consequently allows streams to achieve "maximum flow efficiency."

As has been reported for previous studies, channel surveying revealed that anabranching reaches of Plum Creek had a higher relative sediment transport capacity than single-thread reaches, based on the Rubey sediment transport relation (Fig. 5). The difference is particularly pronounced at bankfull flow. However, sediment transport potential in the anabranching reaches was enhanced not because of a smaller width-to-depth ratio than single-threaded reaches but because the anabranching reaches tended to have higher bed slopes. Indeed, the increased sediment transport efficiency of anabranching reaches at high flows may have reinforced the higher bed slopes through greater scour within the anabranching reach relative to upstream. The persistent anabranching reaches also tended to occur in areas that remained relatively confined by terraces, which may have promoted in-channel deposition and the formation of fluvial islands.

From field studies, Tooth and Nanson (2004) concluded that simple morphometric, hydraulic, and sedimentary variables could not predict the development of anabranching, but that subtle differences in water supply and grain size can differentiate between anabranching and single-thread morphology. In a similar manner, complex differences in valley geometry, grain size, and, roughness (e.g., roughness characterized by snags and/or refuse) differentiate between single-thread and anabranching reaches of Plum Creek. It appears that at Plum Creek, anabranching has persisted not because of increased sediment transport efficiency within the anabranches (indeed, this increased efficiency has moved those reaches farther from equilibrium) but because the processes controlling its elimination (sedimentation, vegetation colonization) have simply not been occurring for long enough.

• Friedman, J. M., W. R. Osterkamp, and W. M. L. Jr. 1996a. The role of vegetation and bed-level fluctuations in the process of channel narrowing. Geomorphology 14:341-351.
• Friedman, J. M., W. R. Osterkamp, and J. W.M. Lewis. 1996. Channel narrowing and vegetation development following a Great Plains flood. Ecology 77:2167-2181.
• Nanson, G. C., and H. Q. Huang. 1999. Anabranching rivers: Divided efficiency leading to fluvial diversity. Pages 477-494 in A. J. Miller and A. Gupta, editors. Varieties of Fluvial Form. John Wiley, Hoboken, NJ.
• Tooth, S., and G. C. Nanson. 2004. Forms and processes of two highly contrasting rivers in arid central Australia, and the implications for channel-pattern discrimination and prediction. GSA Bulletin 116:802-816.