Durability of Pacific Northwest instream structures following floods

Abstract

The durability of 3,946 instream structures in 94 streams that had floods with return intervals exceeding 5 years were assessed. Overall structure durability (defined as the degree to which a structure remained at its original location) was high; less than 20% of the sampled structures had been removed from the site of original placement. The magnitude of flood events had a significant effect on structure durability with higher magnitude floods reducing durability. Stream order also affected structure durability; structures in large streams were 20 times more likely to have been removed from the site of original placement than structures in small streams. Other conditions that affected structure durability included location of the structure within the stream channel, whether the structure was anchored or not, structure material, and upslope landslide frequency. Instream structures are most appropriate when used as short-term tools to improve degraded stream conditions while activities that caused the habitat degradation are simultaneously modified. When instream structures are part of a properly sequenced watershed restoration strategy, they can improve habitat conditions through a range of flow conditions including large floods.

Full text

686
North American Journal of Fisheries Management 18:686–693, 1998
American Fisheries Society 1998
Durability of Pacific Northwest Instream Structures
Following Floods
B
RETT
B. R
OPER
*
U.S. Forest Service, Idaho Panhandle National Forest
3815 Schreiber Way, Coeur d’Alene, Idaho 83815, USA
D
EBORAH
K
ONNOFF AND
D
AVE
H
ELLER
U.S. Forest Service, Pacific Northwest Region Office
333 Southwest 1st Avenue, Portland, Oregon 97208, USA
K
EN
W
IEMAN
U.S. Forest Service, Wind River Ranger District
M.P. 1.23 R Hemlock Road, Carson, Washington 98610, USA
Abstract.—The durability of 3,946 instream structures in 94 streams that had floods with return
intervals exceeding 5 years were assessed. Overall structure durability (defined as the degree to
which a structure remained at its original location) was high; less than 20% of the sampled structures
had been removed from the site of original placement. The magnitude of flood events had a
significant effect on structure durability with higher magnitude floods reducing durability. Stream
order also affected structure durability; structures in large streams were 20 times more likely to
have been removed from the site of original placement than structures in small streams. Other
conditions that affected structure durability included location of the structure within the stream
channel, whether the structure was anchored or not, structure material, and upslope landslide
frequency. Instream structures are most appropriate when used as short-term tools to improve
degraded stream conditions while activities that caused the habitat degradation are simultaneously
modified. When instream structures are part of a properly sequenced watershed restorationstrategy,
they can improve habitat conditions through a range of flow conditions including large floods.
Instream structures, an important and estab-
lished component of stream restoration projects,
have been used throughout the world (Jungwirth
et al. 1995) to alter stream characteristics and im-
prove stream conditions for fish populations
(House and Boehne 1986; Hunter 1991; Rosgen
1996). Although widespread tactical reliance on
instream structures has been criticized recently be-
cause of the lack of strategic planning involved in
their placement and use (National Research Coun-
cil 1992; Doppelt et al. 1993; Roper et al. 1997),
instream structures continue to be an important
component of watershed restoration.
Use of instream structures to restore watersheds
has been prevalent in stream systems of the Pacific
Northwest due to three interacting factors: (1) the
relationship between degraded stream habitats and
the precipitous decline of anadromous salmon pop-
ulations (Nelhsen et al. 1991), (2) better under-
standing of the role that instream large organic
debris plays in maintaining stream habitat condi-
tions and fish populations (Bisson et al. 1987;
Fetherston et al. 1995), and (3) availability of mon-
* Corresponding author: broper/rl
p
ipnf@fs.fed.us
ies generated from harvesting timber on federal
lands to pay for restoration projects. In the past
15 years, federal agencies have placed thousands
of structures to mitigate effects of timber harvests
on streams and their biota and to improve fish
habitat. Most of these structures consist of large
logs (
.
30 cm diameter) or boulders placed by
heavy machinery in stream channels.
Although numerous studies have shown that in-
stream structures increase specific stream biota
(e.g., House and Boehne 1986; Fontaine 1987),
only a limited number of studies have examined
the physical durability of structures (Frissell and
Nawa 1992; House 1996). Structure durability is
an important component of assessing structure ef-
ficiency. Although a structure may increase fish
production in the short term, if it has a short life
span (
,
5 years), its overall benefit may be mini-
mal. Conclusions of past structure durability stud-
ies have been mixed, but the general perception is
that instream structures are not durable through
flood events (Frissell and Nawa 1992). The small
scale of previous studies combined with absence
of data from structures experiencing high-magnitude
floods (
.
25-year return interval) has limited abil-

687
DURABILITY OF INSTREAM STRUCTURES
ity to predict durability of instream structures dur-
ing floods.
As part of a broad, multiphased assessment of
the effects of the 1995–1996 floods on U.S. Forest
Service (USFS) lands in the Pacific Northwest, we
evaluated instream structure durability with three
specific objectives: (1) determine the overall du-
rability of instream structures following floods
with 5–150-year return intervals, (2) relate dura-
bility to geomorphic and stream conditions, and
(3) provide recommendations to improve future
performance of structural instream restoration
treatments.
Methods
During the winter of 1995–1996 streams
throughout much of the Pacific Northwest expe-
rienced two major storms producing floods that
tested the durability of instream restoration projects
constructed by the USFS in Oregon and Washing-
ton. Flood intensities varied across the region,
ranging from events that could be expected to re-
turn once every 5 years to floods that could be
expected to return less than once every 150 years.
The flood magnitude experienced within a water-
shed depended upon proximity to storm cells, el-
evation, and intensity of precipitation. Runoff was
greatest where high-intensity rainfall occurred on
deep snow in mid- and low-elevation zones, re-
sulting in rain-on-snow events (see Swanston
1991).
Following these floods a survey was conducted
on USFS lands to assess the durability of instream
structures in the flood-affected area. Streams with-
in the Puget Sound, Yakima, Lower Columbia,
Middle Columbia, Willamette, Lower Snake, and
the coast range regions of Washington and Oregon
were surveyed. The protocol classified durability
of instream structures into one of three movement
categories: (1) in place, (2) shifted on site, and (3)
removed from site. A structure was categorized as
in place if the majority (
.
50%) of the structural
units (logs, rocks, etc.) remained as placed (within
practical limits of determining such). For example,
for a structure consisting of a single log to be
categorized as in place, the entire log had to have
maintained its exact, original location and orien-
tation. However, if the structure consisted ofthree
logs cabled together, the structurewas categorized
as in place provided two of the logs maintained
their original location and orientation. For a struc-
ture to have been classified as shifted on site, the
majority of structural components were still ba-
sically on site but had shifted in orientation during
the flood. A structure was determined to have been
removed from the site if the majority of the struc-
tural units were no longer on the original location.
Surveyors used original project plans and ob-
served physical evidence (i.e., broken cables, shat-
tered logs) to determine structural movements. A
day of training ensured consistent application of
the survey protocol among the surveyors.
The survey protocol also measured a variety of
basin characteristics and structural attributes
thought to affect structure movement. These data
included flood magnitude, stream order, location
of the structure within the stream channel, stream
channel constrainment, structure material, and
whether the structure was anchored. In addition,
the relative upslope landslide frequency associated
with floods was determined as part of the regional
flood assessment (conducted by Pacific Northwest
Region 6 USFS 1996).
To determine effects through a range of stream
flow conditions, streams with floods ranging from
5- to 150-year return intervals were included in
the sample. Because some sampled streams were
not gauged, the magnitude of the flood in each
watershed was determined by extrapolating flow
data from the nearest U.S. Geological Survey
stream gauge. Because of the lack of precision in
estimating the true magnitude of the discharge us-
ing this method, we stratified streams based on
four broad return-interval categories (
,
15, 15–39,
40–64, and
.
65 years).
Stream order was determined following Strahler
(1957). Map scale for stream order determination
was 1:24,000. Structure location within the stream
channel was described using three categories: (1)
edge—structure built on or associated with one of
the banks of the stream, (2) cross channel—full
spanning structure associated with both banks, and
(3) in-channel—structure not associated with ei-
ther stream bank. Channels were categorized as
either constrained or unconstrained (Rosgen
1996). Four different categories of structural ma-
terial were used: (1) log, (2) boulder, (3) log and
boulder, and (4) gabion. Analysis of structure ma-
terial was limited to the first three groups because
only five gabion structures were evaluated.
Landslide frequency was determined by USFS
aerial observers at a watershed scale (10,000–
20,000 hectares). Three groupings were used to
describe landslide frequency for watersheds: (1)
high—landslide rates in the 90.1–100th percentile,
(2) medium—landslide rates in the 80.1–90th per-

688
ROPER ET AL.
F
IGURE
1.—The relationship between flood magnitude and structure durability (yr is year).
centile, and (3) landslide rates in the 0–80th per-
centile.
The simple format of the survey protocol (lim-
ited categories), combined with surveyor training,
probably resulted in low observer variability
(Roper and Scarnecchia 1995). Independent vari-
ables related to structure durability were examined
by using chi-square goodness-of-fit test (Dowdy
and Wearden 1983). The null hypotheses—that
flood magnitude, stream order, structure location,
structure material, structure anchoring, or land-
slide frequency were independent of structure du-
rability—were tested at
a5
0.05 for all compar-
isons.
Results
Initial surveys included about 4,000 structures
in more than 100 streams, but because of incom-
plete records (i.e., no discharge estimates, multiple
entries in the same field of the database), the data
set used for analysis consisted of 3,946 structures
in 94 streams. Overall durability was high; less
than 20% of the 3,946 instream structures were
removed from site following floods exceeding a 5-
year return interval.
Flood magnitude had a significant (
x
2
5
227.2,
df
5
6, P
,
0.01) effect on structure durability.
Higher flood magnitudes resulted in a higher per-
centage of structures being removed from the site
of original placement (Figure 1). Less than 15%
of the structures were removed from the site of
placement in floods with return intervals less than
65 years. In streams where floods exceeded a 64-
year return interval, structures were almost twice
as likely (25%) to have been removed from the
site of original placement than those that experi-
enced lower intensity floods.
Stream order also significantly affected structure
durability (
x
2
5
315.9, df
5
10, P
,
0.01). Move-
ment of structures was less in low-order streams
than in high-order streams (Figure 2). Structures
in higher order streams were up to 20 times more
likely to have been removed from the site of orig-
inal placement than in lower order streams; in sixth
order streams 63% were removed (N
5
72), where-
as in first order 3% were removed (N
5
35). In
third and fourth order streams, which had the
greatest representation in our survey (N
5
3,054,
77% of the structures sampled), only 13% of the
structures evaluated had been removed from site.
In addition to being affected independently by
both flood magnitude and stream order, structure
durability was also affected by the interaction be-
tween these two variables. When analysis was lim-
ited to second- through fifth-order streams (be-
cause N
.
100 in each group), we found that struc-
ture durability was more affected by flood mag-
nitude in higher order streams than in smaller order
streams (Table 1). Structures were twice as likely
to have been removed from site in a fifth-order

689
DURABILITY OF INSTREAM STRUCTURES
F
IGURE
2.—The relationship between stream order and structure durability.
than a second-order stream when flood magnitude
was less than a 40-year return interval. In contrast,
when flood return intervals exceeded 40 years,
structures in a fifth-order stream were four times
more likely to have been removed from site than
comparable structures in second-order streams.
Both the location of a structure within the chan-
nel and the interaction between structure location
and stream channel constrainment affected struc-
ture movement. Structures not connected to the
channel edge were 50% more likely to have been
removed from the site than structures placed either
on the channel edge or that spanned the entire
channel (
x
2
5
138.9, df
5
4, P
,
0.01; Figure 3).
Channel constrainment had little overall effect on
whether a structure remained in place or not; in
constrained channels (N
5
1,540) 65% of the struc-
tures remained in place, and in unconstrained
channels (N
5
2,332) 61% of the structures re-
mained in place. However, the interaction between
structure location and channel constrainment did
affect structure movement; 58% of structures not
attached to the stream bank in constrained chan-
nels (N
5
208) remained in place, whereas 33%
of the structures not attached to stream banks in
unconstrained channels (N
5
377) remained in
place (
x
2
5
239.9, df
5
2, P
,
0.01).
Structure material and whether the structure was
anchored also had a significant effect on structure
durability. Structures made of logs or boulders
were more likely to have remained in place (67%)
than those made of a combination of logs and boul-
ders (57%) (
x
2
5
73.0, df
5
4, P
,
0.01). Simi-
larly, anchored structures were more durable(15%
were removed) than structures not anchored (22%
were removed;
x
2
5
19.3, df
5
2, P
,
0.01).
Durability of instream structures was related to
upslope landslide frequencies. Structures were less
durable (
x
2
5
103.0, df
5
4, P
,
0.01) in the 20%
of the sub-basins having the highest landslide fre-
quencies (Figure 4). Structures in the basin with
the highest landslide frequencies (top 10%) were
almost three times more likely to have been re-
moved from site (28% removed) than structures in
basins with low landslide frequencies (11% re-
moved).
Discussion
Instream structures sampled during this study
remained in place throughout a range of flood mag-
nitudes. These findings were surprising because
project plans for many instream structures pro-
jected structure failure in a 20-year or greater flood
(House and Crispin 1990) and because of the low
structure durability observed by Frissell and Nawa
(1992). There are several explanations for the dif-
ferences between this study and the work of Fris-
sell and Nawa (1992).
One explanation is geographic variation in stan-
dardized (by basin area) peak flows. Frissell and
Nawa worked in southwestern Oregon where
streams have higher standardized peak flows than
most of the basins we surveyed (see Frissell and
Nawa 1992: Figure 5). They reported that higher

690
ROPER ET AL.
T
ABLE
1.—The proportion (percentage) of structures in
each movement category stratified by stream order and
flood frequency.
Stream
order Sample
size
Flood
frequency
(years)
Structure movement
category
In
place
(%)
Shifted
on site
(%) Removed
(%)
2
2
3
3
4
4
5
5
176
157
794
702
711
847
118
334
,
40
$
40
,
40
$
40
,
40
$
40
,
40
$
40
83
70
70
64
76
51
60
41
8
20
21
19
15
31
20
17
9
10
9
17
9
18
20
42
F
IGURE
3.—The relationship between placement of a structure and structure durability.
standardized peak flows resulted in lower structure
durability, and we observed our lowest structure
durability on streams closest to their study area.
Although we did find that structures were less du-
rable in areas of higher standardized peak flow,
our data indicated that structures performed better
following large floods than was predicted by Fris-
sell and Nawa (1992). Although our results do not
fully support findings presented by Frissell and
Nawa (1992), they are consistent with many un-
published evaluations of structure durability (Ta-
ble 2).
A second reason for differences among studies
could be inconsistent use of terms describing struc-
ture durability. Frissell and Nawa (1992) defined
a structure as impaired if it ‘‘no longer functioned
in the intended mode or appeared to be at least
temporary ineffective.’’ Terms used in other as-
sessments included successful, outstanding,
moved and functioning, and minor faults in orig-
inal design (Table 2). If terms used to definestruc-
ture durability differ among studies, then a portion
of the variability in results of these studies will be
qualitative differences in what was judged to be a
successful or failed structure.
Many terms used to evaluate structure success
integrate two distinct questions regarding instream
structures: Were the structures durable? and Did
the structures meet design objectives? These two
questions may be mutually exclusive. For exam-
ple, a structure can remain in place but not create
the intended stream habitat. In contrast, a structure
could move 100 m downstream but create a de-
sirable deep scour pool. Information related to
both facets is important, so evaluation criteria for
both types of assessments should be quantitative
and repeatable among surveyors. Whereas struc-
ture durability can be easily categorized based on
movement, quantifying whether a structure met its
objective(s) is more difficult (Fontaine 1987). Fu-
ture evaluations of structure durability and deter-
mining whether they met their intended objec-
tive(s) will be easier with enhanced definitions de-
lineating movement and satisfaction of objectives
(Bryant 1995; Kondolf 1995; Hobbs and Norton
1996).
Our research and Frissell and Nawa’s (1992)
both found high variability of structure durability

691
DURABILITY OF INSTREAM STRUCTURES
F
IGURE
4.—The relationship between upslope landslide frequency and structure durability. Landslide frequency cat-
egorization was determined by a 1996 report by B. McCammon (USFS, Pacific Northwest Region, unpublished).
T
ABLE
2.—Estimates of structure durability (percentage)
presented in this and other studies. Categories for deter-
mining if a structure remained on its placement site are
listed in the table footnote.
Study
Stayed
on or
near
site
Re-
moved
from
site
Return
interval
of flood
(years)
This Study 84
a
16 150
1
Frissell and Nawa (1992) 57
b
43 10
Unthank (Willmette Nation-
al Forest)
c
97
d
3 Not recorded
Doyle (Mt. Baker–Snoqual-
mie National Forest)
c
87
e
13 5
Fitch et al. (1994) 64
f
36 Not recorded
Higgins and Forsgren (Mt
Hood National Forest)
c
96
g
420
a
In-place and shifted-on-site.
b
Success and impaired.
c
Monitoring reports (USFS, Pacific Northwest Region, Portland
Oregon).
d
In place and functioning, moved and functioning, and inplace and
not functioning.
e
In-place, moved, and buried.
f
Retained original design and minor faults in original design.
g
Fully functional, damaged, and not functional.
among streams. In some streams all structures re-
mained in place, while in others most structures
were removed from the location they were placed.
Cases of extensive damage and displacement of
structures probably produced the fuel for mass me-
dia articles on the futility of instream structures
(see Anonymous 1996). Although instream struc-
tures moved en masse during large floods, streams
where these events occurred were comparatively
infrequent (
,
15% of the streams we sampled);
nevertheless, these failures are problematic if they
cause resource or property damage.
This study indicated that structure durability is
affected by many stream characteristics and en-
gineering elements. We found structures became
less durable as stream energy increased, as did
Frissell and Nawa (1992). This association was
evident in the decreasing durability of structures
among low-order compared to high-order streams
and for more frequent floods (
,
15-year event)
compared to less frequent floods (
.
40-yearevent).
Structures designed with connections to the stream
bank had greater durability. In addition, structures
could be less durable in unconstrained stream
channels (Kauffman et al. 1997). Consequently,
we suggest that durability is probably greatest for
instream structures used in constrained reaches of
small- to moderate-sized streams (fourth order and
smaller) and placed with a connection(s) to the
stream bank. However, fine-scale variables not
evaluated by us also affect the performance of in-
dividual structures (Fitch et al. 1994).
A factor we found affecting instream structure
durability that was not explored in other studies
was a basin’s landslide frequency. Loss of struc-
tures in this study was sometimes initiated by up-
slope debris torrents and landslides that carried
into the main channel. Consequently, landslide fre-
quencies influence durability of structures. Be-
cause high landslide frequencies are often related
to land management activities conducted in areas
of naturally unstable terrain (Furniss et al. 1991),
stream systems in managed basins may be less

692
ROPER ET AL.
stable (Jones and Grant 1996). As a result, in-
stream structures in managed watersheds with nat-
urally occurring unstable terrain are less likely to
be durable.
The relationship between instream structure du-
rability and upslope variables indicates the im-
portance of a basinwide perspective when imple-
menting stream restoration projects (Hartman et
al. 1996; Roper et al. 1997). Although instream
structures were durable through floods, instream
structures alone cannot restore watersheds where
landscape processes have been altered by land
management activities (Kauffman et al. 1997). We
contend, as did Frissell and Nawa (1992) and oth-
ers (National Resource Council 1992; Doppelt et
al. 1993; Roper et al. 1997), that watershed res-
toration can only be achieved by rectifying upslope
conditions that degrade channel conditions. In
many cases, instream structures placed prior to or
in the absence of stabilization of hill slopes and
reestablishment of riparian zones will fail to meet
objectives of restoring stream conditions (USFS
1993; Hartman et al. 1996; Kauffman et al. 1997).
Instream structures are most appropriate when
used as short-term tools to improve degraded
stream conditions while activities that caused hab-
itat degradation are simultaneously modified
(House 1996). When instream structures are part
of a properly sequenced watershed restoration
strategy they can improve habitat conditions
through a range of flow conditions including large
floods.
Acknowledgments
We thank J. Anderson, C. Clifton, J. Doyle, S.
Kozlowski, K. MacDonald, B. McCammon, R.
Metzger, J. Moreau, and A. Unthank for their role
in collecting the data used for this study as well
as the input they provided in revising the manu-
script. Earlier drafts of this manuscript were im-
proved significantly due to reviews provided by P.
Bisson, M. Clady, D. Hohler, B. House, and L.
Fitch. This study was funded by USFS through the
Pacific Northwest Region.
References
Anonymous. 1996. A place for fish to go. The Oregonian
(February 19):D8.
Bisson, P. A., and eight others. 1987. Large woody de-
bris in forested streams in the Pacific Northwest:
past, present and future. Pages 143–190 in E. Salo
and T. Cundy, editors. Streamsidemanagement: for-
estry and fishery interactions. University of Wash-
ington, Institute of Forest Resources, Contribution
57, Seattle.
Bryant, M. D. 1995. Pulsed monitoring for watershed
and stream restoration. Fisheries 20(11):6–13.
Doppelt, B., M. Scurlock, C. Frissell, and J. Karr. 1993.
Entering the watershed: a new approach to save
America’s river ecosystems. Island Press, Washing-
ton, D.C.
Dowdy, S., and S. Wearden. 1983. Statistics for re-
search. Wiley, New York.
Fetherston, K. L., R. J. Naiman, and R. E. Bilby. 1995.
Large woody debris, physical processes, and ripar-
ian forest development in montane river networks
of the Pacific Northwest. Geomorphology 13:133–
144.
Fitch, L., M. Miles, J. O’Neil, R. Pattenden, and G. Van
Der Vinne. 1994. Defining the variables that influ-
ence success of habitat structures in Southwestern
Alberta. In Proceeding of the 9th International Trout
Stream Habitat Improvement Workshop. Pennsyl-
vania Fish Commission, Harrisburg.
Fontaine, B. L. 1987. An evaluation of the effectiveness
of instream structures for steelhead trout rearing in
Steamboat Creek basin. Master’s thesis. Oregon
State University, Corvallis.
Frissell, C. A., and R. K. Nawa. 1992. Incidence and
causes of physical failure of artificial habitat struc-
tures in streams of Western Oregon and Washington.
North American Journal of Fisheries Management
12:182–197.
Furniss, M. J, T. D. Roelofs, and C. S. Yee. 1991. Road
construction and maintenance. Pages 297–323 in W.
R. Meehan, editor. Influences of forest and range-
land management on salmonid fishes and their hab-
itats. American Fisheries Society, Special Publica-
tion 19, Bethesda, Maryland.
Hartman, G. F., J. C. Scrivener, and M. J. Miles. 1996.
Impacts of logging in Carnation Creek, a high-energy
coastal stream in British Columbia, and their im-
plications for restoring fish habitat. Canadian Jour-
nal of Fisheries and Aquatic Sciences 53(Supple-
ment 1):237–251.
Hobbs, R. J., and D. A. Norton. 1996. Towards a con-
ceptual framework for restoration ecology. Resto-
ration Ecology 4:93–110.
House, R. 1996. An evaluation of stream restoration
structures in a coastal Oregon stream, 1981–1993.
North American Journal of Fisheries Management
16:272–281.
House, R., and V. Crispin. 1990. Economic analyses of
the value of large woody debris as salmonid habitat
in coastal Oregon streams. U.S. Bureau of Land
Management, Technical Note OR-7, 6512, Portland,
Oregon.
House, R. A., and P. L. Boehne. 1986. Effects of in-
stream structures on salmonid habitat and popula-
tions in Tobe Creek, Oregon.North American Jour-
nal of Fisheries Management 6:38–46.
Hunter, C. J. 1991. Better trout habitat: a guide to stream
restoration and management. Island Press, Wash-
ington, D.C.
Jones, J. A., and G. E. Grant. 1996. Peak flow responses
to clearcutting and roads in small and large basins,

693
DURABILITY OF INSTREAM STRUCTURES
western cascades, Oregon. Water Resources Re-
search 32:959–974.
Jungwirth, M. S., R. L. Peters II, and E. B. Allen. 1995.
Effects of recreated instream and ecotone structures
on fish and fauna of an epipotomal river. Hydro-
biologia 30:195–206.
Kauffman, J. B., R. L. Beschta, N. Otting, and D. Lytjen.
1997. An ecological perspective of riparian and
stream restoration in the western United States.
Fisheries 22(5):12–24.
Kondolf, G. M. 1995. Five elements for effective eval-
uation of stream restoration. Restoration Ecology3:
133–136.
National Research Council. 1992. Restoration of aquat-
ic ecosystems: science, technology, and public pol-
icy. National Academy Press, Washington, D.C.
Nehlsen, W., J. E. Williams, and J. A. Lichatowich.
1991. Pacific salmon at the crossroads: stocks at
risk from California, Oregon, Idaho, and Washing-
ton. Fisheries 16(2):2–21.
Roper, B. B., J. J. Dose, andJ. E. Williams. 1997. Stream
restoration: is fisheries biology enough? Fisheries
22(5):6–11.
Roper, B. B., and D. L. Scarnecchia. 1995. Observer
variability in classifying habitat types in stream sur-
veys. North American Journal of Fisheries Man-
agement 15:49–53.
Rosgen, D. 1996. Applied river hydrology. Wildland
Hydrology, Pagosa Springs, Colorado.
Strahler, A. N. 1957. Quantitative analysis of watershed
geomorphology. Transactions of the AmericanGeo-
physical Union 38:913–920.
Swanston, D. N. 1991. Natural process. Pages 139–179
in W. R. Meehan. Influences of forest andrangeland
management on salmonid fishes and their habitats.
American Fisheries Society, Special Publication 19,
Bethesda, Maryland.
USFS (U.S. Forest Service). 1993. Forest ecosystem
management: an ecological, economic, and social
assessment. U.S. Government Printing Office,
1993-793-071, Washington, D.C.
Received June 9, 1997
Accepted February 23, 1998