MONITORING
AMPHIBIANS, TEMPERATURE, CHANNEL STRUCTURE, AND WOOD IN 10 HEADWATER STREAMS
IN THE BLUE RIVER LANDSCAPE STUDY
Matthew
G. Hunter, USFS PNW Research Station, 3200 SW Jefferson Way, Corvallis,
Oregon 97331, 541-758-7791, matt.hunter@orst.edu
20
April 2000
CHANNEL STRUCTURE
AND WOOD FUNCTION: YEAR ONE.
OVERVIEW
The
Blue River Landscape Study
The
Blue River Landscape Study (BRLS) is located in the Blue River watershed,
the core watershed of the Central Cascades Adaptive Management Area (Figure
1). The BRLS is implementing and monitoring a landscape management approach
that uses spatial and temporal aspects of disturbance history as a general
template for future management activities. The overall goal of this approach
is to promote long-term native biodiversity and provide desired forest
products from the same landscape. Forest management in the past tended
to simplify forest structure, eliminate old forest characteristics, and
reduce fire effects. The BRLS attempts to incorporate diverse stand and
landscape structures, old forest components, and fire effects into a workable
landscape management approach. The basic premise guiding this approach
is that managing canopy structure and distribution (and other ecosystem
components and processes) in a manner similar to that expected from natural
disturbances will likely maintain environmental conditions to which native
species are adapted (Cissel 1997, Swanson et al. 1993).
Figure
1. Location of Blue River Watershed.
The
BRLS includes several key elements which differ from a default application
of the Northwest Forest Plan (NFP)(USDA and USDI 1994). In upslope locations,
the NFP calls for 15% retention, whereas the BRLS prescribes 15%, 30%,
or 50%, depending on the location in the watershed. Near streams, the
NFP prescribes riparian reserves along all streams, which predominantly
function as no-harvest buffers. In the BRLS it is assumed that fire did
not regularly skip over Class III and Class IV streamside forests, and
some partial harvest is prescribed in these areas. However, site-specific
areas of concern and stream-bank trees are protected.
Monitoring
Approach
The
management practice of most interest in regard to stream ecosystems was
the harvest or partial harvest of streamside forests near Class IV (intermittent)
and Class III (perennial without fish) streams. Therefore, streams of
this type within harvest units were priority to monitor. We initiated
monitoring in 3 streams in each of 3 timber sale areas. In each timber
sale area, we chose one stream with a prescribed buffer, one with no buffer
prescribed, and one "control." The buffer treatment is a 15-
to 20-m no-harvest buffer around the entire length of stream within the
harvest unit. The no-buffer treatment does not have a no-harvest buffer,
but will have increased canopy retention near the stream, compared to
upslope, and no streambank trees will be cut. The "control"
units are nearby streams of similar size and character where no harvest
is expected to occur.
This
approach is patterned after an experimental block design in which several
treatments are replicated in multiple blocks merely to allow more structured
and intelligent comparison among individual streams and among regions
in the watershed. Because of the low rate of harvest in the watershed,
and the limited number of harvest units containing streams, nearly the
entire population of interest currently is being monitored. Therefore,
sampling statistics which test differences among mean values in treatments
and blocks will not be used. Means will be calculated directly and meaningful
differences assessed directly among treatments and blocks. Sampling statistics
may be used to assess differences among individual streams, and time series
analyses may be used to assess trends over time among streams, treatments,
and blocks.
Monitoring
Status
Monitoring
of stream amphibians and water temperature was begun in 1998 in the Blue
River Face (BRF)(Figure 2) and North Fork Quartz (NFQ)(Figure
3) Timber Sale areas. In 1999, amphibian and water temperature monitoring
was continued in these areas and expanded to a proposed timber sale area
(Wolf-Mann [WM]). Also in 1999, channel structure and wood function monitoring
was initiated in all three timber sale areas, in order to provide information
potentially helpful in interpreting long-term changes in amphibian populations
and water temperature in these streams.
As
of spring 2000, all units have been treated in the North Fork Quartz (NFQ)
Timber Sale area, none have been treated in the Blue River Face (BRF)
Timber Sale area, and the Wolf-Mann (WM) Timber Sale has not yet been
put up for bid. Thus, all data collected to date are pre-treatment. Data
collected in 2000 will represent the first year of post-treatment data
collection in the NFQ area, but will still be pre-treatment for the BRF
and WM areas.
Figure
2. Measurement areas in the Blue River Face Timber Sale area. All 2-m
amphibian search areas are identified by the dots along streams. Temperature
sensors were located approximately near the upper 1st, 3rd,
and 5th dots. Channel structure and wood function transects
ran from the 1st to the 5th uppermost dots.
Figure
3. Measurement areas in the North Fork Quartz Timber Sale area. All 2-m
amphibian search areas are identified by the dots along streams. Temperature
sensors were located approximately near the upper 1st, 3rd,
5th, and bottom dots. Channel structure and wood function transects
ran from the 1st to the 5th uppermost dots.
Figure
4. Measurement areas in the Wolf-Mann Timber Sale area. All 2-m amphibian
search areas are identified by the dots along streams. Temperature sensors
were located approximately near the upper 1st, 3rd,
and 5th dots. Channel structure and wood function transects
ran from the 1st to the 5th uppermost dots.
STREAM AMPHIBIANS: YEAR TWO
OBJECTIVES
The
objective of this monitoring effort is to generate information necessary
to determine the effectiveness of the BRLS in meeting Aquatic Conservation
Strategy objectives, and to test key assumptions in the BRLS regarding
the response of stream amphibians to prescribed streamside treatments.
The most relevant objective from the Northwest Forest Plan Aquatic Conservation
Strategy is objective #9: "Maintain and restore habitat to support
well-distributed populations of native plant, invertebrate, and vertebrate
riparian-dependent species." This objective is non-quantitative and
the extent and resolution undefined, therefore some interpretation was
necessary to develop a monitoring approach that would allow evaluation
of this objective. We take this objective to apply at the extent of the
Blue River watershed, and at a resolution of 1st- to 3rd-order
basins. A key assumption of the BRLS is that prescribed harvest activities
may temporarily reduce local populations of stream amphibians, but that
these populations are expected to recover as the surrounding forest re-establishes
(Cissel 1997, p 60-63). Furthermore, some temporary decrease in local
populations is deemed acceptable as long as the overall watershed population
remains relatively stable. Therefore, our investigation will attempt to
quantify the temporal effects of harvest prescriptions in the BRLS to
aquatic amphibians in 1st-order streams.
METHODS
Sampling
Within a Treatment
At
each harvest unit and "control" 10, 2-m amphibian search areas
were placed along study streams. Five were within the harvest unit, and
five were downstream of the harvest unit. Search areas were placed approximately
20-50 m apart, depending on the length of stream available within the
harvest unit or below it (Figures 2-4), and were randomly located as possible
by the extent of perennial flow.
Detection
of Amphibians
A
"light-touch" method was used to search for amphibians in these
streams, in order to maintain stream integrity for future years. All potential
cover particles that were not key to maintenance of the water level in
the channel unit being searched were temporarily moved or removed. Therefore,
small steps or debris jams were not disassembled, and small pools located
above these structures were not drained. Instead, in such cases, sticks
and twigs were used to probe into all potentially inhabited interstitial
space.
Data
Collection and Management
A
variety of environmental data were collected at each amphibian sample
point, including average stream width, gradient, substrate composition,
canopy cover, and water temperature. Each captured amphibian was identified
to species, age and sex (when possible), and measured (snout-to-vent length
[SVL], and total length [TL]). Data from this project are stored in the
Forest Science Data Bank, Oregon State University. Metadata are viewable
on the web at http://sequoia.fsl.orst.edu/lter/data/studies/we022/we022fmt.htm.
RESULTS
AND DISCUSSION
Habitat
Sample
areas in the BRF and NFQ TS areas remained similar in depth, width, channel
unit composition, and vegetation cover (Table
1). The BRF area remained a lower gradient, slightly higher water
and air temperatures during the survey, more fine particles (silt-litter),
and less medium-sized particles (pebble). There was a general, slight
increase in water depth and a slight decrease in width among all streams.
This could be due to channel structure change, or simply a lowering of
the water level which concentrated surface water over the deepest portions
of the channel. There was also a general increase in percent pool morphology
and a decrease in percent glide, which may also account for increases
in water depth.
Sample areas in the WM TS area were most similar to BRF in substrate composition, having a high average content of fines, and aspect, mainly west-facing; most similar to NFQ in gradient (fairly high on average); but narrower on average than both NFQ and BRF, with greater consistency of surface water present, and a greater percentage of riffle (Table 1).
Amphibians
Species
composition remained notably consistent within individual streams among
years; therefore, differences in species composition among NFQ and BRF
TS areas were consistent among years as well (Tables
2-4). Three species were again found in the NFQ TS area, while only
the Pacific giant salamander was found in the BRF TS streams. In addition,
within the NFQ TS streams, one stream (2e, buffer) had an unusual species
composition: high densities of Cascade torrent salamander larvae, metamorphosed
tailed frogs, and no Pacific giant salamander larvae. In contrast, the
control and no-buffer sites had rather typical numbers of Pacific giant
salamander larvae and Cascade torrent salamander larvae, and no tailed
frogs. This pattern was present both years. Species composition of the
WM sites was similar to BRF streams, consisting solely of Pacific giant
salamander larvae.
Overall,
numbers of amphibians located in 1999 were about 2/3 of 1998 (Table
2); densities were about 1/2 of what they were in 1998 (Table
3); and occupancy of 2-m reaches were about 2/3 of 1998 (Table
4). Reasons for this overall decline, pre-treatment, are not known.
Some of the more likely hypotheses are: 1) 1999 was a drier year than
1998 meaning perhaps fewer animals at the surface, 2) negative impacts
of search effort on animal behavior and suitability of hiding substrate,
3) regular population fluctuation. Hypothesis #1 could certainly have
contributed, as average widths of search areas in 1999 were about 3/4
of what they were in 1998. Hypothesis #2 could be a factor, but there
is almost no way to support or refute it without additional seasons of
data that might support or eliminate other hypotheses. Hypothesis #3 seems
justified only for Pacific giant salamander larvae in NFQ TS area, where
it appears that predominantly a single cohort exists from breeding activity
that took place probably in fall 1997, and straightforward mortality from
1998 to 1999 could account for the changes in numbers there. However,
since all species displayed decreased numbers in all streams, it seems
that hypotheses #1 or #2 are more likely explanations.
If
it were not for knowledge of the 1998 size distribution data for Pacific
giant salamander larvae (Fig. 5), size distributions for 1999 (Fig.
6) may have just looked like a jumbled mess. However, some interesting
observations can be made. First, look at the difference in size distributions
among years for NFQ. In 1998, nearly all individuals were packed tightly
into a single cohort centered on about 30-31 mm SVL (Fig. 5). In 1999,
almost no individuals were found in the range displayed for 1998, but
the distribution has shifted (albeit more fragmented, with fewer individuals)
about 15 mm to the right (Fig. 6). I suspect that
the 1999 individuals are what remains of the 1998 cohort, which were at
that time newly-emerged larvae. It appears, then, that breeding did not
take place in NFQ streams in fall 1998 (which would have produced new
hatchlings in summer 1999). I expect that in the next year or two we may
see another cohort of new hatchlings in one or more of the NFQ streams.
It is interesting that the cohort size appears synchronous for the three
different streams in the NFQ area for both years. In contrast, BRF streams
exhibited their typical scattered size distribution pattern. The WM size
distributions are difficult to interpret because so few individuals were
found. My best guess is that two distinct cohorts of young larvae were
present, plus the one neotene that likely was responsible for some or
all of the young in the stream (Fig. 6). See the
report from the 1998 season (Hunter 1999a) for further discussion of potential
interpretations and implications of size-class distributions in these
streams.
Figure 5. Size distribution of Pacific giant salamander larvae in the Blue River Face and North Fork Quartz Timber Sale areas in 1998 (a single individual in the BRF TS area at 137 mm is not shown).
Figure 6. Size distribution of Pacific giant salamander larvae in the Blue River Face, North Fork Quartz, and Wolf-Mann Timber Sale areas in 1999 (one in the BRF TS area at 137 mm, and one in WM at 127 mm, are not shown).
STREAM TEMPERATURE: YEAR TWO
[Dave Kretzing
and/or Mike Cobb doing this?]
CHANNEL STRUCTURE AND WOOD FUNCTION: YEAR ONE
BACKGROUND
AND OBJECTIVES
The
channel structure and wood function components of this monitoring effort
were initiated to provide information that could potentially assist in
interpretation of stream amphibian and water temperature monitoring results.
Bedrock streambeds in headwater stream channels are typically areas of
non-habitat for stream amphibians and amplified warming for stream water
(especially when combined with direct solar exposure). Therefore, a measure
of bedrock exposure in the stream channel may help explain future trends
in amphibian populations and stream temperatures. Similarly, but in an
opposite fashion to amount of bedrock, wood in headwater stream channels
often creates complex morphologies and hyporheic flow (especially via
step structures), which increase amphibian habitat and potential for cooling
of stream water. Step structures, often created by wood and accumulated
bed material in the active channel, may be especially important refugia
for stream amphibians. Therefore, amount of wood and a specific measure
of step morphology are additional measures that may help interpret future
trends in amphibian populations and water temperature.
METHODS
Establishment of transect locations
Transects
were established over nearly the entire stream length within each harvest
unit (or a comparable length on nearby control streams). Each transect
began at the 5th amphibian search area down from the top of
the stream, and extended up to include at least the lower end of the uppermost
amphibian search area (Figs. 2-4). For control
streams, the 1st and 5th uppermost amphibian search
areas were used as endpoints for channel monitoring transects.
Data
Collection and Management
Measurements of channel structure and wood function
were taken in 5-m intervals along the transect. Two wooden staffs connected
with a relatively non-elastic 5-m chord were used to designate consecutive
5-m measurement areas, from downstream to upstream. Each 5-m measurement
area was not marked in the field. In all cases, each of 5 amphibian search areas (identified by a PVC
stake) was noted when the PVC marker fell within a 5-m measurement area,
to allow comparisons of transect interval locations among years.
Three
sets of data were collected: channel structure, wood tallies, and step
structures. Channel structure measurements included active channel width,
active channel depth, gradient, azimuth, and percent of the active channel
composed of bedrock. Wood tallies involved tallying 10-cm diameter classes
of all wood pieces >2 cm diameter in two location categories: vertically
over the active channel and within the active channel. Step structures
>=30 cm in height were classified based on the key piece(s) forming
the step (rock, wood, or both), the height measured in 10-cm classes,
and if a wood piece was judged as a key piece forming the step, the diameter
class of the piece of wood was recorded.
Metadata
will be viewable by June 2000 on the web at http://sequoia.fsl.orst.edu/lter/data/studies/[????].
Analysis
Initial
years will simply involve summary statistics of results. In 3-5 years,
analyses will include a visual display of percent bedrock over years,
displays of wood amounts, size of steps, and comparisons of contributions
of wood versus rock particles in creating steps in small streams. It will
also be possible to map changes in bedrock and log configurations in channels
within each transect over time.
RESULTS
AND DISCUSSION
Eight
of 10 transects consisted of 36-44, 5-m measurement areas (Table
5). Two transects were substantially shorter, with only 14 and 22
measurement areas. The former was due to a relatively short length of
stream present within the harvest unit, and the latter was simply a result
of the choice of amphibian search areas in this control stream. Active
channel widths were greatest in NFQ streams, smallest on average in WM
streams, with BRF streams in between, but closer to WM streams in size.
Active channel depths were remarkably consistent among all streams with
the exception of the NFQ control stream which was about twice as deep
as other channels. Slopes of NFQ and WM streams were nearly identical
on average, while BRF streams were less than half as steep as other streams.
Bedrock was absent in BRF streams, and present in a majority of NFQ and
WM streams. Bedrock covered substantial portions of the streambed only
in one NFQ and one WM stream, at 33% and 18% respectively. Number of steps
(>=30 cm in height) per 100 m of transect length varied substantially
(1.5 to 18.5) within and among timber sale areas. Percent of steps which
included wood as a key piece in the step formation was consistently high,
but varied from 68% to 100%. Interestingly, the streams with the highest
number of steps had the lowest proportions of steps formed by wood, while
streams with few steps had most or all steps with wood as a key piece.
Most
channel steps had key wood pieces which were fairly small. About 41% of
wood or wood-rock steps >=30 cm in height had key wood pieces 2-<10
cm in diameter, and 91% had key wood pieces <40 cm diameter (Table
6). Taller steps were notÝ necessarily
created by larger logs, but often by a combination of key rock pieces
and small to medium-sized key wood pieces. In fact, logs >=70 cm diameter
(class >=7), were not found as key pieces in any steps (Tables
6, 7). About 30% of all steps >=30 cm in height had a rock as the
key piece either in combination with wood (11%) or without a key wood
piece (19%).
Proportions
of wood pieces in each diameter class were nearly identical for tallies
over and within the active channel (Table
7). Also, quite consistently twice as many pieces were tallied within
as over the active channel, suggesting no particular tendancy for recruitment
of a particular diameter class of wood into the stream channel. The percent
of tallied wood piecesÝ in each
diameter class functioning as key pieces in channel steps (Table
7) seems rather variable and no obvious trends are apparent. What
is surprising though is the absence of large-diameter (e.g. >50-cm)
logs as key pieces in channel steps. One might hypothesize that large
logs did not contribute substantially to existing channel structure in
these streams because large-diameter logs were more often suspended above
the active channel, rather than present within the channel. However, the
proportions of pieces above vs. within the active channel was fairly consistent
at about 1:2 for each size class. Another hypothesis is that large logs
with a diameter 2-3 times that of the average active channel depth (depth),
simply may not often break or lay in such a way as to block the whole
channel; instead they may lay longitudinally in the stream, or one end
of the log present in the channel may act as a channel bank or impediment
that the stream routes around rather than flowing over. Lastly, there
simply may have been so few logs that just by chance none were positioned
properly to function as key pieces in channel steps.
While
large-diameter logs rarely contribute to instream structure in these streams,
they likely provide a much more stable step structure than the smaller
materials. The relative importance of many less-stable step structures
vs. a few very stable step structures is unknown. Even the many less-stable
step structures may be stable long enough to provide adequate hiding or
nesting habitat for amphibians. In contrast, step stability or duration
probably has little or no effect on water temperature, the number and
function of these in promoting hyporheic flow likely being of most importance.
While
pieces were not identified in the field as boles or branches, my observations
suggest that nearly all of the 2- to <10-cm diameter wood pieces tallied
were tree branches, pieces in the 10- to <20-cm diameter class, were
a mixture of branches and small boles, and pieces >=20 cm were almost
exclusively boles. Most Pacific Northwest literature on wood function
in streams addresses pieces >10 cm in diameter, which are are often
tree boles.
LITERATURE CITED
Cissel, J. H. 1997. The Blue River landscape project: landscape
management and monitoring strategy. Unpublished document on file at the
Blue River Ranger District, Blue River, Oregon, 97413, 541-822-3317.
Hunter, Matthew G. 1999a. Blue
River landscape study stream amphibian monitoring: pre-harvest results,
first year. Unpublished report on file at the Blue River Ranger District,
Blue River, Oregon, 97413, 541-822-3317.
Swanson, F.J., J.A. Jones, D.O. Wallin, and J.H. Cissel.
1993. Natural variabilityóimplications for ecosystem management. In: Jensen,
M.E. and P.S. Bourgeron (eds), eastside forest health assessmentóvolume
II: ecosystem management: principles and applications. Portland, OR: USDA
Forest Service, Pacific Northwest Research Station: 89-103.
USDA Forest Service, and USDI Bureau of Land Management.
1994. Record of Decision for amendments to Forest Service and Bureau of
Land Management planning documents within the range of the northern spotted
owl. USDA For. Serv. and USDI Bur. Land Manage. 74p.