J. N. Am. Benthol. Soc., 2001, 20(2):299–310
q 2001 by The North American Benthological Society
Comparison of correlations between environmental characteristics and
stream diatom assemblages characterized at genus and species levels
BRIAN H. HILL1,7, R. JAN STEVENSON2, YANGDONG PAN3, ALAN T. HERLIHY4,
PHILIP R. KAUFMANN5, AND COLLEEN BURCH JOHNSON6
US Environmental Protection Agency, National Center for Environmental Assessment,
26 W. Martin Luther King Dr., Cincinnati, Ohio 45268 USA
2
Department of Zoology, Michigan State University, East Lansing, Michigan 48824 USA
3
Environmental Sciences and Resources, Portland State University, Portland, Oregon 97207 USA
4
Department of Fisheries and Wildlife, Oregon State University, c/o US Environmental Protection Agency,
200 SW 35th Street, Corvallis, Oregon 97333 USA
5
US Environmental Protection Agency, National Health and Environmental Effects Research Laboratory,
200 SW 35th Street, Corvallis, Oregon 97333 USA
6
OAO Corporation, c/o US Environmental Protection Agency, 200 SW 35th Street,
Corvallis, Oregon 97333 USA
1
Biological surveys of stream communities
have long been used to assess the impacts of
human activities on receiving waters (see reviews in Whitton and Kelly 1995, Lowe and Pan
1996, Stoermer and Smol 1999, Stevenson and
Smol 2001). Stream biological integrity reveals
itself in the condition, abundance, and diversity
of its biota. These data may be used to assess
stream condition relative to biological condition
of an unimpaired stream. However, water-quality assessments using biological criteria are less
common than those based on stream chemistry
or toxicology. A recent survey of biological assessment programs in the United States reported that 39 states used biological information in
stream assessments, but only 27 states used this
information in setting water-quality criteria (Davis et al. 1996, Kroeger et al. 1999). Within these
programs, macroinvertebrates were most often
collected (33 states), followed by fish (21 states),
and algae (5 states).
Despite their infrequent use by state monitoring agencies, algae (especially diatoms) are
valuable indicators of stream ecosystem conditions because they: 1) are relatively simple to
collect, 2) respond rapidly and predictably to
changes in stream chemistry and habitat quality,
3) are taxonomically diverse, 4) have short regeneration times, and 5) are ubiquitous, which
allows for comparisons across geographic regions (Round 1991, van Dam et al. 1994, Leland
1995). An obstacle to widespread use of diatoms
for biological monitoring may be the large number of species and the need for taxonomic ex7
E-mail address: [email protected]
pertise (Kelly et al. 1995). To counter this problem, a few researchers have used genus-level
identification for the assessment of stream condition (Prygiel and Coste 1993, Kelly et al. 1995,
Chessman et al. 1999, Wu 1999, Hill et al. 2000).
Kelly et al. (1995) compared 3 species-based indices with 1 genus-based index and reported
that the genus-based index compared favorably
to the species-based indices. Round (1991), however, cautioned that it would be ‘‘dangerous to
compare streams simply on the genera recorded
and using generic identifications in water quality studies is even more dubious.’’
Our objective was to compare genus- versus
species-level identification for assessing the conditions of Mid-Appalachian streams using established relationships between diatoms and
their environments. We based our comparison
on 6 diatom assemblage attributes: taxa richness, taxon dominance, and proportions of acidobiontic, eutraphentic, motile, and pollutiontolerant diatoms. These 6 measures were compared with each other and against 19 stream
chemistry, 13 physical habitat, and 15 catchment
land-use variables collected from 199 streams.
Diatom assemblage attributes
Taxa richness
Diatom species richness usually decreases as
a result of stream contamination by organic enrichment (Nather Khan 1990, Whitton et al.
1991), metals (Crossey and La Point 1988, Sudhakar et al. 1991, Whitton et al. 1991), and pesticides (Kosinski 1984), although some researchers have reported increases in richness under
299
300
B. H. HILL
moderate stress (Archibald 1972, van Dam 1982,
Stevenson 1984, Lobo et al. 1995, Jüttner et al.
1996). We expected taxa richness in Mid-Appalachian streams to be inversely related to environmental stressors.
Taxon dominance
The relative abundance of any taxon of diatom can influence the evenness of diatom taxonomic representation within an assemblage.
Taxa better adapted to unfavorable conditions
(e.g., nutrient enrichment or chemical toxicity)
will have an advantage, resulting in an uneven
distribution of individuals among taxa (Archibald 1972, Sudhakar et al. 1991, Stevenson and
Pan 1999). We expected dominance by a single
taxon to increase with increasing environmental
stress.
Acidobiontic diatoms
Acidobiontic diatoms are those taxa with tolerances for pH ,5.5 (Lowe 1974, van Dam et al.
1994). The % of acidobiontic diatoms in an assemblage has been extensively used in the reconstruction of lake acidification (Charles 1985,
Dixit et al. 1992), but less often in analyses of
streams (Pan et al. 1996). As pH decreases, we
expected the diatom community to shift to
dominance by those taxa that can tolerate pH
,5.5.
Eutraphentic diatoms
Eutraphentic diatoms are those taxa with
preferences for nutrient-enriched, eutrophic waters (Lowe 1974, van Dam et al. 1994). Eutraphentic diatoms have been widely used in the
identification and assessment of sites impacted
by nutrients (Lange-Bertalot 1979, Hall and
Smol 1992, Pan et al. 1996). We expected an increase in the % eutraphentic diatoms with increasing nutrient and organic matter enrichment
of Mid-Appalachian streams.
Motile diatoms
Motile diatoms include all species of Navicula,
Nitzschia, and Surirella (Bahls 1993, Kutka and
Richards 1996, Barbour et al. 1999). Most anthropogenic activities within a catchment generate silt in receiving streams, even if no chem-
ET AL.
[Volume 20
ical signatures of such disturbance are present.
The % of motile diatoms has been used as an
index of siltation in Montana streams (Bahls
1993), and is cited as a potential attribute by
other investigators (Kutka and Richards 1996,
Stevenson and Pan 1999). We expected the % of
motile diatoms to increase with increasing siltation in streams.
Pollution-tolerant diatoms
Several investigators (Descy 1979, Lange-Bertalot 1979, Watanabe et al. 1988, Bahls 1993)
have developed pollution tolerance classifications for diatoms. The various systems have in
common that they all differentiate diatom species as pollution-tolerant, less tolerant, or pollution-intolerant. The pollution tolerances of diatoms have been recognized for decades, but the
use of this information in water-quality assessments has not been widespread. Investigators
assessing diatom pollution tolerances have reported proportional increases of tolerant taxa
with increasing organic pollution (Nather Khan
1990, Bahls 1993, Kwandrans et al. 1998). We
assumed the same in this paper.
Methods
Data used in our analyses were compiled
from the United States Environmental Protection Agency’s Environmental Monitoring and
Assessment Program (EMAP) surveys of MidAppalachian streams conducted from 1993 to
1995 (Hill et al. 2000). Diatoms were sampled
from 45 streams in 1993, 46 in 1994, and 108 in
1995. Approximately 10% of the streams were
sampled both near the beginning and the end
of the sampling period to test for diatom metric
stability. In total, data were collected from 199
streams during 233 site visits.
Characterization of each stream site included
19 stream chemistry variables, 13 stream habitat
variables, and 15 land-use variables. The methods used in, and the results of, these analyses
were previously reported (Hill et al. 2000).
Diatoms were collected from erosional habitats at each of 9 transects along the stream reach,
combined into a single sample for each stream,
and returned to the laboratory for processing
(Hill et al. 2000). Environmental preferences of
the diatom species were taken from published
research (Lowe 1974, van Dam et al. 1994). Di-
2001]
BRIDGES
atom species for which environmental preferences are not reported in the above references
were excluded from calculations. Environmental
preferences of diatom genera were determined
as the mean environmental preference rating for
species within a genus, and were used to determine which genera were included in the determination of the relative abundance of acidobiontic and eutraphentic diatoms. Similarly, pollution tolerances of diatom species were based on
published research (Lange-Bertalot 1979, Bahls
1993), with genus-level classifications based on
the average score for all diatom species within
a genus. Only Nitzschia is classified as a pollution-tolerant genus. Motile diatoms included all
species of Navicula, Nitzschia, and Surirella (Bahls
1993, Barbour et al. 1999).
Spearman correlation analysis (rs) was the initial test for relationships among diatom assemblage attributes and the 47 chemical, physical
habitat, and catchment land-use variables. For
comparisons of genus- and species-level diatom
assemblage attributes with environmental conditions, attributes were correlated with selected
chemical, physical habitat, and land-use variables for which correlations were expected to be
strong. Canonical correlation analyses were
used to further evaluate the associations between diatom assemblage attributes and environmental variables. Twenty-four variables that
were significantly correlated with at least one of
the diatom assemblage attributes were selected
for subsequent canonical correlation analyses.
Significant canonical correlations (p , 0.05)
were estimated as rk 5 (n 2 2)/(1 2 r2) (Rohlf
and Sokal 1969). All statistical analyses were
performed using SAS statistical software for
personal computers (SAS for Windows, release
7.0, SAS Institute, Cary, North Carolina).
Results
Forty-four genera and 522 species of diatoms
were collected from the study streams (Table 1).
Fifteen genera were represented by a single species; another 10 genera had ,5 species. Ten genera had .10 species, with Navicula (119 species),
Nitzschia (66), Achnanthes (48), and Eunotia (44)
being the most species-rich (Table 1).
Canonical correlation analyses of genus- and
species-level attributes and the environmental
variables resulted in 2 significant canonical axes.
The 1st canonical axis (W1) represented a human
301
disturbance/geomorphology gradient having
positive correlations with pH, acid neutralizing
capacity (ANC), specific conductance, base cations (Ca12, Mg12, Na1), agricultural and urban/
suburban land uses, stream sediments ,2 mm
diameter, and riparian zone disturbances, and
negative correlations with riparian canopy cover
(Table 2). A nutrient-enrichment/stream-size
gradient is described by the 2nd canonical axis
(W2), which was positively correlated with nutrients (total P and total N), Cl-, total suspended
solids, fine-grained stream sediments, and agricultural land uses, and negatively correlated
with stream width and depth (Table 2).
Generic richness in any given stream ranged
from 3 to 17, with a mean of 11 genera per
stream, whereas species richness ranged from
10 to 72 with a mean of 34 species per stream.
Genera richness was significantly correlated
with species richness (Fig. 1A). Species richness
exhibited stronger correlations with the selected
environmental variables than generic richness,
especially the catchment-scale land-use variables (last column of Table 3). Both generic and
species richness were significantly correlated
with W2, the nutrient-enrichment/stream-size
canonical axis (Table 4).
Dominance by a single taxon ranged from
16% to 98% for genera and from 12% to 88% for
species. Dominance by a single genus was significantly correlated with dominance by a single
species (Fig. 1B). The correlation of dominance
by a single genus to the selected environmental
variables was similar to that exhibited by species dominance with these same variables (Table
3). Taxon dominance was correlated with W2
(Table 4).
The proportions of acidobiontic diatoms in
the assemblages were nearly identical for both
the genus- and species-level attributes (Fig. 1C).
Acidobiontic diatoms accounted for 0% to 99%
of both the genus- and species-level counts, and
both genus- and species-level attributes were
similarly correlated with the selected environmental variables (Table 3). Both genus- and species-level attributes were significantly correlated
with W1, the human disturbance/geomorphology gradient (Table 4).
The relative abundance of eutraphentic taxa
ranged from 0% to 56% of the genus-level
counts and from 0% to 90% of the species-level
counts, and genus-level counts were significantly correlated with species-level counts (Fig. 1D).
302
B. H. HILL
ET AL.
[Volume 20
TABLE 1. Diatom genera, number of species (including varieties), and environmental preferences (based on
average scores for species within a genus) for Mid-Appalachian streams.
Genus
Achnanthes
Actinella
Amphipleura
Amphora
Anomoeoneis
Asterionella
Aulacoseira
Bacillaria
Caloneis
Cocconeis
Cyclostephanos
Cyclotella
Cymatopleura
Cymbella
Diatoma
Diploneis
Entomoneis
Epithemia
Eunotia
Fragilaria
Frustulia
Gomphoneis
Gomphonema
Gyrosigma
Hantzschia
Mastogloia
Melosira
Meridion
Navicula
Neidium
Nitzschia
Opephora
Pinnularia
Pleurosira
Reimeria
Rhoicosphenia
Rhopaladia
Stauroneis
Stenopterobia
Stephanodiscus
Surirella
Synedra
Tabellaria
Thalassiosira
No. of
species
48
1
1
9
3
1
8
1
3
3
1
8
2
32
5
7
2
2
44
22
5
1
33
5
1
1
2
2
119
12
66
1
23
1
1
1
1
9
1
3
8
18
4
1
Acidobiontic
Eutraphentic
Motile
Pollution-tolerant
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
Yes
No
No
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
Yes
No
No
No
Yes
Yes
No
No
No
No
No
Yes
No
No
No
No
Yes
No
No
No
No
No
No
No
No
Yes
No
No
No
Yes
No
No
Yes
No
No
No
No
No
Yes
No
No
Yes
No
Yes
No
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
Yes
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
No
No
No
The species-level eutraphentic attribute was
more strongly correlated with the selected environmental variables than the genus-level eutraphentic attribute, especially for the land-use
variable (Table 3). The proportion of eutraphentic diatoms based on genus-level counts was not
significantly correlated with either W1 or W2,
whereas the proportion based on species-level
counts was significantly correlated with W1 (Table 4).
Motile diatoms represented 0% to 92% of the
genus-level and 0% to 91% of the species-level
2001]
BRIDGES
303
TABLE 2. Correlations between environmental variables selected and their canonical axes in the genusand species-level diatom attributes. Significant (p ,
0.05) correlations are indicated in bold. W1 5 1st canonical axis, W2 5 2nd canonical axis. ANC 5 acid
neutralizing capacity.
dients (Table 3), although the species-level attribute was more strongly correlated than the genus-level attribute with W2 (Table 4).
Canonical variables
A simple measure of the reliability of genuslevel attributes as an alternative to species-level
attributes is the correlation between attributes
based on diatom counts at the 2 taxonomic levels. Perfect correlations (r 5 1.00) would suggest
that there is no loss of information, for the measures presented, by limiting identifications to
the generic level. Genus- and species-based attributes were significantly correlated, suggesting that for some purposes diatom identification
only to genus may adequately characterize the
diatom assemblage’s response to environmental
conditions.
For those diatom assemblage attributes in
which genus-level counts are highly correlated
with species-level counts (acidobiontic and motile attributes), it appears that genus-level
counts may suffice. Deviations from perfect correlations have simple explanations, and their impacts on our assessments may be small. For example, the close agreement between genus- and
species-level counts of acidobiontic diatoms reflects the fact that most acidobiontic species are
in genera that are also classed as acidobiontic
(Eunotia, Frustulia, Pinnularia, Tabellaria). Exceptions to this fact include acidobiontic species in
genera not classified as acidobiontic: Achnanthes
(1 sp.), Anomoeoneis (2), Aulocoseira (2), Cymbella
(2), Fragilaria (1), Navicula (5), Neidium (1), Stenopterobia (1), and Surirella (1). The impact of excluding the above species from the genus-level
assessment of acidobiontic diatoms can be determined only by species-level counts, but the
near-perfect correlation of the genus- and species-level attributes suggests that this discrepancy is small.
Environmental variables
W1
W2
Dissolved Al
ANC
Ca12
Cl2
Specific conductance
Fe13
Mg12
Na1
Total N
Total P
pH
SiO2
SO422
Total suspended solids
Fine sediment (,0.06 mm)
Sand 1 fine sediments (,2 mm)
Canopy cover
Channel embeddedness
Stream depth
Wetted channel width
Presence of riparian agriculture
Presence of riparian human
disturbance
% of catchment in agricultural
land use
% of catchment in urban/
suburban land use
20.35
0.64
0.65
0.37
0.66
20.04
0.59
0.44
0.26
0.15
0.84
0.02
0.34
0.11
0.33
0.50
20.49
0.24
0.37
0.21
0.40
0.21
0.22
0.32
0.40
0.35
0.19
0.23
0.31
0.43
0.65
0.07
20.03
0.21
0.41
0.44
0.34
20.19
0.17
20.42
20.48
0.32
0.44
0.17
0.61
0.43
0.43
0.30
counts, and the genus-level counts were significantly correlated with species-level counts (Fig.
1E). Both genus- and species-level counts of motile diatoms were similarly correlated with the
selected environmental variables (Table 3), and
both were equally correlated with W2 (Table 4).
The relative abundance of pollution-tolerant
diatoms based on genus-level counts ranged
from 0% to 64%, compared to 0% to 25% of the
assemblages based on species-level counts. The
proportion of pollution-tolerant organisms identified to the genus level was weakly, though significantly, correlated to the counts based on species-level (Fig. 1F). Relative abundance of pollution-tolerant diatoms based on both genusand species-level counts were similarly
correlated with the selected environmental gra-
Discussion
Genus and species comparisons
All correlations of genus- and species-level attributes were significant, but the strengths of
these correlations highlight exceptions to this
generalization. Comparisons of taxa richness attributes, taxon dominance, and the relative
abundances of eutraphentic and pollution-tolerant diatoms suggest significant loss of information attributable to restricting identifications
304
B. H. HILL
ET AL.
[Volume 20
FIG. 1. Comparisons of diatom taxa richness (A), dominance (B), and acidobiontic (C), eutraphentic (D),
motile (E), and pollution-tolerance (F) attributes based on genus and species counts in streams in the MidAppalachian region of the United States.
to the higher taxonomic level. For these poorerperforming attributes we must judge how much
information loss is acceptable without compromising our interpretation of the relationships
between diatom assemblages and stream conditions.
Despite the large numbers of species in some
genera, there is little evidence to suggest that
these genera bias the outcome of genus- and
species-level richness attributes of the diatom
assemblages of Appalachian streams. These results suggest that generic richness may be as
indicative as species richness of environmental
conditions in streams. It might be argued that
generic richness is a more conservative measure
of the diatom assemblage’s response to its environment, given reports of increases in species
richness with moderate environmental stress
(van Dam 1982, Stevenson 1984, Lobo et al.
1995). The conservative nature of generic rich-
2001]
BRIDGES
305
TABLE 3. Comparisons of genus- and species-level diatom assemblage attributes with chemical, physical
habitat, and land-use variables for which strong correlations were expected. Spearman correlation (rs) values
are shown in parentheses. ns 5 p . 0.05, * 5 p , 0.05, ** 5 p , 0.01, *** 5 p , .001.
Attribute
Taxa richness
% dominance
% acidobiontic
% eutraphentic
% motile
% pollution-tolerant
Taxa level
Genus
Species
Genus
Species
Genus
Species
Genus
Species
Genus
Species
Genus
Species
Water chemistry
Total N (0.28)**
Total N (0.42)***
Total P (20.41)**
Total P (20.45)***
pH (20.64)***
pH (20.62)***
Total N (0.32)***
Total N (0.52)***
Cl2 (0.56)***
Cl2 (0.57)***
Total P (0.58)***
Total P (0.48)***
ness may be especially true if richness is the result of increasing numbers of species in only a
few genera (e.g., Navicula and Nitzschia). However, the correlation of generic richness with any
of the environmental variables or canonical axes
was low (r , 0.28), which limits its usefulness
in assessing stream condition.
Dominance of the diatom assemblage by a
single genus was strongly correlated with dominance by a single species. That the agreement
was not stronger underscores the loss of sensitivity of diatom assemblage attributes when
identification is restricted to the genus level. Cumulatively, the species within a species-rich genus would account for a greater dominance by
that genus than would be attributable to any
single species within that genus, resulting in a
poorer-than-expected correlation among the 2
dominance attributes. Ecological information
lost by stopping at the genus level, however, is
not apparent in the relationships of taxon dominance to environmental variables or the nutrient-enrichment/stream-size canonical axis.
Eutraphentic genera were significantly correlated with the abundance of eutraphentic species, although the correlation was far from perfect. The relative lack of agreement among genus- and species-level counts is likely a result of
the number of eutraphentic species that are not
classified as such when counts are made at the
genus level. The following genera are not classified as eutraphentic, but include eutraphentic
species: Achnanthes (3 spp.), Caloneis (1), Cyclotella (3), Cymbella (5), Fragilaria (8), Frustulia (2),
Gomphonema (10), Mastogloia (1), Navicula (69),
Physical habitat
Land use
Sand (0.19)*
Sand 1 fine (0.32)**
Sand 1 fine (20.40)***
Sand 1 fine (20.28)**
Canopy (0.34)**
Canopy (0.40)***
Fines (0.23)*
Fines (0.49)***
Fines (0.52)***
Fines (0.51)***
Fines (0.45)***
Fines (0.40)***
Forest (20.08)ns
Forest (20.27)**
Agriculture (20.37)**
Agriculture (20.31)**
Agriculture (20.38)***
Agriculture (20.29)**
Agriculture (0.12)ns
Agriculture (0.57)***
Forest (20.54)***
Forest (20.52)***
Forest (20.52)***
Forest (20.50)***
Neidium (5), Pleurosira (1), Stauroneis (5), Stephanodiscus (3), and Surirella (6). The relative abundance of species that are excluded from the genus-level counts can be determined only by doing the species-level counts, and these results
suggest that genus-level identifications are inappropriate for determining the abundance of
eutraphentic taxa in streams. This problem is
further indicated by the discrepancies in the correlations of both the genus- and species-level attributes with the environmental variables associated with nutrient enrichment. Our understanding of the relationship between eutraphentic diatoms and their environment is further
complicated by the fact that neither the genusnor the species-level eutraphentic metric was
correlated with the nutrient-enrichment/
stream-size gradient axis.
Pollution-tolerant diatoms counted as genera
were only weakly correlated with the abundance of pollution-tolerant species. Much of the
disagreement between taxonomic levels arises
from the exclusion of Amphora veneta, Cymbella
pusilla, Surirella angusta, 3 species of Gomphonema, and 14 species of Navicula. These species are
classified as pollution-tolerant, but are in genera
that are not similarly classified. The weakness
of the correlation of genus- and species-level
counts suggests that a significant proportion of
the diatom assemblage is in these excluded genera. The impact of excluding these pollution-tolerant species from the genus-level counts is not
clear because both assemblage attributes were
similarly correlated with the selected environ-
306
B. H. HILL
ET AL.
[Volume 20
TABLE 4. Correlations between diatom assemblage attributes and the canonical axes of environmental variables. Significant (p , 0.05) correlations are indicated in bold. W1 5 1st canonical axis, W2 5 2nd canonical
axis.
Canonical variables
Attribute
Taxa richness
% dominance
% acidobiontic
% eutraphentic
% motile
% pollution-tolerant
Eigenvalue
Variance explained (%)
Cumulative variance explained (%)
Significance of the canonical axes (p , rk)
mental variables, although less so with the nutrient-enrichment/stream-size gradient axis.
Our comparison of genus- and species-level
identification yielded mixed results. When there
are few indicator taxa (e.g., acidobiontic diatoms) or when all species are included in the
genus-level counts (e.g., motile diatoms), genuslevel identification appears to adequately describe the assemblage’s response to environmental variables for which those taxa are known indicators. Similarly, dominance of the diatom assemblage by a single genus may be a close
approximation of the species dominating that
assemblage.
However, our data also indicate situations
where genus-level indices yield only a poor representation of the species-level indices. The effect of this poorer representation on assessment
of the stream environment is unclear. In the case
of eutraphentic diatoms, species-level counts
were more strongly correlated with the selected
environmental variables than were the genuslevel counts. The opposite was true, however, for
counts of pollution-tolerant diatoms in which
genus-level counts exhibited stronger correlations with the selected environmental variables
than did the species-level counts. The case for
taxa richness is complicated by conflicting predictions and observations (e.g., greater species
richness with moderate environmental stress),
Taxa level
W1
W2
Genus
Species
Genus
Species
Genus
Species
Genus
Species
Genus
Species
Genus
Species
20.08
20.01
20.21
20.11
20.54
20.53
0.04
0.60
0.36
0.34
0.14
0.15
5.16
32
32
,0.0001
0.36
0.40
20.52
20.55
0.05
0.03
0.11
0.26
0.51
0.51
0.39
0.65
3.31
20
52
,0.0001
and may be of limited value, by itself, in assessing stream condition, a conclusion frequently
reached in similar studies (e.g., Stevenson 1984,
Lobo et al. 1995, Jüttner et al. 1996).
Implications of genus-level identification
Our analyses show that the choice of restricting diatom identifications to the genus level
comes at the expense of the depth and flexibility
of interpretation of the ecological information
contained within the diatom assemblage. A lack
of species-level identification precludes building
databases on species distributions and their tolerances or preferences for chemical and physical
conditions, databases that in the long-run would
facilitate wider use and improved precision of
diatom indices for stream monitoring.
All of these arguments are moot if the person
or agency lacks expertise in diatom identification. The question then becomes not whether
species-level identification is better, but whether
some information on diatoms in streams is better than none. One argument for identifying diatoms only to the genus-level is time saved
when learning the higher-level taxonomy. Students and stream monitoring volunteers usually
learn diatom genera easily and quickly (perhaps
within 1 wk), whereas becoming proficient in
species identification may take months or even
2001]
BRIDGES
years. Identification at the genus level would
also require a smaller set of taxonomic references. Quality assurance would be easier to monitor and errors easier to correct. Therefore, the
real advantage to genus-based attributes may be
realized when starting new stream monitoring
programs with inexperienced researchers and/
or volunteers.
Several researchers have proposed compromises between genus- and species-level identification of diatoms. Round (1991) proposed
working only with the dominant taxa, which
would reduce the list of species to ;10% of the
total richness. Kelly and Whitton (1995) proposed a hybrid of genus- and species-level identification. Taxa that were particularly sensitive
to environmental stressors and were easily identified would be identified to the species level,
whereas all other diatoms would be identified
only to the genus level. Their resulting trophic
diatom index relies on the relative abundance of
3 species, Gomphonema parvulum, Navicula gregaria, and N. lanceolata, combined with genus-level
counts of small Navicula and Sellaphora, and Nitzschia for classifying streams by degree of organic pollution.
Wu (1999) advocated an even more restricted
index of diatoms for pollution monitoring: the
generic index (GI), which compares the abundance of Achnanthes, Cocconeis, and Cymbella to
that of Cyclotella, Melosira, and Nitzschia. Wu’s GI
is significantly correlated with our genus-level
counts of pollution-tolerant (r 5 0.87), motile (r
5 0.75), eutraphentic (r 5 0.64), and dominant
diatoms (r 5 0.57).
Resh and McElravy (1993) examined 34 published studies addressing the question of the appropriate taxonomic level for identification of
macroinvertebrates for monitoring aquatic systems. Half of the studies favored species-level
identification, whereas 26% favored genus-level
identification. Proponents of higher-level identification cited 2 scenarios when genus-level
identification would be appropriate, including 1)
the detection of instances of gross pollution and
2) situations in which taxa included in the higher taxonomic level are consistent in their responses to environmental conditions. We believe
the same to be true for assessments based on
diatom assemblages.
Geographic-scale issues
Geographic scale is perhaps the biggest unknown in predicting biotic responses to envi-
307
ronmental conditions. How is the selection of
genus- or species-level identification affected by
the geographic scale of the study? Our data
come from a regional-scale study designed to
answer 2 questions: 1) What is the condition of
streams in the Mid-Appalachian region? and 2)
Are conditions getting better or worse? Although sampling intensity was sufficient to answer these questions on a regional- or ecoregion- (sensu Omernik 1987) scale, it was not intense enough within an ecoregion to address
catchment-wide variance in diatom assemblages. There are 2 views regarding the appropriateness of genus-level identifications with increasing geographic scale. First, genus-level attributes might be developed, calibrated, and
used successfully at smaller geographic scales,
but become increasingly problematic at broader
geographic scales (Whittier et al. 1988, Pan et al.
1999, 2000). If this view is true, nearby streams
that are environmentally similar should contain
species within genera that have similar environmental preferences, and genus-based assessments should approximate those based on species-level identifications. Second, as geographicscale increases so does environmental heterogeneity, and the reliability of genus-based
assessments should decline. Hierarchy theory
tells us that spatial hierarchies are nested and
that higher levels of organization are composed
of lower levels (O’Neill et al. 1986). Therefore,
assessments at broader geographic scales
should be predictable from smaller geographicscale assessments. If this argument and the previous view are true, then regional-scale, genusbased assessments should be as reliable as sitescale, genus-based assessments. Unfortunately,
data are lacking to support or reject either view.
Conclusions
Our results support 4 general conclusions.
First, richness attributes are not consistently or
predictably related to gradients of human disturbances within a catchment, and should be
used cautiously for environmental monitoring.
Second, diatom assemblage attributes based on
genus or species sensitivities and tolerances to
environmental conditions, even based on European diatom autecologies, are consistently and
reliably related to gradients of human disturbances within a catchment. Third, genus-level
diatom assemblage attributes for some environ-
308
B. H. HILL
mental gradients are predictable and appear to
be relatively precise compared to species-level
attributes. Fourth, the environmental gradients
with which genus-level diatom assemblage attributes are most strongly correlated are those
gradients that involve morphological (motility)
or physiological (pH tolerance) adaptations that
are related to evolved genus-level characteristics.
Acknowledgements
Preparation of this manuscript was a joint effort of the National Exposure Research Laboratory in Cincinnati, Ohio, and National Health
and Environmental Effects Research Laboratory
in Corvallis, Oregon, through a cooperative
agreement with Oregon State University
(CR824682) and contracts with the University of
Louisville and OAO Corporation. It has been
subjected to the Agency’s peer and administrative review and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use. We thank G. Collins, M. Vander
Borgh, M. Gurtz, R. Lowe, and an anonymous
reviewer for their helpful comments on the
preparation of the manuscript. We also thank S.
Paulsen and J. Stoddard for their leadership of
the EMAP Surface Waters Program.
Literature Cited
ARCHIBALD, R. E. M. 1972. Diversity in some South
African diatom associations and its relation to
water quality. Water Research 6:1229–1238.
BAHLS, L. L. 1993. Periphyton bioassessment methods
for Montana streams. Water Quality Bureau, Department of Health and Environmental Sciences,
Helena, Montana. (Available from: Water Quality
Bureau, Department of Health and Environmental Sciences, State of Montana, P. O. Box 200901,
Helena, Montana 59620 USA.)
BARBOUR, M. T., J. GERRITSEN, B. D. SNYDER, AND J. B.
STRIBLING. 1999. Rapid bioassessment protocols
for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates and fish. Second
edition. EPA 841-B-99-002. Office of Water, US Environmental Protection Agency, Washington, DC.
CHARLES, D. F. 1985. Relationship between surface
sediment diatom assemblages and lake water
characteristics in Adirondack lakes. Ecology 66:
994–1011.
CHESSMAN, B., I. GROWNS, J. CURRY, AND N. PLUNKETT-COLE. 1999. Predicting diatom communities
ET AL.
[Volume 20
at the genus level for the rapid biological assessment of rivers. Freshwater Biology 41:317–331.
CROSSEY, M. J., AND T. W. LA POINT. 1988. A comparison of periphyton structural and functional responses to heavy metals. Hydrobiologia 162:109–
121.
DAVIS, W. S., B. D. SNYDER, J. B. STRIBLING, AND C.
STOUGHTON. 1996. Summary of state biological
assessment programs for streams and rivers. EPA
230-R-96-007. Office of Policy, Planning, and
Evaluation, US Environmental Protection Agency,
Washington, DC.
DESCY, J. P. 1979. A new approach to water quality
estimation using diatoms. Nova Hedwigia 64:
305–323.
DIXIT, S. S., J. P. SMOL, J. C. KINGSTON, AND D. F.
CHARLES. 1992. Diatoms: powerful indicators of
environmental change. Environmental Science
and Technology 26:23–33.
HALL, R. I., AND J. P. SMOL. 1992. A weighted-average
regression and calibration model for inferring total phosphorus concentration from diatoms in
British Columbia (Canada) lakes. Freshwater Biology 27:417–434.
HILL, B. H., A. T. HERLIHY, P. R. KAUFMANN, R. J. STEVENSON, AND F. H. MCCORMICK. 2000. The use of
periphyton assemblage data as an index of biotic
integrity. Journal of the North American Benthological Society 19:50–67.
JÜTTNER, I., H. ROTHFRITZ, AND S. J. ORMEROD. 1996.
Diatoms as indicators of river quality in the Nepalese Middle Hills with consideration of the effects of habitat specific sampling. Freshwater Biology 36:475–486.
KELLY, M. G., C. J. PENNY, AND B. A. WHITTON. 1995.
Comparative performance of benthic diatom indices used to assess river water quality. Hydrobiologia 302:179–188.
KELLY, M. G., AND B. A. WHITTON. 1995. The trophic
diatom index: a new index for monitoring eutrophication in rivers. Journal of Applied Phycology
7:433–444.
KOSINSKI, R. J. 1984. The effect of terrestrial herbicides
on the community structure of stream periphyton. Environmental Pollution Series A Ecological
and Biological 36:165–189.
KROEGER, S., E. FENSIN, K. LYNCH, AND M. VANDER
BORGH. 1999. United States water quality programs
that use algae as a biological assessment tool. Division of Water Quality, Department of Environment and Natural Resources, Raleigh, North Carolina. (Available from: http://www.esb.enr.state.
nc.us/phytofolder/downloads/statesum.pdf)
KUTKA, F. J., AND C. RICHARDS. 1996. Relating diatom
assemblage structure to stream habitat quality.
Journal of the North American Benthological Society 15:469–480.
KWANDRANS, J., P. ELORANTA, B. KAWECKA, AND K.
2001]
BRIDGES
WOJTAN. 1998. Use of benthic diatom communities to evaluate water quality in rivers of southern
Poland. Journal of Applied Phycology 10:193–201.
LANGE-BERTALOT, H. 1979. Pollution tolerance of diatoms as a criterion for water quality estimation.
Nova Hedwigia 64:285–305.
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 52:1108–1129.
LOBO, E. A., K. KATOH, AND Y. ARUGA. 1995. Response
of epilithic diatom assemblages to water pollution
in rivers in the Tokyo metropolitan area, Japan.
Freshwater Biology 34:191–204.
LOWE, R. L. 1974. Environmental requirements and
pollution tolerance of freshwater diatoms. EPA670/4–74–005. US Environmental Protection
Agency, Cincinnati, Ohio.
LOWE, R. L., AND Y. PAN. 1996. Benthic algal communities and biological monitors. Pages 705–739
in R. J. Stevenson, M. Bothwell, and R. L. Lowe
(editors). Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, California.
NATHER KHAN, I. S. A. 1990. Assessment of water pollution using diatom community structure and
species distribution: a case study of a tropical river basin. Internationale Revue der gesamten Hydrobiologie 75:317–338.
OMERNIK, J. M. 1987. Aquatic ecoregions of the conterminous United States. Annals of the Association of American Geographers 77:118–125.
O’NEILL, R. V., D. L. DEANGELIS, J. B. WAIDE, AND T.
F. H. ALLEN. 1986. A hierarchical concept of ecosystems. Princeton University Press, Princeton,
New Jersey.
PAN, Y., R. J. STEVENSON, B. H. HILL, AND A. T. HERLIHY. 2000. Ecoregions and benthic diatom assemblages in Mid-Atlantic Highland streams, USA.
Journal of the North American Benthological Society 19:518–540.
PAN, Y., R. J. STEVENSON, B. H. HILL, A. T. HERLIHY,
AND G. B. COLLINS. 1996. Using diatoms as indicators of ecological conditions in lotic systems: a
regional assessment. Journal of the North American Benthological Society 15:481–495.
PAN, Y., R. J. STEVENSON, B. H. HILL, P. R. KAUFMANN,
AND A. T. HERLIHY. 1999. Spatial patterns and
ecological determinants of benthic algal assemblages in Mid-Atlantic streams, USA. Journal of
Phycology 35:460–468.
PRYGIEL, J., AND M. COSTE. 1993. The assessment of
water quality in the Artois-Picardie water basin
(France) by the use of diatom indices. Hydrobiologia 269/ 270:343–349.
RESH, V. H., AND E. P. MCELRAVY. 1993. Contemporary
quantitative approaches to biomonitoring using
benthic macroinvertebrates. Pages 159–194 in D.
309
M. Rosenberg and V. H. Resh (editors). Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York.
ROHLF, F. J., AND R. R. SOKAL. 1969. Statistical tables.
W. H. Freeman and Company, San Francisco, California.
ROUND, F. E. 1991. Diatoms in river-monitoring studies. Journal of Applied Phycology 3:129–145.
STEVENSON, R. J. 1984. Epilithic and epipelic diatoms
in the Sandusky River, with emphasis on species
diversity and water pollution. Hydrobiologia 114:
161–175.
STEVENSON, R. J., AND Y. PAN. 1999. Assessing ecological conditions in rivers and streams with diatoms. Pages 11–40 in E. F. Stoermer and J. P. Smol
(editors). The diatoms: applications to the environmental and earth sciences. Cambridge University Press, Cambridge, UK.
STEVENSON, R. J., AND J. P. SMOL. 2001. Use of algae in
environmental assessments. In J. D. Wehr and R.
G. Sheath (editors). Freshwater algae in North
America: classification and ecology. Academic
Press, San Diego, California (in press).
STOERMER, E. F., AND J. P. SMOL. 1999. The diatoms:
applications for the environmental and earth sciences. Cambridge University Press, Cambridge,
UK.
SUDHAKAR, G., B. JYOTHI, AND V. VENKATESWARLU.
1991. Metal pollution and its impact on algae in
flowing waters in India. Archives of Environmental Contamination and Toxicology 21:556–566.
VAN DAM, H. 1982. On the use of measures of structure and diversity in applied diatom ecology.
Nova Hedwigia 73:97–115.
VAN DAM, H., A. MERTENS, AND J. SINKELDAM. 1994.
A coded checklist and ecological indicator values
of freshwater diatoms from the Netherlands.
Netherlands Journal of Aquatic Ecology 28:117–
133.
WATANABE, T., K. ASAI, AND A. HOUKI. 1988. Numerical water quality monitoring of organic pollution
using diatom assemblages. Pages 123–141 in F. E.
Round (editor). Proceedings of the 9th International Diatom Symposium, Bristol, UK. Biopress
Ltd., Bristol and Koeltz Scientific Books, Koenigstein, Germany.
WHITTIER, T. R., R. M. HUGHES, AND D. P. LARSEN.
1988. The correspondence between ecoregions
and spatial patterns in stream ecosystems in
Oregon. Canadian Journal of Fisheries and Aquatic Sciences 45:1264–1278.
WHITTON, B. A., AND M. G. KELLY. 1995. Use of algae
and other plants for monitoring rivers. Australian
Journal of Ecology 20:45–56.
WHITTON, B. A., E. ROTT, AND G. FRIEDRICH. 1991. Use
of algae for monitoring rivers. E. Rott Publishers,
Institut fur Botanik, Universitaat Innsbruck, Innsbruck, Austria.
310
B. H. HILL
WU, J.-T. 1999. A generic index of diatom assemblages
as bioindicator of pollution in the Keelung River
of Taiwan. Hydrobiologia 397:79–87.
About the Authors
Brian Hill is an ecologist with the US Environmental Protection Agency, and studies algal community structure and stream ecosystem function. Jan
Stevenson is a professor at Michigan State University, where he specializes in the use of diatoms for
monitoring aquatic ecosystem integrity. Yangdong
Pan is an assistant professor at Portland State Uni-
ET AL.
[Volume 20
versity, and studies diatom–environment interactions.
Alan Herlihy is a research assistant professor at
Oregon State University; his research interests are
survey sampling design and ecological indicators.
Philip Kaufmann is an ecologist with the US Environmental Protection Agency, and studies the relationships between biota and physical habitat structure in streams and lakes. Colleen Burch Johnson is
a senior computer specialist for the OAO Corporation,
under contract to the US Environmental Protection
Agency; she evaluates catchment conditions using
geographic information system techniques.