Historical change in channel form and riparian vegetation of the McKenzie River, Oregon

This study examined channel structure and position and riparianvegetation and land use on the upper 70 km of the McKenzie River, Oregon inthe 1940s, compared the 1940s conditions to present conditions, and exploredthe processes driving change in this system and the implications for aquatichabitat. The hydrologic record was analyzed, and field surveys were conductedand compared to historical habitat surveys. Riparian characteristics andchannel features were digitized from aerial photographs from 1945/49 and1986 and imported into ArcInfo GIS for analysis. Types of data digitized fromthe aerial photos included locations and length or area of wetted channel,active channel, tributaries, side channels, large woody debris, exposed gravelbars, roads, and dominant vegetation or land use within 200 m of the activechannel.
Construction of dams on the mainstem McKenzie River and two majortributaries, Blue River and South Fork, in the 1960s has altered the flow regimeand sediment supply to the mainstem McKenzie, decreasing the frequency,mean and variation of peak flows, reducing the competence of flows to moveexisting bedload, and cutting off sediment from over half of the drainage area.Mean peak flows decreased 44% and competence of peak flows with a 2-yrrecurrence interval declined approximately 29% after dams were constructed upriver. Adjustments to reduced sediment supply and flow alteration by dams inthis system included 57% decrease in exposed gravel bars, 40% decrease inside channel length, and possible substrate coarsening (as compared tohistorical estimates).
Channel straightening occurred in each of three instances of channelchange during the study period, and sinuosity decreased one half of the amountneeded to produce a straight channel in the most susceptible, unconstrainedreach. Human actions prior to high flow events played a role in the direction ofchannel change in each case. Over the entire study area, 7°A, of the mainchannel changed position by at 30 m or more and little or no change in channelposition was noted in reaches constrained by valley floors. Additional channelconstraint has been produced by road construction near the channel andriprapping for roads, bridges, and residences.
Less large woody debris was observed in the 1986 channel than in the1949 channel, indicating a reduction in pool-forming agents and channelroughness elements. Frequency of large pools (^2 m depth and >40 m2 area)decreased 19% over the study area. The greatest loss in pools (73%) wasnoted in the unconstrained reach that exhibited two areas of channel changeand an increase in exposed gravel bars.
Increased human use of the riparian area for roads and residentialpurposes has led to an increased fragmentation of the riparian landscape.Density of residential or developed patches within the riparian area hasincreased 215% as more and smaller areas are converted from naturalvegetation to human use. Riparian area devoted to roads and residential useshas nearly doubled since the 1940s. Mean vegetation or land-use patch sizehas decreased from 2.2 ha to 1.6 ha as larger patches have been sub-divided,and patch and edge densities have increased. Agriculture and clearcuts for timber removal have decreased within the riparian area while continuingupslope. Riparian area in mature conifers has decreased 44% from levels inthe 1940s while hardwoods have increased 45% in the riparian area. Futurewood loading to the channel is reduced by a decline in mature riparianvegetation, especially mature conifers.
Channel and riparian changes noted in this study have implications forfish populations. Channel straightening, reduction in side channels, and loss ofpool-forming agents reduce habitat heterogeneity and off-channel refugia.
Ecosystem management of watersheds requires evaluation of conditionsacross scales of time and space. The use of GIS in this study made it possibleto detect changes in channel form and riparian conditions during four decades,along a 70-m channel and 90-m riparian area and to analyze the large datasets relevant to understanding functions and change in channels and riparianareas