Ecological Indicators 84 (2018) 140–151

Contents lists available at ScienceDirect

Ecological Indicators
journal homepage: www.elsevier.com/locate/ecolind

Original Articles

River water quality assessment based on a multi-descriptor approach MARK

including chemistry, diatom assemblage structure, and non-taxonomical
diatom metrics
Lalit K. Pandeya,b, Isabelle Lavoiec, Soizic Morind, Jihae Parke,f, Jie Lyug, Soyeon Choib,
⁎
Hojun Leeb, T. Hanb,g,
a
Institute of Green Environmental Research Center, 169, Gaetbeol-ro, Yeonsu-gu, Incheon 21999, Republic of Korea
b
Department of Marine Sciences, Incheon National University, 119, Academy-ro, Yeonsu-gu, Incheon 22012, Republic of Korea
c
Institut national de la recherche scientifique, centre Eau Terre Environnement, 490 rue de la Couronne, Québec City, Québec G1K 9A9, Canada
d
Irstea, UR EABX, 50 Avenue de Verdun, 33612 Cestas Cedex, France
e
Division of Life Sciences, Incheon National University, 119, Academy-ro, Yeonsu-gu, Incheon 22012, Republic of Korea
f
Lab of Plant Growth Analysis, Ghent University Global Campus, 119, Songdomunwha-ro, Yeonsu-gu, Incheon 21985, Republic of Korea
g
Ghent University Global Campus, 119, Songdomunwha-ro, Yeonsu-gu, Incheon 21985, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Keywords: In the present study, the ecological status of four major South Korean rivers flowing near industrial complexes is
Biomonitoring assessed using physico-chemical analyses and various descriptors based on benthic diatom assemblages.
Diatoms Principal component analysis conducted with physical and chemical variables singled-out four sites as more
Lipid bodies severely impacted (higher concentration of metals and total nitrogen, higher conductivity and biological oxygen
Deformities
demand), while the remaining 12 sites indicated less impacted conditions, although showing nutrient enrich-
Diatom assemblages
ment. Less-contaminated sites based on physico-chemical properties showed higher cell densities, higher rich-
Diversity
Cell size ness and diversity, a larger proportion of live and healthy-looking diatoms, lower production of lipid bodies (in
terms of number and size) and fewer teratologies than more-contaminated sites. Non-taxonomical metrics (cell
heath status, cell size, lipid bodies and valve deformities) were in good agreement with traditional taxonomical
metrics (assemblage structure, richness, diversity). Overall, water quality assessment based on diatom assem-
blages and diatom-based metrics had a good fit with the available physico-chemical data and agreed on the most
impacted sites. However, additional sites showed signs of degradation based on the diatom metrics used. This
suggests that the use of biotic indicators provides useful complementary information on the health status at the
selected sites.

1. Introduction criticized, as they do not provide information on the impacts of en-

vironmental stresses on organisms (Wolska et al., 2007; Lavoie et al.,
Industrialization and urbanization contribute to freshwater eco- 2008). The aquatic biota, however, responds in a complex and dynamic
system degradation worldwide. Alteration and channelization of wa- manner to cumulative effects of various anthropogenic disturbances,
tercourses in urban areas result in progressive hydrological and hy- and thus provides an integrated picture of the biological health of an
drobiological degradation that may have serious impacts on the biotic ecosystem. Bioindicators therefore represent a crucial component for
and abiotic components of these systems. For example, urban water- assessing overall ecological health (Ricciardi et al., 2009), and are a
courses may suffer from elevated and fluctuating concentrations of valuable complement to traditional physico-chemical measurements.
nutrients and contaminants (e.g., metals, pesticides, aromatic poly- There is growing interest in studying the response of aquatic biota to
cyclic hydrocarbons), causing changes in species assemblage composi- various environmental stresses, its resilience, and its recovery trajec-
tion, with increases in the number and abundance of pollution-tolerant tories. Due to its key ecological role at the base of the food web, sessile
species (Morin et al., 2012). mode of life, cosmopolitan nature, and short life cycle, the algal com-
Water quality programs targeting chemical criteria alone have been ponent of biofilms is often selected for its sensitive and rapid response

⁎
Corresponding author at: Department of Marine Science, Incheon National University, 119, Academy-ro, Yeonsu-gu, Incheon 22012, Republic of Korea.
E-mail address: [email protected] (T. Han).

http://dx.doi.org/10.1016/j.ecolind.2017.07.043
Received 21 March 2017; Received in revised form 19 July 2017; Accepted 21 July 2017
Available online 25 September 2017
1470-160X/ © 2017 Elsevier Ltd. All rights reserved.
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Fig. 1. Study area showing the 16 sites selected on four main Korean rivers.

to environmental fluctuations as compared with higher-level organ- used in the development of diatom-based indices (Kelly and Whitton,
isms. In particular, diatoms (Bacillariophyta) have been extensively 1995; Ponader et al., 2007; Coste et al., 2009; Lavoie et al., 2006,
included in water quality assessment for monitoring the biological in- 2014). Multi-metric diatom indices have already been developed
tegrity of lotic ecosystems under various stresses such as eutrophica- (Zalack et al., 2010), but generally include taxonomy-derived metrics
tion, salinization, as well as metal and organic contamination (Gold such as % tolerant species, richness and chlorophyll a (Fore and Grafe,
et al., 2002; Debenest et al., 2008; Lavoie et al., 2014). Diatoms are a 2002). Various non-taxonomic diatom descriptors such as cell health
group of algae exhibiting a large diversity in terms of taxonomy, mor- status, motility, lipid bodies, cell density and teratologies (abnormal
phology and ecology, and are present in most aquatic and humid en- valves) are well documented as biomonitoring tools. For example,
vironments (Round et al., 1990). Diatoms have specific degrees of tol- Morin et al. (2012) reviewed the response of diatoms to metal con-
erance to environmental variables and have a rapid reproduction rate, tamination from the individual level (e.g., size, growth form, and
which are among the characteristics making them excellent indicators morphological abnormalities) to the assemblage structure level (turn-
of aquatic biological integrity (Kelly et al., 2008; Stevenson et al., 2008, over from sensitive to tolerant species). Similarly, Pandey and Bergey
2010). Subtle changes in environmental conditions may cause a shift in (2016) used various metrics to monitor the effect of chronic metal ex-
assemblage structure; sensitive diatoms decrease in abundance, while posure (Cu and Zn) on benthic diatoms, and reported a reduction in
tolerant species increase (Van Dam et al., 1994; Harding et al., 2005). diatom motility and cell size, as well as an increase in lipid body con-
The analysis of the diatom assemblage structure and species’ autecology tent and in teratology frequency.
allows for the characterization of past and present environmental In this study, benthic diatom assemblages from 16 sites selected
conditions and health status in a system (Kelly et al., 2008; Leira et al., along four major South Korean rivers flowing in close proximity to
2015; Schallenber and Saulnier-Tablbot, 2015). Diatom-based mon- industrial complexes and municipal effluents were examined to eval-
itoring is also gaining in popularity for the assessment of ecosystem uate their ecological status. For this purpose, multiple biological de-
rehabilitation to evaluate recovery success (e.g., Coe et al., 2009; Feld scriptors were used. First, more traditional metrics based on diatom
et al., 2011; Arini et al., 2012; Hamilton et al., 2015; Żelazna-Wieczorek assemblage composition, richness, diversity and cell density were ex-
and Nowicka-Krawczyk, 2015). amined. Additional biological metrics were also investigated: cell size
To adequately estimate the biological status of a system before and distribution, diatom health status (live, unhealthy or dead), and the
after alteration or rehabilitation, it is fundamental to develop appro- presence of lipid bodies and teratologies. Finally, we compared the
priate biological metrics. An approach based on multiple descriptors information provided by the various diatom-based metrics examined
may provide a more complete and integrated picture of the environ- with the estimation of the water quality status of each site based on a
mental conditions experienced by the diatoms, as compared with the suite of physico-chemical characteristics.
more traditional taxonomic approach (assemblage structure) generally

141
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

2. Materials and methods intact chloroplasts), unhealthy (with distorted/degraded chloroplasts)

and dead (empty frustules) (Gillet et al., 2009, 2011Gillet et al., 2009,
2.1. Study area and site selection 2011; Pandey et al., 2017). LBs were investigated in the living diatom
cells following the protocol of Pandey et al. (2015). Briefly, a lipophilic
The South Korean Ministry of the Environment (MOE) performs fluorescent dye (4,4-difluoro-1,3,5,7-tetramethyl-4-bora-3a,4adiaza-s-
regular water quality assessment of streams and rivers, including nu- indacene; BODIPY 505/515, Life technologies®, USA) was used to stain
merous variables such as metals, total phosphorus (TP), total nitrogen the lipid bodies inside the diatom cells, and the samples were examined
(TN), biological oxygen demand (BOD5), total dissolved solids (TDS), using fluorescence microscopy (Carl Zeiss/LSM 700; USA, equipped
pH, conductivity and total coliforms. The South Korean government with a WIG filter set for chlorophyll fluorescence). Excitation was at
also recently conducted an extensive study on micro-pollutants (mea- 450–490 nm and emission wavelengths were imaged through a 515nm-
sured at 159 sites impacted by effluents from 34 industrial complexes) long pass filter. Stained lipid bodies (green) are easily distinguished
in four major watersheds, and results from this project are summarized from chloroplasts (red) in the diatom cells. LBs in each individual living
in Cho et al. (2014). Based on the available information (eight years of diatom species were counted and measured at 400× and 1000×
regular monitoring performed by the South Korean government; magnification. Biovolume of individual lipid droplets was calculated
2008–2015), 16 sites were selected on the basis of potential chronic considering that LBs are more or less spherical, thus applying the for-
contamination from industrial effluents (Fig. 1). Sites were chosen in mula V = 4/3 πr3, where V is the volume of the lipid body and r is its
such a manner as to cover various sub-watersheds across South Korea. radius. The total relative contribution of all LBs inside individual
Additional information on the sampling sites and on the four water- diatom cells was estimated by adding the biovolume of each lipid body
sheds is provided as supplementary material (S1 and S2). divided by the biovolume of the frustules. A total of 100 live frustules
were observed in each pseudo-replicate for each of the above-men-
2.2. Surface water sampling and analyses tioned metrics (cells health status, lipid bodies, densities).

On-site water quality measurements and surface water collection 2.4. Taxonomical analysis and assessment of teratologies
were performed in September 2015 at the 16 selected sites following
the guidelines of the National surveys for stream ecosystem health Clean diatom frustules were also examined after digesting cell
(MOE, Korea). Temperature, pH, and dissolved oxygen were measured content to facilitate taxonomical identification and to assess the pre-
directly in the field with a multi-parameter display system (YSI 650, sence of abnormal valves. The fresh material collected was first treated
USA). Conductivity was measured in the field with a portable meter with 90% acetone to remove cytoplasmic content. Subsequently, the
(Milwaukee, USA). Total dissolved solid measurements were performed samples were dried for 30 min at 25 °C and treated with concentrated
in the field with an amperometric graphite electrode (HI98301 DiST® H2SO4 (99.99%, analytical grade, Sigma-Aldrich, St. Louis, USA) fol-
1). Biochemical oxygen demand was measured by dark and light bottle lowed by hydrogen peroxide (30%) (w/w, analytical grade, Sigma-
incubation for 24 h at 20 °C in controlled light conditions. EPA-ap- Aldrich, St. Louis, USA). The cleaned samples were then washed 10
proved Colilert-18 test by IDEXX was carried out to estimate total co- times with deionized water to remove the acid present inside the cells
liforms. A water automatic analyzer (Skalar/Netherlands, SAN ++) and in the liquid fraction. Diluted samples were spread onto cover slips
was used to measure total nitrogen (TN) and total phosphorus (TP). for drying. Permanent slides were mounted in Pleurax mounting
Analyses of metals (As, Ag, Cd, Cr, Cu, Hg, Ni, Pb and Zn) were con- medium (refractive index = 1.73) for microscopic examination.
ducted by inductively coupled plasma-optical emission spectro- Diatoms were counted and identified using images from Algaebase
photometer (ICP-OES; Varian Vista PRO, CA, USA) using 15 ml water (Academy of Natural Sciences of Philadelphia, http://diatom.acnatsci.
samples. Standard solutions were prepared fresh and calibration curves org/AlgaeImage/) and the Algal image database of India (AIDI)
(r2 > 0.995) were performed. To ensure the quality of the measure- (Pandey et al., 2016). Monographs on diatoms by Patrick and Reimer
ments, standard solutions were analyzed after every 10 samples to (1966, 1975) and a manual on stream periphyton by Biggs and Kilroy
verify their concentrations. Measurement precision ranged from 94 to (2000) were also consulted. A total of 500 valves per sample were
107%, and detection limits were calculated based on the standard de- enumerated. Diatom assemblage composition was expressed as percent
viations of blanks triplicates (range: 4–14 μg L−1). Organic compounds relative abundance. Qualitative and quantitative assessments of valve
were analyzed by the South Korean Ministry of the Environment, fol- deformities were conducted using permanent diatom slides (n = 3).
lowing the methods presented in Cho et al. (2014). All samples col- Deformities in diatom frustules were categorized into five types: (1)
lected for analyses of metals and organic compounds were previously deformities in valve outline, (2) deformities in striations, (3) defor-
filtered through 0.45 μm filters. mities in the sternum/raphe structure, (4) mixed deformities and (5)
deformities in the costae (Falasco et al., 2009a; Pandey et al., 2014,
2.3. Benthic diatom sampling and fresh material analyses 2015). Diatom cell biovolumes were classified in three categories based
on Omnidia database (https://hydrobio-dce.irstea.fr/telecharger/
Benthic diatom assemblages were collected in September 2015 from diatomees-ibd/): small (< 400 μm3), intermediate (400–4000 μm3),
hard surfaces (concrete walls, scrapped surface area ∼25 cm2) using a and large (> 4000 μm3).
new toothbrush at each of the 16 selected sites. Six samples from each
site were collected and pooled together to obtain a composite sample 2.5. Statistical analyses
(Winter and Duthie, 2000; Lavoie et al., 2006). The collected material
was placed in a plastic bottle with 5 ml of river water, kept cold and Shannon’s diversity index values were calculated using the freeware
transported to the laboratory, and analyzed within one week. Fresh “PAST” (Hammer et al., 2001) for each pseudo-replicate (n = 3). A
samples were analyzed under a light microscope at 450× and 100× preliminary detrended correspondence analysis was conducted on
magnification (Zeiss, Axioskop 2, equipped with a Nikon camera) for species data and revealed a short gradient length (1.06) suggesting a
live diatom observation. All diatom cells were counted to estimate linear response. Principal component analyses (PCAs; Euclidian dis-
densities with a Spencer’s haemocytometer at 450× magnification tance) were thus conducted to explore site distribution based on (i)
using three aliquots of 1 ml (pseudo-replicates, n = 3) of composite physico-chemical variables and (ii) diatom species assemblages. En-
biofilm sample. Fresh diatom cells were examined to evaluate health vironmental data were centered and standardized. Diatom relative
status (live, unhealthy and dead cells), and for the presence of lipid species abundances were square root transformed. Only the species
bodies (LBs). Diatom cells were categorized as live or healthy (with present at a relative abundance of ≥1% (in at least one sample) were

142
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Table 1
Physical and chemical variables measured in the surface water of the 16 sites. nd = under detection limit.

River Site DO mg BOD5 mg Coli CFU Temp °C Cond pH TN mg TP mg TDS mg Cu mg Ni mg l−1 Zn mg l−1
l−1 l−1 ml−1 μS cm−1 l−1 l−1 l−1 l−1

Nakdong Ayang 11.44 1.51 2 23.4 470 8.8 5.96 0.028 380 nd nd 0.075
Dalseo 11.54 1.93 10 27.3 2895 7.67 10.04 0.084 1704 0.029 0.068 0.017
Keumho 11.55 2.08 9 26.4 2580 8.16 10.93 0.112 2180 0.048 0.045 0.034
Yeongsan Gwangjugong 11.1 2.08 72 22.1 309 8.02 3.96 0.111 184 nd nd 0.071
Gwangju-1 11.65 1.43 6 23.4 204 9.25 2.23 0.013 190 nd nd 0.071
GJ-2 11.75 1.49 5 22.8 321 7.87 3.91 0.111 242 nd nd 0.05
PD-1 11.77 1.66 19 20.6 170 7.67 5.51 0.115 276 nd nd 0.07
Geum Ohryang 11.75 0.71 16 23.4 439 7.6 6.22 0.34 282 nd nd 0.09
Bangchuk 11.12 2.53 8 23.2 438 8.14 16.04 0.194 40 nd nd 0.062
Masan 11.3 3.14 15 23.3 321 8.38 7.23 0.297 302 nd nd 0.057
Sucheol 11.48 1.44 5 22.4 362 8.74 7.14 0.304 302 nd nd 0.053
Miho-5 11.45 2.63 15 21.5 659 7.45 7 0.12 454 nd nd 0.081
Miho-8 11.44 1.22 4 24.0 482 8.15 3.83 0.12 268 nd nd 0.079
Miho-7 11.33 1.25 46 21.8 505 7.72 5.44 0.111 342 0.004 nd 0.075
Han Soyo 10.34 9.25 240 27.6 4007 7.72 14.19 0.098 2132 nd 0.021 0.3
Daejeon 10.86 3.43 64 24.7 2579 8.35 9.88 0.097 898 0.043 0.016 0.44

included in the PCA. Ordinations were performed with the software sampling of September 2015. A PCA was conducted to explore site dis-
CANOCO 5.0 (ter Braak and Šmilauer, 2012). tribution based on surface water characteristics (Fig. 2), and showed distinct
differences among sampling sites. The PCA biplot explained 69% of the
3. Results and discussion variance (PCA axis 1 contributed 50% and PCA axis 2, 19%). The ordination
indicates that Dalseo, Keumho, Soyo and Daejeon cluster on the right-hand
3.1. Environmental characteristics of the study sites side of the biplot. Within the 16 sites, these four sites have higher con-
centrations of Ni/Cu or Zn and nutrients (TN), along with high conductivity,
Values of the chemical and physical variables measured at the 16 sites TDS, BOD5, and coliforms, and were therefore classified as “severely im-
are presented in Table 1. Three metals detected in water samples were pacted”. The 12 other sites occupy the left-hand side of the biplot and
above quantification limits with elevated values: Cu (Keumho - cluster close to the center. These sites generally have high nutrient con-
= 0.048 mg l−1), Zn (Daejeon = 0.44 mg l−1), Ni (Dalseo - centrations potentially reflecting eutrophication, but were classified here as
−1
= 0.068 mg l ). Certain sites showed elevated conductivity values (Soyo, “less-impacted” because chemical analyses did not reveal any metal con-
Dalseo, Keumho, Daejeon > 2000 μS cm−1) and coliforms tamination. Dalseo, Keumho, Soyo and Daejeon cluster apart from the other
(Soyo = 240 CFU ml−1). The data also reflected nutrient enrichment at sites mostly due to high metal concentrations at these sites (values were
most sites with mean TP reaching 0.141 mg l−1 (ran- under detection limits at the other sites). However, a PCA conducted
ge = 0.013–0.34 mg l−1) and mean TN reaching 7.5 mg l−1 (ran- without metal data (results not presented) also revealed that these four sites
ge = 2.23–16.04 mg l−1). Although governmental routine monitoring fre- stood apart from the others, principally along a gradient of total nitrogen,
quently reports the presence of organic contaminants (Cho et al., 2014), coliforms, BOD5, and conductivity. In general, based on the available water
concentrations were found to be under detection limit based on the water quality parameters, the sampling sites on the Nakdong and Han rivers

Fig. 2. Principal component analysis showing site

distribution based on physico-chemical parameters.
Axes 1 and 2 combined explain 69% of the variance.

143
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Fig. 3. Principal component analysis showing site

distribution based on diatom assemblages (including
species with relative abundance > 1%). Axes 1 and 2
combined explain 45% of the variance. The dashed
line is an arbitrary separation to classify the sites as
either severely-impacted or less-impacted. Diatom
species were a posteriori added onto the biplot to
indicate the most important species responsible for
structuring site distribution. Achnanthidium minu-
tissimum (AMIN), Cocconeis placentula (CPLA),
Cymatopleura solea (CSOL), Eolimna minima (EMIN),
Fragilariforma virescens (FVIR), Gomphonema par-
vulum (GPAR), Luticola muticopsis (LMUT), Navicula
capitatoradiata (NCAP), Navicula gregaria (NGRE),
Navicula rostellata (NROS), Nitzschia frustulum
(NFRU), Nitzschia inconspicua (NINC), Nitzschia palea
(NPAL), Pinnularia viridis (PVIR), Surirella linearis
(SLIN).

Fig. 5. Species richness and Shannon diversity index values (Mean ± SE; n = 3) of
diatom assemblages collected at the 16 sites.

Fig. 4. Diatom cell density (Mean ± SE; n = 3) in samples collected at the 16 sites.

ordination, with site distribution mainly structured by the diatoms

appeared to be more impacted than the sites on the Yeongsan and Geum Eolimna minima, Gomphonema parvulum, Nitzschia frustulum, Nitzschia
rivers. inconspicua and Nitzschia palea (Fig. 3). Ayang and Keumho clustered at
the top of the biplot and their position was mainly driven by Ach-
3.2. Diatom assemblage structure nanthidium minutissimum and Cocconeis placentula. The other ten sites
were mainly positioned on the left of the graph and showed higher
A total of 65 diatom species were observed in this study, but only 23 relative abundances of Cymatopleura solea, Fragilariforma virescens, Lu-
were included in the ordination based on the 1% relative abundance ticola muticopsis, Navicula capitatoradiata, Navicula gregaria, Navicula
criterion. This is based on the results of Lavoie et al. (2009), which rostellata, Pinnularia viridis and Surirella linearis. A dashed line was a
show that deleting rare species that occur with relative abundances posteriori arbitrarily added to the graph to separate the sites based on
of < 2% does not affect the general outcome of site distribution using their distribution: most severely impacted sites mainly reflecting metal
ordinations. However, because only 65 species were observed in the contamination, versus less-impacted sites (but still showing relatively
present study, a more conservative criterion of 1% was used. PCA high nutrient concentrations). The diatom-based PCA biplot explained
analysis performed on the diatom relative abundance data showed that 45% of the variance in site distribution (PCA axis 1 contributed 31%
Dalseo, GJ-2, Soyo and Daejeon clustered on the right-hand side of the and PCA axis 2, 14%).

144
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Fig. 6. Diatom assemblages collected at the 16 sites sampled in four major South Korean rivers. Black, grey and white sections show the relative abundances of diatom size classes: small
(< 400 μm3), intermediate (400–4000 μm3), and large (> 4000 μm3).

3.3. Comparing PCAs: one-time water chemistry versus diatom assemblages separated the sites into two main groups that may be coarsely char-
acterized as “severely impacted” and “less-impacted”. However, the
Comparing site distribution based on physico-chemical variables diatom-based PCA suggests that, in addition to Dalseo, Keumho, Soyo
versus diatom assemblages revealed similarities in overall water quality and Daejeon, the assemblages from Ayang and GJ-2 may also reflect
assessment: Dalseo, Soyo and Daejeon clustered apart from the other impacted conditions. This situation may represent a good example in
sites in both approaches, suggesting more contaminated environments. which the biota provides additional information on water quality status
The diatom assemblage structure observed at GJ-2 was similar to the beyond traditional chemistry measurements. The fact that Ayang and
assemblages found at the three above-mentioned sites, which suggests GJ-2 suggest more impacted conditions based on diatom assemblages
that this site may also have impacted conditions. Keumho, classified as compared to chemistry-based assessment may result from temporally
severely impacted based on chemistry, also clustered apart from the fluctuating contamination (e.g., episodic releases). The timing of spot
sites showing less impacted conditions. Ayang may also be severely sampling for water chemistry analyses may not correspond to the re-
impacted as it clustered with Keumho and was apart from the other lease of pollutants, such as metals.
sites. Overall, the two PCAs conducted using chemistry or diatoms

145
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

differences, it is important to keep in mind for data interpretation that

the cell densities are nevertheless within the same order of magnitude.
Soyo, Daejeon and Keumho have water chemistry profiles reflecting
higher degradation levels, which may in part explain the lower cell
densities in the samples. Lower diatom cell counts have been reported
in metal-contaminated water bodies in numerous studies (Gold et al.,
2002; Duong et al., 2010), and namely in India under Cu, Zn and Pb
stress in the river Ganges (Pandey et al., 2014), and under Cu and Zn
contamination in mining areas of Rajasthan (Pandey and Bergey, 2016).
However, Gwangjugong and Gwangju-1, which also show lower cell
densities, did not cluster with the most impacted sites based on water
chemistry. These results suggest that cell density alone may not always
provide a good appreciation of water quality. Indeed, biofilm biomass
(including diatoms) is influenced by numerous other factors, such as
water velocity (Villeneuve et al., 2011) and grazing pressure (Guasch
et al., 2016), and these may have a greater effect on cell density than
water chemistry at certain sites. Furthermore, a high diatom cell count
does not necessarily reflect good conditions (as observed in the sample
collected at Ayang), as it could also be the result of nutrient enrichment.
While this metric is interesting, particularly in the context of metallic or
organic contamination, it should be used as complementary informa-
tion with other biological descriptors. Comparing diatom cell densities
with soft-algae densities might also be a good metric to consider, as the
changes in the proportions of algal groups may provide valuable ad-
ditional information on the ecosystem structure and function. For ex-
Fig. 7. Relative contribution of live frustules, unhealthy frustules, and dead frustules in
the samples collected at the 16 sites. ample, the fact that diatoms represent a high-quality food resource for
invertebrates and their reestablishment after a disturbance, in com-
parison with other algal groups (Steinman and McIntire, 1990; Morin
3.4. Non-taxonomical diatom-based metrics
et al., 2010; Arini et al., 2013), may be a critical factor in the recovery
rates of consumers.
3.4.1. Cell density
Diatom frustule density was estimated at each of the 16 sites
(Fig. 4). Lower diatom cell densities (120,000–180,000 frustules cm−2) 3.4.2. Benthic diatom richness and diversity
were observed at Keumho, Gwangjugong, Gwangju-1, Soyo and Dae- Biodiversity descriptors such as Shannon’s diversity index and spe-
jeon, while relatively higher frustule densities (300,000–570 cies richness are regularly used as tools for investigating the biological
000 cm−2) were estimated at the 11 other sites. Despite the observed integrity status of an ecosystem. For example, Verb and Vis (2000)
reported significantly lower species richness and Shannon’s diversity

Fig. 8. Gradient in lipid body biovolume observed in six commonly occurring diatom species (a) Navicula gregaria, (b) Navicula capitatoradiata, (c) Nitzschia inconspicua, (d) Navicula
rostellata, (e) Nitzschia frustulum and (f) Nitzschia palea. Scale bar = 7 μm.

146
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Table 2a
Overall range in % biovolume contribution of lipid bodies to total cell biovolume at each site.

Rivers Nakdong Yeongsan Geum Han

Sites Ayang Dalseo Keumho Gwangjugong Gwangju-1 GJ-2 PD-1 Ohryang Masan Sucheol Bangchuk Miho-5 Miho-8 Miho-7 Soyo Daejeon

% biovolume 4–7 8–28 5–40 2–5 4–7 2–6 1–3 2–5 2–7 1–5 4–7 2–6 4–7 2–6 5–60 12–55
contribution
of LBs to total
cell volume

Table 2b
Detailed lipid bodies status in six commonly occurring diatom species. Sites have been grouped as severely contaminated (Dalseo, Keumho, Soyo and Daejeon) and less-contaminated
waters bodies (other 12 sites) based on the chemistry-based ordination.

Number of lipid bodies Diameter (μm) of lipid bodies % biovolume contribution of lipid bodies to total cell volume

Less contaminated Contaminated sites Less contaminated Contaminated sites Less contaminated sites Contaminated sites
sites sites

Navicula gregaria 0–2 2–4 0.0–0.5 0.5–3.0 4–7 20–60

Navicula capitatoradiata 0–2 2–4 0.0–3.0 0.2–3.0 2–6 15–60
Navicula rostellata 0–2 2–6 0.0–0.5 0–2.5 2–5 5–20
Nitzschia inconspicua 0–4 4–6 0.0–0.2 0–3.2 1–3 6–35
Nitzschia frustulum 0–4 0–4 0.0–0.3 0.5–2.0 2–7 10–20
Nitzschia palea 0–2 2–8 0.0–0.4 0.5–2.0 1–3 8–25

index at acid mine drainage sites than at reclaimed sites, while Barral- which did not seem to reflect metallic contamination based on the
Fraga et al. (2016) reported lower values for these metrics in diatom- available data. On the other hand, the ordination conducted on diatom
dominated biofilms contaminated with arsenic. Similarly, the diatom assemblages (PCA) indicated severely impacted conditions for these
assemblages collected in the present study showed lower species rich- two sites (Fig. 3) that may not have been noticed based solely on the
ness (10–23) and/or Shannon’s diversity index (1.6–2.0) at polluted available chemistry data. It is also possible that factors other than
sites Ayang, Dalseo, Keumho, GJ-2, Soyo and Daejeon, but also at sites contamination favored small diatoms, including grazing and coloniza-
Gwangjugong and Gwangju-1. PD-1, Ohryang, Bangchuk, Masan, Su- tion phase following a physical disturbance. Gwangjugong, PD-1,
cheol, Miho-5, Miho-8 and Miho-7 showed higher species richness Masan, and Miho-5 showed high proportions of small-sized diatoms,
(30–50) and Shannon diversity index (2.6–3.8) (Fig. 5). As a general although neither of these sites was considered as severely impacted
trend, the sites clustered on the right-hand side of the PCA (Fig. 2) with based on chemistry or diatom assemblage data. This suggests that cell-
more severely degraded conditions were associated with lower richness size alone may not provide a complete picture of an ecosystem’s health
and diversity. However, these metrics also showed lower values for status. However, as a general trend, the sites that were considered less
certain sites that were not considered as severely impacted, based on contaminated showed a dominance of intermediate and large diatom
chemical analyses. This lack of concordance at some sites may suggest species, reaching up to more than 50% of the assemblage.
that the available one-time physico-chemical data do not reflect the
temporally fluctuating nature of these waters and that diatom diversity 3.4.4. Diatom cell health status
brings additional insight on the environmental integrity of the sites. It Investigation of live and dead diatoms in the assemblage may pro-
may also indicate that diversity and richness are not always good vide additional environmental information (Cox, 1998; Stevenson and
proxies of water quality, and perhaps should not be used alone as Pan, 1999; Gillett et al., 2011). Wood et al. (2014, 2016) reported
bioindicators. This aspect is discussed in Blanco et al. (2012), where health status of diatoms (intact vs. unhealthy cells) as a sensitive and
diversity indices are reported to generally exhibit poor linear correla- rapid method for assessing herbicide toxicity. Gillett et al. (2009) ar-
tions with environmental factors indicating ecological status. Further- gued that conventional diatom cell count methods provide taxonomical
more, sampling effort or sample size may strongly affect richness and confidence while the live diatom count offers better ecological relia-
diversity values, thus influencing water quality status comparisons. bility. In a general biomonitoring perspective, both approaches are
However, this does not represent an issue in the present study as the valuable and complementary. Among the 16 sites examined, Gwangju-
same protocol was used for all samples (∼500 frustules per sample). 1, PD-1, Ohryang, Bangchuk, Masan, Sucheol, Miho-5, Miho-7 and
Miho-8 showed a predominance (40–50%) of live frustules (Fig. 7),
3.4.3. Cell size while percent unhealthy frustules showed higher values (> 40%) at
It has been shown in numerous studies that small-sized species tend Dalseo and GJ-2. Percent dead frustules in the diatom assemblages
to dominate in metal-contaminated environments (Ivorra et al., 1999; showed higher values (40–65%) at Ayang, Gwangjugong, Soyo and
Morin et al., 2008, 2012; Tlili et al., 2011; Barral-Fraga et al., 2016). Daejeon. As a general trend, the less-impacted sites clustering on the
Decrease in diatom frustule size within a given species has also been left-hand side of the diatom-based ordination (Fig. 3) also showed
reported as a response to metal stress. For example, Cattaneo et al. higher proportions of live diatoms, suggesting healthier conditions. This
(2004) reported that metal contamination within sediment cores from a trend was not as clear when relating diatom health status and chemistry
Canadian lake was responsible for size reduction. Similarly, Luís et al. data, indicating that diatoms may have responded to other stresses or
(2011) reported size decline in Brachysira vitrea frustules at metal- that there is a time lag between water quality change and cell de-
contaminated sites of Corona stream, Portugal. In the present in- gradation. It is also possible that the presence of dead cells is in-
vestigation, the four sites considered as the most impacted based on dependent of site degradation. For example, very low water velocity
physico-chemical parameters (Dalseo, Soyo, Daejeon, Keumho) were may favor dead cell deposition. However, all sites sampled in this study
characterized by a dominance of small adnate diatom species (Fig. 6). were characterized by relatively low flow, which suggests that the
Ayang and GJ-2 also showed high abundances of small-sized diatoms, variation in diatom health status among sites are more likely to reflect

147
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Fig. 9. Deformed frustules observed in various diatom species (a) Ulnaria ulna, (b) Cymbella turgidula, (c) Cocconeis placentula, (d) Gomphonema parvulum, (e) Achnanthes exigua, (f)
Eolimna minima, (g) Diatoma vulgaris, (h) Tabularia fasciculata, (i) Fragilaria capucina, (j) Staurosira construens and (k) Planothidium frequentissimum. “N” denotes normal frustules. Scale
bar = 5 μm.

water quality than differences in dead cell accumulation. to a metal stress, as LBs were observed with higher abundances and
larger sizes at the sites showing the highest concentrations of metals.
3.4.5. Lipid bodies However, it is important to note that other environmental stresses may
Lipid bodies constitute energy reserves in algae. Under stressed have influenced LB production and that this was not assessed in this
conditions (e.g. nutrient depletion, metal contamination, excess light), study. LB stimulation by metal contamination has been widely reported
these lipid droplets are reported to increase in number and size (Pandey (Einicker-Lamas et al., 2002; Liu et al., 2008; Wang et al., 2009; Pillai
et al., 2015; Yang et al., 2013; Liang et al., 2015; d’Ippolito et al., 2015). et al., 2014; Pandey et al., 2015, 2016; Gautam et al., 2017). LBs are
The number of LBs and their respective diameters were examined in six promising biomarkers of stress although more research is required to
commonly occurring diatom species at the 16 study sites (N. gregaria, N. define their variability under various natural conditions for optimal use
capitatoradiata, N. inconspicua, N. rostellata, N. frustulum and N. palea) in a bioassessment perspective. The approach is relatively easy, as
(Fig. 8). Table 2a presents the range in % biovolume contribution of taxonomical skills are not required (all diatom species are likely to
lipid bodies to total cell volume at each site, while Table 2b presents a produce LBs in larger number and size under stressed conditions).
detailed summary of the LB's characteristics for the six most common
species in which LBs were counted and measured. Looking at the sites 3.4.6. Deformities
individually, the proportion of LBs relative to cell biovolume was Numerous studies relate deformities with environmental dis-
higher for Soyo, Daejeon, Dalseo and Keumho (Table 2a). A marked turbances (see review in Falasco et al., 2009a), particularly in metal-
difference in LBs was observed when comparing contaminated sites contaminated conditions (Morin et al., 2012; Lavoie et al., 2012;
versus less-contaminated sites (based on physico-chemistry data) where Leguay et al., 2016; Pandey and Bergey, 2016). In the present study, 13
more impacted conditions seemed to have stimulated LB production. In species out of the total 65 species identified showed deformities in their
the present study, lipid body analysis suggests that diatoms responded frustules (Fig. 9). Deformed frustules were encountered in small

148
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

unidentified environmental variables or it could indicate that chemistry

values available for this study did not reflect the actual, prevalent en-
vironmental conditions due to temporal fluctuation. Overall, abnormal
valve outline was the most widely occurring aberration and was found
in A. exigua, C. placentula, C. turgidula, D. vulgaris, E. minima, F. capu-
cina, G. parvulum, P. frequentissimum, Staurosira construens, T. fasciculata
and U. ulna. However, the deformity in the striation pattern and mixed
deformity showed the highest proportion, particularly at sites Soyo and
Daejeon (Fig. 11). Deformity types may be metal specific. For example,
Pandey et al. (2014, 2015) also reported prevalence of abnormal
striations and mixed deformities under Zn stress in field and laboratory
studies. Deformed raphes were rarely observed and found only at a few
sites (and only in Navicula recens), while deformed costae were ob-
served only at Miho-8 (and only in D. vulgaris).
Numerous studies, including the present investigation, report de-
formed frustules in adnate and araphid diatoms (Morin et al., 2008;
Falasco et al., 2009a,b; Pandey et al., 2014; Leguay et al., 2016; Pandey
et al., 2016), which suggests that species belonging to these groups may
be more prone to teratologies (also see discussion in Lavoie et al.,
2017). For example, deformities in A. exigua and C. placentula were
frequently observed in this study. Deformed frustules were also noticed
in A. minutissimum, C. turgidula, Eolimna subminusula, F. capucina, G.
parvulum, P. frequentissimum and U. ulna. These are species that have
Fig. 10. Occurrence of deformed valves (% relative abundance, Mean ± SE; n = 3) been regularly reported as presenting aberrations under metal-con-
observed in the samples collected at the 16 sites. taminated environments (Morin et al., 2008; Falasco et al., 2009a,b;
Silva et al., 2009; Lavoie et al., 2012; Cantonati et al., 2014; Pandey
et al., 2014, 2015, 2016).

3.4.7. Water quality status as a function of chemistry and diatom-based

metrics
Previous biomonitoring (algae and macroinvertebrates) assessment
indicated that many South Korean rivers and streams are severely de-
graded, with biological conditions that are much worse than their water
chemistry suggests (Hwang et al., 2011). Biological metrics do not always
correlate well with water chemistry, which is the key argument as to
whether biological indicators should be included in routine water quality
assessment. The present investigation seems to corroborate this finding;
four sites were characterized as more severely impacted based on chem-
istry data alone (mostly metals, TN, salinity/conductivity and coliforms),
while at least two additional sites showed signs of degradation/stress
based on multiple diatom-based metrics. Table 3 presents a simplified
summary of the water quality status of the 16 sites based on the traditional
physico-chemical information, diatom-based ordination, as well as non-
taxonomic descriptors. Most importantly, all metrics used agreed on the
poor status of the Han River sites (Soyo and Daejeon), and on the lower
impacts in the Geum River. Interestingly, along the Nakdong River, the
majority of the descriptors put Dalseo and Keumho into the impacted site
category. This result is not surprising as the biota generally integrates the
environmental conditions well; if the chemistry suggests a degraded
Fig. 11. Relative proportion of each of the five types of deformities observed in the system, the organisms should also reflect this reality. On the other hand,
samples collected at the 16 sites. chemistry-based assessment is subjected to strong variability and may not
always provide a complete and integrated picture of the real conditions
(Lavoie et al., 2008). Because biological metrics integrate environmental
(Achnanthes exigua, E. minima, Fragilaria capucina, G. parvulum, Pla- conditions over time, their response may be delayed compared to che-
nothidium frequentissimum and Staurosira sp.) as well as in larger diatom mical analyses. The time frame within which a biotic response will be
species (C. placentula, Cymbella turgidula, Diatoma vulgaris, Tabularia noted depends on the nature of the biological endpoint assessed (func-
fasciculata and Ulnaria ulna). Mean percent deformities were estimated tional versus structural descriptors). This could explain why several me-
in each of the 16 samples (Fig. 10). Ayang, Dalseo, GJ-2, Soyo and trics calculated for Ayang and GJ-2 suggest a biological impairment, while
Daejeon showed relatively higher percentages of teratologies (2, 2, 2, 6 chemical status does not support such a trend: the diatom assemblages
and 8%, respectively) compared with the other sites (< 0.5%). Ac- may have responded to one or more stresses that were not assessed here,
cording to Morin et al. (2008) and Arini et al. (2012), naturally oc- or it may suggest that the system is in a transitional state towards ecolo-
curring deformities in diatom assemblages are usually rare (below 1%). gical improvement or degradation. Moreover, following the control of
The frequencies of teratologies observed in the present study therefore pollution sources, biological metrics may still indicate impairment of
suggest a response to a stress. While Dalseo, Soyo and Daejeon showed water quality and biological integrity due to sediment contamination and
higher concentrations of metals (Table 1, Fig. 2), the causes of defor- its lingering effects. The diagnostic of the status of Gwangjugong and
mities at Ayang and GJ-2 are unclear. It may be a response to Gwangju-1 based on the biological descriptors used was not as

149
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

Table 3
Summary of the water quality descriptors used in this study and site classification as “severely impacted (grey)” or “less-impacted (white)”
based on each metric.

straightforward because fewer metrics suggested degraded conditions at Appendix A. Supplementary data
these sites. The fact that some diatom metrics are more likely to respond
quickly to changes in chemistry, in contrast to more resilient indicators Supplementary data associated with this article can be found, in the
such as taxonomical composition, may reflect progressive recovery/de- online version, at http://dx.doi.org/10.1016/j.ecolind.2017.07.043.
gradation of sites. This explanation could be of particular interest at
Gwangjugong, where rehabilitation operations have been undertaken. References
This hypothesis agrees with the results of Arini et al. (2012), who showed
that diatom composition failed to demonstrate the recovery, on an inter- Arini, A., Feurtet-Mazel, A., Morin, S., Maury-Brachet, R., Coste, M., Delmas, F., 2012.
mediate time frame, of a mine-impacted site after remediation practices. Remediation of a watershed contaminated by heavy metals: a 2-year field biomoni-
toring of periphytic biofilms. Sci. Total Environ. 425, 242–253.
Thus, the multiple metrics used could also be informative of the “ecolo- Arini, A., Durant, F., Coste, M., Delmas, F., Feurtet-Mazel, A., 2013. Cadmium deconta-
gical trajectory” and time-dependent recovery signs, by highlighting the mination and reversal potential of teratological forms of the diatom Planothidium
first stages of change in environmental conditions. Ongoing ecosystem frequentissimum (Bacillariophyceae) after experimental contamination. J. Phycol. 49,
361–370.
recuperation processes could thus be estimated through improvements in Barral-Fraga, L., Morin, S., Rovira, M.D.M., Urrea, G., Magellan, G., Guasch, H., 2016.
early-responding descriptors (eg., teratologies, cell health status, LB con- Short-term arsenic exposure reduces diatom cell size in biofilm communities.
tent), whereas recovery in diatom taxonomical composition as an end- Environ. Sci. Pollut. R. 23, 4257–4270.
Biggs, B.J.F., Kilroy, C., 2000. Stream Periphyton Monitoring Manual. The New Zealand
point indicator would necessitate a longer time frame. The complementary Ministry for the Environment, Christchurch.
information provided by non-taxonomical descriptors and biomarkers, Blanco, S., Cejudo-Figueiras, C., Tudesque, L., Bécares, E., Hoffmann, L., Ector, L., 2012.
combined with classical assemblage structure analyses, possibly makes the Are diatom diversity indices reliable monitoring metrics? Hydrobiologia 695,
199–206.
overall water quality assessment more powerful and reliable.
Cantonati, M., Angeli, N., Virtanen, L., Wojtal, A.Z., Gabriell, J., Falasco, E., Lavoie, I.,
Morin, S., Marchetoo, A., Fortin, C., Smirnova, S., 2014. Achnanthidium minutissimum
(Bacillariophyta) valve deformities as indicators of metal enrichment in diverse
4. Conclusion widely-distributed freshwater habitats. Sci. Total Environ. 75, 201–215.
Cattaneo, A., Couillard, Y., Wunsam, S., Courcelles, M., 2004. Diatom taxonomic and
This study underscores the importance of assessing ecological integrity morphological changes as indicators of metal pollution and recovery in Lac Dufault
(Québec, Canada). J. Paleolimnol. 32, 163–175.
based on a suite of chemical, physical and biological characteristics of a Cho, E., Khim, J., Chung, S., Seo, D., Son, Y., 2014. Occurrence of micropollutants in four
system. While all metrics examined have the potential to reflect some major rivers in Korea. Sci. Total Environ. 491–492, 138–147.
aspects of biological integrity, the most reliable approach would be to Coe, H.J., Kiffney, P.M., Pess, G.R., Kloehn, K.K., McHenry, M.L., 2009. Periphyton and
invertebrate response to wood placement in large pacific coastal rivers. River Res.
include several descriptors in order to obtain a more complete picture. Appl. 25, 1025–1035.
However, this may not be realistic in a routine monitoring program where Coste, M., Boutry, S., Tison-Rosebery, J., Delmas, F., 2009. Improvements of the biolo-
time and resources are limited. Diatom assemblage structure, widely used gical diatom index (BDI): description and efficiency of the new version (BDI-2006).
Ecol. Indic. 9, 621–650.
in biomonitoring, allowed for a clear separation of the sites and generally Cox, E.J., 1998. A rationale and some suggestions for developing rapid biomonitoring
showed good concordance with other metrics. The size and number of techniques using identification of live diatoms. Proceedings of the 15th International
lipid bodies and abnormal diatom frustules (teratologies) were also pow- Diatom Symposium 43–50.
d’Ippolito, G., Sardo, A., Paris, D., Vella, F.M., Adelfi, M.G., Botte, P., Gallo, C., Fontana,
erful and reliable biomarkers reflecting a response to one or multiple
A., 2015. Potential of lipid metabolism in marine diatoms for biofuel production.
stressors. Supplementary information from cell size and diatom health Biotechnol. Biofuels 8 (28), 1–10.
status provided valuable addition as well. Biomonitoring programs based Debenest, T., Silvestrea, J., Coste, M., Delmas, F., Pinelli, E., 2008. Herbicide effects on
freshwater benthic diatoms: induction of nucleus alterations and silica cell wall ab-
on multiple descriptors would thus be an interesting and promising ap-
normalities. Aquat. Toxicol. 88, 88–94.
proach to develop for the assessment of aquatic ecosystems integrity. Duong, T.T., Morin, S., Coste, M., Herlory, O., Feurtet-Mazel, A., Boudou, A., 2010.
Experimental toxicity and bioaccumulation of cadmium in freshwater periphytic
diatoms in relation with biofilm maturity. Sci. Total Environ. 408, 552–562.
Acknowledgements Einicker-Lamas, M., Mezian, G.A., Fernandes, T.B., Silva, F.L.S., Guerra, F., Miranda, K.,
Attias, M., Oliveira, M.M., 2002. Euglena gracilis as a model for the study of Cu+2?
and Zn+2? toxicity and accumulation in eukaryotic cells. Environ. Pollut. 120,
This work was supported by a Grant of Incheon National University 779–786.
Research (2013). We are grateful to Dr. J.C. Taylor (North-West Falasco, E., Bona, F., Badino, G., Hoffmann, L., Ector, L., 2009a. Diatom teratological
University, South Africa) for his generous gift of Pleurax, and to Dr. É. forms and environmental alterations: a review. Hydrobiologia 623, 1–35.
Falasco, E., Bona, F., Ginepro, M., Hlúbiková, D., Hoffmann, L., Ector, L., 2009b.
Saulnier-Talbot for providing language revision and comments on an Morphological abnormalities of diatom silica walls in relation to heavy metal
earlier version of the text.

150
L.K. Pandey et al. Ecological Indicators 84 (2018) 140–151

contamination and artificial growth conditions. Water SA 35, 595–605. Pandey, L.K., Kumar, D., Yadav, A., Rai, J., Gaur, J.P., 2014. Morphological abnormalities
Feld, C.K., Birk, S.D., Bradley, C., Hering, D., Kail, J., Marzin, A., Melcher, A., Nemitz, D., in periphytic diatoms as a tool for biomonitoring of heavy metal pollution in a river.
Pedersen, M.L., Pletterbauer, F., Pont, D., Verdonschot, P.F.M., Friberg, N., 2011. Ecol. Indic. 36, 272–279.
From natural to degraded rivers and back again: a test of restoration ecology theory Pandey, L.K., Han, T., Gaur, J.P., 2015. Response of a phytoplanktonic assemblage to
and practice. Adv. Ecol. Res. 44, 119–209. copper and zinc enrichment in microcosm. Ecotoxicology 24, 573–582.
Fore, L.S., Grafe, C., 2002. Using diatoms to assess the biological condition of large rivers Pandey, L.K., Ojha, K.K., Singh, P., Singh, C.S., Dwivedi, S., Bergey, E.A., 2016. Diatoms
in Idaho (USA). Freshw. Biol. 47, 2015–2037. image database of India (DIDI): a research tool. Environ. Technol. Innov. 5, 148–160.
Gautam, S., Pandey, L.K., Vinayak, V., Arya, A., 2017. Morphological and physiological Pandey, L.K., Bergey, E.A., 2016. Exploring the status of motility, lipid bodies, deformities
alterations in the diatom Gomphonema pseudoaugurdue to heavy metal stress. Ecol. and size reduction in periphytic diatom community from chronically metal (Cu, Zn)
Indic. 72, 67–76. polluted waterbodies as a biomonitoring tool. Sci. Total Environ. 550, 372–381.
Gillett, N., Pan, Y., Parker, C., 2009. Should only live diatoms be used in the bioassess- Pandey, L.K., Bergey, E.A., Jie, L., Park, J., Choi, S., Lee, H., Depuydt, S., Oh, Y.T., Lee,
ment of small mountain streams? Hydrobiologia 620, 135–147. S.M., Han, T., 2017. The use of diatoms in ecotoxicology and bioassessment: insights
Gillett, N.D., Pan, Y., Manoylov, K.M., Stevenson, R.J., 2011. The role of live diatoms in advances and challenges. Water Res. 118, 39–58.
bioassessment: a large-scale study of Western US streams. Hydrobiologia 665, 79–92. Patrick, R., Reimer, C.W., 1966. The Diatoms of the United States, vol. I. Monograph of
Gold, C., Feurtet-Mazel, A., Coste, M., Boudou, A., 2002. Field transfer of periphytic the Academy of Natural Sciences of Philadelphia, Pennsylvania.
diatom communities to assess short-term structural effects of metals (Cd, Zn) in Patrick, R., Reimer, C.W., 1975. The Diatoms of the United States, vol. II. Monograph of
rivers. Water Res. 36, 3654–3664. the Academy of Natural Sciences of Philadelphia, Pennsylvania.
Guasch, H., Ricart, M., López-Doval, J., Bonnineau, C., Proia, L., Morin, S., Muñoz, I., Pillai, S., Behra, R., Nestler, H., Suter, M.J.F., Sigg, L., Schirmer, K., 2014. Linking toxicity
Romaní, A.M., Sabater, S., 2016. Influence of grazing on triclosan toxicity to stream and adaptive responses across the transcriptome, proteome, and phenotype of
periphyton. Freshw. Biol. 61 (12), 2002–2012. Chlamydomonas reinhardtii exposed to silver. Proc. Natl. Acad. Sci. 111, 3490–3495.
Hamilton, P.B., Lavoie, I., Alpay, S., Ponader, K., 2015. Using diatom assemblages and Ponader, K.C., Charles, D.F., Belton, T.J., 2007. Diatom-based TP and TN inference
sulfur in sediments to uncover the effects of historical mining on LakeArnoux models and indices for monitoring nutrient enrichment of New Jersey streams. Ecol.
(Quebec, Canada): a retrospective of economic benefits vs. environmental debt. Indic. 7, 79–93.
Front. Ecol. Evol. 3 (99), 1–16. Ricciardi, F., Bonnineau, C., Faggiano, L., Geiszinger, A., Guasch, H., Lopez-Doval, J.,
Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. Past: paleontological statistics software Muñoz, I., Proia, L., Ricart, M., Romaní, A., Sabater, S., 2009. Is chemical con-
package for education and data analysis. Palaeontol. Electron. 4 (1) art. 4: 9pp., tamination linked to the diversity of biological communities in rivers? Trends Anal.
178 kb. Chem. 28, 592–602.
Harding, W.R., Archibald, C.G.M., Taylor, J.C., 2005. The relevance of diatoms for water Round, F.E., Crawford, R.M., Mann, D.G., 1990. The Diatoms: Biology and Morphology of
quality assessment in South Africa. Water SA 31, 41–46. the Genera. Cambridge University Press, Cambridge, U.K.
Hwang, S.-J., Lee, S.-W., Park, Y.-S., 2011. Ecological monitoring, assessment, and re- Schallenber, M., Saulnier-Tablbot, É., 2015. Trajectory of an anthropogenically inducted
storation of running waters in Korea. Ann. Limnol.—Int. J. Limnol. 47, S1–S2. ecological regime shift in a New Zealand shallow coastal lake. Mar. Freshw. Res.
Ivorra, N., Hettelaar, J., Tubbing, G.M.J., Kraak, M.H.S., Sabater, S., Admiraal, W., 1999. http://dx.doi.org/10.1071/MF15211.
Translocation of microbenthic algal assemblages used for in situ analysis of metal Silva, E.F., Almeida, S.F.P., Nunes, M.L., Luís, A.T., Borg, F., Hedlund, M., Sá, C.M.,
pollution in rivers. Arch. Environ. Con. Tox. 37, 19–28. Patinha, C., Teixeira, P., 2009. Heavy metal pollution downstream the abandoned
Kelly, M.G., Whitton, B.A., 1995. The trophic diatom index: a new index for monitoring Coval da Mó mine (Portugal) and associated effects on epilithic diatom communities.
eutrophication in rivers. J. Appl. Phycol. 7, 433–444. Sci. Total Environ. 407 (21), 5620–5636.
Kelly, M., Juggins, S., Guthrie, R., Pritchard, S., Jamieson, J., Rippey, B., Hirst, H., Yallop, Steinman, A.D., McIntire, C.D., 1990. Recovery of lotic periphyton communities after
M., 2008. Assessment of ecological status in U.K. rivers using diatoms. Freshw. Biol. disturbance. Environ. Manage. 14, 589–604.
53, 403–422. Stevenson, R.J., Pan, Y., 1999. Assessing environmental conditions in rivers and streams
Lavoie, I., Grenier, M., Campeau, S., Dillon, P.J., 2006. A diatom-based index for water with diatoms. In: Stoermer, E.F., Smol, J.P. (Eds.), The Diatoms: Applications for the
quality assessment in eastern Canada: an application of Canonical Analysis. Can. J. Environmental and Earth Sciences. Cambridge University Press, Cambridge, pp.
Fish. Aquat. Sci. 63, 1793–1811. 11–40.
Lavoie, I., Darchambeau, F., Cabana, G., Dillon, P., Campeau, S., 2008. Are diatoms good Stevenson, R.J., Pan, Y., Manoylov, K.M., Parker, C.A., Larsen, D.P., Herlihy, A.T., 2008.
integrators of temporal variability in stream water quality? Freshw. Biol. 53, Development of diatom indicators of ecological conditions for streams of the western
827–841. US. J. N. Am. Benthol. Soc. 27, 1000–1016.
Lavoie, I., Dillon, P.J., Campeau, S., 2009. The effect of excluding diatom taxa and re- Stevenson, R.J., Pan, Y., Vandam, H., 2010. Assessing environmental conditions in rivers
ducing taxonomic resolution on multivariate analysis and stream bioassessment. and streams with diatoms. In: Smol, J.P., Stoermer, E.F. (Eds.), The
Ecol. Indic. 9, 213–225. Diatoms:applications for the Environmental and Earth Sciences, 2nd ed. Published by
Lavoie, I., Lavoie, M., Fortin, C., 2012. A mine of information: benthic algal communities Cambridge University Press.
as biomonitors of metal contamination from abandoned tailings. Sci. Total Environ. ter Braak C.J.F., Šmilauer, P., 2012. Canoco Reference Manual and User’s Guide: Software
425, 231–241. for Ordination (version 5.0). Microcomputer Power www.canoco.com.
Lavoie, I., Campeau, S., Zugic-Drakulic, N., Winter, J.G., Fortin, C., 2014. Using diatoms Tlili, A., Corcoll, N., Bonet, B., Morin, S., Montuelle, B., Bérard, A., Guasch, H., 2011. In
to monitor stream biological integrity in Eastern Canada: an overview of 10 years of situ spatio-temporal changes in pollution-induced community tolerance to zinc in
index development and ongoing challenges. Sci. Total Environ. 475, 187–200. autotrophic and heterotrophic biofilm communities. Ecotoxicology 20, 1823–1839.
Lavoie, I., Hamilton, P.B., Morin, S., Kim Tiam, S., Kahlert, K., Gonçalves, S., Falasco, E., Van Dam, H., Mertens, A., Sinkeldam, J., 1994. A coded checklist and ecological indicator
Fortin, C., Gontero, B., Heudre, D., Kojadinovic-Sirinelli, M., Manoylov, D., Pandey, values of freshwater diatoms from the Netherlands. Neth. J. Aquat. Ecol. 28,
L.K., Taylor, J., 2017. Diatom teratologies as biomarkers of contamination: are all- 117–133.
deformities ecologically meaningful? Ecol. Indic. 82, 539–550. Verb, R.G., Vis, M.L., 2000. Comparison of benthic diatom assemblages from streams
Leguay, S., Lavoie, I., Levy, J.L., Fortin, C., 2016. Using biofilms for monitoring metal draining abandoned and reclaimed coal mines and non-impacted sites. J. N. Am.
contamination in lotic ecosystems: the protective effects of hardness and pH on metal Benthol. Soc. 19, 274–288.
bioaccumulation. Environ. Toxicol. Chem. 35, 1489–1501. Villeneuve, A., Bouchez, A., Montuelle, B., 2011. In situ interactions between the effects
Leira, M., Fillipi, M.L., Cantonati, M., 2015. Diatom community response to extreme of season, current velocity and pollution on a river biofilm. Freshw. Biol. 56,
water level fluctuations in two Alpine lakes: a core case study. J. Paleolimnol. 53, 2245–2259.
289–307. Wang, Z.T., Ullrich, N., Joo, S., Waffenschmidt, S., Goodenough, U., 2009. Algal lipid
Liang, Y., Osada, K., Sunaga, Y., Yoshino, T., Bowler, C., Tanaka, T., 2015. Dynamic oil bodies: stress induction, purification and biochemical characterization in wild- type
body generation in the marine oleaginous diatom Fistulifera solaris in response to and starchless Chlamydomonas reinhardtii. Eukaryot. Cell 8, 1856–1868.
nutrient limitation as revealed by morphological and lipidomic analysis. Algal Res. Winter, J.G., Duthie, H.C., 2000. Epilithic diatoms as indicators of stream total N and total
12, 359–367. P concentrations. J. N. Am. Benthol. Soc. 19, 32–49.
Liu, Z.Y., Wang, G.C., Zhou, B.C., 2008. Effect of iron on growth and lipid accumulation in Wolska, L.W., Sagajdakow, A., Kuczyńska, A., Namieśnik, J., 2007. Application of eco-
Chlorella vulgaris. Bioresour. Technol. 99, 4717–4772. toxicological studies in integrated environmental monitoring: possibilities and pro-
Luís, A.T., Teixeira, P., Almeida, S.F.P., Ector, L., Matos, J.X., Ferreira da Silva, E.A., blems. Trends Anal. Chem. 26, 332–344.
2011. Environmental impact of mining activities in the Lousal area (Portugal): che- Wood, R.J., Mitrovic, S.M., Kefford, B.J., 2014. Determining the relative sensitivity of
mical and diatom characterization of metal-contaminated stream sediments and benthic diatoms to atrazine using rapid toxicity testing: a novel method. Sci. Total
surface water of Corona stream. Sci. Total Environ. 409, 4312–4325. Environ. 485–486, 421–427.
Morin, S., Duong, T.T., Dabrin, A., Coynel, A., Herlory, O., Baudrimont, M., Delmas, F., Wood, R.J., Mitrovic, S.M., Lim, R.P., Kefford, B.J., 2016. How benthic diatoms within
Durrieu, G., Schäfer, J., Winterton, P., Blanc, G., Coste, M., 2008. Long-term survey of natural communities respond to eight common herbicides with different modes of
heavy-metal pollution, biofilm contamination and diatom community structure in the action. Sci. Total Environ. 557–558, 636–643.
Riou Mort watershed. South-West France. Environ. Pollut. 151, 532–542. Yang, Z.K., Niu, Y.F., Ma, Y.H., Xue, J., Zhang, M.H., Yang, W.D., Liu, J.S., Lu, S.H., Guan,
Morin, S., Pesce, S., Tili, A., Coste, M., Montuelle, B., 2010. Recovery potential of per- Y., Li, H.Y., 2013. Molecular and cellular mechanisms of neutral lipid accumulation
iphytic communities in a river impacted by a vineyard watershed. Ecol. Indic. 10, in diatom following nitrogen deprivation. Biotechnol. Biofuels 6 (1), 1–14 (67).
419–426. Zalack, J.T., Smucker, N.J., Vis, M.L., 2010. Development of a diatom index of biotic
Morin, S., Cordonier, A., Lavoie, I., Arini, A., Blanco, S., Duong, T.T., Tornés, E., Bonet, B., integrity for acid mine drainage impacted streams. Ecol. Indic. 10, 287–295.
Corcoll, N., Faggiano, L., Laviale, M., Pérès, F., Becares, E., Coste, M., Feurtet-Mazel, Żelazna-Wieczorek, J., Nowicka-Krawczyk, P., 2015. The cascade construction of artifi-
A., Fortin, C., Guasch, H., Sabater, S., 2012.. Consistency in diatom response to metal- cial ponds as a tool for urban stream restoration—the use of benthic diatoms to assess
contaminated environments. In: Guasch, H., Ginebreda, A., Geiszinger, A. (Eds.), the effects of restoration practices. Sci. Total Environ. 538, 591–599.
Emerging and Priority Pollutants in Rivers. Springer- Verlag, Berlin, pp. 117–146.

151