Diatom Assemblages and Their Associations With Environmental Variables in Oregon Coast Range Streams, USA
Diatom Assemblages and Their Associations With Environmental Variables in Oregon Coast Range Streams, USA
Diatom Assemblages and Their Associations With Environmental Variables in Oregon Coast Range Streams, USA
Springer 2006
R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe
DOI 10.1007/s10750-005-1615-1
Key words: periphyton, nonmetric multidimensional scaling (NMDS), cluster analysis, TWINSPAN, bioassessment,
Oregon Coast Range
Abstract
The Oregon Coast Range, rich in natural resources, is under increasing pressure from rapid development.
The purpose of this study was to examine diatom species patterns in relation to environmental variables in
streams of this region. Diatoms, water quality, physical habitat and watershed characteristics were assessed
for 33 randomly selected stream sites. Watershed size, elevation, geology, vegetation and stream morphology varied substantially among sites. Streams were characterized by dilute water chemistry and a low
percent of ne substrate. A total of 80 diatom taxa were identied. Taxa richness was low throughout the
region (median 15, range 1026). Assemblages were dominated by two adnate species, Achnanthidium
minutissimum and Achnanthes pyrenaicum. Diatoms sensitive to organic pollution dominated the assemblages at all sites (median 85%). Non-metric multidimensional scaling (NMDS) and correlational analysis
showed quantitative relationships between diatom assemblages and environmental variables. NMDS axes
were signicantly correlated with watershed area, watershed geology, conductivity, total nitrogen, total
solids and stream width. Diatom-based site classication (Two-way Indicators Species Analysis, (TWINSPAN)) yielded 4 discrete groups that displayed weak correlations with environmental variables. When
stream sites were classied by dominant watershed geology, overall diatom assemblages between groups
were signicantly dierent (Analysis of Similarity (ANOSIM) global R = 0.19, p < 0.05). Our results
suggest that streams in the coastal region are in relatively good condition. High natural variability in stream
conditions in the Oregon Coast Range ecoregion may obscure quantitative relationships between environmental variables and diatom assemblages. A bioassessment protocol that classies sites by major
landscape variables and selects streams along the major human disturbance gradient might allow for
detection of early signs of human disturbance in environmentally heterogeneous regions, such as the Pacic
Northwest.
Introduction
Streams are a distinctive feature in the Oregon
Coast Range ecoregion, with a typical density of
12 km of perennial streams per square km in the
mountains (Omernik & Gallant, 1986). The
declining status of stream biota, particularly salmonid sh, has been of interest in recent years,
208
the utility of stream periphyton assemblages is
only beginning to be explored (e.g., Carpenter &
Waite, 2000).
Periphyton assemblages have frequently been
used in stream bioassessment because they respond
rapidly to changes in stream conditions (Leland,
1995; Pan et al., 1996; 2000). Streams in the Coast
Range ecoregion have naturally dilute water
chemistry, with low alkalinity and nutrient levels,
due to regional bedrock (Welch et al., 1998).
Consequently, their periphyton may be highly
sensitive to anthropogenic disturbances and serve
as a good bioindicator of stream conditions. For
example, periphyton respond positively to nutrient
enrichment in mountain streams of British
Columbia (Perrin et al., 1987).
The major anthropogenic disturbances in the
Coast Range are activities associated with forest
management practices (McClain et al., 1998).
However, the watersheds vary widely in elevation,
size, geology and vegetation cover, leading to a
high natural variability in water chemistry and
physical habitat conditions. This high natural
variability in stream conditions may interfere with
the use of diatoms in bioassessment by obscuring
relationships between pollution or habitat alterations and species assemblages. The main objective
of this study was to characterize diatom species
composition in streams of the Oregon Coast
Range ecoregion. The diatom assemblages of these
streams were then related to environmental variables, including water chemistry, physical habitat,
watershed vegetation and watershed geology, in an
attempt to explain the observed patterns. A better
understanding of the role of natural landscape
variability in inuencing diatom assemblages may
improve the utility of diatom-based bioassessment
in the region.
Methods
All sites were sampled once for biological, chemical and physical variables during June through
September of 1999. The sampling unit was the
stream reach, dened as 40 times the average
wetted width measured at 3 locations within 10 m
209
Figure 1. Sampling locations and streams (5th order or higher) in the Oregon Coast Range ecoregion.
210
(47 mm Millipore type HA lters, 0.45 lm pore
size), and the other left unltered and stored on ice
until returned to the laboratory for nutrient analysis. Water samples were analyzed for concentrations of total and dissolved nutrients according to
standard EPA methods (Clesceri et al., 1998).
Periphyton were sampled by randomly selecting 10 cobbles within rie habitats of the stream
sampling reach. A known area (7.1 cm2) of each
rock was scraped using a toothbrush and delimiter
and combined into 1 composite sample per site.
The composite sample was homogenized and split
into 3 subsamples for identication and enumeration of diatom species and other assays. Samples
for diatom identication and enumeration were
preserved with formalin for a nal formalin concentration of 4%.
211
sites were considered to have sedimentary rock
dominated watersheds for analysis while the
remaining 12 sites were considered to have nonsedimentary dominated watersheds. Rock types
other than sedimentary and volcanic (primarily
alluvial/unconsolidated material) were found at 6
sites, with the relative abundance ranging from <1
to 9%. To simplify data analysis, these rock types
were grouped with volcanic-dominated sites.
Average % sedimentary rock in the sedimentarydominated group was 91% while average % volcanic rock in the volcanic-dominated group was
78%. The % sedimentary rock and % volcanic/
other rock varied signicantly between the two
groups (ANOVA p < 0.001, in both cases). An
ANOSIM was used to detect dierence in diatom
assemblages between the 2 geology-based groups
(ANOSIM, PRIMER-4, BrayCurtis distance
measure, 999 permutations). Dierences in environmental variables and relative abundances of
dominant diatom species between geology-based
site groups were examined with t-tests.
Two-way indicator species analysis (TWINSPAN, PC-ORD), a divisive cluster method using
a topdown approach, was used to classify stream
sites into several groups based on diatom relative
abundance data (Hill et al., 1975). To determine
the number of TWINSPAN groups to retain,
dierences in species composition between the
groups were tested using ANOSIM (PRIMER-4;
BrayCurtis similarity measure, 999 permutations). A separate ANOSIM was performed on the
groups created by the rst 4 TWINSPAN breaks.
Groups were retained if the probability of sites
being members of a TWINSPAN specied group
was greater than the probability of sites being
members of randomly created groups. A set of
indicator species for each of the retained TWINSPAN groups was characterized using Indicator
Species Analysis using PC-ORD (Dufrene &
Legendre, 1997). Statistical signicance of each
species indicator value was tested using a MonteCarlo permutation test (999 permutations,
p<0.05). Dierences in environmental variables
and diatom indices between TWINSPAN groups
were detected using an Analysis of Variance
(ANOVA, SigmaStat v. 1.00, Jandel Scientic).
Multiple comparisons between TWINSPAN
groups were performed with a StudentNewman
Keuls test using a Bonferroni-corrected alpha.
Results
Environmental factors
Watershed area and sampling point elevation
varied widely among sites. Watershed area ranged
between 1 and 169 km2 (median = 15 km2;
Table 1). Elevation ranged between 3 and 707 m
(median = 171 m). Sedimentary rocks were present in the watersheds of all but 3 sites and 13
watersheds were composed entirely of sedimentary
rock. Rock of volcanic origin was present in the
watersheds of 18 sites and only 3 watersheds were
composed entirely of volcanic rocks. Sites with
volcanic-dominated watersheds tended to be in the
northern part of the Coast Range ecoregion.
Conifer forest was the most common vegetation
type within the watersheds (median coverage = 63%; Table 1). Mixed forest covered
between 1 and 36% of the watersheds, while
coverage by broadleaf vegetation ranged between
0 and 25%. Spatial vegetation coverages were
similar between the sedimentary and volcanicdominated watersheds, however, the age structure
for conifer forest is unknown. Correlations
between watershed vegetation type (conifer, mixed
and broadleaf forest) and water quality, physical
habitat and watershed geology variables were
weak overall (|r| < 0.5).
Stream morphology varied substantially among
sites. Mean bankful width varied between 3 and
52 m, with a median of 9 m. Mean thalweg depth
ranged between 7 and 37 cm. Channel slope was
low overall all, but varied between 0 and 29% of
mean reach length (Table 1). Riparian canopy
cover ranged between 13 and 88%. In general, the
percent of ne substrate was low at most sites
(median 8%); only 2 sites had greater than 50%
percent ne substrate. Percent of ne substrate was
the only environmental variable that was signicantly dierent between sedimentary- and
volcanic-dominated watersheds (median = 11 c.f.
2%; p = 0.003).
The streams in the study had low stream ionic
strength and nutrient levels (Table 1). Median
total nitrogen (TN) concentration was 0.4 mg l)1
and median dissolved inorganic nitrogen (DIN)
was 0.2 mg l)1. Both total phosphorus (TP) and
soluble reactive phosphorus (SRP) concentrations
ranged from below detection limits (0.01 mg l)1)
212
Table 1. Minimum, median and maximum values for landscape, water chemistry and physical habitat variables for Oregon Coast
Range streams (n=33)
Minimum
Median
Maximum
1
3
15
171
169
707
75
100
25
100
23
63
98
25
12
36
17
45
Conductivity (lS/cm)
Total Nitrogen (mg/l)
38
0.2
66
0.4
146
0.8
0.0
0.2
0.6
0.01
0.02
0.10
0.01
0.01
0.02
41
56
93
52
15
37
0
0
2
8
29
69
13
51
88
to 0.1 mg l)1 (TP) and 0.02 mg l)1 (SRP). Alkalinity, conductivity and total solids varied
throughout the study area, but were low overall.
Alkalinity ranged between 6 and 45 mg
CaCO3 l)1, conductivity ranged between 38 and
146 lS cm)1 and total solids ranged between 41
and 93 mg l)1. Conductivity, alkalinity and total
solids were positively correlated for all sites (r >
0.6 for all).
Diatom assemblages
A total of 80 diatom species, from 22 dierent
genera, were identied from the 33 sites. Nitzschia,
Navicula and Achnanthes/Achnanthidium were the
most common genera with 13, 10 and 9 species,
respectively. Overall species richness was low,
ranging between 10 and 26 species per site (median = 15; Table 2). Shannon diversity index values ranged from 0.92.4 (median = 1.8; Table 2).
Over 50% of the total diatoms counted were from
the genus Achnanthes/Achnanthidium. Achnanthidium minutissimum (Kutz.) Czarnecki was present
at all sites and was the most abundant species
throughout the study area (median relative abun-
dance = 19%, range = 364%; Table 2). Achnanthes pyrenaicum (Hustedt) Kobayasi was also
common. Diatoms sensitive to organic pollution
dominated the assemblages at all sites
(median = 85%). Overall siltation index was low;
however, there were a few sites with high siltation
index values (median = 8%, range = 082).
A three-dimensional solution was obtained for
the NMDS ordination that explained 87% of the
variance in the diatom distance matrix (Figure 2).
Seventy percent of the variance was partitioned
between the second (r2 = 0.28) and third
(r2 = 0.42) axes. Axis 1 explained 17% of the
variance in the diatom data. The axes explained
signicantly more variance than would be
expected by chance based on Monte-Carlo permutation tests (p = 0.03). Axis 3 was primarily
driven by relative abundance of A. pyrenaicum
(r = )0.69; Table 3) and secondarily by N.
inconspicua Grunow (r = 0.55) and Rhoicosphenia
abbreviata (Agardh) Lange-Bertalot (r = 0.61).
Axis 2 was driven primarily by the relative abundances of A. minutissimum (r = 0.79) and Cocconeis placentula Ehrenberg (r = )0.56). Axis 1
was driven by the relative abundance of Cocconeis
213
Table 2. Relative abundance of the most common diatom species and diatom autecological indices values (median and ranges) for all
sites and sites grouped by dominant watershed geology. Bold values indicate signicant dierences between the two watershed geology
groupings. Pollution sensitivity values are from Lange-Bertalot (1979). Siltation index values are from Bahls (1993)
All Sites (n=33)
Sedimentary-dominated (n=21)
Volcanic-dominated (n=12)
Achnanthidium minutissimum
19 (364)
15 (464)
38 (364)
Achnanthidium pyrenaicum
13 (064)
21 (064)
5 (047)
Cocconeis placentula
6 (052)
11 (046)
1 (052)
Nitzschia inconspicua
3 (034)
2 (019)
10 (034)
Rhoicosphenia abbreviata
2 (032)
2 (032)
3 (032)
Shannon diversity
1.8 (0.92.4)
1.8 (0.92.4)
1.5 (1.22.0)
Species richness
% Pollution sensitive taxa
15 (1026)
85 (2499)
17 (1025)
85 (2899)
14 (1026)
84 (2497)
Siltation Index
8 (082)
4 (037)
15 (082)
Table 3. Pearsons correlation coecients (r) for environmental variables and common diatom species with signicant correlations to
NMDS axes. Bold values indicate signicant correlations (a = 0.05)
Axis 1
Axis 2
Axis 3
0.52
0.08
)0.31
Elevation (m)
0.00
)0.35
0.24
)0.07
0.05
)0.57
0.58
0.00
0.03
Conductivity (lS/cm)
)0.26
)0.15
0.52
)0.39
0.41
0.17
)0.34
0.52
0.08
)0.24
)0.21
0.50
0.42
0.11
)0.30
)0.38
)0.04
0.44
Achnanthidium pyrenaicum
Achnanthidium minutissimum
0.37
)0.12
)0.48
0.79
)0.69
)0.40
0.25
Cocconeis placentula
0.59
)0.56
Nitzschia inconspicua
)0.22
0.46
0.55
Planothidium lanceolata
)0.60
)0.35
0.37
Rhoicosphenia abbreviata
)0.44
)0.02
0.61
(a)
Axis 3
214
Group IIA:
Achnanthidium minutissimum
Nitzschia fonticola
Nitzschia frustulum
Axis 2
Group IC:
Axis 3
(b)
Amphora pediculus
Rhoicosphenia abbreviata
Axis 2
Discussion
on diatom assemblages between the 2 geologic
groups was signicant (global R = 0.19, p
<0.05). The relative abundance of A. minutissimum was signicantly greater at sites with volcanic-dominated watersheds ( p = 0.01), while that
of A. pyrenaicum was signicantly lower
( p = 0.05).
TWINSPAN produced four groups with statistically signicant site membership based on
ANOSIM results (Figure 3). Indicator species
analysis showed that Group I was characterized
by Amphora pediculus (Kutzing) Grunow and
R. abbreviata. Group II was characterized by
A. minutissimum, Nitzschia fonticola Grunow and
215
Table 4. Landscape, water quality, physical habitat variables and diatom metrics for diatom-based TWINSPAN groups for Oregon
Coast Range streams. Superscripts indicate signicant dierences among groups (ANOVA, alpha = 0.05). Pollution sensitivity values
are from Lange-Bertalot (1979)
Variables
TWINSPAN Group
I (n=6)
II (n=6)
III (n=18)
IV (n=3)
11
17
33
Elevation (m)
27a
320b
175b
207b
100
28
83
100
72
15
28a
14ab
17b
14ab
Conductivity (lS/cm)
Nitrate + Nitrite (mg/l)
82
0.3
66
0.2
63
0.1
64
0.5
0.5ab
0.4ab
0.3a
0.7b
59
55
53
62
7ab
15a
10a
ab
23
b
16
10b
1ab
10
45b
49
18
50
15
56
14
18
20
84
84
87
76
ab
216
of relatively homogeneous ecosystems based on
landscape factors (soils, vegetation, climate, geology and physiography; Omernik & Gallant, 1986).
The use of ecoregions to classify sites has produced
mixed results. Invertebrate communities of New
Zealand streams diered signicantly among ecoregions (Harding et al., 1997). In Oregon, sh and
macroinvertebrates diered signicantly only
between mountain and valley ecoregions (Whittier
et al., 1988). Pan et al. (2000) also found signicant
dierences in diatom assemblages only between
montane and valley ecoregions. The weak relationships between diatom assemblages and environmental variables found in this study point to
the fact that Level III ecoregion classication
might be too coarse to detect the human disturbance signal in this moderately impacted area. The
Oregon Coast Range ecoregion encompasses both
the coastal mountain range and the at, coastal
lowlands with a mixture of geology, soil, vegetation and landuse/land cover (Clarke et al., 1991).
In this study, when sites were stratied by dominant watershed geology, relationships between
diatom assemblages and environmental variables
became stronger. Diatom assemblages were signicantly dierent between sedimentary and volcanic-dominated watersheds. In addition, the
relative abundance of dominant species, A. minutissimum and A. pyrenaicum were signicantly
dierent between the two groups.
A major criticism of the ecoregion classication
approach is that it places too much emphasis on
landscape features. In summarizing several studies
on the relationship between landscape classications and stream biota, Hawkins et al. (2000a)
found that the amount of variance explained by
landscape variables was low. The ecoregion
approach is non-hierarchical and therefore smaller-scale variation (e.g., reach-level) is overlooked.
In a study of stream macroinvertebrates, reachscale physical properties were more predictive of
species assemblages than were landscape-scale
properties (Richards et al., 1997). In regions with
high environmental heterogeneity, biota-based
stream classication may delineate discrete stream
types with more clear relationships between biota
and environmental conditions. The RIVPACS
assessment approach rst classies sites based on
biota and then uses a predictive model based on
environmental variables to provide a list of taxa to
217
studies covering small spatial extents. Hawkins
et al. (2000b) found that rare taxa responded
positively to impacts and their inclusion in data
analysis improved precision and sensitivity of the
bioassessment. The BrayCurtis distance measure
used in our study is insensitive to rare taxa.
However, we did not include rare taxa in our
analysis due to taxonomic uncertainty of taxa
encountered only a few times. This resulted in the
exclusion of 35% of the taxa encountered in the
study. Conventional xed-count methods (counting 500600 diatom valves per sample) may not be
adequate to characterize rare species. Employment
of a stratied counting method that stops counting
dominant taxa after their abundance stabilizes but
continues to count rarer taxa until taxonomic
precision is established is one possible approach to
adequately identify and enumerate rare species.
In summary, diatom-based assessment suggests
that Coast Range ecoregion streams are in good
condition overall. The high natural variability in
stream condition and short human disturbance
gradients in this region might obscure the patterns
between diatoms and environmental variables. In
minimally impacted, complex landscapes, diatombased bioassessment might benet from: (1) classifying sites by either major landscape variables
(i.e., geology) or stream biota, (2) selecting sites
that cover the entire human disturbance gradient
of interest, (3) sampling dominant in-stream habitat and (4) careful statistical treatment of rare and
cosmopolitan taxa.
Acknowledgements
Stream sampling was conducted by the Oregon
Department of Environmental Quality as part of
the Salmon Plan for Oregon Watersheds. We
thank Rick Hafele, Mike Mulvey, Doug Drake,
Paul Gill and Shannon Hubler for their assistance.
Watershed geology and landscape calculations
would not have been possible without the CLAMS
dataset. Jesse Naymik and Brian Bowder provided
considerable assistance with GIS calculations. An
EPA STAR Graduate Fellowship provided funding for C. L. Weilhoefer during the writing of this
manuscript. The comments of Jan Stevenson and
an anonymous reviewer greatly improved the
quality of the manuscript.
References
Bahls, L. L., 1993. Periphyton bioassessment methods for Montana streams. Water Quality Bureau, Department of Health
and Environmental Sciences, Helena, Montana., 23 pp.
Barbour, M. T., J. Gerritsen, B. D. Snyder & J. B. Stribling,
1999. Rapid Bioassessment Protocols for Use in Streams and
Wadeable Rivers: Periphyton, Benthic Macroinvertebrates
and Fish, 2nd edn. Report EPA 841-B-99-002. U. S. Environmental Protection Agency, Oce of Water, Washington,
DC.
Cao, Y., D. P. Larsen & R. St-J. Thorne, 2001. Rare species in
multivariate analysis for bioassessment: some considerations. Journal of the North American Benthological Society
20: 144153.
Carpenter, K. D. & I. R. Waite, 2000. Relations of habitatspecic algal assemblages to land use and water chemistry in
the Willamette Basin, Oregon. Environmental Monitoring
and Assessment 64: 247257.
Chessman, B., I. Growns, J. Currey & N. Plunkett-Cole, 1999.
Predicting diatom communities at the genus level for the
rapid biological assessment of rivers. Freshwater Biology 41:
317331.
CLAMS, 1996. Coastal Landscape Analysis and Modeling
Study. www.fsl.orst.edu/clams.
Clarke, K. R. & R. H. Greene, 1988. Statistical design and
analysis for a biological eects study. Marine Ecology
Progress Series 46: 213222.
Clarke, S. E., D. White & A. L. Schaedel, 1991. Oregon, USA,
ecological regions and subregions for water quality management. Environmental Management 15: 847856.
Clesceri, L. S., A. E. Greenberg & A. D. Eaton, 1998. Standard
Methods for the Examination of Water and Wastewater.
American Public Heath Association, Washington, DC 1325
pp.
Dufrene, M. P. & P. Legendre, 1997. Species assemblages and
indicator species: the need for a exible asymmetrical
approach. Ecological Monographs 67: 345366.
Ford, J. & C. E. Rose, 2000. Characterizing small subbasins: a
case study from coastal Oregon. Environmental Monitoring
and Assessment 64: 359377.
Fore, L. S., J. R. Karr & R. W. Wisseman, 1996. Assessing
invertebrate responses to human activities: evaluating alternative approaches. Journal of the North American Benthological Society 15: 212231.
Frenkel, R. E., 1993. Vegetation. In Jackson, P. L. & A. J.
Kimmerling (eds), Atlas of the Pacic Northwest. Oregon
State University Press: 5865.
Hansmann, E. W. & H. K. Phinney, 1973. Eects of logging on
periphyton in coastal streams of Oregon. Ecology 54:
194199.
Harding, J. S., M. J. Winterbourn & W. F. McDiett, 1997.
Stream faunas and ecoregions in South Island, New Zealand: do they correspond?. Archiv fur Hydrobiologie 140:
289307.
Hawkins, C. P. & M. R. Vinson, 2000. Weak correspondence
between landscape classications and stream invertebrate
assemblages: Implications for bioassessment. Journal of the
North American Benthological Society 19: 501517.
218
Hawkins, C. P., R. H. Norris, J. Gerritsen, R. M. Hughes, S. K.
Jackson, R. K. Johnson & R. J. Stevenson, 2000a. Evaluation of the use of landscape classications for the prediction
of freshwater biota: Synthesis and recommendations.
Journal of the North American Benthological Society 19:
518540.
Hawkins, C. P., R. H. Norris, J. H. Hogue & J. W. Feminella,
2000b. Development and evaluation of predictive models for
measuring the biological integrity of streams. Ecological
Applications 10: 14561477.
Herlihy, A. T., D. P. Larsen, S. G. Paulsen, N. S. Urquhart &
B. J. Rosenbaum, 2000. Designing a spatially balanced,
randomized site selection process for regional stream surveys: The EMAP Mid-Atlantic pilot study. Environmental
Monitoring and Assessment 63: 95113.
Hill, M. O., R. G. H. Bunce & M. W. Shaw, 1975. Indicator
species analysis, a divisive polythetic method of classication
and its application to a survey of native pinewoods in
Scotland. Journal of Ecology 63: 597613.
Johnson, B. R. & G. L. Raines, 1995. Digital map of major
bedrock lithologic units for the Pacic Northwest: A Contribution to the Interior Columbia River Basin Ecosystem
Management Project. U. S. Geological Survey, Open File
Report 95680.
Krammer, K. & H. Lange-Bertalot, 1986. Bacillariophyceae, Teil
1. Naviculaceae . VEB Gustav Fisher Verlag, Jena, 876 pp.
Krammer, K. & H. Lange-Bertalot, 1988. Bacillariophyceae,
Teil 2. Epithemiaceae, Bacillariophyceae, Surirellaceae. VEB
Gustav Fisher Verlag, Jena 596 pp.
Krammer, K. & H. Lange-Bertalot, 1991a. Bacillariophyceae,
Teil 3. Centrales, Fragilariaceae, Eunotiaceae, Achnanthaceae. VEB Gustav Fisher Verlag, Jena, 576 pp.
Krammer, K. & H. Lange-Bertalot, 1991b. Bacillariophyceae,
Teil 4. Achnanthaceae, Kritische Erganzungen zu Navicula
(Lineolate) und Gomphonema. VEB Gustav Fisher Verlag,
Jena, 437 pp.
Krammer, K. & H. Lange-Bertalot, 2000. Bacillariophyceae,
Part 5. English and French Translation of the Keys. VEB
Gustav Fisher Verlag, Jena, 311 pp.
Lange-Bertalot, H., 1979. Pollution tolerance of diatoms as a
criterion for water quality estimation. Nova Hedwigia 64:
285305.
Leland, H. V., 1995. Distribution of phytobenthos in the
Yakima River Basin, Washington, in relation to geology,
land use, and other environmental factors. Canadian Journal
of Fisheries and Aquatic Sciences 52: 11081129.
McClain, M. E., R. E. Bilby & F. J. Triska, 1998. Nutrient
cycles and responses to disturbance. In Naiman, R. J. & R.
E. Bilby (eds), River Ecology and Management: Lessons
from the Pacic Coastal Ecoregion. Springer-Verlag, New
York: 347372.
Mrazik, S., 1999. Reference Site Selection: A Six-Step
Approach for Selecting Reference Sites for Biomonitoring
and Stream Evaluation Studies. Oregon Department of
Environmental Quality Technical Report BIO99-03. 14 pp.
Naymik, J., Y. Pan & J. Ford, 2005. Diatom assemblages as
indicators of timber harvest eects in coastal Oregon
streams. Journal of North American Benthological Society:
accepted.
219
Welch, E. B., J. M. Jacoby & C. W. May, 1998. Stream quality.
In Naiman, R. & R. E. Bilby (eds), River Ecology and
Management: Lessons from the Pacic Coastal Ecoregion.
Springer-Verlag, New York, NY: 6985.
Whittier, T. R., R. M. Hughes & D. P. Larsen, 1988. Correspondence between ecoregions and spatial patterns in stream
ecosystems of Oregon. Canadian Journal of Fisheries and
Aquatic Sciences 45: 12641278.
Wigington, P. J. Jr., M. R. Church, T. C. Strickland, K. N.
Eshleman & J. Van Sickle, 1998. Autumn chemistry of
Oregon Coast Range Streams. Journal of the American
Water Resources Association 34: 10351049.