CHANNEL NARROWING AND VEGETATION DEVELOPMENT FOLLOWING A GREAT-PLAINSFLOOD

Citation
Jm. Friedman et al., CHANNEL NARROWING AND VEGETATION DEVELOPMENT FOLLOWING A GREAT-PLAINSFLOOD, Ecology, 77(7), 1996, pp. 2167-2181
Citations number
73
Categorie Soggetti
Ecology
Journal title
ISSN journal
00129658
Volume
77
Issue
7
Year of publication
1996
Pages
2167 - 2181
Database
ISI
SICI code
0012-9658(1996)77:7<2167:CNAVDF>2.0.ZU;2-4
Abstract
Streams in the plains of eastern Colorado are prone to intense floods following summer thunderstorms. Here, and in other semiarid and arid r egions, channel recovery after a flood may take several decades. As a result, flood history strongly influences spatial and temporal variabi lity in bottomland vegetation. Interpretation of these patterns must b e based on understanding the long-term response of bottomland morpholo gy and vegetation to specific floods. A major flood in 1965 on Plum Cr eek, a perennial sandbed stream, removed most of the bottomland vegeta tion and transformed the single-thread stream into a wider, braided ch annel. Channel narrowing began in 1973 and continues today. In 1991, w e determined occurrences of 150 vascular plant species in 341 plots (0 .5 m(2)) along a 7-km reach of Plum Creek near Louviers, Colorado. We related patterns of vegetation to elevation, litter cover, vegetative cover, sediment particle size, shade, and year of formation of the und erlying surface (based on age of the excavated root flare of the oldes t woody plants). Geomorphic investigation determined that Plum Creek f luvial surfaces sort into five groups by year of formation: terraces o f fine sand formed before 1965; terraces of coarse sand deposited by t he 1965 flood; stable bars formed by channel narrowing during periods of relatively high bed level (1973-1986); stable bars similarly formed during a recent period of low bed level (1987-1990); and the present channel bed (1991). Canonical correspondence analysis indicates a stro ng influence of elevation and litter cover, and lesser effects of vege tative cover, shade, and sediment particle size. However, the sum of a ll canonical eigenvalues explained by these factors is less than that explained by an analysis including only the dummy variables that defin e the five geomorphically determined age groups. The effect of age gro up is significant even when all five other environmental variables are specified as covariables. Therefore, the process of postflood channel narrowing has a dominant influence on vegetation pattern. Channel nar rowing at Plum Creek includes a successional process: annual and peren nial plants become established on the channel bed, sediment accretes a round the vegetation, and increasing litter cover, shade, and scarcity of water eliminate species that are not rhizomatous perennials. Howev er, successional trajectories of individual surfaces are modified by f low-related fluctuations of the bed level; surfaces deposited by the 1 965 flood have had distinct sediment and vegetation since their format ion. Species richness is highest on surfaces dating to 1987-1990; the many species restricted to this transitory assemblage are perpetuated by flood-related fluctuations in channel width. Since the 1965 flood, seedling establishment of the dominant trees (genus Populus) has occur red only on low surfaces formed during channel narrowing. Thus, the fl ood has indirectly promoted Populus establishment over a 26-yr period.