Watershed-scale landuse is associated with temporal and spatial compositional variation in Lake Michigan tributary bacterial communities

Abstract

Populations of stream organisms across trophic levels, including microbial taxa, are adapted to physical and biotic stream features, and are sentinels of geological and hydrological landscape processes and anthropogenic disturbance. Stream bacterial diversity and composition can have profound effects on resident and migratory species in Great Lakes tributaries. Study objectives were to characterize and compare the taxonomic composition and diversity of bacterial communities in 18 rivers of the Lake Michigan basin during April and June 2019 and to quantify associations with stream and watershed physical features and dominant landuse practices. River water was filtered, and genomic DNA was extracted from filtrate using antiseptic techniques. We performed high-throughput amplicon sequencing using the highly variable V4 region of the 16S rRNA gene to characterize microbial community composition and diversity. Effects of landscape-scale landuse, environmental variables and dispersal predictors (e.g., inter-stream distance) on community compositional differences were quantified. Greater than 90% of variation in bacterial relative abundance between rivers and time were attributed to 11 phyla representing 10,800 operational taxonomic units. Inter-stream geographic distance, stream hydrology, and variation in stream properties that were tied to patterns of watershed landuse were significantly associated with differences in bacterial community composition among streams at both sampling time periods. based on Bray-Curtis distances. Understanding how environmental characteristics and watershed-scale landuse influence lower trophic level stream communities such as bacteria will inform managers as biological indicators of ecosystem health, sources of disturbance, and current and future bottom-up trophic changes in coupled tributary-Great Lakes ecosystems.

Full text

Watershed-scale landuse is associated with temporal and spatial
compositional variation in Lake Michigan tributary bacterial
communities
Gabrielle E. Sanfilippo
a
, Jared J. Homola
a
, Jared Ross
a
, Jeannette Kanefsky
a
, Jacob Kimmel
a
,
Terence L. Marsh
b
, Kim T. Scribner
a,c,
⇑
a
Department of Fisheries and Wildlife, 480 Wilson Road, 13 Natural Resources Building, Michigan State University, East Lansing, MI 48864, United States
b
Department of Microbiology and Molecular Genetics, 567 Wilson Road, 2215 Biomedical Physical Sciences Building, Michigan State University, East Lansing, MI 48864, United States
c
Department of Integrative Biology, 288 Farm Ln, Natural Sciences Building, Michigan State University, East Lansing, MI 48864, United States
article info
Article history:
Received 1 September 2020
Accepted 10 February 2021
Available online xxxx
Communicated by Wendylee Stott
Keywords:
Bacterial diversity
Great Lakes tributaries
Landuse
Landscape ecology
Watershed ecology
Riverscapes
abstract
Populations of stream organisms across trophic levels, including microbial taxa, are adapted to physical
and biotic stream features, and are sentinels of geological and hydrological landscape processes and
anthropogenic disturbance. Stream bacterial diversity and composition can have profound effects on res-
ident and migratory species in Great Lakes tributaries. Study objectives were to characterize and compare
the taxonomic composition and diversity of bacterial communities in 18 rivers of the Lake Michigan basin
during April and June 2019 and to quantify associations with stream and watershed physical features and
dominant landuse practices. River water was filtered, and genomic DNA was extracted from filtrate using
antiseptic techniques. We performed high-throughput amplicon sequencing using the highly variable V4
region of the 16S rRNA gene to characterize microbial community composition and diversity. Effects of
landscape-scale landuse, environmental variables and dispersal predictors (e.g., inter-stream distance)
on community compositional differences were quantified. Greater than 90% of variation in bacterial rela-
tive abundance between rivers and time were attributed to 11 phyla representing 10,800 operational tax-
onomic units. Inter-stream geographic distance, stream hydrology, and variation in stream properties that
were tied to patterns of watershed landuse were significantly associated with differences in bacterial com-
munity composition among streams at both sampling time periods. based on Bray-Curtis distances.
Understanding how environmental characteristics and watershed-scale landuse influence lower trophic
level stream communities such as bacteria will inform managers as biological indicators of ecosystem
health, sources of disturbance, and current and future bottom-up trophic changes in coupled tributary-
Great Lakes ecosystems.
Ó2021 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.
Introduction
River ecosystems are spatially complex and temporally
dynamic (Vannote et al., 1980). The dendritic structure that char-
acterizes stream habitats flows through riparian and terrestrial
landscapes composed of different landcover types that often have
been altered to varying degrees by human occupation (Tonkin
et al., 2018). Metacommunity composition and diversity can reflect
landscape-level inputs from the watershed’s terrestrial habitats
(mass effects), including anthropogenic stressors (Esselman et al.,
2013;Riseng et al., 2010). Within the Great Lakes region, ecological
conditions within tributaries spanning multiple trophic levels,
including diverse microbial metacommunities are likely to vary
temporally and spatially among watersheds.
Widely observed properties of aquatic microbial communities
(e.g., Langenheder et al., 2012; Tonkin et al., 2018; Viana et al.,
2015) include high taxonomic diversity, co-occurrence of closely
related species and functional groups, functional stability despite
large species turnover spatially and temporally, and spatial and
temporal differences in the relative influences of community
assembly mechanisms owing to neutral and deterministic pro-
cesses (Vellend and Agrawal, 2010). Neutral processes include
advection from adjoining terrestrial landscape and stream hydrol-
ogy aiding in dispersal (Crump et al., 2012; Isabwe et al., 2018;
Lindstrom et al., 2005). Deterministic processes include natural
https://doi.org/10.1016/j.jglr.2021.02.009
0380-1330/Ó2021 International Association for Great Lakes Research. Published by Elsevier B.V. All rights reserved.
⇑
Corresponding author at: Department of Fisheries and Wildlife, 480 Wilson
Road, 13 Natural Resources Building, Michigan State University, East Lansing, MI
48864, United States.
E-mail address: scribne3@msu.edu (K.T. Scribner).
Journal of Great Lakes Research xxx (xxxx) xxx
Contents lists available at ScienceDirect
Journal of Great Lakes Research
journal homepage: www.elsevier.com/locate/ijglr
Please cite this article as: G.E. Sanfilippo, J.J. Homola, J. Ross et al., Watershed-scale landuse is associated with temporal and spatial compositional variation
in Lake Michigan tributary bacterial communities, Journal of Great Lakes Research, https://doi.org/10.1016/j.jglr.2021.02.009

variation in soil and subsurface geochemistry (Kärnä et al., 2019)
and influences of human disturbance. Convergence in microbial
taxonomic composition and function due to deterministic pro-
cesses is likely to occur over larger spatial scales owing to com-
monalities in landuse and degree of anthropogenic disturbance in
different streams. For example, phylogenetic variation in bacterial
communities across ecosystems (Wang et al., 2013) has been asso-
ciated with hydrological fragmentation (Fazi et al., 2013; Ruiz-
González et al., 2015), agricultural practices (Derose et al., 2020),
and effluent from urban sources (Verhougstraete et al., 2015).
Aquatic microbial communities play critical roles in a wide
range of natural processes in aquatic ecosystems (Zeglin, 2015),
from animal development and host health (e.g., fish mortality dur-
ing the egg stage; Forsythe et al., 2013; Fujimoto et al., 2020, and
colonization of internal and external surfaces; Abdul Razak and
Scribner, 2020; Llewellyn et al., 2014) to biogeochemical processes
(Cotner and Biddanda, 2002) such as organic matter decomposi-
tion, and stream respiration (Gessner et al., 2010; Hieber and
Gessner, 2002). Stream food web structure is also affected by
microbial community composition and diversity (Hieber and
Gessner, 2002; Tank et al., 2010), including phytoplankton and
zooplankton community composition and diversity (Parveen
et al., 2011). Bacterial taxonomic composition and diversity in
waterbodies have been used as indicators of environmental health
and water quality (Colford et al., 2007; Fan et al., 2015). Microbial
taxa associated with wastewater, septic effluent and other indica-
tors of urban landuse are widely used as biological indicators of
water quality that are of importance to human health (e.g.,
Verhougstraete et al., 2015). Microbial community composition
and diversity have been shown to be indicators of ecosystem stress
associated with anthropogenic disturbance as documented by
effects of urban runoff and nutrient levels. For example, anthro-
pogenic point-source inputs including pollutants from urban run-
off and sewage can also greatly affect microbial community
composition (Mansfeldt et al., 2020; Xu et al., 2018).
Knowledge of associations between stream physical and biotic
components and inter-relationships with landscape level processes
is critical to understand patterns of species composition, diversity
and relative abundance, and is key for effective management in
times of increasing anthropogenic change (Allan, 2004; Fausch
et al., 2002; Soranno et al., 2010). Processes that underlie microbial
community assembly and affect the taxonomic composition of
microbial communities on a coarse (i.e., watershed) spatial scale
are governed by many factors (Goldford et al., 2018). Due to their
small size, microbes are subject to advection and dispersal
(Beisner et al., 2006) from terrestrial sources from landscapes dom-
inated by different landuse. Mass effects associated with transport
of materials from terrestrial landscapes into streams contribute to
levels of salinity, pH, and nutrients that collectively affect micro-
bial community structure (Huggett et al., 2017). Stream hydrologic
forces can influence dispersal and taxonomic homogenization by
overwhelming local selective processes that could otherwise lead
to sorting associated with stream conditions (Nino-Garcia et al.,
2016) such as local stream microchemistry (Kaestli et al., 2017).
Because water and bacteria move directionally, riverine ecosys-
tems act as both bacterial sources and recipients. Therefore, the
pool of taxa present at any single point in a river will depend on
the characteristics of the surrounding landscape and the specific
location within the river continuum (Nelson, 2017). Samples col-
lected in downstream regions of watersheds can therefore expect
to capture constituent taxa transported from large areas of water-
sheds that are likely influenced by upstream environmental char-
acteristics throughout the watershed.
Aquatic microbial communities are under-studied components
of coupled Great Lakes-tributary ecosystems. Recently reported
spatial and temporal variation in microbial community composi-
tion from open waters of the Great Lakes were associated with
hydrological connectivity and environmental conditions (largely
temperature; Paver et al., 2020). However, knowledge of temporal
variation in microbial communities is also important (Hermans
et al., 2019). Microbial communities elsewhere have been shown
to be important to many levels of stream community organization
and function, including primary productivity (Linz et al., 2020),
organic matter decomposition (Mlambo et al., 2019), and denitrifi-
cation (Mulholland et al., 2008).
Recent advances in culture-independent methods of microbial
taxonomic identification, including high throughput amplicon
DNA sequencing, have facilitated characterizations of microbial
communities in field situations that were previously impossible.
These widely used methods have facilitated characterizations of
the diversity and taxonomic composition of microbial communi-
ties and causal relationships with physical and biotic processes
(Linz et al., 2020) and are the basis for data produced in this study.
The goal of this study was to characterize and compare bacterial
taxonomic composition and diversity within and among a geo-
graphically dispersed and compositionally heterogeneous set of
18 tributaries of the Lake Michigan basin during each of two time
periods. Additionally, we quantified associations between tempo-
ral and spatial community compositional differences with
watershed-level and local landscape features and measures of dis-
turbance. We hypothesized that we would document community
compositional heterogeneity across time and space at a scale
reflecting geographic variation in landuse.
Methods
Field collections
Water samples were collected from 18 streams within the Lake
Michigan basin during two time periods (April 20–21, 2019 and
June 15–16, 2019; Fig. 1). The timing of the April and June sam-
pling periods corresponds to the beginning and end of spawning
for many river-resident and migratory adfluvial fishes. Microbial
community interactions with fish hosts have important impacts
because of the high bacterial abundance in aquatic habitats (De
Schryver and Vadstein, 2014; Verschuere et al., 2000). For example,
gut and skin microbiota are associated with multiple physiological
and immunological functions (review in Llewellyn et al. (2014)).
Microbial colonization and community successional changes on
fish egg surfaces can also be a significant source of mortality
(Forsythe et al., 2013; Fujimoto et al., 2020). April and June also
represent periods of substantial differences in temperature and
discharge (Electronic Electronic Supplementary Material (ESM)
Table S1), and concomitantly sources and levels of material trans-
port from adjoining terrestrial habitats into streams.
Samples were collected within 0.32–141.7 km of the river
mouth (mean ± SE = 21.3 ± 8.1 km), down river from the first
upstream barrier, which was typically a dam nearest to the river
mouth. One-liter wide-mouth plastic sampling bottles were used
to collect a one liter sample of surface water from the shore of each
sampling location. Water samples were immediately filtered in the
field through 0.2-mm Corning disposable vacuum filter systems
using a vacuum filtration pump. Filter membranes were removed
with a sterile scalpel and forceps and stored at ambient tempera-
ture in 95% ethanol in 50 mL conical tubes.
Laboratory methodology
DNA extractions on materials adhered to filter membranes were
performed on a dedicated sterile (using 10% bleach) workbench
using sterile pipets, filter tips and microcentrifuge tubes. Before
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
2

extraction, ethanol in which the filters were stored was diluted to
80% using distilled water. To prepare for extraction of bacterial
cells, filter samples were vortexed vigorously for five minutes to
agitate cells bound to the filter. Filters were removed using sterile
forceps and stored at 20 °C in a new sterile conical tube. The con-
ical tubes containing the remaining ethanol were centrifuged at
4°C for 40 min at 11,500 rpm to pellet cells. Ethanol was then aspi-
rated from the sample tubes without disturbing the pellets. The
remaining ethanol that could not be removed via aspiration was
used to resuspend the pellets, and the pellet suspensions were
transferred to new sterile 1.5 mL microcentrifuge tubes. The
1.5 mL tubes were centrifuged for 10 min at 14,000 rpm at 4 °C
to concentrate the cells again, and the excess ethanol was again
aspirated, and pellets were dried at room temperature for up to
15 min. Pellets containing fine particulate sediment that adhered
to the filters, biological cells, and DNA were small in size
(<250 mg). Pellets were stored at 20 °C until DNA extraction
was performed. DNA extraction from the pellets was carried out
using the DNeasy PowerSoil kit (QIAGEN, Inc., Germantown, MD)
following the manufacturer’s protocols.
PCR amplification and Illumina sequencing were conducted at
the Michigan State University (MSU) Research Technology Support
Facility (RTSF). The V4 region of the prokaryotic 16 s rRNA gene
was amplified from the samples using dual indexed, Illumina com-
Fig. 1. Sampled streams and network catchment area in the Lake Michigan basin and visualization of spatial variation in land use across study sites. Numerical designations
correspond to stream numbers that are used in the tables, and do not represent sampling locations.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
3

patible primers 515F (5
0
– GTGCCAGCMGCCGCGGTAA – 3
0
) and
806R (5
0
– GGACTACHVGGGTWTCTAAT – 3
0
)(Caporaso et al.,
2011) following the protocol outlined by Kozich et al. (2013). After
PCR, the products were batch normalized using an Invitrogen
SequalPrep DNA Normalization plate (Invitrogen, Carlsbad, CA)
and the recovered products were pooled. AmpureXP magnetic
beads (Beckman Coulter, Indianapolis, IN) were used to clean and
concentrate the pooled amplicons. DNA concentration was quanti-
fied for each sample using a combination of the Qubit dsDNA High
Sensitivity assay (Invitrogen, Carlsbad, CA), the Agilent Advanced
Analytical Fragment Analyzer High Sensitivity DNA assay (Agilent,
Santa Clara, CA) and the Kapa Illumina Library Quantification qPCR
assay (Roche, Pleasanton, CA). The pooled amplicons were loaded
onto a single Illumina MiSeq v2 Nano flow cell, and sequencing
was conducted in a 2 250 bp paired end format (Illumina, San
Diego, CA). MSU Genomics RTSF provided standard Illumina qual-
ity control. Illumina RTA (v1.17.21.3) and Illumina Bcl2Fastq
(v1.8.4) were used to perform base calling and initial processing
(demultiplexing, barcode removal and RTA conversion to FastQ
format).
Sequence processing
Sequencing data was processed using program Mothur version
1.41.3 (Schloss et al., 2009), using a SILVA-based bacterial refer-
ence alignment (version 132), and a Mothur-formatted version of
the 16s rRNA RDP (version 16) reference file. The processing proto-
col followed the Mothur MiSeq SOP (Kozich et al., 2013). Samples
were initially rarefied to 1636 sequences per sample, which was
the lowest number of reads among retained samples. Two samples
were excluded from further analysis due to low read counts
(Boardman River sampled in April and Muskegon River sampled
in June).
PAST (PAleontological STatisics) software version 4.0 (Hammer
et al., 2001) was used to characterize measures of bacterial com-
munity diversity within streams. PAST was used to calculate Simp-
son’s diversity index (Simpson, 1949), the Shannon-Weaver
diversity index (Peet, 1975), richness, and taxonomic evenness
(Peet, 1975).
Summaries and analyses of spatial and temporal variation in
diversity and community variation among streams and between
time periods were based on taxonomic classifications at the OTU
level (mostly genera). However, visualizations of community com-
position and summaries of alpha diversity measures were also con-
ducted based on phyla-level classifications. The R (version 3.6.2)
package ‘‘vegan” was used to create a Bray-Curtis dissimilarity
matrix (based on OTU-based classification) for statistical analyses
using the ‘vegdist’ function of the package ‘vegan’ (Oksanen
et al., 2019). To create community visualizations, OTUs were clas-
sified to phylum level and read counts for each phylum for each
sample were summed and associated with the appropriate sample
names in a single file using a custom R (version 3.6.2) script. Sin-
gleton OTU reads were eliminated to truncate data and simplify
further analyses.
Visualization of inter-sample relationships describing differ-
ences in microbial community composition among sampling loca-
tions and between collection times, summarized by Bray-Curtis
distance, was carried out using Unweighted Pair Group with Arith-
metic Averaging (UPGMA). Hierarchical clustering was performed
using the ‘hclust’ function of the ‘stats’ R package (version 3.6.2).
Clustering was performed using the ‘‘average” argument, to calcu-
late the mean of all pairwise distances between samples. The final
visual UPGMA tree was then created using R package ‘‘dendex-
tend” to convert the clustering to a dendrogram (Galili, 2015).
Multi-scale bootstrapping was conducted in PAST to quantify sup-
port for notes in the UPGMA tree. Bootstrap values greater than
70% are shown based on 1000 bootstrap replicates. The R graphing
package ‘‘ggplot2” was used to create visualizations of raw phyla
classification data (Wickham, 2016).
Characterization of landscape data
River mouth locations were linked to stream reaches from the
National Hydrography Dataset Plus Version 2 (NHDPlus V2;
McKay et al., 2012) to help characterize landscape information
occurring within the watershed of each sampling site. All stream
sampling sites and watershed boundaries are visualized in Fig. 1,
including patterns of local land use (forest, urban, agriculture).
River mouths are marked as the sampling points in Fig. 1, despite
sampling having occurred at varying distance from the river
mouths. Local catchment area was also estimated for all streams.
Stream reaches from the NHDPlus V2 are typically defined as
inter-confluence stretches of stream, each having an associated
local catchment that outlines the land area draining to any partic-
ular stream reach. Local catchments could then be characterized by
landscape condition such as percentage of various land cover
classes or density of dams occurring within that catchment. The
hierarchical nature of this framework allowed for a network catch-
ment to be defined for any particular reach by aggregating all
upstream local catchments (Tsang et al., 2014; Wang et al., 2011).
We characterized river watersheds using a range of biotic and
abiotic variables. The percent of estimated annual streamflow vol-
ume stored in upstream reservoirs (Cooper et al., 2017) was used
to characterize potential influences of large dams on microbial
community composition (Wu et al., 2019) within each river net-
work. Percentages of landuse/cover classes from the 2011 National
Land Cover Database (MRLC, 2018) were summarized in network
catchments, and we retained forest (i.e., the sum of deciduous,
evergreen, and mixed), urban (i.e., the sum of open space, low
intensity, medium intensity, and high intensity urban), and agri-
culture (i.e., the sum of cultivated crops and pasture/hay). Some
NLCD categories, such as woody wetlands, emergent herbaceous
wetlands, and shrub/scrub were excluded from analyses due to
perceived insignificant contributions to microbe community
composition.
Statistical analyses
River mouth locations, rather than sampling sites, were used to
calculate straight-line distances between rivers to simplify geo-
graphic location within the Lake Michigan basin. Straight-line geo-
graphic distance between each river was calculated using the
‘distGeo’ function of the R package ‘‘geosphere” (Hijmans et al.,
2019). A Mantel test for the significant correlation between geo-
graphic distance of river mouths and Bray-Curtis distances was
conducted using the program PASSaGE (Rosenberg and Anderson,
2011). Inter-sample relationships were also characterized based
on geographic location and time using non-metric multidimen-
sional scaling (NMDS; Cox and Cox, 2000) using the ‘metaMDS’
function of R package ‘‘vegan”. We analyzed differences in stream
microbial community diversity between April and June sampling
times using a two-tailed t-tests assuming unequal variances,
implemented in Microsoft Excel.
A heatmap of OTU relative abundances was constructed to visu-
alize community composition patterns among taxa and samples.
Included taxa were represented by at least 0.2% of the total number
of reads. Relative abundance data was log10 transformed following
addition of a pseudocount of 0.001 that allowed values of 0 to be
transformed. The heatmap was constructed using the R package
‘pheatmap’ (Kolde, 2019). Multiscale bootstrapping was performed
to calculate p-values for hierarchical clustering of the taxa and
samples included in the heatmap. This process used 1000 boot-
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
4

strap resamples of a Euclidean distance measure and was executed
using the R package ‘pvclust’ (Suzuki et al., 2019).
We used distance-based redundancy analysis (db-RDA;
Legendre and Andersson, 1999) to evaluate the influence of envi-
ronmental variables on beta diversity, performed on microbial taxa
identified to the OTU level. Normality of explanatory landscape
variables used in analyses of associations with stream microbial
community compositional differentiation was evaluated visually
by examining distributions of each variable. Because data were
often non-normally distributed, we performed transformations
using the natural log (X + 0.01; N_areasqkm, UMD, latitude, and
longitude) or arcsine square root (UDOR, N_urban, N_forest,
N_agriculture) of each variable. Collinearity among explanatory
variables was evaluated by calculating Pearson product-moment
correlation coefficients. One variable was removed from analyses
anytime correlation coefficients between explanatory variables
exceeded 0.8. We also calculated variance inflation factors for all
variables, removing variables with values greater than 10. After fil-
tering for multicollinearity, we retained watershed area (N_ar-
easqkm), upstream mainstem dam density per unit upstream
mainstem length (UMD), proportion of estimated annual discharge
stored in upstream reservoirs (UDOR), urban (N_urban) and agri-
cultural (N_agriculture) land cover in network catchment, latitude,
and longitude. Db-RDA was performed with the ‘dbRDA’ function
of ‘‘vegan” using Bray-Curtis distances as response variables and
the retained transformed environmental measures as explanatory
variables. Separate analyses were conducted for communities sam-
pled in April and June.
Results
Watershed physical characteristics
Network (watershed) catchment area varied considerably
across rivers. The Fox River (16,642 km
2
Þand Grand River
(14,450 km
2
) encompassed large areas compared to smaller rivers,
such as the Kewaunee River (369 km
2
Þand Millecoquins River
(263 km
2
;Table 1;Fig. 1). Watershed landuse also covaried with
latitude (e.g., urban and agriculture more prominent in more
southern tributaries) and in the eastern (e.g., generally more heav-
ily forested) and western portions of the Lake Michigan basin
(Fig. 1;Table 1).
Untransformed landscape data (Table 1) showed large variation
in spatial extent and dispersion of landuse within watershed
basins. Tributaries occurring at similar latitudes and on either side
(east or west) of Lake Michigan had similar landuse patterns
(Table 1). For example, the Millecoquins, Manistique, Escanaba,
Cedar, and Whitefish rivers occur along a similar latitude in Michi-
gan’s Upper Peninsula and were characterized by similar
watershed-scale landuse percentages (low urban land cover = 2.
2–3.8%; high forest land cover = 36.2–44.4%; low to moderate agri-
cultural land cover = 1.0–12.2%; Table 1). The Grand, Macatawa
and Kalamazoo rivers in the south eastern portion of the study area
(Fig. 1) pass through major urban centers within watersheds
adjoining large agricultural regions (Table 1). Rivers occurring on
the west side of Lake Michigan had similar urban land cover
(range = 2.24–8.38%), which was generally lower than rivers occur-
ring on the east side of the lake (range = 5.94–34.42%; Table 1).
Some rivers were characterized by both zero and non-zero values
for dam-related metrics (UMD and UDOR) in Table 1. A zero value
in upstream mainstem dam density (UMD) and a non-zero value in
estimated annual discharge stored in upstream reservoirs (UDOR)
was possible because rivers may lack upstream mainstem dams
yet still have dams on non-navigable tributaries that are storing
water and contributing to the UDOR values. Mainstem reaches
were defined as the longest navigable path upstream of a given
river.
Differences in measured environmental variables including dis-
charge and stream temperature were considerable between collec-
tion times and among streams (ESM Table S1). Within streams. The
mean difference (±stdev) in temperature on the day of sampling
was 8.42 ± 2.54 °C warmer in June than in April. The lowest and
highest temporal difference in stream temperature was recorded
in the Pere Marquette (3.3 °C) and Escanaba rivers (13.2 °C),
respectively (ESM Table S1). Within streams, the mean difference
(±stdev) in discharge based on U.S. Geological Survey gauging sta-
tions was 75.3 ± 101.2 cu m/s lower in June than in April (ESM
Table S1). The lowest and highest temporal difference in stream
discharge were recorded in the Boardman (0.028 cu m/s) and
Menominee rivers (346.7 cu m/s), respectively.
Characteristics of microbial communities within and among
tributaries and collection times
A total of 51,572 16S sequence reads were classified from the
truncated data set, with 25,823 reads contributed by April samples
and 25,749 reads from June samples. Following removal of single-
ton OTUs, a total of 26 unique phyla were present across all tribu-
taries and time periods (ESM Table S2), including a composite
phyla category containing unclassified bacteria. 10,800 operational
taxonomic units (OTUs) remained after subsampling.
Bacterial community alpha diversity measures estimated using
all non-singleton OTUs including taxonomic richness (OTU num-
ber), evenness, Shannon diversity, Simpson’s diversity indices,
and Chao’s I were calculated for each tributary and time period
(Table 2). Considerable variation was observed in the number of
taxa, evenness, and Chao’s I between April and June sampling peri-
ods (Table 2), however there was no evidence of a consistent tem-
poral trend. Little variation was observed between sampling
periods for Simpson’s and Shannon’s diversity indices (Table 2).
Two-tailed t-tests indicated no significant difference between sam-
pling times for any alpha diversity measure (P > 0.05).
At the phyla level of classification (ESM Table S2), measures of
bacterial taxonomic diversity estimated from water collected in
April and June were comparable. Little variation in Simpson and
Shannon diversity was observed. Community evenness was
slightly higher in June relative to April. Taxonomic richness was
slightly higher in April than June (ESM Table S2). Two-tailed t-
tests indicated no significant difference between sampling times
for any alpha diversity measure (P > 0.05).
Lack of sampling replication at each site precluded statistical
analyses of differences in community diversity among streams.
However, there were qualitative patterns exhibited by alpha diver-
sity measures of certain streams. For example, the Peshtigo and
Oconto River samples from the Green Bay area of Wisconsin and
the Michigan Grand River during both times had consistently high
measures of alpha diversity relative to other rivers (Table 2). A sim-
ilar trend was observed to phyla-level diversity (ESM Table S2).
Streams with comparatively low levels of microbial community
diversity included the Kewaunee, and Kalamazoo (Table 2), which
are located within highly developed and industrialized watersheds.
Phyla-level diversity was also low in the Kewaunee and Macatawa
Rivers (ESM Table S2).
Several diversity indices were correlated with untransformed
landscape variables measured at the entire watershed spatial scale.
Positive correlation was also observed between Shannon and
Simpson diversity indices during both sampling times and percent-
age of forested land cover in the watershed (0.60 r0.62 for
these parameters). Pearson product moment correlations between
watershed catchment area (Table 1) and measures of Simpson and
Shannon diversity (Table 2) were not high or significant in either
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
5

sampling period (R = 0.154 and 0.152 in April and 0.214 and 0.198
in June, respectively). The proportion of watershed area in agricul-
ture was higher in the south eastern basin streams (Grand, Maca-
tawa, and Kalamazoo rivers) and in the Kewaunee River in
Wisconsin (Table 1).
Each tributary was characterized by a different relative abun-
dance of bacteria at the phyla level between April and June sam-
pling times (Fig. 2). For example, the samples collected from the
Manistee River were characterized by large differences in the rela-
tive frequency of phylum Verrucomicrobia (7.4%–24.5%; Fig. 2;
Table 1
Untransformed environmental explanatory variables describing watershed-level features for 18 Lake Michigan tributaries predicted to be associated with stream bacterial
community composition and diversity within and among tributaries during two periods (April and June) in 2019. Variables include network catchment area (km2), upstream
mainstem dam density per unit upstream mainstem length (UMD, #/100 km) and proportion of estimated annual discharge stored in upstream reservoirs (UDOR, %; Cooper et al.,
2017), total watershed area in urban, forest and agricultural land cover (%).
Lake Michigan
tributary
Numeric stream
designation
a
Region of Lake
Michigan
Latitude
(°)
Network catchment
area (km
2
)
UMD (#/
100 km)
UDOR
(%)
Urban land
cover (%)
Forest land
cover (%)
Agricultural land
cover (%)
Millecoquins 1 North (MI Upper
Peninsula)
46.096 263.54 0 0.14 3.85 36.23 12.24
Manistique 2 North (MI Upper
Peninsula)
45.949 3809.95 0.60 4.88 2.93 37.37 0.96
Whitefish 3 North (MI Upper
Peninsula)
45.906 800.13 0 0.02 2.24 44.43 3.53
Escanaba 4 North (MI Upper
Peninsula)
45.777 2504.33 4.15 10.78 3.99 45.09 2.51
Cedar 5 North (MI Upper
Peninsula)
45.41 720.75 0 0.02 4.33 28.30 9.60
Menominee 6 West (WI) 45.095 10538.19 4.05 8.65 3.90 53.47 4.33
Peshtigo 7 West (WI) 44.975 2920.18 2.46 6.83 4.44 43.07 15.38
Oconto 8 West (WI) 44.894 2530.04 2.06 4.63 5.10 40.98 21.31
Fox 9 West (WI) 44.538 16642.56 3.10 25.69 8.38 25.11 40.94
Kewaunee 10 West (WI) 44.459 369.96 0 0 6.65 8.18 77.53
Boardman 11 East (MI) 44.764 735.14 1.31 4.25 12.55 44.06 9.81
Manistee 12 East (MI) 44.249 5050.09 1.03 3.91 5.99 55.38 9.02
Pere
Marquette
13 East (MI) 43.952 1957.59 0 0.18 5.94 56.98 12.28
White 14 East (MI) 43.475 1393.75 1.57 0.35 7.29 49.55 18.23
Muskegon 15 East (MI) 43.228 7061.86 1.06 2.53 8.52 39.47 19.55
Grand 16 East (MI) 43.057 14450.87 2.98 1.53 14.82 16.64 53.44
Macatawa 17 East (MI) 42.773 442.37 0 0 34.42 7.38 50.64
Kalamazoo 18 East (MI) 42.676 5272.18 3.65 2.13 13.68 21.43 48.03
a
Numerical values correspond to numerical stream designations presented in Fig. 1.
Table 2
Measures of bacterial community diversity for 18 Lake Michigan tributary samples collected during April and June 2019.
Lake Michigan
tributary
Numeric
stream
a
Region of Lake Michigan Latitude
(°)
No. Taxa
b
Simpson Shannon Evenness Chao 1
April June April June April June April June April June
Millecoquins 1 North (MI Upper
Peninsula)
46.096 338 273 0.98 0.99 4.98 4.96 0.43 0.52 593.1 346.9
Manistique 2 North (MI Upper
Peninsula)
45.949 248 260 0.96 0.99 4.55 4.88 0.38 0.51 296.4 355.2
Whitefish 3 North (MI Upper
Peninsula)
45.906 218 301 0.99 0.98 4.99 4.77 0.67 0.39 235.2 497.1
Escanaba 4 North (MI Upper
Peninsula)
45.777 282 238 0.99 0.98 5.19 4.60 0.64 0.42 379.1 301
Cedar 5 North (MI Upper
Peninsula)
45.41 344 333 0.99 0.99 5.24 5.07 0.55 0.48 447.1 494.3
Menominee 6 West (WI) 45.095 366 378 0.99 0.99 5.36 5.33 0.58 0.55 553.7 613.7
Peshtigo 7 West (WI) 44.975 371 397 0.99 0.99 5.24 5.51 0.60 0.62 488 540.4
Oconto 8 West (WI) 44.894 386 351 0.99 0.99 5.49 5.22 0.62 0.53 559.5 522
Fox 9 West (WI) 44.538 232 323 0.92 0.99 3.99 5.22 0.23 0.52 339.3 468.3
Kewaunee 10 West (WI) 44.459 207 212 0.95 0.96 3.86 4.13 0.23 0.29 294.1 306.5
Boardman 11 East (MI) 44.764 – 203 – 0.99 – 4.81 – 0.60 – 225.5
Manistee 12 East (MI) 44.249 299 298 0.98 0.96 4.92 4.62 0.46 0.34 451.9 413.7
Pere Marquette 13 East (MI) 43.952 233 375 0.98 0.99 4.81 5.29 0.53 0.53 272.1 635.9
White 14 East (MI) 43.475 249 315 0.98 0.98 4.71 4.98 0.45 0.46 301.7 505.8
Muskegon 15 East (MI) 43.228 331 – 0.98 – 5.17 – 0.53 – 420.8 –
Grand 16 East (MI) 43.057 332 364 0.98 0.98 4.75 5.02 0.35 0.41 485 599.1
Macatawa 17 East (MI) 42.773 298 311 0.98 0.99 4.77 5.06 0.40 0.51 443.2 487.2
Kalamazoo 18 East (MI) 42.676 278 201 0.99 0.98 5.02 4.72 0.54 0.56 358.9 227.3
Average 294.82 301.94 0.98 0.98 4.89 4.94 0.48 0.48 407.01 443.52
SE of
mean
57.18 62.67 0.02 0.01 0.45 0.33 0.13 0.09 108.58 129.66
a
Numerical vales correspond to numeric stream designations presented in Fig. 1.
b
Number of OTUs classified to the lowest taxonomic resolution based on the SILVA-based reference alignment after singleton OUT observations were removed.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
6

ESM Table S3). The relative abundance of Verrucomicrobia also
varied between sampling periods in the Peshtigo River (1.7%–
5.4%), Oconto River (4.1%–6.7%), Menominee River (1.6%–8.3%),
and Kalamazoo River (4.1%–7.2%; Fig. 2). The relative abundance
of taxa on the phylum Acidobacteria decreased between sampling
period in all but three tributaries (Fig. 2).
SIMPER tests for dissimilarity evaluated beta diversity among
rivers and between April and June sampling times. The overall
average dissimilarity among samples collected in April (30.18%)
was slightly higher than dissimilarity among samples collected in
June (26.24%). 86.5% of dissimilarity among tributaries sampled
in April was contributed by five bacteria phyla: Firmicutes
(29.8%), unclassified bacteria (18.9%), Proteobacteria (17.1%), Acti-
nobacteria (12.2%) and Bacteroidetes (8.5%). 88.9% of dissimilarity
among samples collected in June was contributed by five top phyla
including Firmicutes (26.9%), unclassified bacteria (21.6%), Pro-
teobacteria (16.8%), Bacteroidetes (14.7%) and Verrucomicrobia
(8.9%). Dissimilarity in Bray-Curtis distances between stream bac-
Fig. 2. A visual representation of variation in bacterial community diversity (taxa classified to phyla) from samples of 18 Lake Michigan tributaries across two time periods in
2019. Pies represent the proportion of each phylum observed at each sampling period in each river.
Fig. 3. UPGMA tree based on Bray-Curtis distances (taxa classified to phyla) describing the microbial community compositional dissimilarity between Lake Michigan
tributaries sampled during April and June 2019. Values in parentheses following stream names indicate the numeric stream designation. Values greater than 70% are shown
based on 1000 bootstrap replicates.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
7

terial communities were also documented at the genus level. In the
southeastern portion of the lake basin, several taxa including
Acinetobacter and obligate anaerobic genera Romboutsua, Turicibac-
ter and Clostridium were present in >10-fold higher abundance in
both sampling periods relative to other streams.
Bray-Curtis dissimilarity provided a measure of bacterial com-
munity beta-diversity and was used to produce a visual represen-
tation showing heterogeneity in tributary bacterial community
composition based on UPGMA clustering (Fig. 3). UPGMA tree
topology, levels of bootstrap support, and branch lengths indicate
bacterial community composition differed based on collection time
(April vs June; Fig. 3). Clustering was observed at a macro-spatial
scale, based on Lake Michigan basin region (e.g., Manistique,
Escanaba, Whitefish and Cedar rivers in the northern sub-basin
vs the Grand, Macatawa and Kalamazoo rivers in the south eastern
portion of the sub-basin. Across the entire basin, Mantel tests indi-
cated significant correlation between tributary geographical dis-
tance and microbial community Bray-Curtis dissimilarity for both
sampling times (April: p = 0.0015, correlation = 0.302, t = 3.169;
June: p = 0.021, correlation = 0.239, t = 2.304).
Compositional heterogeneity in stream bacterial communities
was further seen based on clustering using a different ordination
method (non-metric multidimensional scaling; NMDS, stress
value = 0.121; Fig. 4) that shows differences in community compo-
sition between sampling periods (Fig. 4, panel A and panel B). This
result is similar to that of the UPGMA dissimilarity tree, where
clustering based on sampling time was commonly seen (Fig. 3).
Hierarchical clustering of the samples shown in the heatmap
(Fig. 5) indicates modest influences of both watershed landcover
and seasonality, although these relationships did not receive
strong statistical support (ESM Fig. S2). Samples from watersheds
with relatively high levels of urban and agricultural landcover such
as the Kewaunee, Grand, Kalamazoo, and Macatawa rivers clus-
tered together regardless of whether samples were collected in
April or June. Rivers with a more forest dominated watershed in
the northern portion of the basin, such as the Menominee, Escan-
aba, Peshtigo, and Millecoquins rivers, were more likely to cluster
by sampling time. Microbial taxa included in the heatmap and
hierarchical clustering analysis were broadly divided into two
groups based on their relative abundances, however, the bootstrap
resampling suggested little statistical support for the split (ESM
Fig. S3).
Distance-based redundancy analysis models indicated signifi-
cant associations between microbial community composition and
landscape features in April (p = 0.007) and June (p = 0.01) sample
sets (Table 3). For each time period, significant effects of urban
(April: p < 0.001; June: p = 0.032) and agricultural (April:
p = 0.026; June: p = 0.0004) land cover were observed (Table 3).
Fig. 4. Non-metric multidimensional scaling plots with accompanying Shepard’s plot for residual goodness of fit for relationships of clustering based on sample time period
and stream location for 2019 bacterial communities (taxa classified to phyla) of Lake Michigan tributaries. Panel A illustrates clustering based on seasonality, with stress value
0.121 (April: solid line/triangles; June: dotted line/circles). Panel B illustrates clustering based on geographic location with stress value 0.121 (East Lake Michigan [state of
Michigan]: triangles; West Lake Michigan [state of Wisconsin]: squares; North Lake Michigan [Upper Peninsula of Michigan]: circles). Panel C contains the Shepard’s plot of
goodness of fit regressions for both time and state clustering.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
8

Discussion
Influence of landuse on bacterial community spatial compositional
structure
Results from this study identified a significant positive rela-
tionship between Great Lakes tributary microbial community
compositional dissimilarity (Bray-Curtis distance) and geographic
distance between rivers. Community taxonomic dissimilarity was
likely not related to geographic distance-dependent exchange of
waters among tributaries per se but was likely associated with
geographic patterns in subsurface geology (and water chemistry
that was not measured) and geographic variation in landuse in
terrestrial landscapes adjoining the rivers surveyed (Fig. 1; see
results above). We documented north to south latitudinal varia-
tion in landuse across the Lake Michigan basin. Landscapes asso-
ciated with northern basin watersheds and rivers were
dominated by forest cover, while rivers and watersheds in south-
ern regions of the survey area were composed largely of land-
scapes converted to agriculture and urban landuse (Fig. 1).
Differences in subsurface geology exists between eastern and
western basin streams as well as proportional contributions to
hydrological flow due to surface run-off in the west (Wisconsin)
and groundwater in the east (Michigan).
Fig. 5. Heatmap of analyzed microbial community samples including only taxa represented by a minimum of 0.2% of total collected reads. Relative abundances following a
log10 transformation with an additional pseudocount of 0.001 are shown. Categorical sample descriptors shown in rows above the heatmap include sum of urban and
agricultural National Land Cover Database (NLCD) categories in the watershed (low: <35%, high: 35%), proportion of NLCD forest categories in the watershed (low: <35%,
high: 35%), Subbasin of Lake Michigan, and sampling time. Asterisks on taxon and sample labels indicate 90% bootstrap support at the finest hierarchical clustering level.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
9

Studies of microbial compositional variation (e.g., Mulholland
et al, 2008; Philippot et al., 2010), have described the conservation
of microbial community functions of phylogenetically related taxa
to ecologically similar environmental conditions. Past studies have
reported associations between bacterial community composition
and specific stream chemical and hydrological features. For exam-
ple, analyses of microbial community spatial compositional
heterogeneity have highlighted several assembly mechanisms
including hydrological processes (Langenheder et al., 2012) and
stream discharge (Chen et al., 2018). Data in this study are novel
in the Great Lakes and elsewhere by highlighting effects of
broad-scale watershed-level landuse features, specifically percent-
ages of watershed landscapes in forest, agriculture, and urban lan-
duse on microbial community composition and diversity (Table 3,
Fig. 5). Levels of beta diversity characterized as Bray-Curtis dis-
tances can be attributed to niche or selective processes (Viana
et al., 2015), as demonstrated in this study based on associated
with watershed-level landuse patterns and correlated environ-
mental covariates (ESM Fig. S1).
Deforestation associated with land conversion to agriculture
and urban/suburban uses in watersheds causes a shift in microbial
biomass as well as a reduction in leaf decomposition rates
(Mlambo et al., 2019). The phylogenetic structure of stream com-
munities has also been linked to urbanization (Hose et al., 2016).
Anthropogenic point-source inputs including pollutants from
urban runoff and sewage have been shown to affect microbial com-
munity composition (Mansfeldt et al., 2020; Xu et al., 2018). Urban
storm-water run-off and snowmelt pollution from impervious sur-
faces significantly degrades surface waters quality (Müller et al.,
2020) that can affect biotic communities. High levels of urban
activity as measured by percentages of landscapes devoted to
urban activities shifts the microbial community to greater denitri-
fication, and anammox activities (Caillon and Schelker, 2020). In
this study, we documented high relative sequence abundance of
anaerobic bacteria in southeastern basin streams that are associ-
ated with high levels of urban landuse.
Effects of sampling locations on microbial community composition
We sampled below the first stream barrier relative to river con-
fluence with Lake Michigan. Therefore, we were unable to project
community compositional difference in upstream locations. How-
ever, considerable portions of all rivers were impacted by dams
(UDOR; Table 1), which can impact stream chemistry and particu-
late matter (Bilby, 1981), and concomitantly the taxonomy of
stream microbial communities (Ruiz-González et al., 2013) in sim-
ilar ways. Water retained above a dam increases surface area and
temperature, while concomitantly increasing photosynthetic rate
and productivity. Dams have also been found to decrease diversity
of planktonic bacteria and reduce carbon and nitrogen in down-
stream sediments (Mao et al., 2019). Other investigators docu-
mented a decrease of microbial phyla Beta- and Gamma-
Proteobacteria and Bacteroidetes downstream from dams and a
concomitant increase in alpha-Proteobacteria and Actinobacteria
(Ruiz-González et al., 2015).
Temporal differences in microbial community composition
Clustering using community dissimilarity (Bray-Curtis distance)
examined using NMDS plots (Fig. 4) and a UPGMA tree (Fig. 3)
revealed that bacterial community composition in the surveyed
Lake Michigan tributaries varied greatly between April and June
sampling periods (Figs. 3 and 5). With a two-month difference
between sampling times, temporal changes in weather patterns
occurred, such as changes in rainfall. Absence of snow melt
reduced June discharge levels from April conditions. Water tem-
peratures also increased considerably from April to June (ESM
Table S1). Snowmelt and variation in precipitation can lead to ter-
restrial bacterial washing into waterbodies, altering overall aquatic
bacterial community composition (Mohanta and Goel, 2014).
Bacteria community composition has been observed to change
based on temporal variation in weather patterns that affect stream
temperature and discharge (Crump and Hobbie, 2005; Hullar et al.,
2006; Leff et al., 1998). Temporal changes in aquatic microbial
communities can be substantial even over short time periods
(Zhang et al., 2020), as communities change with increasing tem-
perature and photosynthetic activity. Temporal variability in com-
munity composition has been described as being associated with
differences in primary productivity that is impacted by water tem-
perature (Langenheder and Ragnarsson, 2007). Variation in micro-
bial community composition and diversity in space and time has
been attributed to responses to differences in nutrient availability
(Dodds et al., 2000), amount of organic matter (Nelson, 2017; Tank
et al., 2010), temperature (Boyero et al., 2011), and hydrology
(Valett et al., 1997).
In the context of watershed-wide scales surveyed in this study,
watersheds provide a source of dissolved organic carbon and soil
microorganisms to streams, most notably during high flow events
(Caillon and Schelker, 2020) such as our April sampling period
(ESM Table S1).
The two time points (April and June) were selected because of
large hydrological and temperature changes known to occur
(ESM Table S1), and another recent Great Lakes lake study had
demonstrated the importance of hydrology and temperature to dif-
ference in microbial community composition (Paver et al., 2020). In
addition, the sampling dates encompass the spawning and rearing
conditions of early fish life stages that have been shown to be
widely affected by microbe-induced fish egg mortality (Forsythe
et al., 2013; Fujimoto et al., 2020) and colonization of internal
(Abdul Razak and Scribner, 2020; Abdul Razak et al., 2019) and
external (Llewellyn et al., 2014) fish surfaces.
Table 3
Distance-based redundancy analysis using Bray-Curtis distances and transformed environmental variables from sampled Lake Michigan watershed streams with significant
model effects in bold.
Landscape/spatial variable April June
Df Sum of Squares F Pr(>F) Df Sum of Squares F Pr(>F)
Network catchment area 1 0.316 1.123 0.225 1 0.311 1.133 0.186
Upstream mainstem dam density per unit upstream mainstem length 1 0.299 1.062 0.313 1 0.327 1.190 0.129
Proportion of est. annual discharge stored in upstream reservoirs 1 0.327 1.164 0.173 1 0.276 1.004 0.450
Urban land cover in network catchment 1 0.471 1.676 0.001 1 0.365 1.330 0.028
Agricultural land cover in network catchment 1 0.397 1.414 0.025 1 0.426 1.552 0.004
Latitude of river mouth 1 0.330 1.173 0.167 1 0.280 1.020 0.406
Longitude of river mouth 1 0.309 1.100 0.257 1 0.307 1.118 0.212
Residual 9 2.529 NA NA 9 2.471 NA NA
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
10

Conclusions
Several important findings from this study contribute to under-
standing stream bacterial metacommunity organization, which
have relevance to the management of stream ecosystems. Associa-
tions established here between bacteria community composition
and landuse, specifically the proportions of river watersheds in
urban, forest, and agriculture, demonstrates the importance of
landscapes as deterministic sources of variation (mass effects) over
space and time. Urban and agricultural landuse practices have sig-
nificant impacts on Lake Michigan tributary bacterial communities,
as quantified using by distance-based redundancy analysis of bac-
terial community sequences (Table 3) obtained from high-
throughput metabarcoding. The magnitude of differences in bacte-
rial communities between tributaries that was quantitatively asso-
ciated with watershed-scale land use proportions was a novel
finding. Mass effects of landscape-scale environmental factors
shaped spatial and temporal bacterial community composition.
Microbial communities represent the lower trophic level building
blocks on which stream ecosystems are based, emphasizing the
importance of micro-organism communities to aquatic ecosystem
organization and function, including primary productivity (Linz
et al., 2020), organic matter decomposition (Mlambo et al., 2019),
and denitrification (Mulholland et al., 2008). Microbial communi-
ties thus can be used in Great Lakes aquatic ecosystem manage-
ment as biological indicators of human impacts and landscape
level. Further studies are warranted that tie spatial and temporal
landuse-based changes in community composition to ecosystem
function.
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgements
Funding for this project was provided by the Great Lakes Fish-
ery Trust, the Fisheries Division of the Michigan Department of
Natural Resources, the U.S. Geological Survey Aquatic GAP Pro-
gram, and Michigan State University. We thank Dana Infante and
Arthur Cooper for help obtaining variables associated dams and
landscape condition from our study watersheds. Members of the
Scribner Lab provided useful comments on earlier drafts of the
paper.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jglr.2021.02.009.
References
Abdul Razak, S., Griffin, M.J., Mischke, C.C., Bosworth, B.G., Waldbieser, G.C., Wise, D.
J., Marsh, T.L., Scribner, K.T., 2019. Biotic and abiotic factors influencing channel
catfish egg and gut microbiome dynamics during early life stages. Aquaculture
498, 556–567. https://doi.org/10.1016/j.aquaculture.2018.08.073.
Abdul Razak, S.A., Scribner, K.T., 2020. Ecological and ontogenetic components of
larval Lake Sturgeon gut microbiota assembly, successional dynamics, and
ecological evaluation of neutral community processes. J. Appl. Environ
Microbiol. 86. https://doi.org/10.1128/AEM.02662-19.
Allan, J.D., 2004. Landscapes and Riverscapes: the influence of land use on stream
ecosystems. Annu. Rev. Ecol. Evol. Syst. 35, 257–284. https://doi.org/10.1146/
annurev.ecolsys.35.120202.110122.
Beisner, B.E., Peres-Neto, P.R., Lindstrom, E.S., Barnett, A., Longhi, M.L., 2006. The
role of environmental and spatial processes in structuring lake communities
from bacteria to fish. Ecology 87, 2985–2991. https://doi.org/10.1890/0012-
9658(2006)87[2985:troeas]2.0.co;2.
Bilby, R.E., 1981. Role of organic debris dams in regulating the export of dissolved
and particulate matter from a forested watershed. Ecology 62, 1234–1243.
https://doi.org/10.2307/1937288.
Boyero, L., Pearson, R.G., Gessner, M.O., Barmuta, L.A., Ferreira, V., Graca, M.A.S.,
2011. A global experiment suggests climate warming will not accelerate litter
decomposition in streams but might reduce carbon sequestration. Ecol. Lett. 14,
289–294. https://doi.org/10.1111/j.1461-0248.2010. 01578.x.
Caillon, F., Schelker, J., 2020. Dynamic transfer of soil bacteria and dissolved organic
carbon into small streams during hydrological events. Aquat. Sci. 82, 41. https://
doi.org/10.1007/s00027-020-0714-4.
Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-Lyons, D., Lozupone, C.A., Turnbaugh,
P.J., Fierer, N., Knight, R., 2011. Global patterns of 16S rRNA diversity at a depth
of millions of sequences per sample. PNAS 108 (Suppl 1), 4516–4522. https://
doi.org/10.1073/pnas.1000080107.
Chen, W., Wilkes, G., Khan, I., Katarina, D.M., Pintar, J., Thomas, L., Lévesque, A.,
Chapados, J.T., Topp, E., Lapen, D.R., 2018. Aquatic bacterial communities
associated with land use and environmental factors in agricultural landscapes
using a metabarcoding approach. Front. Microbiol. 9. https://doi.org/10.3389/
fmicb.2018.02301.
Colford, J.M., Wade, T.J., Schiff, K.C., Wright, C.C., Griffith, J.F., Sandhu, S.K., Burns, S.,
Sobsey, M., Lovelace, G., Weisberg, S.B., 2007. Water quality indicators and the
risk of illness at beaches with nonpoint sources of fecal contamination.
Epidemiology. 18 (1), 27–35. https://doi.org/10.1097/01.
ede.0000249425.32990.b9.
Cooper, A.R., Infante, D.M., Daniel, W.M., Wehrly, K.E., Wang, L., Brenden, T.O., 2017.
Assessment of dam effects on streams and fish assemblages of the
conterminous USA. Sci. Total Environ. 586, 879–889. https://doi.org/10.1016/j.
scitotenv.2017.02.067.
Cotner, J.B., Biddanda, B.A., 2002. Small players, large role: microbial influence on
biogeochemical processes in pelagic aquatic ecosystems. Ecosystems 5, 105–
121. https://doi.org/10.1007/s10021-001-0059-3.
Cox, T., Cox, M., 2000. In: Multidimensional Scaling. Chapman and Hall, New York.
https://doi.org/10.1201/9781420036121.
Crump, B.C., Hobbie, J.E., 2005. Synchrony and seasonality in bacterioplankton
communities of two temperate rivers. Limnol. Oceanogr. 50, 1718–1729.
https://doi.org/10.4319/lo.2005.50.6.1718.
Crump, B.C., Amaral-Zettler, L.A., Kling, G.W., 2012. Microbial diversity in arctic
fresh waters is structured by inoculation of microbes from soils. ISME J. 6,
1629–1639. https://doi.org/10.1038/ismej.2012.9.
Dodds, W.K., López, A.J., Bowden, W.B., Gregory, S., Grimm, N.B., Hamilton, S.K.,
2000. N uptake as a function of concentration in streams. J. North Am. Benthol.
Soc. 21, 206–220. https://doi.org/10.2307/1468410.
Derose, K.L., Roche, L.M., Lile, D.F., Eastburn, D.J., Tate, K.W., 2020. Microbial water
quality conditions associated with livestock grazing, recreation, and rural
residences in mixed-use landscapes. Sustainability 12. https://doi.org/
10.3390/su12125207.
De Schryver, P., Vadstein, O., 2014. Ecological theory as a foundation to control
pathogenic invasion in aquaculture. ISME J. 8, 2360–2368. https://doi.org/
10.1038/ismej.2014.84.
Esselman, P.C., Infante, D.M., Wang, L., Cooper, A.R., Wieferich, D.L., Tsang, Y.P.,
Thornbrugh, D.J., Taylor, W.W., 2013. Regional fish community indicators of
landscape disturbance to catchments of the conterminous United States. Ecol.
Indic. 26, 163–173. https://doi.org/10.1016/j.ecolind.2012.10.028.
Fan, J., Ming, H., Li, L., Su, J., 2015. Evaluating spatial-temporal variations and
correlation between fecal indicator bacteria (FIB) in marine bathing beaches. J.
Water Health 13, 1029–1038. https://doi.org/10.2166/wh.2015.031.
Fausch, K.D., Torgersen, C.E., Baxter, C.V., Li, W.H., 2002. Landscapes to riverscapes:
bridging the gap between research and conservation of stream fishes.
Bioscience 52, 483–495. https://doi.org/10.1641/0006-3568(2002)052[0483:
LTRBTG]2.0.CO;2.
Fazi, S., Vázquez, E., Casamayor, E.O., Amalfitano, S., Butturini, J., 2013. Stream
hydrological fragmentation drives bacterioplankton community composition.
PLoS One 8. https://doi.org/10.1371/journal.pone.0064109.
Forsythe, P.S., Scribner, K.T., Crossman, J.A., Ragavendran, R., Baker, E.A., 2013.
Experimental assessment of the magnitude and sources of lake sturgeon egg
mortality in a natural stream setting. Trans. Am. Fish. Soc. 142, 1005–1011.
https://doi.org/10.1080/00028487.2013.790847.
Fujimoto, M., Marsh, T., Scribner, K.T., 2020. Effects of filtration and temperature on
microbial colonization of lake sturgeon eggs, survival, and development in
aquaculture. N. Am. J. Aquat. 83, 26–37. https://doi.org/10.1002/naaq.10169.
Galili, T., 2015. dendextend: an R package for visualizing, adjusting and comparing
trees of hierarchical clustering. Bioinformatics 31, 3718–3720. https://doi.org/
10.1093/bioinformatics/btv428.
Goldford, J.E., Lu, N., Bajic
´, D., Estrela, S., Tikhonov, M., Sanchez-Gorostiaga, A., Segrè,
D., Mehta, P., Sanchez, A., 2018. Emergent simplicity in microbial community
assembly. Science 361, 469–474. https://doi.org/10.1126/science:aat1168.
Gessner, M.O., Swan, C.M., Dang, C.K., McKie, B.G., Bardgett, R.D., Wall, D., 2010.
Diversity meets decomposition. Trends Ecol. Evol. 25, 372–380.
Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: Paleontological statistics
software package for education and data analysis. Palaeontol. Electron. 4
http://palaeo-electronica.org/2001_1/past/issue1_01.htm.
Hermans, S.M., Buckley, H.L., Case, B.S., Lear, G., 2019. Connecting through space
and time: catchment-scale distributions of bacteria in soil, stream water and
sediment. Environ. Microbiol. 22, 1000–1010. https://doi.org/10.1111/1462-
2920.14792.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
11

Hieber, M., Gessner, M.O., 2002. Contribution of stream detritivores, fungi, and
bacteria to leaf breakdown based on biomass estimates. Ecology 83, 1026–1038.
https://doi.org/10.1890/0012-9658(2002)083[1026:COSDFA]2.0.CO;2.
Hijmans, R.J., Williams, E., Vennes, C., 2019. geosphere: Spherical Trigonometry. R
Package Version 1.5-10. CRAN https://cran.r-project.org/
web/packages/geosphere/index.html.
Hose, J.D., Febria, C.M., Crump, B.C., Palner, M.A., 2016. Watershed urbanization
linked to differences in stream bacterial commuity composition. Front.
Microbiol 8, 1452. https://doi.org/10.3389/fmicb.2017.01452.
Huggett, M.J., Kavazos, C.R.J., Bernasconi, R., Czarnik, R., Horwitz, P., 2017.
Bacterioplankton assemblages in coastal pond reflect the influence of
hydrology and geomorphological setting. FEMS Microbiol. Ecol. 93. https://
doi.org/10.1093/femsec/fixo67.
Hullar, M.A.J., Kaplan, L.A., Stahl, D.A., 2006. Recurring seasonal dynamics of
microbial communities in stream habitats. Appl. Environ. Microbiol. 72, 713–
722. https://doi.org/10.1128/AEM.72.1.713-722.2006.
Isabwe, A., Yang, J.R., Wang, Y., Liu, L., Chen, H., Yang, J., 2018. Community assembly
processes underlying phytoplankton and bacterioplankton across a hydrologic
change in a human-impacted river. Sci. Total Environ. 630, 658–667. https://doi.
org/10.1016/j.scitotenv.2018.210.
Kaestli, M., Skillington, A., Kennedy, K., Majid, M., Williams, D., McGuinness, K.,
Munksgaard, N., Gibb, K., 2017. Spatial and temporal microbial patterns in a
tropical macrotidal estuary subject to urbanization. Front. Microbiol. 8. https://
doi.org/10.3389/fmicb.2017.01313.
Kärnä, O., Heino, J., Laamanen, T., 2019. Does catchment geodiversity foster stream
biodiversity? Landsc. Ecol. 34, 2469–2485. https://doi.org/10.1007/s10980-
019-00901-z.
Kolde, R., 2019. pheatmap: Pretty heatmaps. R Package Version 1.0.12. CRAN
https://CRAN.R-project.org/package=pheatmap.
Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K., Schloss, P.D., 2013.
Development of a dual-index sequencing strategy and curation pipeline for
analyzing amplicon sequence data on the MiSeq Illumina sequencing platform.
Appl. Environ. Microbiol. 79, 5112–5120. https://doi.org/10.1128/AEM.01043-
13.
Langenheder, S., Ragnarsson, H., 2007. The role of environmental and spatial factors
for the composition of aquatic bacterial communities. Ecology 88, 2154–2161.
https://doi.org/10.1890/06-2098.1.
Langenheder, S., Berga, M., Ostman, O., Szekely, A.J., 2012. Temporal variation of
beta-diversity and assembly mechanisms in a bacterial metacommunity. ISME
J. 6, 1107–1114. https://doi.org/10.1038/ismej.2011.177.
Leff, L.G., Leff, A.A., Lemke, M.J., 1998. Seasonal changes in planktonic bacterial
assemblages of two Ohio streams. Freshw. Biol. 39, 129–134. https://doi.org/
10.1046/j.1365-2427.1998.00269.x.
Legendre, P., Andersson, M.J., 1999. Distance-based redundancy analysis: testing
multispecies responses in multifactorial ecological experiments. Ecol. Monogr.
69. https://doi.org/10.1890/0012-9615. 512–512.
Lindstrom, E.S., Kamst-Van Agterveld, M.P., Zwart, G., 2005. Distribution of typical
freshwater bacterial groups is associated with pH, temperature, and lake water
retention time. Appl. Environ. Microbiol. 71 (12), 8201–8206. https://doi.org/
10.1128/AEM.71.12.8201-8206.2005.
Linz, A.M., Aylward, F.O., Bertilsson, S., McMahon, K.D., 2020. Time-series
metatranscriptomes reveal conserved patterns between phototrophic and
heterotrophic microbes in diverse freshwater systems. Limnol. Oceanogr. 65,
S101–S112. https://doi.org/10.1002/Ino.11306.
Llewellyn, M.S., Boutin, S., Hoseinifar, S.H., Derome, N., 2014. Teleost microbiomes:
the state of the art in their characterization, manipulation and importance in
aquaculture and fisheries. Front. Microbiol. 5. https://doi.org/10.3389/
fmicb.2014.00207.
Mansfeldt, C., Deiner, K., Machler, E., Fenner, K., Eggen, R.I.L., Stamm, C.,
Schonenberger, U., Walser, J.-C., Altermatt, J., 2020. Microbial community
shits in streams receiving treated wastewater effluent. Sci. Total Environ. 709.
https://doi.org/10.1016/j.scitotenv.2019.135727.
Mao, Y., Liu, Y., Li, H., He, Q., Ai, H., Gu, W., Yang, G., 2019. Distinct responses of
planktonic and sedimentary bacterial communities to anthropogenic activities:
case study of a tributary of the Three Gorges Reservoir, China. Sci. Total Environ.
682, 324–332. https://doi.org/10.1016/j.scitotenv.2019.05.172.
McKay, L., Bondelid, T., Dewald, T., Johnston, J., Moore, R., Rea, A., 2012. NHDPlus
Version 2: User Guide (accessed 11 Feb 2021) https://nhdplus.com/NHDPlus/
NHDPlusV2_documentation.php#NHDPlusV2%20User%20Guide.
Mlambo, M.C., Paavola, R., Fritze, H., Louhi, P., Muotka, T., 2019. Leaf litter
decomposition and decomposer communities in streams affected by intensive
forest biomass removal. Ecol. Indic. 101, 364–372. https://doi.org/10.1016/j.
ecolind.2019.01.035.
Mohanta, T., Goel, S., 2014. Prevalence of antibiotic-resistant bacteria in three
different aquatic environments over three seasons. Environ. Monitor. Assess.
186, 5089–5100. https://doi.org/10.1007/s10661-014-3762-1.
Mulholland, P., Helton, A., Poole, G., 2008. Stream denitrification across biomes and
its response to anthropogenic nitrate loading. Nature 452, 202–205. https://doi.
org/10.1038/nature06686.
Müller, A., Österlund, H., Marsalek, J., Viklander, M., 2020. The pollution conveyed
by urban runoff: a review of sources. Sci. Total Environ. 709. https://doi.org/
10.1016/j.scitotenv.2019.136125.
[dataset] National Land Cover Database 2011 (NLCD 2011), Multi-Resolution Land
Characteristics Consortium, 2018, 1f6db7c0-808f-48c5-810c-8b9d1e11aa02.
Nelson, E.B., 2017. The seed microbiome: origins, interactions, and impacts. Plant
Soil 422, 7–34. https://doi.org/10.1007/s11104-017-3289-7.
Nino-Garcia, J.P., Ruiz-González, C., del Giorgio, P.A., 2016. Interactions between
hydrology and water chemistry shape bacterioplankton biogeography across
boreal freshwater networks. 2016. ISME J. 10, 1755–1766. https://doi.org/
10.1038/ismej.2015.226.
Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin,
P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Szoecs, E., Wagner,
H., 2019. vegan: Community Ecology Package. R Package Version 2.5-2. CRAN
http://CRAN.Rproject.org/package=vegan.
Parveen, B., Reveilliez, J.-P., Mary, I., Ravet, V., Bronner, G., Mangot, J.-F., Domaizon,
E., Debroas, D., 2011. Diversity and dynamics of free-living and particle-
associated Betaproteobacteria and actinobacteria in relation to phytoplankton
and zooplankton communities. FEMS Microb. Ecol. 2011, 1–16. https://doi.org/
10.1111/j.1574-6941.2011.01130.x.
Paver, S.F., Newton, R.J., Coleman, M.L., 2020. Microbial communities of the
Laurentian Great Lakes reflect connectivity and local biogeochemistry.
Environ. Microbiol. 22, 433–446. https://doi.org/10.1111/1462-2920.14862.
Peet, R.K., 1975. Relative diversity indices. Ecology 56, 496–498. https://doi.org/
10.2307/1934984.
Philippot, L., Andersson, S.G.E., Battin, T.J., Prosser, J.I., Schimel, J.P., Whitman, W.B.,
2010. The ecological coherence of high bacterial taxonomic ranks. Nat. Rev.
Microbiol. 8, 523–529.
Riseng, C.M., Wiley, M.J., Seelbach, P.W., Stevenson, R.J., 2010. An ecological
assessment of Great Lakes tributaries in the Michigan Peninsulas. J. Great Lakes
Res. 36, 505–519. https://doi.org/10.1016/j.jglr.2010.04.008.
Rosenberg, M.S., Anderson, C.D., 2011. PASSaGE: pattern analysis, spatial statistics
and geographic exegesis. Version 2. Methods Ecol. Evol. 2, 229–232. https://doi.
org/10.1111/j.2041-210X.2010.00081.x.
Ruiz-González, C., Proia, L., Ferrera, I., Gasol, J.M., Sabater, S., 2013. Effects of large
river dam regulation on bacterioplankton community structure. FEMS Microb.
Ecol. 84, 316–331. https://doi.org/10.1111/1574-6941.12063.
Ruiz-González, C., Niño-García, J.P., del Giorgio, P.A., 2015. Terrestrial origin of
bacterial communities in complex boreal freshwater networks. Ecol. Lett. 18,
1198–1206. https://doi.org/10.1111/ele.12499.
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B.,
Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B.,
Thallinger, G.G., Van Horn, D.J., Weber, C.F., 2009. Introducing mothur: open-
source, platform-independent, community-supported software for describing
and comparing microbial communities. Appl. Environ. Microbiol. 75, 7537–
7541. https://doi.org/10.1128/AEM.01541-09.
Simpson, E.H., 1949. Measurement of diversity. Nature 163, 688. https://doi.org/
10.1038/163688a0.
Soranno, P.A., Cheruvelil, K., Webster, K.E., Bremigan, M., Wagner, T., 2010. Using
landscape limnology to classify freshwater ecosystems for multi-ecosystem
management and conservation. BioScience. 60, 440–454. https://doi.org/
10.1525/bio.2010.60.6.8.
Suzuki, R., Terada, U., Shimodaira, H., 2019. pvclust: Hierarchical Clustering with p-
values via Multiscale Bootstrap Resampling. R Package Version 2.2-0. CRAN
https://CRAN.R-project.org/package=pvclust.
Tank, J.L., Rosi-Marshall, E.J., Griffiths, N.A., Entrekin, S.A., Stephen, M.L., 2010. A
review of allochthonous organic matter dynamics and metabolism in streams.
Freshw. Sci. 29, 118–146. https://doi.org/10.1899/08-170.1.
Tonkin, J.D., Altermatt, F., Finn, D.S., Heino, J., Older, J.D., Pauls, S.U., Lytle, D.A., 2018.
The role of dispersal in river network metacommunities: patterns,
processes, and pathways. Freshw. Biol. 63, 141–163. https://doi.org/10.1111/
fwb/13037.
Tsang, Y.P., Wieferich, D., Fung, K., Infante, D.M., Cooper, A.R., 2014. An approach for
aggregating upstream catchment information to support research and
management of fluvial systems across large landscapes. SpringerPlus 3, 589.
https://doi.org/10.1186/2193-1801-3-589.
Valett, H.M., Dahm, C.N., Campana, M.E., Morrice, J.A., Baker, M.A., Fellows, C.S.,
1997. Hydrologic influences on groundwater surface water ecotones:
heterogeneity in nutrient composition and retention. J. North Am. Benthol.
Soc.16, 239–247. https://doi:10.2307/1468254.
Vannote, R.L., Minshall, G.W., Cummins, K.W., Sedell, J.R., Cushing, C.E., 1980. The
river continuum concept. Can. J. Fish. Aquat. Sci. 37, 130–137. https://doi.org/
10.1139/f80-017.
Vellend, M., Agrawal, A., 2010. Conceptual synthesis in community ecology. Q. Rev.
Biol. 85, 183–206. https://doi.org/10.1086/652373.
Verhougstraete, M.P., Martin, S.L., Kendall, A.D., Hyndman, D.W., Rose, J.B., 2015.
Linking fecal bacteria in rivers to landscape, geochemical, and hydrological
factors and sources at the basin scale. Proc. Natl. Acad. Sci. U.S.A. 112, 10419–
10424. https://doi.org/10.1073/pnas.1415836112.
Verschuere, L., Rombaut, G., Sorgeloos, P., Verstraete, W., 2000. Probiotic bacteria as
biological control agents in aquaculture. Microbiol. Mol. Biol. Rev. 64, 655–671.
https://doi.org/10.1128/MMBR.64.4.655-671.2000.
Viana, D.S., Figuerola, J., Schwenk, K., Manca, M., Hobaek, A., Mjelde, M., Preston, C.
D., Gornall, R.J., Croft, J.M., King, R.A., Green, A.J., Santameria, L., 2015. Assembly
mechanisms determine high species turnover in aquatic communities over
regional and continental scales. Ecography 38, 1–18. https://doi.org/10.1111/
ecog.01231.
Wang, L., Infante, D., Esselman, P., Cooper, A., Wu, D., Taylor, W., Beard, D., Whelan,
G., Ostroff, A., 2011. A hierarchical spatial framework and database for the
national river fish habitat condition assessment. Fisheries 36, 436–449. https://
doi.org/10.1080/03632415.2011.607075.
Wang, J., Shen, J., Wu, Y., Tu, C., Soininen, J., Stegen, J.C., He, J., Liu, X., Zhang, L.,
Zhang, E., 2013. Phylogenetic beta diversity in bacterial assemblages across
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
12

ecosystems: deterministic versus stochastic processes. ISME J. 7, 1310–1321.
https://doi.org/10.1038/ismej.2013.30.
Wu, H., Chen, J., Xu, J., Zeng, G., Sang, L., Liu, Q., Yin, Z., Dai, J., Yin, D., Liang, J., Ye, S.,
2019. Effects of dam construction on biodiversity: a review. J.Clean. Prod. 221,
480489. https://doi.org/10.1016/j.jclepro.2019.03.001.
Wickham, H., 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag,
New York. https://ggplot2.tidyverse.org.
Xu, Z., Jiang, Y., Te, S.H., He, Y., Gin, K.Y-H., 2018. The effects of antibiotics on
microbial community composition in an estuary reservoir during spring and
summer seasons. Water 10, 154. https://doi.org/10.3390/w10020154.
Zeglin, L.H., 2015. Stream microbial diversity in response to environmental
changes: review and synthesis of existing research. Front. Microbiol. 6.
Zhang, L., Zhao, F., Li, X., Lu, W., 2020. Contribution of influent rivers affected by
different types of pollution to the changes of benthic microbial community
structure in a large lake. Ecotoxicol. Environ. Saf. 198, 110657. https://doi.org/
10.1016/j.ecoenv.2020.110657.
G.E. Sanfilippo, J.J. Homola, J. Ross et al. Journal of Great Lakes Research xxx (xxxx) xxx
13