| Abstract |
- Channel evolution and influences of changing floodplain
characteristics, heterogenous bank materials, and altered flow regimes were
examined along the Willamette River, a large alluvial river in northwestern
Oregon. The Willamette River is composed of a series of geomorphically
diverse reaches, which have each evolved uniquely in the century following
Euro-American settlement. The river was divided into three large (30-50 km)
alluvial reaches according to physiographic characteristics. The historically
anastomosing and relatively steep McKenzie Reach (uppermost study reach),
extends between the confluences of the McKenzie and Long Tom Rivers. The
Willamette along the lower-gradient Santiam Reach (between the confluences
of the Santiam and Yamhill Rivers is primarily contained within a single channel and has experienced lower rates of erosion than upper reaches. The
Long Tom Reach (extending between the confluences of the Long Tom and
Santiam Rivers), acts as a transition between the upper and lower Willamette,
as it is here that the channel adopts a single-thread planform and becomes more
stable.
To assess the role of bank materials on bank-erosion rates, a method
for detecting relative differences in erodibility between bank materials along
large floodplains was developed. Coupling historic patterns of channel change
with a simple model of bank erodibility enabled tracking of relative changes in
bank erodibility among time intervals and bank materials. The analysis was
applied to the McKenzie Reach for three time periods: 1850-1895, 1895-1932,
and 1972-1995, and relative differences in bank erodibility were calculated for
Holocene alluvium, partially cemented Pleistocene gravels, and revetments
constructed in the 20th century. This simple model of bank erodibility reveals
that, for all three periods, banks composed of Holocene alluvium were at least
2-5 times more erodible than banks composed of Pleistocene gravels.
Revetment installed in the twentieth century was highly resistant to erosion and
was at least 10 times less erodible than Pleistocene gravels.
To examine larger-scale controls of geology, flooding, and human
intervention on channel stability, rates and styles of historic channel change
were determined for the McKenzie, Long Tom and Santiam Reaches and were
linked with events or factors that may have triggered the observed patterns of
channel change. Effects of anthropogenic activities on channel change were assessed by reviewing historic documents describing settlement patterns,
riparian deforestation, channel improvements, and other actions. The role of
flooding was assessed by compiling gauge records, anecdotal accounts of
flooding, and by comparing stream-power distributions of large historic floods
against smaller, post-dam floods with a 2-D flood model.
Analyses of these larger-scale controls revealed that between 1850 and
1895, a period marked by the 3 largest floods of record, all reaches
experienced numerous avulsions, increases in channel width, and decreases in
centerline length. During the interval 1895-1932, a period with frequent,
moderate-sized floods, migration rates increased by 50-300%, sinuosity
increased and channel width decreased. The interval 1932-1995 was initially
marked by rapid migration, but channel stabilization and dam building slowed
erosion rates, causing the Long Tom and Santiam Reaches to display similar
migration rates as those recorded for 1850-1895. Along the upper Willamette
(McKenzie Reach), channel change during 1972 to 1995 was primarily limited
to lateral migration along areas unrestricted by revetments and occurred at
rates similar to 1850-1895 levels. Channel width decreased along all reaches
during the 20th century.
It is hypothesized that flooding may have been the primary factor
responsible for the large-scale straightening and widening that occurred during
1850-1895. Actions taken to reduce streamside wood and side-channels along
the McKenzie and Long Tom Reaches may have also contributed to widening.
Along some areas of the floodplain, where the largely straightened and widened 1895 channel flowed through Holocene alluvium, the channel
developed small bends that subsequently migrated rapidly downstream, and
triggered rapid migration of adjacent bends. This concurrence of events and
conditions suggests that accelerated erosion during the period 1895-1932
results from a combination of a "primed" planform, highly erodible bank
materials, and a highly erosive flow regime with many moderate-sized floods.
Migration rates 1895-1932 may have also increased as a result of land clearing
and snag removal, as increasing numbers of settlers occupied floodplain lands
in this interval. Anthropogenic activities have no clear effect on planform or
erosion rates until the 193 D's, when widespread bank stabilization and dam
construction resulted in diminished migration rates, fewer avulsions, and
channel narrowing. By the late century, 30-45% of each reach was
stabilized with revetments, while naturally resistant bank materials bordered an
additional 13-30% of the channel length.
Results indicate that revetments, naturally resistant bank materials, and
flow regulation restrict migration and channel movement along the modern
Willamette River. Efforts aiming to increase lateral migration on the
Willamette River might consider removing revetment from bends bordered by
Holocene alluvium along higher-gradient areas of the floodplain. However,
such efforts may not create the suite of floodplain dynamics displayed by the
historic Willamette, as much of the rapid migration, side channel maintenance
and avulsions were related to flooding, channel change along adjacent bends,
and large wood; all of which are largely absent from the modern floodplain.
|
|---|