Cooperative Science and Monitoring Initiative (CSMI)
Lake Michigan 2015 Report
Edited by Carolyn Foley and Paris Collingsworth, Illinois-Indiana Sea Grant
Contact:
Carolyn Foley
Research Coordinator
Illinois-Indiana Sea Grant
Purdue University
195 Marsteller Street
West Lafayette, IN 47907-2033
765-494-3601 | cfolev@purdue.edu
Paris Collingsworth
Great Lakes Ecosystem Specialist
Illinois-Indiana Sea Grant
Purdue University
195 Marsteller Street
West Lafayette, IN 47907-2033
312-886-7449 | pcolling@purdue.edu
IISG18-HCE-RLA-013
-------
Table of Contents
Executive Summary iii
List of Tables xiv
List of Figures xv
Describing the Distribution and Productivity of Biota Along a Nearshore to Offshore Gradient 1
Temporal and Spatial Coupling of Nutrients and Food Web—Microbes to Fish 16
Major Findings from the CSMI Benthic Macroinvertebrate Survey in Lake Michigan in 2015 With
an Emphasis on Temporal Trends 44
A Summary of Mid-Continent Ecology Division Efforts Associated With the 2015 Lake Michigan
Cooperative Science Monitoring Initiative (CSMI) 81
Water Quality and Lower Trophic Level Summary from the 2015 Lake Michigan CSMI 82
Application of a Nutrient Model to Address Nearshore Phosphorus Levels in Lake Michigan .. 96
"Data in Motion" - Continuous Water Sensor Data Collection for the 2015 Lake Michigan CSMI
102
Atrazine Concentrations in Lake Michigan: Investigating Causes of the Recent Decline 117
Examining Legacy and Emerging Contaminants in Lake Michigan Tributaries 121
CSMI Lake Michigan 2015 Report
ii
-------
Key Findings
Dreissenids and Diporeia. Lakewide quagga mussel density declined between 2010 and 2015, but
biomass increased slightly over the same period. Discrepancies between the trends in density and
biomass are due to population dynamics in the mid-depth (31-90 m) regions of the lake, where
most quagga mussels reside. In these regions quagga mussel densities are decreasing but the
remaining population consists of larger individuals. Meanwhile, in deeper regions of the lake (> 90
m), quagga mussel populations continue to expand, both in terms of density and biomass, but there
are still fewer mussels overall in this region. Diporeia populations continue to decline in Lake
Michigan. In the 2015 benthic survey, Diporeia were only found in 10 sites, most of which were
deeper than 90 m.
Pelagic food web. In 2015, CSMI sampling was undertaken to address the hypothesis that primary
and secondary production are higher in regions of the lake adjacent to high loading tributaries,
particularly in nearshore areas. Results show very little support for this hypothesis. Nearshore total
phosphorus was significantly higher at only one shallow (18 m) station off the Muskegon River,
whereas chlorophyll a concentrations were higher along transects adjacent to high loading
tributaries at shallow stations during May and July and at offshore stations (> 90 m) during July.
Zooplankton density was higher at nearshore stations only during the month of July, and no
differences were found in larval and adult fish density at different sampling depths. Finally, larval
fish growth rates did not differ across transect or sampling depths, but larval fish are growing about
half as fast as they did prior to the establishment of quagga mussels in the early 2000s.
Collectively, these results suggest that resource limitation is prevalent across the pelagic food web
in Lake Michigan. Further, the 2015 study design may not have been adequate to fully address this
CSMI Lake Michigan 2015 Report
iii
-------
this hypothesis. Future studies should consider sampling areas that are shallower than 18 m.
Contaminants. Atrazine in open lake water and PCB concentrations throughout the Lake Michigan
ecosystem are declining at a rate that exceeds the predictions of the Lake Michigan Mass Balance
Study from the 1990s. In addition, tributary loadings of PCBs and mercury were approximately 50-
70% lower in 2015 relative to loads calculated for 1994-1995. Collectively, the results from CSMI
contaminant monitoring efforts suggest that remediation efforts in Lake Michigan are meeting or
exceeding their goals.
Lake Michigan CSMI 2015 Executive Summary
The role of the Cooperative Science and Monitoring Initiative (CSMI) is to provide
enhanced monitoring and research activities that provide relevant information to address the science
priorities of the Lake Partnerships (established under the Lakewide Management Annex of the 2012
Great Lakes Water Quality Agreement) across the Laurentian Great Lakes. The Lake Michigan
Partnership, a collaborative team of natural resource managers led by the U.S. Environmental
Protection Agency with participation from federal, state, tribal, and local governments or agencies,
uses the information collected through CSMI to help develop long term ecosystem-based
management strategies for protecting and restoring Lake Michigan's water quality. On a practical
level, CSMI is an intensive effort to collect information on the health of each lake, rotating to one
Great Lake each year. In 2015, it was Lake Michigan's turn. The following is an executive
summary of the 2015 research results and the associated white paper containing more specific
information.
One of the primary ecosystem stressors in Lake Michigan is the proliferation of invasive
species. Many invasive species have entered the Great Lakes since the 1800s, but none have been as
CSMI Lake Michigan 2015 Report
iv
-------
prolific as dreissenid mussels, the collective term for zebra mussels (Dreissenapolymorpha) and
quagga mussels {Dreissena rostriformis bugensis). Quagga mussels were first detected in Lake
Michigan in 1997 and became well established by 2004. Since their introduction, quagga mussel
populations have widely and rapidly expanded, blanketing a huge portion of the lake bottom such
that they are now one of the most abundant organisms in the lake. As quagga mussels filter water,
they remove nutrients, bacteria, and phytoplankton from the lake, which means that much of the
nutrients that would have been available at the base of the food web under pre-invasion conditions
is now bound up in quagga mussel tissue and shells. In this way, the proliferation of quagga
mussels has disrupted the flow of energy up the food chain—from small zooplankton to top
predator fish—and therefore, quantifying the effects of quagga mussels on energy flow throughout
the lake is an important concern for the management community of Lake Michigan.
The most prevalent mechanistic hypothesis invoked to describe the negative effects of
quagga mussels on the Lake Michigan food web emphasizes their influence on the distribution of
nutrients in the water column. According to this hypothesis, commonly referred to as the nearshore
shunt hypothesis, nutrients entering Lake Michigan, coming primarily from agricultural watersheds
and large urban population centers around the southern basin, remain shunted in nearshore areas
due to the filtering activity of quagga mussels, which, in turn, leads to high primary production in
nearshore areas. While adequate levels of primary production are necessary to support a vibrant
Lake Michigan food web, nuisance levels of algal production, including harmful algal blooms in
productive embayments and Cladophora stands in coastal areas, represent a problem for the Lake
Michigan management community. Despite the long-term changes in food web structure that have
been documented since the establishment of quagga mussels in the early 2000s, the hypothesized
CSMI Lake Michigan 2015 Report
v
-------
effects of quagga mussel grazing, and subsequent localized increases in primary production, on
the food web had not been fully tested. In response to this pressing management concern, many of
the enhanced monitoring and research activities undertaken during the 2015 CSMI field year in
Lake Michigan sought to explicitly test this hypothesis on a lake wide scale.
The current status of the lake drove the following research priorities developed by the
Lake Michigan Partnership:
1) What is the status of the lower food web and can the lower food web be an indicator for
detecting ecological change?
2) What is the distribution, abundance, and movement of nutrients and biota across a
nearshore-offshore gradient?
3) What are the nearshore water-quality effects from tributary nutrient loading?
4) What is the current status of contaminant loads and cycling in the ecosystem?
The scientific approaches to addressing these priorities were not mutually exclusive, and in this
report the enhanced monitoring and research activities undertaken during the 2015 Lake Michigan
CSMI will be presented in three categories:
1) The status of the lower food web today versus the historic structure.
2) The status of the pelagic food web with respect to the timing of delivery and spatial
distribution of nutrients in Lake Michigan.
3) The status of loading and in-lake concentrations of contaminants in Lake Michigan.
The main results from this effort are summarized in the following sections. In addition, full
reports from each research group that participated in 2015 field year sampling are included after
this executive summary.
CSMI Lake Michigan 2015 Report
vi
-------
Lower Food Web and Ecological Change
Scientists working with the National Oceanic and Atmospheric Administration (NOAA)
and the U.S. Environmental Protection Agency (EPA) have been monitoring components of the
lower Lake Michigan food web, including zooplankton and benthic macroinvertebrates, since the
mid-1980s. Sampling schemes for 2015 used historical methods with a few key additions. In 2015,
CSMI partners from EPA, NOAA, Cornell University, the University at Buffalo, and University of
Michigan, monitored the whole benthic community rather than only zebra mussels, quagga
mussels, and Diporeia spp. (as had been done previously), and used additional sampling techniques
such as a benthic sled and optical plankton counters to complement historic sampling methods. The
technological advances in the 2015 CSMI sampling year greatly improved the scientists' abilities
to assess trends in Lake Michigan. However, standard techniques, such as PONAR grabs for
benthic invertebrates and zooplankton net tows, are important to maintain, not only for their ability
to assess trends over time, but to complement and validate the newer methods (e.g., using empty
shell estimates from PONAR samples to calibrate video captured by the benthic sled, and using
traditional net tows to calibrate zooplankton density estimates from towed laser optical particle
counters).
Benthic trends included in this report are summarized from the 1990s through 2015.
Sampling sites varied slightly by sampling year. Benthic samples were collected at 140 stations in
July of 2015 (135 in the main basin of the lake, and five in the outer portion of Green Bay).
Dreissenid density declined significantly between 2010 and 2015 (and dreissenids are now
comprised of quaggas mussels only). Dreissenid densities declined 79%, 56%, and 40% at < 30 m,
31-50 m, 51-90 m intervals, respectively, but increased 37% at depths > 90 m. Despite general
declines in density, overall biomass of dreissenids increased in 2015 over 2010. The discrepancy
CSMI Lake Michigan 2015 Report
vii
-------
between the trends in density and biomass of dreissenids can generally be attributed to a shift
toward fewer but larger-bodied individuals. In 2015, much of the dreissenid biomass was found at
depths of 30-90 m. Diporeia and Sphaeriidae continue to decline in 2015 from previous time
periods sampled. Currently, Diporeia are only found in very few areas (n = 10 sites), and are
generally found deeper than 90 m. Oligochaeta have progressively increased in shallower and mid-
depth regions between 1992 and 2015, especially in southeastern Lake Michigan.
Scientists from EPA, the US Geological Survey (USGS), NOAA, and Cornell University
assessed zooplankton community dynamics. Measurements of zooplankton biomass along transects
off Frankfort (Michigan), and Sturgeon Bay (Wisconsin), suggest that total zooplankton biomass
did not differ between 2010 and 2015 (although biomass of Cladocera such as Daphnia galeata
mendotae and Bythotrephes longimanus declined, and Mysis relicta biomass increased over this
time period). In June, dreissenid veligers were the dominant zooplankton in numbers and biomass
all along the Muskegon transect. Zooplankton were also found deeper in the water column in 2015
as compared to previous years. Quagga mussel filtration activities have led to increased levels of
water clarity throughout the photic zone (i.e., surface layer of the lake that receives sunlight) over
time, and zooplankton may be seeking refuge in murkier waters near the metalimnion. Both larval
fish and another relatively recent invader, Bythotrephes longimanus, visually prey on zooplankton.
Vertical distribution of fish larvae have indeed shifted from the top meter of the water column in
1983 down to the metalimnion in 2015. This supports the notion that the fish are visually tracking
the downward movement of zooplankton prey.
Stable isotope analyses of lower food web members suggest that quagga mussels and
zooplankton may be competing for resources at a different time of year than previously assumed.
13 15
Quagga mussels and zooplankton exhibited well-differentiated 8 C and 8 N isotopic composition
CSMI Lake Michigan 2015 Report
viii
-------
in May, indicating that they were not feeding on similar algae or other particles during winter or
spring. The reason for isotopic differentiation of mussels and zooplankton during unstratified
conditions is not known, however, the two organismal groups became increasingly isotopically
similar through summer and fall, such that they were not different from each other by September.
These results suggest that quagga mussels are competing with zooplankton for food resources
during stratified lake conditions. Direct competition between quagga mussels and zooplankton
throughout the summer growing season could ultimately reduce zooplankton production in the
offshore food web, thereby reducing the amount of food available for prey fish and sportfish alike.
Pelagic food web and nutrients
Uncertainty in exactly how hypothesized nearshore increases in nutrient concentrations
affect in-lake processes such as water quality, primary productivity, and fish production led to the
development of two 2015 CSMI research priorities. To address these uncertainties, EPA, USGS
and NOAA designed a large-scale collaborative field effort to examine trends in nutrients and biota
(from micro-plankton to fish) along nearshore to offshore gradients in Lake Michigan and
formulated two specific hypotheses:
1) Nearshore-offshore transects located near high phosphorus (P) loading tributaries have
higher productivity levels than those located near low P-loading tributaries.
2) Nearshore areas have higher productivity than offshore areas.
Scientists from USGS and the EPA established eight fixed location sampling transects
based on their proximity to tributaries. Based on previous modeling efforts, the tributaries were
CSMI Lake Michigan 2015 Report
ix
-------
designated: "no" loading—Waukegan (Illinois), Frankfort (Michigan), and Sturgeon Bay
(Wisconsin); "low" loading—Root River/Racine (Wisconsin) and Pere Marquette River/Ludington
(Michigan); or "high" loading—St. Joseph River (Michigan), Kalamazoo River (Michigan), and
Manitowoc River (Wisconsin). The distance from the tributary varied slightly between transects,
and three fixed sampling locations along each transect were designated nearshore (18 m), mid-
depth (45 m), or offshore (> 90 m). Most transects were sampled once each season (spring,
summer, and fall). One transect, off Muskegon, Michigan, relatively close to both Muskegon Lake
and the mouth of the Grand River, was sampled by NOAA scientists almost monthly to capture
localized fine-scale temporal food-web dynamics.
Results of the fixed sampling transect work showed little evidence for a positive
relationship between proximity to high loading tributaries and the shoreline and the spatial
distribution of total phosphorus (TP) and chlorophyll a concentrations in Lake Michigan. Broadly
speaking, TP and chlorophyll a were more variable along the western shore of the lake (versus the
eastern), and in the northern basin (versus the southern). However, at the individual transect level,
nearshore TP was significantly higher at only one shallow (18 m) station off the Muskegon River.
When averaging across months, mean chlorophyll a concentrations were consistently higher for
stations that were near high-loading tributaries, but these differences were not statistically
significant from lower-loading tributaries. Looking at specific sampling months, chlorophyll a
concentrations were higher along transects adjacent to high loading tributaries at shallow stations
during May, and at both shallow and offshore stations (> 90 m) during July.
The spatial and temporal relationships between TP, chlorophyll a, and productivity at
higher trophic levels is even less clear. For example, along the Muskegon transect, nearshore and
offshore zooplankton had distinct differences, with smaller, epilimnetic species found nearshore.
CSMI Lake Michigan 2015 Report
x
-------
Lake-wide, stable isotope results for benthic macroinvertebrates and zooplankton demonstrate a
15 13 15 13
shift from relatively low 8 N and high 8 C values nearshore to higher 8 N and lower 8 C
13
offshore. The 8 C patterns may suggest a greater contribution of nearshore carbon sources (i.e.,
benthic algae) to the food web, higher production in nearshore versus offshore waters, or both. The
S15N patterns are consistent with previous observations regarding nitrogen cycling in Lake
Michigan. Relatively low S15N values of animals collected in the nearshore areas (i.e., 18 m depth)
suggest that animals in these areas are not strongly reliant on anthropogenic inputs of nitrogen.
Fish data suggest that nearshore areas, and potentially the entire Lake Michigan ecosystem,
may be resource-limited. High mortality rates during the larval period have long been believed to
be a bottleneck to fish production. Along the Muskegon transect, decreasing cyclopoid copepod
biomass is coincident with increased abundance of Dreissena veligers, and changing prey types
and availability may have a negative effect on larval fish growth. In offshore areas, densities and
growth rates of larval bloater remained low compared to previous years. Conversely, in nearshore
areas, larval densities of alewife and yellow perch increased in 2015 over previous years. Along
the Muskegon transect, larval alewife growth rates and condition were extremely low as compared
to the early 2000s. Catches of adult pelagic prey fish declined by more than 91% between 2010 and
2015 along the Frankfort and Sturgeon Bay transects (declines by species: alewife 90%, rainbow
smelt 99%, bloater 86%). Adult alewife energetic condition was similar in 2015 to values
measured in the early 2000s when quagga mussel proliferation was underway; however, alewife
production continues to be "squeezed" from both limited prey resources and the well-documented
predation pressure from piscivorous salmon and trout.
CSMI Lake Michigan 2015 Report
xi
-------
Loading and in-lake concentrations of contaminants
Human activities have led to contamination of Lake Michigan water, sediment, and biota.
Pollutants found include certain chemicals used as pesticides on agricultural fields and
polychlorinated biphenyls (PCBs) used in industrial practices. Much of our current knowledge
about the dynamics of pollutants in Lake Michigan comes from the Lake Michigan Mass Balance
(LMMB) study conducted by EPA in the 1990s. The LMMB research team measured common
pollutants and developed models to predict the fate of these pollutants under different management
scenarios. To this day, the study results are the baseline for measuring progress. Scientists from
EPA, USGS, and Indiana University specifically addressed the loading of PCBs and mercury and
estimated in-lake concentrations of atrazine during the 2015 CSMI field year. They also estimated
the loadings of organophosphate esters (OPEs), brominated flame retardants (BFRs), and
dechlorane-related compounds, many of which are chemicals of emerging concern.
Atrazine concentrations measured in Lake Michigan in 2015 were lower than what had
been forecasted by the LMMB study. Water samples collected from EPA Great Lakes National
Program Office (GLNPO) Lake Michigan open water stations in August of 2015, at both the mid-
epilimnion and mid-hypolimnion depths, reflected a mean atrazine concentration of 36 ng/L. This
concentration fell between the LMMB model predictions for 100% reduction of tributary loading
and 100% reduction of total loading scenarios. Literature reviews, plus examination of atrazine use
in the Lake Michigan basin, suggest that a combination of decreased atrazine use on land and
higher than originally estimated degradation rates resulted in lower atrazine concentrations than
those forecasted.
PCB, mercury, and flame retardants were analyzed from water samples collected every
three weeks from April through December in 2015. Five tributaries, namely the Grand,
CSMI Lake Michigan 2015 Report
xii
-------
Kalamazoo, St. Joseph, and Lower Fox rivers, plus the Indiana Harbor and Ship Canal (IHSC),
were selected for sampling because they showed the highest loads of PCBs in the LMMB study.
The geometric mean concentrations of EPCB (sum of 85 congeners) ranged from 1.52 to 22.4
ng/L. The highest concentrations of PCBs were generally found in the Lower Fox River and the
IHSC. The highest BFR concentrations were measured in either the IHSC or the St. Joseph River.
OPEs were the most abundant among the targeted compounds with geometric mean concentrations
ranging from 20 to 54 ng/L, and concentrations were similar across tributaries. BFR concentrations
were about 1 ng/L while dechlorane-related compounds were less than 0.001 ng/L. Mercury loads
from all five tributaries were on the order of 50% to 75% lower in 2015 relative to loads calculated
for 1994-1995. The Lower Fox River remains the highest loading tributary of mercury into Lake
Michigan (43 kg/yr in 2015), with other tributary loadings ranging from 1.7 kg/yr to 12.6 kg/yr.
PCB data from this study were combined with open-lake water, air, and sediment PCB
concentration data from other studies to calculate an updated mass budget for Lake Michigan. The
estimated net transfer of PCBs out of the lake is 1240 ± 531 kg yr-1, which is 46% lower than that
estimated in the 1994-1995 LMMB study. In most lake matrices, PCB concentrations are
decreasing, which drives the decline of all the individual input and output flows. Tributary loads at
the Lower Fox River and the Indiana Harbor and Ship Canal both decreased substantially relative
to 1994-1995 loads. Atmospheric deposition to Lake Michigan has become negligible, while
volatilization from the water surface is still a major route of loss, releasing PCBs from the lake into
the air. Large PCB masses remain in the water column and surface sediments and are likely to
contribute to future efflux of PCBs from the lake to the air.
CSMI Lake Michigan 2015 Report
xiii
-------
List of Tables
GLERL sampling activities during the 2015 Lake Michigan CSMI field season 43
Location, depth, and described substrate of sites sampling in Lake Michigan in 2015 (benthos)... 61
Sites where additional Dreissena was collected for determination of length-weight relationships in
2010 and 2015 65
Relationship between shell length (SL in mm) and tissue ash-free dry weight (AFDW in mg) for D.
polymorpha and D. r. bugensis at various depth intervals in Lake Michigan in 2004, 2008, 2010,
and 2015 66
Mean (± SE) density (no./m ) of Diporeia, Dreissena polymorpha, and Dreissena r. bugensis at
four depth intervals (< 30 m, 31-50 m, 51-90 m, and > 90 m) in each survey year 67
Percentage of measured D. r. bugensis within various size categories at four depth intervals (< 30
m, 31-50 m, 51-90 m, and > 90 m) in 2010 and 2015 68
Mean (± SE) density (no./m ) of major macroinvertebrate taxa at four depth intervals (< 30 m, 31-
50 m, 51-90 m, and > 90 m) at 40 sites in the southern basin of Lake Michigan 69
Mean (± SE) biomass (gAFDW/m2) of Dreissena at < 30 m, 31-50 m, 51-90 m, and > 90 m depth
intervals based on the latest lake-wide surveys in Lake Michigan, Lake Ontario, and Lake Huron
70
Data collection dates for 2015 Lake Michigan CSMI (transect and glider) 108
Sensors on various sampling platforms used for CSMI 108
CSMI Lake Michigan 2015 Report
xiv
-------
List of Figures
Expected loading level for different CSMI sites based on historical models 9
Mean chlorophyll a concentrations estimated at nearshore sampling sites categorized as "no" (i.e.,
Waukegan, Frankfort, Sturgeon Bay), "low" (i.e., Root, Pere Marquette), and "high" (i.e., St.
Joseph, Kalamazoo, Manitowoc) total phosphorus loading sites 10
Mean chlorophyll a concentrations estimated by depth zone for three different seasons in 2015 ... 11
Mean zooplankton biomass (natural-log transformed) estimated by depth zone for three different
seasons in 2015 12
Mean Mysis biomass estimated by depth zone for three different seasons in 2015 13
Mean energetic condition for commonly sized large alewife (135-165 mm) and commonly sized
small round goby (45-95 mm) at two different depth zones during the fall of 2015 near Waukegan
14
Location of sampling sites for 2015 CSMI (transects) 15
Sampling sites along the Muskegon transect in Lake Michigan 22
Daytime PSS long transect results from M10-M110 on May 20, 2015 23
Seasonal zooplankton composition at Ml5 24
Seasonal zooplankton composition at Ml 10 25
Density of predatory cladocerans at Muskegon transect 15, 45, and 110 m sites during 2015 26
Fine scale spatial distribution of zooplankton from diel sampling at night on June 25, 2015 at
M110 27
Daytime PSS long transect results from M10-M110 on June 23, 2015 along with corresponding net
tows 28
Larval densities pre (2001-2002) and post (2010-2014) quagga mussel invasion for alewife,
yellow perch and bloater 29
Larval alewife and bloater growth rates along Muskegon transect pre- and post-quagga mussel
invasion 30
Nearshore/offshore diet contents of alewife, bloater and yellow perch 31
Regression analysis of factors influencing larval alewife daily growth rates from 2001-2002, 2010-
2011, 2013-2015 32
PSS plots of temperature (A) and chlorophyll (B) show an upwelling event on July 13, 2015 that
displaced nearshore larval alewife offshore 33
Attenuation coefficients of nearshore, mid-depth and offshore sites in Lake Michigan 34
Comparing 1983 to 2010, a decline in offshore bloater larvae density, and change in vertical
distribution 35
Density of Mysis at Muskegon transect 45 and 110 m sites during 2015 36
Total seasonal phosphorous concentration along the Muskegon transect in 2015 37
CSMI Lake Michigan 2015 Report
xv
-------
Acoustics transects show the seasonal succession of planktivore overlap with their zooplankton
prey 38
UV Irradiance at Ml 10 on July 25, 2015. UV-B (305 nm) radiation could have a potentially
negative effect on some zooplankton and larval fish in the upper 5 m of the water column 39
Dominant nano and micro-plankton grazers in Lake Michigan 40
Flow cytometric analysis of the bacterial community across the Muskegon transect 41
Abundance-weighted bacterial community composition dynamics 42
Location of sampling sites in the southern region of Lake Michigan in 2015 (benthos) 71
Location of sampling sites in the central region of Lake Michigan in 2015 (benthos) 72
Location of sampling sites in the northern region of Lake Michigan in 2015 (benthos) 73
Density (no. per m ) of Dreissenapolymorpha in Lake Michigan based on lake-wide surveys in
1994/1995, 2000, 2005, 2010, and 2015 74
Density (no. per m ) of Dreissena r. bugensis in Lake Michigan based on lake-wide surveys in
1994/1995, 2000, 2005, 2010, and 2015 75
Long-term trends of total Dreissena in Lake Michigan in 1994/1995, 2000, 2005, 2010, and 2015
76
Ash free dry weight (AFDW, mg) of a standard 15 mm I). r. bugensis at four depth intervals
intervals in Lake Michigan between 2004 and 2015 77
Relationship between ash free dry weight (AFDW) and total wet weight (TWW, whole mussel,
tissue and shell) of D. r. bugenisis at each sampling site in the main basin of Lake Michigan in
2015 78
Density (no. per m ) of Diporeia spp.in Lake Michigan based on lake-wide surveys in 1994/1995,
2000, 2005, 2010, and 2015 79
Density (no. per m ) of total Dreissena at < 30 m, 31-90 m, and > 90 m in Lake Ontario, Lake
Michigan, and Lake Huron 80
Transects sampled by CSMI in 2015 and associated rivers and loading categories, overlain on a
map of TP loading differences among shoreline locations 88
Relationship between surface-water TP measured on the May or July 2015 CSMI cruises vs. 2002-
2008 avg annual TP loading from the adjacent tributary (the three no-tributary transects were
assigned to zero load) 89
Relationship between surface-water planktonic CHLA measured on the May or July 2015 CSMI
cruises vs. 2002-2008 avg annual TP loading from the adjacent tributary (the three no-tributary
transects were assigned to zero load) 90
1982-2015 time-series of upper-water column total phosphorous concentrations (mean ± 1
standard deviation) from a set of long-term monitoring stations (i.e., GLNPO monitoring data)
measured in either spring (April or May) or summer (August or September) 91
CSMI Lake Michigan 2015 Report
xvi
-------
1982 to 2015 time-series of upper-water column chlorophyll a (CHLA) concentration (mean ± 1
standard deviation) from a set of long-term monitoring stations (i.e., GLNPO monitoring data)
measured in either spring (April or May) or summer (August or September) 92
Box plots comparing levels of TP (left panels) and CHLA (right panels) across depths in 1983
(GLNPO monitoring data) relative to 2015 CSMI stations 93
Carbon and nitrogen stable isotope ratios of quagga mussel, zooplankton (Zoop.), and oligochaetes
(Oligo.) at the 18 m, 46 m, and 110 m stations 94
The difference between zooplankton and quagga mussel carbon and nitrogen stable isotope ratios
by month (May, July, September) 95
Computational grid (1km x 1km) of the study area (nutrient model) 99
Model phosphorus results versus observational data (in (J,g/L): Base TP load 100
Model phosphorus results versus observational data (in (J,g/L): 30% reduction of the base TP load
100
Nearshore spatial patterns of TP in the Lake Michigan for two dates in summer (2015) 101
Lake Michigan 2015 CSMI transects with dots indicating vertical CTD casts 109
Clockwise from top left: Photos of Slocum Glider, tow body, CTD Rosette glider, Research vessel
the Lake Explorer II and tow body ready to be deployed 109
Screen shot of Cesium software animation of glider mission track displaying chlorophyll sensor
data 110
Screen shot of ESRI Story Map developed for Lake Superior glider data 110
Dates of glider deployment segments Ill
Cross section of Lake Michigan temperature and chlorophyll-a from 2015 Deployment 2, segment
7 112
Comparison of wind data measured by different platforms 113
Lake Michigan surface temperature on Sept. 26, 2015 and Oct. 10, 2015 114
Lake Michigan Upwelling index summed from 1994 through 2013 using methods developed by
Plattner et al 2006 114
Vertical CTD Cast for the Ludington, MI transect at depths of 110, 46 and 18 meters collected on
Sept. 21, 2015 (before the upwelling event) 115
Tow temperature data from all four transects during September 2015 116
Lake Michigan atrazine concentration prediction modeling (Kreiss, Rygwelski) 118
Sampling locations related to the 2015 CSMI Lake Michigan contaminant study 125
Tributary loading of organophosphate esters (OPEs), non-BDE novel flame retardants (nFRs), and
polybrominated diphenylethers (PBDEs) to Lake Michigan 126
Box and whisker plots of concentrations of individual congeners and EPCBs in 2015 tributary
water samples 127
CSMI Lake Michigan 2015 Report
xvii
-------
Tributary EPCB flows to Lake Michigan for 1994-1995, 2005-2006, and 2015 128
Estimated total PCB mass budget flows (kg/yr) and inventories (kg) for 2010-2015 and
comparison to the 1994-1995 mass balance results based on the MICHTOX modell6 in Lake
Michigan 129
Box and whisker plots of concentrations of organophosphate esters (OPEs) in tributary water
samples 130
Box and whisker plots of concentrations of polybrominated diphenyl ethers (PBDEs), non-BDE
novel flame retardants (nFRs), and dechlorane related compounds (Decs) in tributary water
samples 131
CSMI Lake Michigan 2015 Report
xviii
-------
Report: Describing the Distribution and Productivity of
Biota Along a Nearshore to Offshore Gradient
Authors:
David Bunnell, USGS Great Lakes Science Center
Patricia Armenio, USGS Great Lakes Science Center
David Warner, USGS Great Lakes Science Center
Lauren Eaton, University of Toledo
Drew Eppehimer, University of Arizona
Contact:
David Bunnell
Email: dbunnell@usgs.gov
Phone: 734-214-9324
Address:
Great Lakes Science Center
1451 Green Rd
Ann Arbor, MI 48105
CSMI Lake Michigan 2015 Report
1
-------
Background
The Lake Michigan Lakewide Action and Management Plan (LAMP) proposed adding nutrients
(phosphorus) to its "pollutant of concern" list in 2002, given that excessive nutrients were
causing impairments in nearshore waters. Since that time, scientists have highlighted the
"shunting" of nutrients to the nearshore (Hecky et al. 2004), owing to the ability of invasive
dreissenid mussels to capture some portion of allochthonous phosphorus that enters the lake
through tributaries. These changes are believed to increase productivity in the nearshore,
reflected in increased benthic and pelagic primary production and nuisance Cladophora (Auer et
al. 2010). Whether increases in primary productivity lead to concomitant increases for
secondary (by zooplankton) and tertiary (by fish) production remains largely untested. Hence,
understanding the distribution and abundance of nutrients and biota (e.g., zooplankton, fish)
across a nearshore to offshore gradient was identified as a Cooperative Science and Monitoring
Initiative (CSMI) priority in 2015. Increased understanding of the Lake Michigan nearshore will
also facilitate the development of a Nearshore Strategy by the LAMP, which is called for in the
2012 Great Lakes Water Quality Agreement.
Working collaboratively with Environmental Protection Agency (EPA) and National Oceanic
and Atmospheric Administration (NOAA), United States Geological Survey (USGS) described
the distribution of nutrients and biota across nearshore to offshore transects in 2015 (see
Appendix 1). At each transect, we sampled the food web at three sites with differing bottom
depths: 18 m, 46 m, and 91-110 m. We purposefully chose transects near tributaries of varying
total phosphorus (TP) input (see Figure 1, Dolan and Chapra 2012): three transects that were not
associated with any large tributary where total phosphorus would be loaded (Waukegan IL,
Frankfort MI, Sturgeon Bay WI), three transects adjacent to tributaries presumed to be relatively
low loaders of TP (Pere Marquette MI, Root WI, Muskegon MI), and three transects adjacent to
tributaries presumed to be relatively high loaders of TP (St. Joseph MI, Kalamazoo MI,
Manitowoc WI). USGS estimated chlorophyll concentrations, zooplankton, Mysis, larval fish,
and juvenile and adult fish seasonally (April/May, July, October/November) at eight of these
transects (all but Muskegon).
From this design, we tested two hypotheses that organize our results below. 1) Among nearshore
sites, productivity should be greater at sites near tributaries with higher TP loading. Because we
do not have 2015 estimates of TP loading, we substituted chlorophyll a, the most relevant index
of productivity from the next lower trophic level, when seeking to explain variation in
productivity across nearshore sites. 2) The nearshore should be more productive than the
offshore. We used several different response variables for productivity in these hypotheses,
including chlorophyll a, zooplankton biomass and production, larval fish density and growth (for
alewife only), and prey fish biomass and energetic condition. We hypothesized that enhanced
productivity for lower trophic levels would lead to greater prey resources to support fish
production or fish condition. We also provide results for Mysis diluviana, a large native
zooplankton that is omnivorous but is well known to reach higher densities in deeper waters. We
acknowledge that benthic invertebrates (e.g., quagga mussels) would be another important
response variable to evaluate for these hypotheses, but these data were collected by Buffalo State
University and NOAA and should be integrated in future analyses. We used Pearson correlation
coefficient to determine whether variation in nearshore productivity corresponded with lower
2
CSMI Lake Michigan 2015 Report
-------
trophic level indicators of productivity. We also used Analysis of Variance to determine whether
productivity response variables differed between nearshore and offshore depths. Where
differences were found, we used Tukey's multiple comparison test to determine which means are
different from one another.
Hypothesis 1. Among nearshore sites, productivity should be greater at sites near
tributaries with higher TP loading.
Chlorophyll a: For each sampling event in each season, we averaged the estimated chlorophyll
concentration for the entire water column or euphotic zone, which ever was deeper. The
maximum depth of the euphotic zone was 65 m, so we only averaged chlorophyll over this depth
range for the offshore sites. Water column chlorophyll was predicted based on fluorescence data
calibrated with at least three in situ estimates of chlorophyll throughout the water column. We
categorized the sites into one of three groups (no loading, low TP loading, high TP loading)
based on historical TP loadings from the adjacent tributaries. Chlorophyll concentrations did not
differ among groupings for any of the seasons (spring: F = 0.86, P = 0.48; summer: F = 0.24, P =
0.80; fall: F = 2.20, P = 0.21). The high TP loading category always had the highest mean
chlorophyll, but the variation surrounding the mean of each group was relatively high (Figure 2).
Hence there was no support for the hypothesis that chlorophyll concentrations increased with
expected increases in TP loading.
Zooplankton biomass and production: Zooplankton biomass and production were lower in
spring, compared to summer and fall. Zooplankton production was estimated through
temperature and size-based regressions (Shuter and Ing 1997; Stockwell and Johannsson 1997).
For each season, we correlated zooplankton with coincident chlorophyll a estimates. During
spring, neither zooplankton biomass (r = -0.46, P = 0.26) nor zooplankton production (r = -0.49,
P = 0.21) was positively correlated with chlorophyll concentrations. Similarly, during summer,
neither zooplankton variable (biomass: r = 0.18, P = 0.68; production: r = 0.41, P = 0.31) was
positively correlated with chlorophyll. Finally, chlorophyll was not positively correlated with
zooplankton biomass (r = 0.38, P = 0.36) or production (r = 0.46, P = 0.25) in the fall either.
Hence there was no support for the hypothesis that zooplankton biomass or production increased
with nearshore productivity.
Larval fish densities: The majority of nearshore larvae sampled in late April and early May were
deepwater sculpin (42%), followed by lake whitefish (41%), and round goby (16%). The highest
densities of larvae (all species) were sampled at Frankfort (4.2/100m3) and Sturgeon Bay
(2.5/100m3). Lower densities (
-------
(1.3/100m3). The lowest densities (
-------
Hence there was no support for the hypothesis that fish energetic condition increased with
nearshore productivity.
Hypothesis 2. The nearshore should be more productive than the offshore.
Chlorophyll a: Chlorophyll concentrations did not differ among nearshore, intermediate depth,
and offshore sites in spring (F= 2.75, P = 0.09), summer (F= 1.34, P = 0.28) or fall (F= 1.16, P
= 0.34). The variation among the sites within a depth zone was relatively high, and the mean
chlorophyll was only highest in the nearshore zone in the fall (Figure 3). Hence there was no
support for the hypothesis that chlorophyll concentrations were higher in the nearshore zone than
in the other zones.
Zooplankton biomass and production: Zooplankton biomass (|ig/m3) was lower in spring than in
summer and fall (Figure 4). Zooplankton biomass did not differ among nearshore, intermediate
depth, and offshore sites in spring (F= 1.69, P = 0.21). During summer, however, zooplankton
biomass was significantly higher at 18 m than at 110 m (F = 5.62, P = 0.01); biomass at 46 m
was similar to the other two depth zones. During fall, only two of the offshore sites were
sampled. Hence we limited the statistical comparison to 18 m and 46 m, and found no difference
in zooplankton biomass between these two depth zones (F = 0.55, P = 0.47). Zooplankton
production results revealed the exact same pattern as zooplankton biomass: during summer
zooplankton production average 706.2 |ig/m:Vd in the nearshore, which was significantly higher
than the average of 209.2 |ig/m:Vd in the offshore. Hence there was seasonal support (i.e..
summer only) for the hypothesis that zooplankton biomass or production was higher in the
nearshore than in other zones. It is noteworthy that the summer season can be a critical period
for many larval fishes that hatch in the nearshore zone (e.g.. alewife. yellow perch).
Larval fish densities: Density (#/m3) of all larval fish did not differ among nearshore,
intermediate depth, and offshore sites in spring (F = 0.74, P = 0.49) or summer (F = 0.99, P =
0.39); larvae were not sampled in fall. Because more than 88% of all larval fish sampled in
summer were alewife, and alewife hatch in the nearshore, it was surprising that larval fish
densities were not highest in this zone. Larval alewife were likely advected offshore through
currents and may even navigate into deeper waters when they reach sizes large enough to be
active swimmers. The distance between the nearshore and offshore sites ranged from 7 km
(Frankfort) to 36 km (Kalamazoo), with the average distance across all transects being 19 km.
Hence there was no support for the hypothesis that larval fish density was higher in the nearshore
than in other zones.
Larval alewife growth rates: Larval alewife growth rates in the summer were slower for those
sampled in the nearshore (mean = 0.47 mm/d) compared to those sampled in the offshore (0.52
mm/d; F= 5.12, P = 0.006); those sampled in the intermediate depths were not different from
any other depth (mean = 0.50 mm/d). Hence there was no support for the hypothesis that larval
fish grow faster in the nearshore than in other zones, despite zooplankton biomass (and
production) being higher in the nearshore than the offshore during summer.
Mysis biomass: Mysis biomass is commonly known to increase with depth in the Great Lakes.
Hence, we did not hypothesize that Mysis biomass would be higher in the nearshore than in the
5
CSMI Lake Michigan 2015 Report
-------
offshore in 2015. During spring, summer, and fall, we observed the expected result thatMysis
biomass (mg/m2) was higher in the offshore than in the nearshore (Figure 5, spring: F= 11.18, P
< 0.001); summer: F= 10.54, P < 0.001; fall: F= 7.07, P = 0.007). Fall offshore collections
were limited to two sites with high variability between sites.
Fish biomass: Biomass of fish was never highest in the nearshore. During spring, nearshore fish
biomass was lower than the biomass in the intermediate or offshore depth zones (F = 6.92, P =
0.005). During summer, there was no difference in fish biomass across the three zones (F =
1.28, P = 0.30). During fall, there was no difference between fish biomass at 18 and 46 m depth
zones (F= 0.16, P = 0.70; only 2 offshore sites were sampled). Hence there was no support for
the hypothesis that fish biomass was higher in the nearshore than in other zones.
Fish energetic condition: To compare alewife energetic condition across depth zones, we only
had one sampling occasion with similar-sized fish caught at both 18 m and 46 m depths. In the
fall at Waukegan, alewife energetic condition was 26% greater at 46 m (i.e., 8066 J/g wet
weight, see Figure 6) than at 18 m (i.e., 6403 J/g wet weight), a significant difference (F= 76.8,
P < 0.0001). For a round goby comparison across depths, we also only had one comparative
opportunity. At Waukegan in the fall, round goby energetic condition was not different (F =
3.86, P = 0.052) between those sampled at 46 m (i.e., 4227 J/g wet weight) and those at 18 m
(i.e., 4352 J/g wet weight). Hence there was no support for the hypothesis that fish energetic
condition was higher in the nearshore than in other zones. Round goby, a fish that relies on more
benthic prey resources, was much closer to supporting the hypothesis than the more pelagic-
oriented alewife.
Summary. Contrary to the hypotheses, our nearshore sampling sites were not sensitive to
expected variable inputs of TP from tributaries, nor were they generally "hot spots" of
productivity for chlorophyll, zooplankton, or larval, juvenile, and adult fish. From a whole lake
perspective, the "nearshore" zone of the lake is widely considered to include depths 30 m and
shallower, and this region only represents about 20% of the lake surface area (far less of a
percentage by volume). Hence our 18 m nearshore sites were relatively centrally located within
the nearshore habitat. Nonetheless, it remains possible that if we had sampled in even shallower
waters or more immediately "downstream" of the tributary inputs, we might have detected higher
productivity within an even shallower portion of the nearshore zone. The only support for
nearshore "hot spot" in our study was for nearshore zooplankton biomass and production in the
summer. Given that the eggs of many important fish species are hatching in the nearshore during
summer, this could be advantageous for first-feeding fish larvae. Surprisingly, however, larval
fish densities were just as high in the offshore zones as in the nearshore zones- even for
nearshore spawning fish like alewife. This advection of young larvae to the offshore waters is
likely driven by circulation and upwelling, which contribute to a nearshore habitat that is likely
less stable in water temperatures than in the offshore. Furthermore, larval alewife growth rates
were lower in nearshore waters than in offshore waters.
For the other instances where CSMI sampling revealed differences in biota along a nearshore to
offshore gradient, the nearshore was typically the site where biomass (juvenile and adult fish in
the spring fish, Mysis in all seasons) or energetic condition (large alewife) was lowest. Contrary
6
CSMI Lake Michigan 2015 Report
-------
to our hypothesis, the offshore habitat was generally more productive than the nearshore habitat
across trophic levels.
Temporal comparisons to other studies. Two of the nearshore to offshore transects sampled in
2015 were the same transects (Frankfort, Sturgeon Bay) that were sampled monthly in 2010.
When we compared different trophic levels between the two years at these sites we found: total
zooplankton biomass did not differ between years (although cladocerans such as Daphnia
galeata mendotae and Bythotrephes declined in 2015), My sis biomass was higher in 2015 than
2010, and pelagic prey fish biomass declined by more than 91% in 2015 relative to 2010 (90%
decline for alewife, 99% decline for rainbow smelt, and 86% decline for bloater).
The 2015 intensive effort also afforded an opportunity to measure key variables for alewife, a
prey fish species of high management concern given its importance to sustaining the
recreationally important Chinook salmon fishery and the recent alewife collapse in adjoining
Lake Huron. Larval alewife growth rates in 2015 were 43% and 38% slower than the rates
estimated from comparable studies in 2001-2002 and 2004-2006, respectively (Hook et al. 2007;
Weber et al. 2015). This slower growth rate, coupled with our finding that 67% of larval alewife
stomachs in 2015 were devoid of prey suggests that the larval period could be a bottleneck to
production of new alewife. At the same time, we determined that the energetic condition of
juvenile and adult alewife in 2015 (that have survived a potential larval bottleneck) were in no
worse condition than a study that measured alewife energetic condition in 2002-2004 (Madenjian
et al. 2006). This result suggests that larger alewife appear to be doing just as well exploiting
prey resources in 2015 as in the early 2000s when quagga mussel proliferation was underway.
At the same time, alewife energetic condition in 2002-2004 and in 2015 was still about 25%
lower than was measured in 1979-1981 when the lake was more productive (and when alewife
densities were considerable higher). These data suggest that alewife production continues to be
"squeezed" from both limiting prey resources, as well as the well documented predation pressure
from piscivorous salmon and trout.
Implications for future 2020 CSMI: Should describing the productivity of the nearshore remain
a goal in future CSMI efforts, 1) additional sampling efforts could be directed in this dynamic
habitat, 2) hydrological models could be developed to place the data collected from discrete
sampling events in greater context, and 3) remote sensing tools (e.g., profiling buoys/drifters,
gliders, satellites) that were tangentially applied in 2015 could be more carefully considered for
their role in 2020 to try and describe this dynamic environment between sampling events by
research boats. Additional research into a potential "bottleneck" for larval fish survival may also
continue to provide insights into how the changing lower trophic levels are influencing prey fish
densities.
References:
Auer, M.T., L.M. Tomlinson, S.N. Higgins, S.Y. Malkin, E.T. Howell, andH.A. Bootsma. 2010.
Great Lakes Cladophora in the 21st century: same algae- different ecosystem. Journal of
Great Lakes Research 36: 248-255.
Dolan, D.M., and S.C. Chapra. 2012. Great Lakes total phosphorus revisited: 1. Loading
analysis and update (1994-2008). Journal of Great Lakes Research 38: 730-740.
7
CSMI Lake Michigan 2015 Report
-------
Hecky R.E., R.E.H. Smith, D.R. Barton, S.J. Guildford, W.D. Taylor, M.N. Charlton, and T.
Howell. 2004. The nearshore phosphorus shunt: A consequence of ecosystem
engineering by dreissenids in the Laurentian Great Lakes. Canadian Journal of Fisheries
and Aquatic Sciences 61: 1285-1293.
Hook, T.O., E.S. Rutherford, D.M. Mason, and G.S. Carter. 2007. Hatch dates, growth,
survival, and overwinter mortality of age-0 alewives in Lake Michigan: implications for
habitat-specific recruitment success. Transactions of the American Fisheries Society
136: 1298-1312.
Madenjian, C. P., S.A. Pothoven, J.M. Dettmers, and J.D. Holuszko. 2006. Changes in seasonal
energy dynamics of alewife (Alosapseudoharengus) in Lake Michigan after invasion of
dreissenid mussels. Canadian Journal of Fisheries and Aquatic Sciences 63: 891-902.
Shuter, B.J., and K.K. Ing. 1997. Factors affecting the production of zooplankton in lakes.
Canadian Journal of Fisheries and Aquatic Sciences 54: 359-377.
Stockwell, J.D., and O.E. Johannsson. 1997. Temperature-dependent allometric models to
estimate zooplankton production in temperature freshwater lakes. Canadian Journal of
Fisheries and Aquatic Sciences 54: 2350-2360.
Weber, M.J., B. C. Ruebush, S. M. Creque, R. A. Redman, S. J. Czesny, D. H. Wahl, and J. M.
Dettmers. 2015. Early life history of alewife Alosa pseudoharengus in southwestern
Lake Michigan. Journal of Great Lakes Research 41: 436-447.
CSMI Lake Michigan 2015 Report
8
-------
U)
c
T3
ns
o
n 1.0 -
W
3
a
CO
O
0.8 -
0.6 -
a.
re
¦4-"
o
in 0.4 -
ro
-------
Summer
3.0
2.5
2.0
1.5
1.0 J
No
Low
High
Fall -1
<
~ *
I J
No Low High
Loading category of sites
Figure 2. Mean chlorophyll a concentrations estimated at nearshore sampling sites categorized
as "no" (i.e., Waukegan, Frankfort, Sturgeon Bay), "low" (i.e., Root, Pere Marquette), and
"high" (i.e., St. Joseph, Kalamazoo, Manitowoc) total phosphorus loading sites. Mean
chlorophyll did not differ across loading category for any season.
CSMI Lake Michigan 2015 Report
10
-------
18 m
18 m
46 m
110 m
Summer
_o go
46 m
110 m
o o)
18 m 46 m
Depth zone
110 m
Figure 3. Mean chlorophyll a concentrations estimated by depth zone for three different seasons
in 2015. Mean chlorophyll did not differ across depth zones for any season. Fall 110 m was
only at the Frankfort and Sturgeon Bay sites (other offshore sites were unable to be sampled).
11
CSMI Lake Michigan 2015 Report
-------
10.25
10.00
8.75 -
£ LU
05
TO +I
Ffe
O LU
O)
ro +l
o~E
N
cf ^
10.25
10.00
9.75
9.50
9.25
9.00
8.75
8.50
10.25
10.00
9.75
9.50
9.25
9.00
8.75
8.50
18 m
46 m
110 m
Summer j
I
$
18 m
46 m
110 m
Fall
o
{
I
18 m 46 m
Depth zone
110 m
Figure 4. Mean zooplankton biomass (natural-log transformed) estimated by depth zone for
three different seasons in 2015. Zooplankton biomass was higher at 18 than at 110 m only in
summer. Fall 110 m was only at the Frankfort and Sturgeon Bay sites (other offshore sites were
unable to be sampled).
12
CSMI Lake Michigan 2015 Report
-------
.52 +1
18 m
46 m
110 m
LU
W
+l
$S>CN
O)
500
400
300
200
100
0
Summer
<
~
i
~
18 m
46 m
110 m
Cf. 300
JO CM
^ E 200
18 m
46 m
110 m
Depth zone
Figure 5. Mean Mysis biomass estimated by depth zone for three different seasons in 2015
Mysis biomass was higher at 110 m than at 18 m in all seasons. Fall 110 m was only at the
Frankfort and Sturgeon Bay sites (other offshore sites were unable to be sampled).
13
CSMI Lake Michigan 2015 Report
-------
CO
c
CD
"O
LLI
•£
0)
<
8500
8000
o)W 7500
o3 +1
7000 -
O)
6500
6000
18 m
46 m
18m 46 m
Depth zone
Figure 6. Mean energetic condition for commonly sized large alewife (135-165 mm, top panel)
and commonly sized small round goby (45 - 95 mm, bottom panel) at two different depth zones
during the fall of 2015 near Waukegan. Energetic condition was higher at 46 m than at 18 m for
alewife, but there was no difference in energetic condition between depths for round goby.
CSMI Lake Michigan 2015 Report
14
-------
Manistee
Pere Marquette
ao Kilometers
• NOAA
¦ USGS/EPA
Appendix 1. Location of sampling sites for 2015 CSMI. At each nearshore to offshore transect,
the food web was sampled at three depths: 18, 46, and 91-110 m.
15
CSMI Lake Michigan 2015 Report
-------
Report: Temporal and Spatial Coupling of Nutrients and
Food Web—Microbes to Fish
Authors:
Henry Vanderploeg, NOAA Great Lakes Environmental Research Laboratory
Edward Rutherford, NOAA Great Lakes Environmental Research Laboratory
Steven Pothoven, NOAA Great Lakes Environmental Research Laboratory
Joann Cavaletto, NOAA Great Lakes Environmental Research Laboratory
James Liebig, NOAA Great Lakes Environmental Research Laboratory
Doran Mason, NOAA Great Lakes Environmental Research Laboratory
Hunter Carrick, Central Michigan University
Vincent Denef, University of Michigan
Paul Glyshaw, Cooperative Institute for Great Lakes Research
David Wells, Cooperative Institute for Great Lakes Research
Contact:
Henry Vanderploeg
Email: henry.vanderploeg@noaa.gov
Phone: 734-741-2284 / 2292
Address:
NOAA GLERL
4840 S. State Rd.
Ann Arbor MI 48108
CSMI Lake Michigan 2015 Report
16
-------
Brief Project Description
In support of EPA/USGS efforts to sample food web components at multiple transects around the
periphery of Lake Michigan during two or three seasons, we conducted intensive temporal
(including diel sampling) and fine-scale spatial sampling across seasons (April-October) in the
Muskegon/Grand River Region. Moreover, we measured the microbial food web (MFW) as part
of our study to describe the often ignored component of the food web. Our approach was similar
to Year of Lake Michigan 2010 and Lake Huron 2012, where we examined, in detail, the
seasonal changes in diel spatial coupling (April, July, and September) of the food web using a
variety of advanced tools including plankton survey system (PSS) with Laser Optical Plankton
Counter (LOPC) and fisheries acoustics. Further, a variety of collection methods to quantify
larval fish and Mysis abundance, size structure, spatial distribution, and critical rates (growth,
mortality), were utilized. Also, we measured depth profiles of UV radiation to examine the
relationship between UV radiation penetration vs. plankton and larval fish vertical distribution.
There is little previous knowledge of spatial coupling during the time of early stratification,
which is critical for deep chlorophyll layer formation as well as zooplankton and larval fish
production. Therefore, we examined spatial coupling of the food web monthly April - October to
understand nutrient flow and trophic dynamics from inshore to offshore. Our transect location
was particularly important since it is near the Muskegon and Grand Rivers, the largest combined
source of nutrient loading going directly into Lake Michigan (Figure 1).
Sampling (Day and Night)
Table 1 shows the sampling scheme for major variables collected in the Muskegon region. In
summary, the following major sampling activities took place:
• Nearshore-offshore PSS/acoustics transects were completed and processed (April to
October).
• Full column net zooplankton tows (64-|im and 153-|im mesh), at the end of all transects (15-
m and 110-m depth stations), were completed and processed.
• Depth-stratified vertical opening-closing zooplankton net tows (to determine species-specific
vertical depth distributions) were made at M45 or Ml 10 or at both stations during both day
and night in all months with the exception of August and October due to severe weather
conditions. PSS transects that gave fine-scale vertical distribution of chlorophyll,
photosynthetically active radiation (PAR), temperature and occasional measures of UV
radiation complimented the net tow data.
• Abundance of pico- to microplankton were measured, and primary production and microbial
grazing was estimated for most months (Table 1).
• Nutrients including total P, dissolved P, and particulate C, N, and P were measured in all
months.
17
CSMI Lake Michigan 2015 Report
-------
Findings
Mesozooplankton
• The PSS tows coupled with back-up information from net tows showed that chlorophyll
levels and zooplankton were highest in the shallow epilimnion during both day and night
during early stratification (Figure 2).
• We found distinct differences in nearshore and offshore zooplankton, with smaller
epilimnetic species found nearshore. (Figure 3, Figure 4, Figure 5).
• We were able to specify both fine-scale horizontal and vertical distribution of some
important species by combining net tow observations with LOPC size category data. (Figure
6). We are continuing to make progress in developing a method to define fine-scale spatial
structure of zooplankton using the coarse vertical distribution from nets and fine-scale spatial
data from the LOPC.
• Net tows (Figs. 3 and 4) and LOPC data revealed that during June Dreissena veligers were
the dominant zooplankton (both number and biomass) from nearshore out to the 80-m depth
contour, i.e., out to 14 km offshore. (Figure 7)
• Most of the summer-fall zooplankton biomass was found at night in the metalimnion and not,
as expected, in the epilimnion. The exception was during October, when Daphnia galeata
dominated the epilimnion in the offshore plankton.
Larval fish
• In the nearshore zone, density of alewife and yellow perch larvae in 2015 (Figure 8) was
higher than in previous years (2010-11, 2013-14) but alewife growth and condition were
extremely low, nearly half the rates as seen in 2001-2002 before quagga mussels irrupted
(Figure 9). Diet analysis indicated alewife and yellow perch larvae consumed high
proportions of mussel veligers which were extremely abundant in the epilimnion in June and
July (Figure 10). Regression analysis of factors influencing larval alewife daily growth rates
from 2001-2002, 2010-2011, 2013-2015 indicated cyclopoid copepod biomass, and not mean
June-July temperature, positively affected alewife growth rate (Figure 11). These results
suggest that the decline in cyclopoid copepod biomass, caused by reduced productivity in the
nearshore zone and increased abundance of dreissena veligers, may have a negative effect on
larvae growth. Given that larval consumption of veligers is increasingly common in studies
of larval diets in the Great Lakes (Marrin Jarrin et al. 2014, Withers et al. 2015), and if larval
growth rate is correlated with probability of their survival to the juvenile stage, then
abundance of veligers in larvae diets may indicate how Dreissena mussels could disrupt the
lower food web and effect fish recruitment. Currently, we are estimating alewife larvae
survival for 2015 and comparing it to rates in 2010 and 2001-2002 (pre-quagga irruption) to
see if there is a relationship between larval growth and survival, and size of juvenile
alewives.
18
CSMI Lake Michigan 2015 Report
-------
• In the offshore, bloater larvae density and growth rate in 2015 remained low compared to
prior years (Figure 9). Bloater larvae consumed mainly small copepods and copepod eggs.
• We also are examining effects of an upwelling event in July 2015 on displacement of alewife
larvae from nearshore to offshore (Figure 12). Preliminary results indicate the upwelling has
had no effect on alewife survival or growth.
• Finally, we are comparing species composition and vertical distributions of fish larvae in
2015 with compositions and distributions of zooplankton and fish larvae measured during
1983 off Grand Haven by Richard Nash and Audrey Geffen. This comparison indicates a
shift in vertical distribution of larvae from the top meter of the water column in 1983 to the
metalimnion in 2015, a likely response to quagga mussel filtration increasing light levels
(Figure 13), and foraging efficiency by the visually predaceous cladoceran Bythotrephes, that
have resulted in deeper vertical distributions of zooplankton (Figure 14).
Mysis
• Mysis were collected at 45 and 110 m sites April-November; offshore abundance was
higher than at the 45m transitional site (Figure 15). Abundance of Mysis in 2015 was
relatively high compared to previous 3 years.
Nutrients
• C:N:P ratios of seston suggest only mild to moderate P limitation at most stations and
depths along the transect. Some even show no P-limitation.
• Phosphorous was highest nearshore in the spring, reaching a maximum of 10.27 jagL"1 in
late June, and then decreased to levels comparable to offshore beginning in July.
Offshore phosphorous was lowest in late April, and increased to a maximum of 6.03 |igL~
1 in late September. (Figure 16)
Fishery acoustics
• Preliminary examination of acoustic target strength (mostly small planktivorous fishes)
showed not much overlap between fish and zooplankton until late summer and fall (Figure
17).
UV radiation and PAR
• UV depth profiles showed that UV-B (305 nm) radiation could have a potentially negative
effect on some zooplankton and larval fish in the upper 5 m of the water column. There is no
useful information as to effect of UV-B radiation on Great Lake zooplankton or fish larvae.
Longer wavelengths of UV radiation and PAR can penetrate to greater depths than UV-B,
and PAR levels remain relatively high to our near bottom sampling depths (Figure 18). PAR
profiles indicate that Bythotrephes could visually detect prey down to depths of 40 m or
greater (10 |imol m'V) during midday. Most planktivorous fishes should be able to visually
19
CSMI Lake Michigan 2015 Report
-------
detect prey throughout the water column during midday as PAR levels remain above
threshold limits of 0.2 |imol m'V1 to depths greater than 85 m.
Microbial food web (Carrick)
• Lake Michigan has experienced recent changes in the phytoplankton assemblage coinciding
with reductions in watershed nutrient loadings and the introduction of invasive species. As
such, we evaluated the population dynamics of key plankton components in Lake Michigan
along a series of near to offshore transects in southern Lake Michigan (2013-
16). Chlorophyll analysis revealed that the picoplankton fraction (Ppico, < 2 (j,m) contributed
an average of >50% to total phytoplankton biomass. Particulate P made up nearly all of the
TP in the water column; this pool was composed of poly-P in the pico-sized particles (>80%
of total) and bioassay experiments revealed that phytoplankton growth was limited primarily
by P. The abundance of Ppico (5,200 to >70,000 cells/mL) was considerable and the
assemblage was dominated by cyanobacteria taxa and pico-eukaryotes. The occurrence of
diatoms (mainly Cyclotella and Discotella taxa) was limited to the nearshore region during
the spring and early stratification periods. We estimated growth and grazing losses
attributable to small grazers (microzooplankton, protists; Fig. 19) and large grazers
(mesozooplankton, crustaceans) from enclosure experiments. Ppico had lower growth (0.19
+/- 0.27) relative to grazing losses by microzooplankton (-0.33 +/- 0.37), indicating tight
coupling with small grazers. These results suggest that carbon (and phosphorus) flow from
Ppico to metazoa may dominate the current, trophic dynamics in the lake.
Microbial food web (omics, Denef)
• Bacterial abundance. Though bacteria play a fundamental role in freshwater
biogeochemical cycling and community dynamics, they are often left out of Great Lakes
studies. We monitored total bacterial numbers and the fraction of cells in the so-called High-
nucleic acid content fraction (HNA), as defined by flow cytometric analysis. The HNA
fraction is thought to be representative of more active bacterial cells. Cell numbers declined
from the estuary to offshore locations, increased during the year in the estuary, while being
more variable over time in Lake Michigan itself, with a marked minimum in July (Figure
20A). The fraction of all cells that were deduced to be more active (HNA fraction) showed
an opposite trend to total cell abundance in the estuary, while it more closely tracked cell
abundance in Lake Michigan (Figure. 20B).
• Bacterial community composition. Typically, bacterial community composition is censused
via high throughput sequencing of the 16S rRNA gene, providing insights into the diversity
of all potentially active and dormant bacteria at the DNA level. However, the influence of
bacteria on ecosystem function is most likely predominated by active organisms. One proxy
for activity is the use of RNA to assess community composition rather than DNA. During the
CSMI 2015 work, we high throughput sequenced the 16S rRNA gene using simultaneously
20
CSMI Lake Michigan 2015 Report
-------
extracted DNA and RNA to evaluate the DNA versus the RNA community in samples
collected monthly in 2015 from April until October along a eutrophic (Muskegon Lake) to
oligotrophic (offshore regions) transect in Lake Michigan. In addition, we monitored DNA-
level community composition of the sediment at each station. Richness and evenness of both
the water and sediment bacterial communities consistently decreased from the estuary to the
offshore station. Across seasons, diversity tended to increase from spring to fall. The
sediment communities significantly differed from water column communities (Figure 21 A),
and strongly clustered by station with limited variability by month, except at the 45 m depth
station (Figure 2 IB). Location along the transect explained the largest fraction of the
observed variance is water column bacterial species presence/absence (22%), whereas season
and whether DNA and RNA samples were used affected communities in similar ways (13
and 11% variance explained, respectively). When relative abundance of each species was
taken into consideration, DNA vs RNA explained 30% of the variance, while station (14%)
and season (9%) explained lower fractions of the observed variance in community
composition (Figure 21 A, C, D). Several microbial phyla differed between DNA and RNA
samples and included five of the most abundant groups. Verrucomicrobia, Alpha-, and Beta-
proteobacteria were overrepresented in the RNA pool while Actinobacteria and
Bacteriodetes were overrepresented in in the DNA pool. These taxonomic differences
between in the potentially active and total bacterial communities indicate that there may be
multiple layers to the microbial communities that underpin lake community and ecosystem
processes.
• Dreissenid mussel feeding impacts on bacteria. In a companion study [not GLRI funded]
we showed that Dreissena grazed more heavily on the species that appear to be more
metabolically active based on the RNA and DNA-based analyses during the CSMI work,
which has potential implications for nutrient cycling {Denef et al., 2017; Props et al, 2018}.
• References.
Props, R., M.L. Schmidt, J. Heyse, H.A. Vanderploeg, N. Boon, and V.J. Denef. (2018) Flow
cytometric monitoring of bacterioplankton phenotypic diversity predicts high population-
specific feeding rates by invasive dreissenid mussels. Environ Microbiol 20 (2), 521-534.
Denef V.J., CarrickH.J., Cavaletto J., Chiang E., Johengen T.H., Vanderploeg H.A. 2017.
Lake bacterial assemblage composition is sensitive to biological disturbance caused by an
invasive filter feeder. mSphere 2 (3), e00189-17.
CSMI Lake Michigan 2015 Report
21
-------
B7°30'
S7°00'
90
85°30' • Station locations
M45
» Muakegon
M110 M15
43°001
Waukegan •
Chicago •
Study
Site
90
SO
30
42°10'
St. Joseph
^3U
• Michigan Qity
ft"
!
42o00"
0 20 40 60 km
1 I I I I l_l
Figure 1: Sampling sites along the Muskegon transect in Lake Michigan.
CSMI Lake Michigan 2015 Report
22
-------
v \5
cl -60
86.55
86.50
86.45
86.40
86.35
E
-40
£
- (
Q.
O)
Q
-60
r\ /
f) \
-80
F J
: ^
100
-oS-
^ —
86.55
86.50
86.45
86.40
86.35
9- -60
¦
r 30
125 g
-20 |
TO
-15
E
-10 H
L.
¦
a3
- 1 o
¦
L o
¦
6.0
I
-5.0_
1
-4.o|
CO
-3.0E,
en
-2.0ro
E
o
1.0cq
.
¦
-0.0
86.55 86.50 86.45
West Longitude
86.40
86.35
Figure 2: Daytime PSS long transect results from M10-M110 on May 20, 2015 show that
chlorophyll and zooplankton were highest in the epilimnion and nearshore during early
stratification.
23
CSMI Lake Michigan 2015 Report
-------
70
Dreissena veligers
¦ Other Cladocerans
Bosmina
¦ Other Calanoid copepods
¦ Diaptomid copepods
¦ Cyclopod copepods
¦ nauplii
Apr. May June July Aug. Sept. Oct.
Figure 3: Seasonal zooplankton composition at M15.
24
CSMI Lake Michigan 2015 Report
-------
70
60
50
V* dO
Dreissena veligers
Daphnia mendotae
Bosmina
¦ Other Calanoid copepods
Diaptomid copepods
¦ Cyclopod copepods
¦ nauplii
Figure 4: Seasonal zooplankton composition at Ml 10.
25
CSMI Lake Michigan 2015 Report
E
&o
E
— 30
CD
E
.2 20
00
10
_
¦ ¦
n
i
Apr. May June July Aug. Sept. Oct.
-------
7000 n
6000
5000 -
4000 -
3000
2000
1000 H
0
Nearshore (15 m)
~ Bythotrephes
~ Cercopagis
~ Leptodora
~ Polyphemus
83 96 104 135 152 166 182 202 217 230 244 264 278 295 307 341
7000 -|
6000 -
5000
4000 -
3000 -
2000 -
1000
0
Transitional (45 m) _
¦=¦ n
H_CL
83 96 104 135 152 166 182 202 217 230 244 264 278 295 307 341
7000
6000 -
5000 -
4000 -
3000
2000 -
1000 -
Offshore (110 m)
o
H H n
n
83 96 104 135 152 166 182 202 217 230 244 264 278 295 307 341
Day of year
Figure 5: Density of predatory cladocerans at Muskegon transect 15, 45, and 110 m sites during
2015. High density of Polyphemus was very unusual.
CSMI Lake Michigan 2015 Report
26
-------
0.2
0.4
Chlorophyll a (|ig/L)
Tempterature (10° C)
0.6 0.8 1
1.2
1.4
1.6
200
400 600 800
Zooplankton biomass (mgL1)
1000
1200
Chlorophyll —•—Temperature —•—Bin 1 - * - Bin 2 —©— Bin 3 Bin 4 -h - Bin 5
Figure 6: Fine scale spatial distribution of zooplankton from diel sampling at night on June 25,
2015 at Ml 10. Bin 1=91-255 |im, Bin 2=256-495 |im, Bin 3=496-750 |im, Bin 4=751-1500 |im,
Bin 5=1501-4005 |im. When the lake is stratified organisms found in each bin typically include:
Bin 1= nauplii, Bosmina, Dreissena veligers; Bin2= copepodites, small copepods, Bosmina\ Bin
3= Daphnia, large copepods; Bin 4= large Daphnia, Cercopagis, Limnocalanus; Bin 5=
Bythotrephes, Leptodora, Mysis. We are currently working to assign species-specific biomass by
combining this data with depth stratified zooplankton net tows.
CSMI Lake Michigan 2015 Report
27
-------
86.55
86.50
86.45
86.40
86.50 86.45 86.40
J I ¦ I < I I I I I I L.
86.50 86.45 86.40
West Longitude
Ql -60
Q)
Q
-80
0
-20
gT -40
If
Q- -60
0)
Q
-80
-100
86.55
86.35
86.35
86.35
¦
30
-25 q
o
-20 jj
03
-15 g.
E
in
-10 h-
i—
0)
Q
5 3?
5
¦
- 0
¦
- 6.0
-5.0_
4.0^
O)
-3-0E,
CO
-2.0ra
E
o
1.0CD
¦
-0.0
86.55
Figure 7: Daytime PSS long transect results from M10-M110 on June 23, 2015 along with
corresponding net tows show that Dreissena veligers were the dominant zooplankton (both
number and biomass) from nearshore out to the 80-m depth contour, i.e., out to 14 km offshore.
28
CSMI Lake Michigan 2015 Report
-------
300
O 200
Pre quagga
Post quagga
2015
Alewife
Bloater
Yperch
Figure 8: Larval densities pre (2001-2002) and post (2010-2014) quagga mussel invasion for
alewife, yellow perch and bloater.
CSMI Lake Michigan 2015 Report
29
-------
K 0.4
O 0.2
Pre quagga
(2001, 2002)
Post quagga
(2010-2014)
2015
Alewife
Bloater
Figure 9: Larval alewife and bloater growth rates along Muskegon transect pre and post quagga
mussel invasion. Yellow perch data not available.
CSMI Lake Michigan 2015 Report
30
-------
Alewife
Bloater
Yellow Perch
0% 2% 0% 2%
5%
19%
0%
2%.
0%
6%
63%
9%
, 2%
¦ Egg
¦ Calanoid Copepod
¦ Cyclopoid
¦ Veliger
¦ Bosmina
¦ Unid. Copepod
¦ Daphnia
¦ Unid.
nearshore
offshore
nearshore
Figure 10: Nearshore/offshore diet contents of alewife, bloater and yellow perch.
CSMI Lake Michigan 2015 Report
31
-------
0.9
y = 0.0432x + 0.4459
R2 = 0.74, n=7, p< 0.02
ro
0.2
0.1
0 H 1 1 1 1 1
0 2 4 6 8 10
Cyclopoid Biomass (mg/L)
Figure 11: Regression analysis of factors influencing larval alewife daily growth rates from
2001-2002, 2010-2011, 2013-2015 indicated cyclopoid copepod biomass, and not mean June-
July temperature, positively affected alewife growth rate.
CSMI Lake Michigan 2015 Report
32
-------
25 0 -20
Lake Michigan, July 13 2015, Night, 01:09-04:24
Zooplanktori Biomass Total [mg wet/L]
Lake Michigan, July 13 2015, Night, 01:09-04:24
Acoustic Biomass [log re g'mA3]
88.55
08.5-0
88.45
West Longitude
80.43
86.35
86.55
8S.50
B6.-45
'/test Longitude
80.4C
83.35
Figure 12: PSS plots of temperature (A) and chlorophyll (B) show an upwelling event on July
13, 2015 that displaced nearshore larval alewife offshore. Also shown are potential prey (C) and
predator (D) fields for larval fish during the upwelling event.
CSMI Lake Michigan 2015 Report
33
-------
0.7
0.6
0.5
Eoa
cc
<
2 0.3
T3
0.2
0.1
0
1983
2013-2016
15 m
50/45 m 100/110 m
Figure 13: Attenuation coefficients of nearshore, mid-depth and offshore sites in Lake
Michigan. 1983 data is from a transect off Grand Haven with depths of 15 m, 50 m, and 100 m;
2013-2016 data is from our 25 m, 45 m, and 90 m sites off Muskegon.
CSMI Lake Michigan 2015 Report
34
-------
100,000
"E 10,000
o
o
o
tt
>
+¦»
"(75
c
CL)
Q
CL)
(0
>
fO
1,000
100
10
Neuston
Epi
Meta
1983
2010
Hypo
Figure 14: Comparing 1983 to 2010, a decline in offshore bloater larvae density, and change in
vertical distribution.
CSMI Lake Michigan 2015 Report
35
-------
200 n
150 -
100
50 -
CD
_Q
E
3
Transitional (45 m)
JUL
104 146 152 202 230 244 278 307 341
£ 200
CD
Q
150 -
100
50 -
Offshore (110 m)
104 146 152 202 230 244 278 307 341
Day of year
Figure 15: Density otMysis at Muskegon transect 45 and 110 m sites during 2015.
CSMI Lake Michigan 2015 Report
36
-------
2015 LTR Transect Stations Total P
12
10
8
4
2
0
*
J FMAMJ J ASO ND
1 1 1 1 1 1 1 1 1 1 1 r
0 30 60 90 120 150 180 210 240 270 300 330 360
Julian Day
M15 Epi
MHO Epi
M110 DCL
M110 NBot
Figure 16: Total seasonal phosphorous concentration along the Muskegon transect in 2015.
CSMI Lake Michigan 2015 Report
37
-------
Lake Michigan. August 12 2015, Night. 00:19-03:24
Zoo plankton Biomass Total [mg wet-'L]
Lake Michigan. August 12 2015, Night. 00:19-03:24
Acoustic Biomass [log re g/mA3]
•DC -
= <: c_
86.40
80.35
86.55
86.50
80.35
Lake Michigan, September 23 2015. Night, 23:50-03:02
Zooplankton Biomass Total [mg wet'L]
Lake Michigan, September 23 2015, Night, 23:50-03:02
Acoustic Biomass [log re g/mA3]
100 - ¦
86.55
86.50
86.45
36.35
Lake Michigan, October 14 2015. Night, 20:35-23:25
Zooplankton Biomass Total [mg wet'L]
pv
86.55
86.50
~r
86.45
West Longitude
86.40
86.35
Lake Michigan, October 14 2015. Night, 20:35-23:25
Acoustic Biomass [log re g.'mA3]
ioc -
86.55
86.50
80.45
West Longitude
86.4C
86.35
Figure 17: Acoustics transects show the seasonal succession of planktivore overlap with their
zooplankton prey. Note, that the overlap with prey becomes greater as we progress through the
seasons from mid-summer to late fall. The degree of predator and prey overlap appears to occur
with the progression and strength of the thermocline where in July it's the weakest and in
September it's the greatest. Note the lessening of the overlap in October as the thermocline
begins to degrade.
CSMI Lake Michigan 2015 Report
38
-------
0.01%
0.10%
Percent of Surface Irradiance
1.00%
10.00%
100.00%
~ 305 nm (UV-B)
A 380 nm (UV-A)
PAR
CDOM (ng/L)
Chlorophyll a (ng/L)
0.2
0.4
0.6
pg/L
0.8
1.2
305 nm
380 nm
PAR
1% Attenuation Depth (m)
4.5
30.8
52.2
Kh (/m)
0.94
0.13
0.08
Surface Irradiance
2.32
(jW/(cm2nm)
43.86
|jW/(cm2nm)
1100
|jE/(m2sec)
Figure 18: UV Irradiance at Ml 10 on July 25, 2015. UV-B (305 nm) radiation could have a
potentially negative effect on some zooplankton and larval fish in the upper 5 m of the water
column. UV-A (380 nm) and PAR levels penetrate much further and remain relatively high
throughout the water column allowing visual predators such as Bythotrephes and planktivorous
fishes to detect prey at greater depths.
CSMI Lake Michigan 2015 Report
39
-------
Key Nano and Micro-plankton Grazers
o.Jjp
Ochtomooos tp. 1
Pa/t>b6do ip.l fth&t)
Ufotrietio spJ
, «>
hatoblpphai nuolt*.'
Peltrg
Figure 19: Dominant Nano and Micro-plankton grazers in Lake Michigan.
CSMI Lake Michigan 2015 Report
40
-------
10000
7500
5000
2500
6000
5000
4000
3000
2000
§_ 1000
ffl 2000
o
1500
1000
4000
3000
2000
1000
Whole samples
B
Whole samples
_l_
Cflfir
50-i
45-
2
40-
CD
35-
30-
25-
50-
2
40-
at
£ 30-
8
©
a
NA
o
x 35-
cn
30-
25-
20-
50-
40-
o
30-
20-
2
r~
m
Ul
c^eW
month
month
Figure 20: Flow cytometric analysis of the bacterial community across the transect. (A) total cell
numbers per microliter of water sampled over time, and (B) the fraction of all cells that were in
the High-nucleic acid content fraction (HNA). MLB = Muskegon Lake estuary, GVSU buoy
location; M15, M45, Ml 10 = Lake Michigan 15 m, 45 m, and 110 m depth stations.
CSMI Lake Michigan 2015 Report
41
-------
A Lake Michigan Water & Sediment Samples
0.4-
0.2
x 0.0
<
-0.2-
0.2
fraction
® Particle
® Free ~
• Whole §
® Sediment co o.O
CM
lakesite
• MLB
* M15
¦ M45
-t-110
-0.50 -0.25 0.00 0.25
Axis.1 [21.6%]
c Lake Michigan Water Samples Only
0.4
0.2
LO
— 0.0
CM
c/i
-0.2-
-0.4-
cg
w
X
<
-0.2
Lake Michigan Sediment Samples
x:
+
I
month
9 April
May
• June
July
® August
t September
# October
lakesite
:MLB
M15
¦ M45
+ 110
-0.4 -0.2 0.0 0.2
Axis.1 [29.8%]
0.4
Lake Michigan Water Samples Only
0.4
0 2
fraction
# Particle
>P
0s-
# Free
• Whole
ib
x-
— 0.0
Cn|
nuc_acid_type
&
• DNA
X
<
~ rna
-0.2
-0.4
* *v
•.V«
i
u
month
• April
May
June
• July
• August
® September
® October
nuc_acid_type
• DNA
Arna
-0.2 0.0 0.2
Axis.1 [19.3%]
-0,2
0.0
Axis.1 [19.3%]
0-2
Figure 21: Abundance-weighted bacterial community composition dynamics. Principal
coordinates ordination of (A) all columns samples, which were fractionated as 0.22-3 (free), 3-20
(particle), and 0.22-20 (whole) micrometer fractions during sampling, and detailed analyses for
(B) sediment samples, and (C, D) water samples. The whole fraction was analyzed at the DNA
and RNA level, while other fractions and the sediment was only analyzed at the DNA level.
CSMI Lake Michigan 2015 Report
42
-------
Table 1: GLERL sampling activities during the 2015 Lake Michigan CSMI field season.
X=M15, *=M45, += Ml 10. PSS Long Transect sampling runs from sites M15 to Ml 10 while
passing through M45; Diel sampling in a short PSS run along the depth isobaths at either M45 or
Ml 10 during the day and night.
April
May
June
July
August
September
October
Acoustics
x+*
x+*
x+*
x+*
x+*
x+
x+
Bottom Trawl
x+*
x+*
x+*
x+*
x+*
x+*
Chlorophyll
x+*
x+
x+
x+*
x+
x+
x+
CTD
x+*
x+*
x+*
x+*
x+*
x+*
x+*
Fluoroprobe
x+
x+*
x+*
x+*
x+*
x+*
Larval Fish
x+
x+*
x+*
x+*
x+*
x+
Microbes
x+
x+*
x+*
x+*
x+*
x+*
x+*
Mysis
+*
+*
+*
+*
+*
+*
+*
Nutrients
x+*
x+
x+
x+*
x+
x+
x+
PSS:
Long Transect
x+*
x+*
x+*
x+*
x+*
x+
x+
Diel
+
+
*
UV Radiometer
+
+
*
+
Zooplankton:
Full
x+
x+
x+
x+*
x+*
x+
x+
Depth Stratified
+
+
+
+*
+
CSMI Lake Michigan 2015 Report
43
-------
Report: Major Findings from the CSMI Benthic
Macroinvertebrate Survey in Lake Michigan in 2015 With
an Emphasis on Temporal Trends
Authors:
Thomas F. Nalepa, University of Michigan
Lyubov E. Burlakova, SUNY Buffalo State
Ashley K. Elgin, NOAA Great Lakes Environmental Research Laboratory
Alexander Y. Karatayev, SUNY Buffalo State
Gregory A. Lang, NOAA Great Lakes Environmental Research Laboratory
Knut Mehler, SUNY Buffalo State
Contact:
Alexander Karatayev
Email: karataay@buffalostate.edu
Phone: 716-878-4329
Address:
Great Lakes Center
Buffalo State, The State University of New York
1300 Elmwood Avenue
Buffalo, NY 14222
CSMI Lake Michigan 2015 Report
44
-------
Introduction
As part of the Coordinated Science and Monitoring Initiative (CSMI) in Lake Michigan in 2015,
a lake-wide benthic survey was conducted to assess the status of the benthic macroinvertebrate
community, with a primary focus on the invasive mussels Dreissena rostriformis bugensis and Dreissena
polymorpha, and the native amphipod Diporeia. Similar lake wide surveys were conducted to assess the
status of these three taxa beginning in 1994/1995 and repeated every five years through 2010 (Nalepa et
al. 2014). Based on previous surveys, major changes in population abundances of all three taxa were
observed over this 15-year period. D. polymorpha was first reported in Lake Michigan in 1989 (Marsden
et al. 1993) and densities subsequently increased to reach a peak in 2000. Thereafter, densities declined
to such an extent that by 2010 it was rarely found. Over the entire period, D. polymorpha was mainly
found at depths < 50 m. After D. r. bugensis was first reported in the lake in 1997 (Nalepa et al. 2001),
densities have mostly continued to increase at all depths through 2010, attaining densities that exceeded
those of D. polymorpha even at depths < 50 m. Lastly, the amphipod Diporeia has been in a steady state
of decline ever since Dreissena became established. Lower densities relative to those in pre- Dreissena
years were first observed in the early 1990s (Nalepa et al. 1998), and declines continued at all depths from
1994/1995 through 2010. In 2010, it had mostly disappeared at depths < 50 m and had declined by 95%
at > 50 m.
Both Dreissena and Diporeia play key roles in the ecosystem of Lake Michigan and the other
Great Lakes. Dreissena has a great capacity to filter particulate material from the water column and
excrete metabolic by-products (biodeposits, nutrients). As a result, Dreissena has dramatically
restructured food webs and altered spatial patterns of energy and nutrient flow (Vanderploeg et al. 2002,
Hecky et al. 2004). Specific impacts of Dreissena on the Lake Michigan ecosystem have been well-
documented, including reduction of the spring phytoplankton bloom and alteration of benthic-pelagic
processes (Fahnenstiel et al. 2010, Cuhel and Aguilar 2013, Vanderploeg et al. 2015). Before it declined,
Diporeia was a keystone species in the offshore food web, accounting for over 70% of benthic biomass
45
CSMI Lake Michigan 2015 Report
-------
and serving as an energy-rich food source for many fish species. As a detritivore that feeds on freshly-
settled material in the upper sediments, Diporeia was an important pathway by which energy was cycled
from the benthic to the pelagic region (Nalepa 1989, Nalepa et al. 2000, 2009). The decline of Diporeia
has led to large changes in the relative health, growth, and community structure of fish communities in
the lake (Pothoven et al. 2001, Hondorp et al. 2005, Bunnell et al. 2009). Because of these key
ecosystem roles and population shifts through 2010, the current status of Dreissena and Diporeia were of
particular interest in 2015.
For the first time since the lake-wide surveys were initiated in 1994/1995, the entire benthic
community (i.e., all benthic taxa) was examined in 2015, which allowed an assessment of other taxa
besides Dreissena and Diporeia, and provided a baseline to examine future changes of the entire benthic
community in Lake Michigan. Lake-wide trends in the entire benthic community have recently been
examined in Lake Huron (Nalepa et al. 2007), Lake Ontario (Birkett et al. 2015), and Lake Erie
(Burlakova et al. 2014).
This report provides a summary of recent trends of Dreissena, Diporeia, and other major taxa
based on results of the 2015 survey. In addition, it also gives a synopsis of other, ancillary data collected
during the survey, such as length-weight relationships and size frequencies of the dreissenid population.
The primary focus is to present major findings and to place these findings into a historic perspective.
More detailed analyses and discussion of trends, spatial patterns, and community composition, including
comparisons to lake-wide surveys in the other Great Lakes, will be provided in other publications.
Methods
Benthic samples were collected at 140 stations in Lake Michigan, July 20-29, 2015 (Table 1). Of
these, 135 were located in the main basin of the lake, and 5 were located in the outer portion of Green
Bay (Table 1, Fig. la, b, c). The number and location of stations have generally remained consistent
since 2000. For the complete list and locations of stations sampled in all previous surveys see Nalepa et
al. (2014).
46
CSMI Lake Michigan 2015 Report
-------
Sampling procedures in 2015 were the same as in previous surveys. In brief, benthic samples
were taken in triplicate at each site with a Ponar grab (area in 2015 = 0.048 m2). Collected material was
washed through an elutriation device fitted with a 0.5-mm mesh net, and retained residue was preserved
in 5-10% buffered formalin containing rose bengal stain. Sample jars were labeled with the station
designation, replicate number, and date. Sampling depth and a general description of the sediments at
each station were recorded (Table 1).
As noted, only Dreissena and Diporeia were counted and identified in surveys prior to 2015,
whereas all organisms were counted and identified in 2015. Details of laboratory procedures and
protocols will not be provided here. Procedures prior to 2015 are given in Nalepa et al. (2014).
Procedures in 2015 followed those in the EPA Standard Operating Procedure (SOP) LG407 "Standard
Operating Procedure for Benthic Invertebrate Laboratory Analysis" (Revision 09, April 2015) as given in:
https://www.epa.gov/sites/production/files/2017-01/documents/sop-for-benthic-invertebrate-lab-analvsis-
201504-13pp.pdf
Methods to determine densities were straight-forward and similar across all survey years. All
organisms were picked and counted under low magnification, with dreissenids proportionally split when
numbers were high. In 2015, biomass of Dreissena was determined as both ash-free dry weight (AFDW,
soft tissue) and total wet weight (TWW, shell included). Surveys prior to 2015 reported dreissenid
biomass as AFDW, which was calculated by first determining relationships between AFDW and shell
length, and then applying these relationships to size frequencies (Nalepa et al. 2014). As given in EPA's
SOP, dreissenid biomass is determined as TWW, which is determined directly by blotting dry all
dreissenids in a sample and then weighing. For consistency, dreissenid biomass was determined by both
methods in 2015.
Length-weight relationships were derived from individuals freshly-collected with a Ponar grab
from 22 sites during the 2015 survey (Table 2). While priority was given to sites where individuals for
length-weight relationships were collected in 2010, the ultimate criteria for site selection depended on the
number of mussels found at the time of sampling, and by a visual estimate of the size range (shell lengths)
47
CSMI Lake Michigan 2015 Report
-------
of the population. For the latter, a broad size range of individuals was a requirement (10 mm to > 20 mm)
so that a representative relationship could be obtained. Also, an effort was made to collect at sites located
throughout the lake and at various depths. Immediately after collection of mussels, soft tissues of about
25 individuals between 10 mm and > 20 mm were removed from the shells, placed individually into pre-
weighed aluminum planchets, and dried at 60 C° for at least 48 h. After drying, the planchets were placed
and kept in a dessicator. Upon completion of the survey cruise and return to the laboratory, soft tissues
were weighed, ashed at 550 C° for 1 h, and then re-weighed. AFDW was then calculated as the difference
between dry weight and post-ashed weight. Corresponding shell lengths were measured to the nearest 0.5
mm. Overall, a total of 569 individuals from the 22 sites were weighed and measured (Table 3). All
individuals for length-weight determinations were D. r. bugensis since D. polymorpha was not found.
Measured AFDWs and shell lengths (SL) were used to develop length-weight relationships according to
the allometric equation: logt AFDW (mg) = b + a*logc,SL (mm). Relationships were developed for pooled
sites within four different depth intervals: < 30 m, 31-50 m, 51-90 m, and > 90 m (Table 3, also see
below). For size frequencies, shell lengths of all mussels in each replicate sample were measured and
then binned into 1-mm size categories. In prior surveys, individuals < 5 mm were not individually
measured and were therefore binned into one category (0-5 mm). In 2015, these small individuals were
measured and binned into 1-mm size categories. Further, mussels < 1 mm were not included in biomass
calculations.
To determine AFDW biomass, the number of individuals in each size category was multiplied by
the AFDW of an individual in that category as derived from the length-weight regression (calculated from
the mid-shell length of that category). All size-category weights were then summed.
For analysis of trends, sites in the main lake were divided into the same four depth intervals as in
previous surveys: < 30 m, 31-50 m, 51-90 m, and > 90 m. These intervals define distinct physical
habitats that result in distinguishable benthic communities (Alley and Mozley 1975, Nalepa 1989).
Because physicochemical conditions in Green Bay are so different than in the main lake, results for the 5
48
CSMI Lake Michigan 2015 Report
-------
sites located in the bay are given separately. All values were loge +1 transformed before any statistical
tests.
Results and Discussion
The 2015 survey extended the assessment of lake-wide trends in D. polymorpha, D. r. bugensis,
and Diporeia that were previously defined between 1994/1995 and 2010 (Nalepa et al. 2014). For D.
polymorpha, no individuals were found in any of the samples collected in 2015 (Table 4, Fig. 2). This
species peaked in 2000 at depths < 50 m and has steadily declined since. Only a few individuals were
found at just one station in 2010, thus it is not surprising that no individuals were collected in 2015. The
decline of D. polymorpha coincided with the rapid expansion of D. r. bugensis between 2000 and 2005
(Fig. 3). Both species are filter-feeders and compete for the same food resources. Because D. r. bugensis
has a lower respiration rate and a higher assimilation rate than D. polymorpha (Baldwin et al. 2002,
Stoeckmann 2003), it is more efficient in allocating resources to growth and reproduction and thus has a
competitive advantage when available food resources are limited. Further, D. r. bugensis has a lower
temperature threshold of reproduction compared to D. polymorpha and therefore is able to colonize to
deeper depths (Karatayev et al. 2015).
For D. r. bugensis, some important temporal patterns emerged in 2015 that perhaps signaled a
shift in population dynamics. Most notably, when compared to densities in 2010, densities in 2015
declined at all depth intervals except at the deepest (> 90 m) (Table 4, Fig. 3). In prior surveys through
2010, densities of D. r. bugensis have generally increased at all depth intervals. The exception was at the
31-50 m interval where densities peaked in 2005 and have declined since (Table 4). In 2015, mean
densities declined by 79%, 56%, and 40% at the < 30 m, 31-50 m, 51-90 m depth intervals, respectively.
These declines were significant for each interval (P < 0.05, t-test). With these declines, densities of D. r.
bugensis have seemingly peaked at depths < 90 m. The only depth interval where densities of D. r.
49
CSMI Lake Michigan 2015 Report
-------
bugensis were not lower in 2015 compared to 2010 was > 90 m. Mean density at this interval increased
from 2,037/m2 to 2,797/m2; this difference, however, was not significant (P > 0.05).
It is worth noting that the number of sites at < 30 m was far lower in 2015 than in 2010 (n = 29
and 38, respectively; see Table 4). Many sites are located around 30 m, and in 2015 some sites were
recorded as a few meters deeper than in 2010, placing them into the 31-50 m interval. Also, two sites in
the < 30 m interval were not sampled in 2015 but were sampled in 2010. To be certain that declines in
D. r. bugensis in 2015 at the < 30 m and 31-50 m intervals were not a result of sites changing depth
categories, means were again determined after placing these sites into the same category as in 2010.
Mean densities in 2015 thus determined were 2,780 ± 661/m2 for < 30 m (n=36) and 5,817 ± 707/m2 for
31-50 m (n=41). Both densities were still significantly lower than in 2010.
Trends in dreissenid AFDW biomass were similar to trends in density at < 30 m and > 90 m.
That is, mean biomass in 2015 declined at the former interval and increased at the latter interval when
compared to 2010 (Fig. 4), and these year-to year differences were significant at both depth intervals (P <
0.05). Mean biomass at 31-50 m and 51-90 m did not decline like density (Fig. 4), and differences
between 2010 and 2015 were not significant (P> 0.05). Considering biomass on a lake-wide basis, the
mean, depth-weighted biomass for Dreissena in 2000, 2005, 2010, and 2015 was 0.30 g/m2, 8.9 g/m2,
13.7 g/m2, and 16.5 g/m2, respectively. Thus, total depth-weighted biomass was greater in 2015 than in
2010, which can mainly be attributed to increased biomass at > 90 m, a depth interval that comprises
41.5% of the main-lake area.
The divergence of trends in dreissenid density and AFDW biomass at the 31-50 and 51-90 m
intervals between 2010 and 2015 can either be attributed to differences in length-weight, or to differences
in size frequencies (or to both). With a decline in density in 2015, weight per unit shell length
(AFDW/SL) must have increased, or the average size of individuals in the population must have
increased. To assess differences in AFDW/SL, the AFDW of a standard 15-mm mussel was calculated
and compared between the two years based on regressions given in Table 3. AFDW of a 15-mm mussel
at 31-50 m was 5.46 mg and 5.17 mg in 2010 and 2015, respectively, and AFDW at 51-90 m was 6.07 mg
50
CSMI Lake Michigan 2015 Report
-------
and 5.78 mg. Thus, AFDW/SL at both intervals was lower in 2015 than in 2010, and hence cannot
account for mean biomass being higher in 2015. Size frequencies in the two years were examined by
placing individuals into 5-mm size categories and then determining the proportion of all mussels in each
category for each depth interval. At both the 31-50 m and 51-90 m intervals, the proportion of the
population <10 mm decreased, while the proportion >10 mm increased in 2015 compared to 2010 (Table
5). Individuals >10 mm increased from 30.4% to 58.4% at 31-50 m, and increased from 27.9% to 40.3%
at 51-90 m. These increases in the proportion of larger-sized individuals in 2015 compared to 2010
appear to be the likely reason for biomass not declining despite significant declines in density. Since
tissue weight increases exponentially with shell size, even a modest increase in the proportion of larger
individuals greatly affects biomass. For the other two intervals, the proportion of individuals >10 mm
declined at < 30 m (37.1% to 30.2%), but increased at > 90 m (26.5% to 45.6%). Increased biomass at >
90 m in 2015 relative to 2010 can thus be attributed to not only an increase in density in 2015, but also to
a greater proportion of larger individuals. An increase in AFDW/SL may also have played a role (see
below).
Besides using length-weight relationships to determine dreissenid biomass, these relationships are
also useful to assess the relative health of the population. For Dreissena, the amount of tissue per unit
shell length is directly related to food availability (Walz 1978, Sprung and Borchering 1991, Nalepa et al.
1995). This relationship holds true for molluscs in general (Russell-Hunter 1985). Given this, a lower
AFDW/SL over time would indicate that tissue loss or tissue "degrowth" has occurred, a sign that
individuals are catabolizing soft tissue while under nutritional stress. Ultimately, lower tissue weight can
hinder survival (Karatayev et al. 2010) and lead to lower reproduction (Bielefeld 1991, Sprung 1995).
Temporal trends in AFDW/SL can thus be a broader indicator of future population growth. As noted, the
AFDW of a standard 15-mm mussel was lower in 2015 than in 2010 at 31-50 m and 51-90 m. To further
explore trends at all depth intervals, AFDW of a standard 15-mm mussel was determined from
regressions for D. r. bugensis in Lake Michigan going back to 2004 (see Table 3). Trends varied widely
between the depth intervals (Fig. 5). AFDW of a 15-mm mussel was consistently greatest at the < 30 m
51
CSMI Lake Michigan 2015 Report
-------
interval over the 11-year period, but because of great variation between years a clear temporal trend was
not readily discernable. On the other hand, the most defined temporal trend occurred at 31-50 m. At this
interval, the AFDW of a 15-mm mussel steadily declined between 2004 and 2015; by 2015 it was 30.8%
lower than in 2004. For the two deeper intervals, 51-90 m and > 90 m, regressions were only available in
2010 and 2015. At the 51-90 m interval, the AFDW of a 15-mm mussel declined by 4.8% over the 5-year
period, while at the > 90 m interval it increased 6.0%. Based on these trends, and the fact that relative
values in 2015 were lowest at 31-50 m and 51-90 m, it appears that D. r. bugensis populations at these
two intervals may be under nutritional stress.
Biomass estimates of Dreissena populations in the Great Lakes have been reported in a number
of different units including AFDW, dry weight (DW), and TWW. Of these, dried mass (AFDW or DW)
of mussel tissue most accurately reflects functional mass, and hence estimates of dreissenid metabolic
functions such as filtering, respiration, and excretion rates are generally provided as per unit AFDW or
DW (Vanderpoeg et al. 2010, Johengen et al. 2014, Tyner et al. 2015). These metabolic rates along with
population biomass provided as AFDW or DW have been used to estimate lake-wide ecosystem impacts
(Nalepa et al. 2009, Vanderploeg et al. 2010, Rowe et al. 2015, Tyner et al. 2015). In 2015, dreissenid
biomass was determined as both AFDW and TWW. To examine the relationship between AFDW and
TWW, biomass estimated by both methods was plotted for each station (Fig. 6). A regression through the
origin between the two values was significant (R2 = 0.92) and defined by: TWW = 50.09*AFDW. Given
this strong relationship between AFDW and TWW, the equation given above may be useful in converting
from one biomass estimate to the other. One caveat, however, is the wide variation between the two
estimates when values of AFDW are greater than about 40 g/m2 (Fig. 6). Reasons for this variation are
unclear. One potential reason is that at sites with a greater number/biomass of mussels, any differences
between the TWW/SL relationship at that one site and the generalized depth-specific length-weight
relationship used to calculate AFDW are compounded and therefore results in a greater discrepancy
between the two methods. Also, at high mussel numbers/biomass, shell weight per unit shell length may
be more inconsistent, the amount of water retained in the shell cavity may be more variable, or more
52
CSMI Lake Michigan 2015 Report
-------
debris might be found on shells. Regardless, at high numbers/biomass AFDW is both lower and higher
relative to TWW which complicates any potential theory.
Based on the 2015 survey, the amphipod Diporeia continued to decline (Table 4, Fig. 7). In
2015, Diporeia was collected at only one site that was < 90 m, and at 9 sites that were > 90 m. In
comparison, in 2010 Diporeia was collected at 13 sites < 90 m and 11 sites > 90 m. This depth-defined
pattern of decline, with densities declining first and most rapidly in nearshore, shallow regions and more
slowly with increased depth, has been apparent since the decline of Diporeia was first reported in the lake
in the early 1990s (Nalepa et al. 1998). Such a spatial pattern coincides directly with the depth-related
expansion of Dreissena. D. polymorpha increased mostly in the nearshore region (< 30 m) until 2000,
and subsequently D.r. bugensis increased rapidly in nearshore regions and more slowly in deeper,
offshore regions (> 90 m). The exact reason for the negative response of Diporeia to Dreissena has not
been determined but, with the exception of Lake Superior where the Dreissena population is very limited,
the decline of Diporeia has consistently occurred in all the Great Lakes within a few years after Dreissena
became established (Nalepa et al. 2006).
Although mean density of Diporeia at > 90 m was not lower in 2015 than in 2010, the continued
increase of D. r. bugensis at this depth interval would suggest that densities of Diporeia will most likely
decrease, or the population will be completely gone, in future surveys. In 2015, not only were densities of
D. r. bugensis greater at sites > 90 m compared to 2010, the spatial extent of the population had expanded.
D. r. bugensis was present at 9 of 10 sites where Diporeia was collected in 2015. In Lake Ontario, D.r.
bugensis expanded to deeper depths (> 90 m) sooner than in Lake Michigan, and in a lake-wide survey of
Lake Ontario in 2013, only one Diporeia was collected at sites > 90 m, and no individuals were collected
at sites < 90 m (Nalepa and Elgin unpublished). Mean density of D. r. bugensis at > 90 m was 2,044/m2
in Lake Ontario in 2013, which is comparable to the mean density of 2,797/m2 found in Lake Michigan at
this depth interval in 2015. Thus, if such a density of D. r. bugensis nearly extirpated Diporeia at this
depth interval in Lake Ontario, a similar outcome might be expected in Lake Michigan.
53
CSMI Lake Michigan 2015 Report
-------
Since 2015 was the first survey year in which the entire benthic community was examined, lake-
wide temporal trends in taxa other than Dreissena and Diporeia could not be assessed. However, a more
limited assessment of changes in these other benthic taxa can be derived by comparing 2015 results to
benthic data collected in the 1990s in just the southern basin. As part of a NOAA monitoring program,
benthic samples have been collected at 40 sites in the southern basin for 2 consecutive years every 5 years
beginning in 1980-1981 (Nalepa 1987, Nalepa et al. 1998). The two most recent years in which data are
entirely available are in 1998-1999 (Nalepa and Elgin, unpublished). Since the same 40 sites were
sampled in 2015 (see Table 1), densities of Oligochaeta, Sphaeriidae, and Chironomidae in 2015 were
compared to densities in 1992-1993 and 1998-1999 at just these 40 sites. The 1992-1993 period was just
after D. polymorpha became established in the southern basin, and the 1998-1999 period was about when
D. polymorpha peaked and just before D. r. bugensis spread into the basin (about 2001). For
oligochaetes, mean densities progressively increased in each of the three sampling periods (that is, 1992-
1993, 1998-1999, and 2015) at the three depth intervals < 90 m (Table 6). These increases, particularly
apparent at the < 30 m and 31-50 m intervals, may be a result of a dreissenid impact known as the
"nearshore shunt" (Hecky et al. 2004). In brief, this is the process by which organic material is retained
for a longer period of time in nearshore regions by the activities of Dreissena. Dreissena filters
particulate material (mainly phytoplankton) from the water column and subsequently deposits this organic
material in the benthic zone in the form of feces and pseudofeces. These biodeposits would then serve as
an added food source for benthic detritivores. Most all oligochaetes are detritivores and thus populations
would benefit from these added food inputs. Benthic inputs of organic material are more pronounced in
nearshore regions since primary production is greatest in this region, and because the water column is
well-mixed giving Dreissena access to all phytoplankton present. Most chironomids are also detritivores
but, although mean densities of chironomids were greater in 2015 than in the 1990s at the two shallowest
intervals, variation was too great to state with certainty that densities increased. Oligochaetes did not
increase at the deepest interval (> 90 m). Although Dreissena in deeper, offshore waters also deposit
organic material, these biodeposits would have less of an impact on detritivores. Benthic food availability
54
CSMI Lake Michigan 2015 Report
-------
in offshore regions is greatly diminished compared to nearshore regions not only because primary
production in the upper water column is less, but also because this organic matter is fed upon by
organisms (bacteria, protozoans, etc.) as it settles downward through a longer water column to ultimately
reach the benthic region.
In contrast to increased densities of oligochaetes in the shallower depth intervals, densities of
sphaeriids were lower at all depth intervals in 2015 compared to the 1990s (Table 6). A decline in
sphaeriids at all depths was first observed soon after Dreissena became established in the southern basin
(Nalepa et al. 1998). Reasons for the negative response of sphaeriids to Dreissena are not clear. Since
sphaeriids are filter-feeders, it is presumed that they are being outcompeted by Dreissena for available
food. Yet the dominant sphaeriid in the Great Lakes is Pisidium spp., a genus that filters bacteria in
benthic interstitial waters and therefore should benefit from increased bacteria associated with dreissenid
biodeposits.
The dominance of Dreissena in the benthic community of Lake Michigan and the other Great
Lakes has clear implications for other benthic taxa. While detailed comparisons of benthic community
trends between lakes will not be provided here, a general overview of between-lake trends in Dreissena
puts Lake Michigan results into a broader perspective. A comparison of density trends of Dreissena in
Lakes Michigan, Ontario, and Huron at the < 30 m, 31-90 m, and > 90 m intervals is given in Fig. 8. To
make this comparison, densities at 31-50 m and 51-90 m were combined (interval becomes 31-90 m) for
Lakes Michigan and Huron since these two depth intervals were not reported separately for Lake Ontario
in previous studies (Watkins et al. 2007, Birkett et al. 2015). Density trends at < 30 m are difficult to
compare between lakes since high variation in physical drivers (i. e., substrate heterogeneity, wave-
induced disturbance) strongly influence dreissenid estimates. This is evident in the wide year-to-year
variation at this depth interval in Lake Ontario (Fig. 8). Physical conditions become more stable as depth
increases, and population trends at depths > 30 m are better suited for lake-to-lake comparisons. The
decline of D. r. bugensis in Lake Michigan in2015at31-90mis similar to an ongoing decline in Lake
Ontario that has been evident since 2008 (Fig. 8). If populations in both lakes have indeed peaked at this
55
CSMI Lake Michigan 2015 Report
-------
depth, a greater peak density was attained in Lake Michigan. In both lakes, densities increased sharply
and then gradually declined. In contrast, densities at 31-90 m in Lake Huron have increased gradually
and, as of 2012, do not yet appear to have peaked. Densities at > 90 m are still increasing in all three
lakes (Fig. 8). Similar comparisons of temporal trends in dreissenid biomass are not possible since
biomass was not historically measured in each lake. However, the most recent lake-wide survey in each
lake determined and reported biomass using the same methods, and values in the four depth intervals are
given in Table 7. Again, considering biomass only at depths > 30 m, mean biomass in Lakes Michigan
and Ontario were generally comparable at 31-50 m, 51-90 m, and > 90 m, whereas biomass in Lake
Huron was about 50%, 78%, and 38% lower than in Lakes Michigan and Ontario at these three depth
intervals, respectively.
Summary
A lake-wide benthic survey was conducted in Lake Michigan in 2015 to assess the current status
of the macroinvertebrate community. Similar lake-wide surveys have been conducted in the lake at 5-
year intervals beginning in 1994/1995. These previous surveys only examined populations of Dreissena
polymorpha, Dreissena r. bugensis, and Diporeia, whereas the 2015 survey examined the entire benthic
community. Perhaps the most significant finding in 2015 was the decline in densities of D. r. bugensis at
depths < 90 m. Compared to densities in 2010, densities in 2015 declined 79%, 56%, and 40% at the < 30
m, 31-50 m, 51-90 m intervals, respectively. In contrast, densities at > 90 m increased 37%. Because of
a greater proportion of larger individuals in the population, biomass at 31-50 m and 51-90 m remained
stable or slightly increased in 2015 compared to 2010. Overall, depth-weighted biomass increased from
13.7 g/m2 in 2010 to 16.5 g/m2 in 2015, largely due to increased biomass at sites > 90 m. The other
dreissenid species, D. polymorpha, was not collected at any of the sites in 2015, indicating it has
essentially been displaced by D. r. bugensis. Also, the amphipod Diporeia continued to disappear. It was
collected at only one site < 90 m and at 9 sites > 90m. Lake-wide temporal trends in other major benthic
taxa such as Oligochaeta, Sphaeriidae, and Chironomidae could not be assessed since 2015 was
56
CSMI Lake Michigan 2015 Report
-------
the first year the entire benthic community was sampled. However, based on comparisons to data
collected in just the southern basin in 1992-1993 and 1998-1999, oligochaetes have progressively
increased in shallower and mid-depth regions between 1992-1993 and 2015. A likely reason is an
increased amount of potential food resulting from the biodeposition of organic material by Dreissena.
contrast, sphaeriids progressively declined all depth intervals between 1992-1993 and 2015.
CSMI Lake Michigan 2015 Report
57
-------
References
Alley, W. P. and S. C. Mozley. 1975. Seasonal abundance and spatial distribution of Lake Michigan
macrobenthos, 1964-67. Great Lakes Res. Division Spec. Publ. 54, University of Michigan, Ann
Arbor, MI.
Baldwin, B. S., M. S. Mayer, J. Dayton, N. Pau, J. Mendilla, M. Sullivan, M. Moore, A. Ma, and E. L.
Mills. 2002. Comparative growth and feeding in zebra amd quagga mussels (Dreissena
polymorpha and Dreissena bugensis): implications for North American lakes. Can. J. Fish.
Aquat. Sci. 59: 680-694.
Bielefeld, U. 1991. Histological observation of gonads and digestive gland in starving Dreissena
polymorpha (Bivalvia). Malacologia 33: 31-42.
Birkett, K., S. J. Lozano, and L. G. Rustram. 2015. Long-term trends in Lake Ontario's benthic
macroinvertebrate community. Aquat. Ecosys. Health & Manag. 18: 78-88.
Burlakova, L. E., A. Y. Karatayev, C. Pennuto, and C. Mayer. 2014. Changes in Lake Erie benthos over
the last 50 years: Historical perspectives, current status, and main drivers. J. Great Lakes Res. 40:
560-573.
Bunnell, D. B., S. R. David, C.P. Madenjian. 2009. Decline of bloater fecundity in southern Lake
Michigan after decline of Diporeia. J. Great Lakes Res. 35: 45-49.
Cuhel, R. L. and C. Aguilar. 2013. Ecosystem transformations of the Laurentian Lake Michigan by
nonidigenous biological invaders. Ann. Rev. Mar. Sci. 5: 289-320.
Fahnenstiel, G., T. Nalepa, S. Pothoven, H. Carrick, and D. Scavia. 2010. Lake Michigan lower food
web: Long-term observations and Dreissena impact. J. Great Lakes Res. 36 Suppl. 3): 1- 4.
Hecky, R. E., R. E. H. Smith, D. R. Barton, et al. 2004. The nearshore phosphorus shunt: a consequence
of ecosystem engineering by dreissenids in the Laurentian Great Lakes. Can. J. Fish. Aquat. Sci.
61:1285-1293.
Hondorp, D. W., S. A. Pothoven, and S. A. Brandt. 2005. Influence of Diporeia density on diet
composition, relative abundance, and energy density of plantivorous fishes in southeast Lake
Michigan. Trans. Am. Fish. Soc. 134: 588-601.
Johengen, T. H., H. A.Vanderploeg, and J. R. Liebig. 2014. Effects of algal composition, seston
stoichiometry, and feeding rate on zebra mussel (Dreissena polymorpha) nutrient excretion in
two Laurentian Great Lakes. In Quagga and Zebra Mussels: Biology, Impacts, and Control,
Second Edition. T. F. Nalepa and D. W. Schloesser, eds., pp. 445-459. CRC Press: Boca Raton.
Fl.
Karatayev, A. Y., S. E. Mastitsky, D. K. Padilla, L. B. Burlakova, and M. M. Hajduk. 2010. Differences
in growth and survivorship in zebra and quagga mussels: size matters. Hydrobiologia 668: 183-
194.
Karatayev, A. Y., L. E. Burlakova, and D. K. Padilla. 2015. Zebra versus guagga mussels: a review of
their spread, population dynamics, and ecosystem impacts. Hydrobiologia 745: 97-112.
Marsden, J. E., N. Trudeau, and T. Keniry. 1993. Zebra mussels study in Lake Michigan. Aquat. Ecol.
Rept 93/14. Illinois Natural History Survey.
Nalepa, T. F. 1987. Long term changes in the macrobenthos of southern Lake Michigan. Can. J. Fish.
Aquat. Sci. 44:515-524.
58
CSMI Lake Michigan 2015 Report
-------
Nalepa, T. F. 1989. Estimates of macroinvertebrate biomass in Lake Michigan. J. Great Lakes Res. 15:
437-443.
Nalepa, T. F., J. A. Wojcik, D. L. Fanslow, and G. A. Lang. 1995. Initial colonization of the zebra
mussel (Dreissena polymorpha) in Saginaw Bay, Lake Huron: Population recruitment, density,
and size structure. J. Great Lakes Res. 21: 417-434.
Nalepa, T. F., D. J. Hartson, D. L. Fanslow, G. A. Lang, and S. J. Lozano. 1998. Declines in benthic
macroinvertebrate populations in southern Lake Michigan, 1980-1993. Can. J. Fish. Aquat. Sci.
55: 2402-2413.
Nalepa, T. F., D. J. Hartson, J. Buchanan, J. F. Cavaletto, G. A. Lang, and S. J. Lozano. 2000. Spatial
variation in density, mean size and physiological condition of the holarctic amphipod Diporeia
spp. in Lake Michigan. Freshwater Biology 43: 107-119.
Nalepa, T. F., D. W. Schloesser, S. A. Pothoven, D. W. Hondorp, D. L. Fanslow, M. L. Tuchman, and G.
L. Fleischer. 2001. First finding of the amphipod Echinogammarus ischnus and the mussel
Dreisssena bugensis in Lake Michigan. J. Great Lakes Res. 27: 384-391.
Nalepa, T. F., D. C. Rockwell, and D. W. Schloesser. 2006. Disappearance of the amphipod Diporeia
spp. in the Great Lakes. Workshop Summary, Discussion, Recommendations. NOAA Technical
Memorrandum GLERL-136. NOAA, Great Lakes Environmental Research Laboratory, Ann
Arbor, MI.
Nalepa, T. F., D. L. Fanslow, S. A. Pothoven, A. J. Foley III, and G. A. Lang. 2007. Long-term trends in
benthic macroinvertebrate populations in Lake Huron over the past four decades. J. Great Lakes
Res. 33: 421-436.
Nalepa, T. F., D. L. Fanslow, G. A. Lang. 2009. Transformation of the offshore benthic community in
Lake Michigan: recent shift from the native amphipod Diporeia spp. to the invasive mussel
Dreissena rostriformis bugensis. Freshwater Biology 54:466-475.
Nalepa, T. F., D. L. Fanslow, and S. A. Pothoven. 2010. Recent changes in density, biomass,
recruitment, size structure, and nutritional state of Dreissena populations in southern Lake
Michigan. J. Great Lakes Res. 36 (Suppl.3): 5-19.
Nalepa, T. F., D. L. Fanslow, G. A. Lang, K. Mabrey, and M. Rowe. 2014. Lake-wide benthic surveys in
Lake Michigan in 1994-1995, 2000, 2005, and 2010: abundances of the amphipod Diporeia spp
and abundances and biomass of the mussels Dreissena polymorpha and Dreissena rostriformis
bugensis. NOAA Technical Memorandum GLERL-164. NOAA Great Lakes Environmental
Research Laboratory, Ann Arbor, MI.
Pothoven, S. A., T. F. Nalepa, P. J. Schneeberger, and S. B. Brandt. 2001. Changes in diet and body
condition of lake whitefish in southern Lake Michigan associated with changes in benthos. N.
Amer. J. Fish. Manag. 21: 876-883.
Russell-Hunter, W. D. 1985. Physiological, ecological, and evolutionary aspects of molluscan tissue
degrowth. Amer. Malacol. Bull. 3: 213-221.
Rowe, M. D., D. R. Obenour, T. F. Nalepa, H. A. Vanderploeg, F. Yousef, and W. C. Kerfoot. 2015.
Mapping the spatial distribution of the biomass and filter-feeding effect of invasive dreissenid
mussels on the winter-spring phytoplankton bloom in Lake Michigan. Freshwater Biology 60:
2270-2285.
Sprung, M. 1995. Physiological energetic of the zebra mussel Dreissena polymorpha in lakes I. Growth
and reproductive effort. Hydrobiologia 304: 117-132.
59
CSMI Lake Michigan 2015 Report
-------
Sprung, M. and J. Borcherding. 1991. Physiological and morphometric changes in Dreissena
polymorpha (Mollusca; Bivalvia) during a starvation period. Malacologia 33: 179-191.
Stoeckmann, A. 2003. Physiological energetics of Lake Erie dreissenid mussels: a basis for the
displacement of Dreissena polymorpha by Dreissena bugensis. Can. J. Fish. Aquat. Sci. 60:126-
134.
Tyner, E. H., H. A. Bootsma, and B. M. Lafrancois. 2015. Dreissenid metabolism and ecosystem effects
as revealed by oxygen consumption. J. Great Lakes Res. 41 (Suppl. 3): 27-47.
Vander ploeg, H. A., T. F. Nalepa, D. J. Jude, et al. 2002. Dispersal and emerging ecological impacts of
Ponto-Caspian species in the Laurentian Great Lakes. Can. J. Fish. Aquat. Sci. 59: 1209-1228.
Vanderploeg, H. A., J. R. Liebig, T. F. Nalepa, G.. L. Fahnenstiel, and S. A. Pothoven. 2010. Dreissena
and the disappearance of the spring phytoplankton bloom in Lake Michigan. J. Great Lakes Res.
36 (Suppl.3): 50-59.
Vanderploeg, H. A., D. B. Bunnell, H. J. Carrick, and T. O. Hook. 2015. Complex interactions in Lake
Michigan's rapidly changing ecosystem. J. Great Lakes Res. 41 (Suppl. 3): 1-6.
Walz, N. 1978. The energy balance of the freshwater mussel Dreissena polymorpha Pallas in laboratory
experiments and in Lake Constance. IV. Growth in Lake Constance. Arch. Hydrobiol./Suppl.
55: 142-156.
Watkins, J. M., R. Dermott, S. J. Lozano, E. L. Mills, L. R. Rudstram, and J. V. Scharold. 2007.
Evidence for remote effects of dreissenid mussels on the amphipod Diporeia: Analysis of Lake
Ontario benthic surveys, 1972-2003. J. Great Lakes Res. 33: 642-657.
CSMI Lake Michigan 2015 Report
60
-------
Table 1. Location, depth, and described substrate of sites sampling in Lake Michigan in 2015. *Stations
that were originally part of NOAA's benthic monitoring program in the southern basin in the 1990s
(Nalepaetal. 1998). See text for details.
Region/Station
Depth
Latitude
Longitude
Substrate
South
A-l*
17.3
42°06.5530
086°31.9709
sand
A-2*
29.9
42°06.0153
086°36.9776
silt and clay
A-4
72.4
42°03.4904
087°06.5073
100% mud
B-2*
49.6
42°23.9931
086°27.0413
100% mud
B-3*
62.0
42°23.9757
086°35.4838
100% mud
B-4*
126.0
42°23.5103
087°00.9441
silty clay
B-5*
102.7
42°22.5024
087°20.9581
silt and clay
B-6*
82.4
42°22.5274
087°29.9469
silt
B-7*
43.7
42°21.9742
087°39.9606
silty sand
C-l*
17.7
42°49.6624
086°14.8867
sand
C-2
45.0
42°49.6581
086°18.1607
silt, clay
C-3*
77.3
42°49.1494
086°28.4125
silt
C-45
45.2
42°09.5638
087°30.1969
silty sand
C-5*
129.0
42°48.9918
086°49.9923
silty clay
C-6*
98
42°47.6759
087°26.7942
95% silt over loam, 5% sand
C-7*
58.5
42°47.5263
087°34.4815
90% sand, 10% mud
EG-12*
54.0
42°20.8597
087°36.9207
sandy silt
EG-14*
93.3
42°22.6546
086°46.4204
100% silt
EG-18*
55.3
42°17.6162
086°38.5844
100% silt
EG-22*
46.4
43°06.1985
086°21.9813
silt
F-2
44.3
42°30.0489
086°21.8592
100% mud
F-3
71.6
42°30.1042
086°31.4951
silty mud
G-45
43.3
41°56.9564
087°13.4598
variable, mostly sand, some gravel & mud
H-8*
17.8
42°23.9597
087°46.2676
silt over loam, no Dreissena
H-9*
39.8
42°26.7390
087°42.3416
80% silt, some loam and sand
H-l 1*
69.9
42°33.2505
087°35.8191
80% silt, 20% sand
H-13*
17.9
41°55.5694
087°29.4711
90% sand, 10% shells
H-14*
34.9
42°04.3359
087°27.2110
sand
H-15*
56.2
42°09.5212
087°26.0221
silty sand
H-18*
19.8
41°58.9774
086°36.0354
silty sand
H-l 9*
34.8
42°00.0033
086°41.0855
silty ooze
H-20*
53.6
42°00.8410
086°45.1599
silty mud, ooze
H-21*
72.0
42°02.4175
086°53.0036
silty fine sediment, ooze like
H-22*
51.3
42°08.3490
086°39.8233
silt, soft
H-24*
19.0
42°23.2856
086°20.0614
100% sand
H-28*
22.3
42°37.7982
086°15.9440
100% sand
H-29*
37.1
42°37.8117
086° 18.3111
silty sand
61
CSMI Lake Michigan 2015 Report
-------
H-30*
73.5
42°37.8048
086°25
H-31*
43.0
43°02.4984
086°19
M-25
26.0
43°12.0097
086°22
M-45
42.5
43°11.4208
086°25
N-2
37.0
41°53.5031
086°52
N-3
60.1
41°57.9916
086°59
Q-13
14.2
42°50.6140
087°47
Q-30
31.0
42°50.5888
087°39
R-20
22.4
42°45.0562
087°41
R-45
47.3
42°45.0205
087°36
S-2*
10.3
41°45.9239
087°23
S-3*
26.5
41°50.9822
087°19
S-4*
40.2
41°56.0843
087°15
SAU-45
43.5
42°41.1347
086°18
T-3
71.6
42°10.0378
086°43
V-l*
17.5
41°41.7981
087°00
V-2*
28.4
41°48.9911
087°02
X-l*
35.6
43°08.2531
086°21
X-2*
100.6
43°11.9988
086°31
Central
E-l
44.9
44°37.5016
086°18
K-2
46.8
43°20.2260
086°30
KE-1
22.4
44°23.3271
087°27
KE-2
31.7
44°23.3271
087°27
KE-3
48.1
44°23.3037
087°26
KE-5
78.5
44°23.3123
087°24
L-220
21.2
43°30.0506
086°30
L-230
33.4
43°30.0446
086°31
L-245
44.0
43°30.0491
086°31
L-260
60.4
43°30.0629
086°33
L-280
80.5
43°30.0621
086°36
LU-1
22.0
43°56.6498
086°32
LU-3
44.0
43°56.6455
086°36
LU-4
62.5
43°56.6250
086°37
LU-5
78.0
43°56.6410
086°39
MAN-1
20.9
44°24.7956
086°16
MAN-2
35.9
44°24.7813
086°17
MAN-3
44.8
44°24.7729
086°19
MAN-4
58.6
44°24.8098
086°20
MAN-5
74.0
44°24.7721
086°20
PW-2
32.0
43°26.8258
087°46
PW-3
44.9
43°26.8217
087°46
PW-4
59.5
43°26.8348
087°43
.9938 black silt
.9544 silty clay
.6710 sand
.7241 50% sand, 50% mud
.0062 silt
.0004 silt
.9134 sand
.2398 90%clay, 10% sand
.7560 100% sand
.3117 90% sand, rest dressenid shells
.4838 100% fine sand
.2111 90% fine sand, 10% siltS
.1277 sand and gravel
.8971 silty ooze
.0227 silt, some sand
.7974 variable, clay, sandy silt
.9051 thick silt
.6891 variable, silt/clay, some sand
.0275 85% silt, 15% sand
.2152 85% sand, 15% mud
.0222 80% mud, 20% sand
.6720 80%sand, 10% silt, 10% dreissenid shells
.6720 Variable, mostly sand, some silt
.2201 80% sand, 20% silt
.0022 50% sand, 50% silt
.1907 sand
.1570 50% mud, 50% sand
.8934 85% mud, 15% sand
.3126 100% dark mud
.1907 100% dark mud
.1102 sand
.4846 silty sand
.6144 silty sand
.0196 70% silt, 30% sand
.8948 100% sand
.1189 80% mud, 20% sand
.8942 silty clay, sand
.3585 silty sand, clay
.8248 sandy silt, clay
.9135 80% silt, 20% fine sand
.1627 80% silt, 20% fine sand
.9985 silty clay, sand
62
CSMI Lake Michigan 2015 Report
-------
PW-5
79.0
43°26.8325
087°41.8609
silty clay, sand
SY-1
22.5
43°55.0747
087°39.8279
silty sand
SY-2
31.0
43°55.0780
087°38.8513
silt
SY-4
59.0
43°55.0786
087°30.2854
sand
SY-5
77.0
43°55.1038
087°22.5379
85% sand, 15% silt
9552
83.3
43°11.1025
087°12.5799
mud over loam
9554
109.0
43°14.2628
086°53.1725
100% mud
9556
72.9
43°18.3335
087°46.3070
silty sand
9561
130.0
43°28.2513
086°47.0433
100% mud
9562
123.0
43°29.9922
087°37.0272
silt
9564
133.0
43°36.0367
087°20.4315
silty clay
9570
165.0
43°53.1746
086°54.4904
silty mud
9574
139.0
44°04.1020
087°08.8314
tin layer mud over loam
9576
164.0
44°09.0855
086°37.2796
70% silt, 30% clay
9577
78.1
44°14.6051
087°22.4592
silty sand
9582
120.0
44°24.5028
086°22.1030
silt, detritus
9587
196.0
44°37.2816
086°21.1621
100% mud
78110
33.0
43°56.6170
086°34.7150
sand, some silt
82882
58.6
44°23.3560
087°25.3558
89% fine sand, 20% silt
82902
40.0
43 55.0900
087 37.4400
silt, fine sand
82922
17.7
43°26.8127
087°47.7663
50% fine sand, 50% silt
North
EA-7
40.0
45°16.8126
085°26.1806
silty, clay, sand
FR-1
20.0
44°48.9956
086°08.3822
mostly Dreissena druses, some sand
FR-2
32.0
44°49.0038
086°09.3452
sand
FR-3
44.0
44°49.0065
086°10.1009
mostly silt, some sand
FR-4
56.4
44°48.9911
086°11.1107
60% silt, 40% sand
FR-5
78.8
44°48.9811
086°11.7992
70% mud, 30% sand
PET-2
38.5
45°26.7409
085°04.5516
silty sand
PET-3
39.0
45°26.7319
085°11.1409
silt, clay, sand
SB-2
35.0
44°51.7024
087°09.7100
sand
SB-3
47.6
44°51.4571
087°09.0359
sand, some clay
SB-4
60.0
44°51.4272
087°08.1949
70% sand, 30% silt
SB-5
79.9
44°51.4479
087°05.1681
silt, mud
SB-6
154.0
44°51.4508
086°55.3928
80% clay, 20% silt
SC-2
29.0
45°50.4724
086°06.3233
coarse sand
SC-3
43.5
45°49.0404
086°06.3392
silt, dreissenid shells
SC-4
60.0
45°47.3931
086°06.3204
silt
SC-5
83.0
45°45.3760
086°06.3413
silty ooze
WI-1
17.4
45°14.8408
086°54.2876
sand
WI-2
31.3
45°14.8303
086°52.5656
sand
WI-3
45.4
45°14.8570
086°49.8001
sand
WI-5
85.0
45°14.8361
086°38.2513
60% silt, 40% sand
63
CSMI Lake Michigan 2015 Report
-------
9597
162.0
44°58
.3213
086°22.1965
silt with clay
74880
24.0
45°54
.5117
085°01.4952
90% mud, 10% fine sand
74900
54.3
45°26
.7280
085°13.2994
silty sand, some clay
76442
19.3
46°00
.0540
085°24.5721
dark silt
76462
64.0
45°32
.0863
085°38.1520
variable, mostly silt, some and rock
76471
31.5
45°14
.5004
085°33.3449
silty sand
76482
28.6
\l"
o
o
\l"
.1289
085°51.4266
sand
78030
33.5
Lh
o
00
.7051
085°43.0632
70% silt, 30% sand
79612
20.5
45°54
.0042
086°06.3019
coarse sand
81220
37.0
45°42
.6096
086°24.5279
sand
81240
56.0
45°14
.8459
086°40.1503
60% sand, 40% silt
82851
80.0
45°03
.0013
086°55.3601
60% clay, 40% silt
82862
13.3
44°51
.4530
087°11.3734
sand
95120
134.0
44°58
.3213
086°22.1965
silt
Green Bay
BBN-1
11.8
45°41
.9760
086°44.5177
rock and sand
BBN-2
25.0
45°37
.2398
086°44.5132
silt
BBN-3
28.6
45°32
.5008
086°44.5119
silt, alga present
LBDN-3
23.3
45°30
.0167
087°05.7984
90% sand, 10% silt
84450
10.2
45°36
.1817
087°05.7656
sand
CSMI Lake Michigan 2015 Report
64
-------
Table 2. Sites where additional Dreissena was collected for determination of length-weight relationships
in 2010 and 2015.
Depth
Interval
Year
< 30 m
2010
2015
31-50 m
2010
2015
51-90 m
2010
2015
> 90 m
2010
2015
H-18, MAN-2, PW-2, SB-2, SC-2
FR-1, H-28, M-25
B-7, H-19, MAN-3, PW-3, SB-3, SC-3
82902, B-2, B-7, FR-3, LU-3, M-45,SB-3, SC-3
FR-5, H-21, LU-5, SB-5, SC-5, SY-5
CSMI Lake Michigan 2015 Report
65
-------
Table 3. Relationship between shell length (SL in mm) and tissue ash-free dry weight (AFDW in mg) for
D. polymorpha and D. r. bugensis at various depth intervals in Lake Michigan in 2004, 2008, 2010, and
2015. Regression constants (a, b) derived from the linear regression: LogeAFDW = a+b*LogeSL; n =
total number of mussels used to derive the relationship. Also given is the AFDW of a standard 15-mm
individual as derived from the given regression. Regressions in 2004 and 2008 were from Nalepa et al.
(2010), and regressions in 2010 were from Nalepa et al. (2014). #AFDWs in 2010 were likely
underestimated by 15 % (Nalepa et al. 2014).
Year/Depth No. of
Interval (m)
Stations
Species
a
b
n
R2
15 mm
2004
<30
2
D. polymorpha
-5.256
2.672
242
0.76
7.24
31-50
2
D. polymorpha
-5.255
2.652
242
0.80
6.87
<30
2
D.
r.
bugensis
-6.095
2.968
244
0.85
6.98
31-50
2
D.
r.
bugensis
-6.969
3.316
247
0.90
7.47
2008
<30
1
D.
r.
bugensis
-6.299
3.193
199
0.92
10.46
31-50
1
D.
r.
bugensis
-5.469
2.659
193
0.93
5.65
2010#
<30
5
D.
r.
bugensis
-5.857
2.814
122
0.63
5.83 (6.70)
31-50
6
D.
r.
bugensis
-5.528
2.617
172
0.85
4.75 (5.46)
51-90
12
D.
r.
bugensis
-5.601
2.683
269
0.87
5.28 (6.07)
>90
1
D.
r.
bugensis
-5.993
2.854
24
0.98
5.67 (6.52)
2015
<30
3
D.
r.
bugensis
-5.608
2.879
77
0.92
8.92
31-50
8
D.
r.
bugensis
-5.793
2.746
211
0.88
5.17
51-90
6
D.
r.
bugensis
-5.392
2.639
153
0.91
5.78
>90
5
D.
r.
bugensis
-5.259
2.656
128
0.85
6.91
CSMI Lake Michigan 2015 Report
66
-------
Table 4. Mean (± SE) density (no./ m2) of Diporeia, Dreissenapolymorpha, and Dreissena r. bugensis at
four depth intervals (< 30 m, 31-50 m, 51-90 m, and > 90 m) in each survey year, n = number of stations
sampled, t-tests were used to determine differences between 2010 and 2015: * significant at P< 0.05, **
significant at P < 0.01. Note: Values for 2010 are slightly different than values given in Table 5 of
Nalepa et al. (2014) as some stations in Table 5 were placed into the wrong depth interval.
Year
Depth Interval/Taxa
1994-95
2000
2005
2010
2015
< 30 m
n = 16
n = 38
n = 41
n = 38
n = 291
Diporeia
3,907 ± 1,005
853 ±315
104 ± 88
1 ± 1
0±0
D. polymorpha
730 ±509
2,113 ±539
258 ± 86
0±0
0±0
D. r. bugensis
0±0
51 ±26
7,547 ± 1,566
9,254 ± 1,689
2,052 ±697**
31-50 m
n= 11
n = 36
n = 36
n = 41
n = 462
Diporeia
6,111 ± 1,377
2,116 ± 563
24 ± 16
<1 ±<1
0±0
D. polymorpha
252 ±239
1,021 ±511
427 ±109
1 ± 1
0±0
D. r. bugensis
0±0
11 ±9
15,838 ±2,860
13,133 ± 1,086
5,800 ±640**
51-90 m
n = 32
n = 41
n = 41
n = 39
n = 423
Diporeia
6,521 ±562
3,469 ± 464
548 ±131
103 ±51
1 ±<1
D. polymorpha
< 1 ±<1
16 ± 8
38 ±29
0±0
0±0
D. r. bugensis
0±0
0±0
6,472 ± 1,704
14,846 ± 1,335
8,955 ± 762*
>90 m
n = 25
n= 13
n= 13
n= 19
n= 18
Diporeia
4,547 ±385
2,804 ±453
1,244 ±217
406 ±117
528± 186
D. polymorpha
0±0
0±0
<1 ±<1
0±0
0±0
D. r. bugensis
0±0
0±0
12 ±7
2,037 ± 872
2,797 ± 824
Green Bay (< 30 m)
n = 6
n = 5
Diporeia
26 ±25
0±0
0±0
0±0
D. polymorpha
820 ± 444
80 ±53
0±0
0±0
D. r. bugensis
1 ± 1
6,640 ±3,637
5,990 ±2,140
3,797 ± 1,270
1 n=26 for Diporeia
2n=38 for Diporeia
3n=37 for Diporeia
CSMI Lake Michigan 2015 Report
67
-------
Table 5. Percentage of measured D. r. bugensis within various size categories
at four depth intervals (< 30 m, 31-50 m, 51-90 m, and > 90 m) in 2010 and 2015.
Categories based on shell length (mm). All collected mussels were measured in 2015.
In 2010, mussels were measured from representative sites (details for 2010 are given
in Nalepa et al. 2014).
Shell Length (mm)
Interval/Year ~^5 5^10 10-15 15-20 20-25 25-30 >30
< 30 m
2010
62.0
19.4
12.4
5.1
1.0
<0.1
0.0
2015
69.7
6.7
6.3
10.2
5.6
0.8
<0.1
31-50 m
2010
41.1
29.6
16.8
8.9
3.3
0.3
<0.1
2015
21.1
21.1
24.8
21.3
9.8
1.7
0.2
51-90 m
2010
55.1
17.0
17.2
8.5
1.9
0.2
<0.1
2015
38.6
21.6
18.7
15.5
4.9
0.7
<0.1
> 90 m
2010
73.5
13.9
7.7
4.5
0.3
<0.1
0.0
2015
54.4
21.1
12.2
7.9
3.8
0.7
<0.1
CSMI Lake Michigan 2015 Report
68
-------
Table 6. Mean (± SE) density (no./ m2) of major macroinvertebrate taxa at
four depth intervals (< 30 m, 31-50 m, 51-90 m, and > 90 m) at 40 sites in the
southern basin of Lake Michigan. n= 12, 10, 12, and 6 for the four intervals,
respectively.
Year
Depth
Interval/Taxa
1992-1993
1998-1999
2015
< 30 m
Diporeia
2,624 ± 568
183 ± 125
0±0
Dreissena
1,159 ± 855
1,521 ±524
627 ± 284
Oligochaeta
1,684 ±430
1,965 ±355
4,087 ± 1,265
Chironomidae
187 ±29
297 ± 46
531 ±431
Sphaeriidae
900 ±287
330± 139
87 ±45
31-50 m
Diporeia
7,857 ± 852
1,425 ± 450
0±0
Dreissena
16 ± 6
955 ±333
7,076 ± 1,639
Oligochaeta
3,050 ±315
4,077 ± 762
6,031 ± 1,248
Chironomidae
100 ± 18
52 ± 12
202± 156
Sphaeriidae
1,677 ±304
1,069 ± 181
7 ± 7
51-90 m
Diporeia
5,911 ±385
3,487 ±616
0±0
Dreissena
1 ±<1
3 ± 1
8,753 ± 1,591
Oligochaeta
1,693 ± 125
2,019 ±244
2,924 ± 650
Chironomidae
66 ± 12
28 ±7
6 ± 3
Sphaeriidae
597± 139
620 ± 68
12 ± 8
>90 m
Diporeia
3,201 ±477
3,314 ±597
207 ± 207
Dreissena
0±0
2 ± 2
5,644 ± 1,712
Oligochaeta
1,124 ± 141
996± 131
887± 196
Chironomidae
45 ± 10
26 ±7
7 ± 6
Sphaeriidae
106 ±36
175 ± 62
15 ±8
CSMI Lake Michigan 2015 Report
69
-------
Table 7. Mean (± SE) biomass (gAFDW/m2) of Dreissena at < 30 m,
31-50 m, 51-90 m, and > 90 m depth intervals based on the latest lake-wide
surveys in Lake Michigan, Lake Ontario, and Lake Huron. Given in parenthesis
is the number of stations sampled.
Dreissena Biomass (gAFDW/m2)
Depth
Lake Michigan
Lake Ontario
Lake Huron
Interval
in 2015
in 2013
in 2012
< 30 m
7.93 ± 3.30 (29)
21.53 ±7.92 (8)
2.65 ± 1.77(19)
31-50 m
26.44 ± 3.05 (46)
28.79 ±9.63 (8)
13.91 ±4.43 (30)
51-90 m
28.39 ± 1.98 (42)
20.86 ± 1.82 (8)
5.43 ±2.45 (26)
> 90 m
6.81 ±2.23 (18)
7.08 ±2.16 (21)
4.32 ±3.97 (8)
CSMI Lake Michigan 2015 Report
70
-------
Mustegan
«-=W.XP
Milwaukee*
}/
¦r
V
C-7 C-6
* •
R-20 R4S
C3 /V
SAU-45
¦
H-30 H-29/C
* * *H'28
« a
irVaufcegan ~ * 7 6^ 0
f • EG42 * *
(
M F-2
• •
•h U
Chicago
T-3
Hi/
•A "#A"1 Hf St, Joseph
H-21 /
• H.20;
' •M"1* M,<8
0-45
V
a. Southern Region
Figure la. Location of sampling sites in the southern region of Lake Michigan in 2015.
CSMI Lake Michigan 2015 Report
71
-------
1
9587 t-\
wv
WE-1 Kfr2 3 82882KE 5
MAN W AN -4VAN 3MAN /MAN-1
• • •
9582
95M
8292^W-2#PW-3 PW-4 PW 5
f\
U^r Ludington
tU-S UJ-4 IU-3 ?8110lti-l
•••• • \
(.2801-2601 ?45l 230L-220
/
Milwaukee ¦
Wf
Y—' Muskegon
b. Central Region
Figure lb. Location of sampling sites in the central region of Lake Michigan in 2015.
CSMI Lake Michigan 2015 Report
72
-------
Manistique
ruT'
/
/
>J
#76442 W
1 A s\
AS
796)2
¦4
74880 (
ft
> 1 / • / 5^
^Vr c*
f C r*
j
81220
SC-2.
SC }•
SC-*'
•
scs
780 JO'
•
CT )
1 t/ y
•
" vxi>
CN ft •
U ' ex
^ ""**V
) y
# BBN-2 •
J 84450 (v
BBN-3 •
9S120
/
76462
•
LBDIM-1*
y « r\
N
74900PET-3 PJT-J
IY ,
k/
WltWI 2W1-3 8I240WI S
-fyT
•
Petoskey
s- xr
EA-7
• ' '
76471 / % \
\ lr 4
/ / •
J 1
J ft
76482
•
fs
#828S1
f \
9597
•
J An
[ Y\
-n j \\
'( /A Vs
; 82$*ftfr 2 p-i se-4 sb s ie-«
\
FRS FR 4 FR^FR-2 FR-I
r 0 /
I A
1! j
? AV )
7%
/
Traverse City
c. Northern Region
Figure lc. Location of sampling sites in the northern region of Lake Michigan in 2015.
73
CSMI Lake Michigan 2015 Report
-------
1994/95
if
I
v
10 1®' 10' 10* 10'
Density (No. in:)
Figure 2. Density (no. per m2) of Dreissencipolymorph® in Lake Michigan based on lake-wide surveys in
1994/1995, 2000, 2005, 2010, and 2015. Small red dots indicate location of sampling sites.
Density (No. m*)
Density (No. mi
Density (No. m;)
Density (No.
CSMI Lake Micliigan 2015 Report
74
-------
1994/95
Density (No. m-'>
Density (No. m:>
Density (No. m:)
Figure 3. Density (no. per m2) of Dreissenci r. bugensis in Lake Michigan based on lake-wide surveys in
1994/1995, 2000, 2005, 2010, and 2015. Small red dots indicate location of sampling sites.
CSMI Lake Michigan 2015 Report
75
-------
19 90 19 9 5 20 0 0 2 00 5 2010 2015 2020
Year
20 0 00
CM
E
i_ 15000
4>
Q.
o
c
—' 10000
£•
u>
c
a>
d 5000
<30 m
31-50 m
rj
E
i_
4>
Q.
<
3
(A
(A
n
E
o
m
35 ¦
30
25
20 -
15
10
5 -
<30 m
31-50 m
51-90 m
> 90 m
1990 1995 2000 2005 2010 2015
Year
2020
Figure 4. Long-term trends of total Dreissena in Lake Michigan in 1994/1995, 2000, 2005, 2010, and
2015. Values given are lake-wide means (± SE) at four depth intervals: < 30 m (black, circles), 31-50 m
(red, triangles), 51-90 m (blue, squares), and > 90 m (green, diamonds). Upper panel = density; lower
panel = biomass.
76
CSMI Lake Michigan 2015 Report
-------
Year
Figure 5. Ash free dry weight (AFDW, mg) of a standard 15-mm D. r. bugensis at four depth intervals
intervals in Lake Michigan between 2004 and 2015. Values derived from regressions given in Table 4.
Depth intervals: < 30 m (black, circles), 31-50 m (red, triangles), 51-90 m (blue, squares), and > 90 m
(green, diamonds).
CSMI Lake Michigan 2015 Report
77
-------
6000
CN
"qj 5000 -
4000 -
3000 -
2000 -
••
1000 -
0
20
40
60
80
100
Ash-free dry weight (g/m2)
Figure 6. Relationship between ash free dry weight (AFDW) and total wet weight (TWW, whole mussel,
tissue and shell) of D. r. bugenisis at each sampling site in the main basin of Lake Michigan in 2015
(n=135). The regression through the origin was defined as: TWW = 50.09*AFDW (R2 = 0.92)
CSMI Lake Michigan 2015 Report
78
-------
1994/95
2000
2005
2010
2015
0 ) t » « 15
Density I No. »m! X 10')
i),
0 J 6 9 12 IS
D#n$ttyfN0. m! x 10')
D
Figure 7 Density (no. per m ) of Diporeia spp.in Lake Michigan based on lake-wide surveys in
1994/1995, 2000, 2005, 2010, and 2015. Small red dots indicate location of sampling sites.
CSMI Lake Michigan 2015 Report
79
-------
V)
y>
£
Q
14000
12000 -
lOCflfl ¦
0000 -
6000 -
4000
2000
<30 ill
A
/ \r \
» / A v
// \ x
\ // \ A.
\ / \ / '¦¦¦
Jim \ / h
¦"
\y . l
1900
1095
1990
1995
3000
2500
2000
1500
1000
500
2000
2005 2010 2015
2020
14000
31-90m
12000 ¦
10000 •
r~\
£000 ¦
J \
6000 ¦
// *
/1 ¦
40 00 ¦
/ 1 *
J i
2000 ¦
0 ¦
\ /
1 1 ^ * A
Af 1 1 1
2000
2005
2010
2015
2020
> 90 ni
Lake Ontario
Lake Michigan
Lake Hui oil
1990
1995 2000
2005 2010 2015
Year
2020
Figure 8. Density (no. per m2) of total Dreissena at < 30 m (upper panel), 31-90 m (middle panel), and >
90 m (lower panel) in Lake Ontario (black, circle), Lake Michigan (blue, square), and Lake Huron (red,
triangle). Values taken from the following sources: Lake Ontario ( Birkett et al. 2015, Nalepa and Elgin
unpublished), Lake Michigan (Nalepa et al. 2014, this study); Lake Huron (Nalepa et al. 2007, Nalepa
unpublished). Note the different scale for the > 90 m interval.
80
CSMI Lake Michigan 2015 Report
-------
A summary of Mid-Continent Ecology Division Efforts Associated With the 2015 Lake
Michigan Cooperative Science Monitoring Initiative (CSMI)
• Water Quality and Lower Trophic Level Summary from the 2015 Lake Michigan CSMI -
Anett Trebitz, Anne Cotter, Joel Hoffman
USEPA Mid-Continent Ecology Lab, Duluth, MN
• Application of a Nutrient Model to Address Nearshore Phosphorus Levels in Lake
Michigan - James Pauer, Terry Brown, Tom Hollenhorst
USEPA Mid-Continent Ecology Lab, Duluth, MN
• "Data in Motion" - Continuous Water Sensor Data Collection for the 2015 Lake
Michigan CSMI - Tom Hollenhorst1, Laura Fiorentino2, Paul McKinney1, Terry Brown1,
Anett Trebitz1, Joel Hoffman1
1USEPA Mid-Continent Ecology Lab, Duluth, MN
2NOAA - Center for Operational Oceanographic Products and Services, Chesapeake, VA
CSMI Lake Michigan 2015 Report
81
-------
Water quality and lower trophic level summary from the 2015 Lake Michigan CSMI
Anett Trebitz, Anne Cotter, Joel Hoffman,
U.S. EPA Office of Research and Development, Mid-Continent Ecology Division, Duluth MN.
Background
Among the major research questions identified for the 2015 Lake Michigan Cooperative Science
Monitoring Initiative (CSMI) was to better understand the distribution, abundance, and movement of
nutrients and biota across nearshore-offshore gradients. To address this question, a comprehensive
suite of biota and water quality data were collected according to a sampling design that consisted of
eight onshore-offshore transects representing a gradient in nutrient loading (Fig. 1). Three of the
transects were placed away from tributaries (e.g., "no load"), two were positioned at "low load"
tributaries, and three were positioned at "high load" tributaries. The loads were determined by
averaging, across 2002 through 2004, the annual total phosphorus loading estimates generated by
Dolan and Chapra (2012). To resolve spatial and temporal dimensions of this nearshore-offshore
gradient, each transect was sampled at a shallow (~18m) mid (~46m) and deep (~110m) station with
sampling repeated in each of three seasons (May, July, September). Sampling was accomplished via an
interagency collaboration that included U.S. Environmental Protection Agency (EPA), National Oceanic
and Atmospheric Administration (NOAAO, and the United States Geological Survey (USGS) as well as
various academic partners.
Data collection under CSMI included both traditional station-based sampling (e.g., for water and biota)
as well as continuous transect-based sampling (e.g., with towed sensor arrays or glider technology).
Here, we examine station-based data from the 2015 Lake Michigan CSMI with respect to water quality
patterns (e.g., nutrients, planktonic chlorophyll) and lower trophic level patterns (carbon and nitrogen
stable isotopes as food-web tracers). For context, we also draw on long-term monitoring data from Lake
Michigan that is collected annually by the EPA Great Lakes National Program Office ("GLNPO monitoring
data", hereafter). This long-term monitoring focuses on mid-lake locations, in contrast to the CSMI 2015
effort which sought specifically to elucidate the nearshore-to-offshore pattern.
Spatial patterns and temporal trends in water quality
The 2015 CSMI sampling design was intended to contrast nearshore vs. further offshore water quality
across transects, and permit examining the influence of differing tributary loading (including absence of
tributaries) on open-lake water quality. Since watershed landuse is a major driver of nutrient loading to
the Great Lakes (Han etal. 2011, Robertson and Saad 2011) and since hydrodynamic features such as
shore-parallel thermal bars and longshore currents tend to "trap" watershed-derived inputs relatively
close to shore (e.g., Bolgrien and Brooks 1992, Beletsky and Schwab 2008, Yurista et al. 2015), we had
expected to find among-transect differences in productivity (e.g., nutrients and planktonic chlorophyll)
to be associated with tributary loading, and to find an onshore-offshore gradient in which shallow
stations had higher nutrient and chlorophyll concentrations than deeper stations.
82
CSMI Lake Michigan 2015 Report
-------
There was a suggestion of a positive relationship across transects between tributary phosphorus load
and in-lake surface water total phosphorus (TP) concentration in spring (May data, Fig 2a) but none of
the lines had significant linear regression slope (see Fig. 2 caption), and in summer (July) TP showed no
relationship to tributary TP load at any station depth (Fig. 2b). In contrast, the slope of chlorophyll a
(CHLA) in spring relative to tributary load was flat for all stations except the shallow one (Fig. 3a), but in
summer CHLA had a positive relationship to tributary load at all station depths (Fig. 3b). The 2015 CSMI
data did not support the expectation that TP would be higher at nearshore relative to deeper stations;
rather the shallowest stations (in green in Fig. 2) had slightly lower TP levels than the deep station in
both May and July. However, CHLA concentrations were slightly higher at shallow compared to mid-
depth and deep stations, and the slope of the relationship to tributary loading was steeper at the
shallow vs. the mid-depth or deep stations in both spring and summer (Fig. 3).
We did not present data for September in Figures 2 and 3 because occurrence of wind-driven upwellings
complicate the interpretation of nearshore to offshore gradients (and because logistic constraints
prevented sampling of two transects). Four of the six transects that were sampled in September had
substantially colder surface water at the shallow than at the deep station, indicating cold benthic water
being "pulled" up-slope behind surface waters that is being driven away from shore by strong lateral
winds. Such upwelling events can be frequent and prolonged in Lake Michigan (Plattner et al. 2006),
and the September CSMI data, in which three of the four transects affected by upwelling are on the
western side, is consistent with the Plattner et al. finding upwelling more prevalent along the western
than eastern shoreline.
As mentioned above, EPA's GLNPO collects data annually for a set of offshore Lake Michigan stations,
comparable to the deep CSMI stations but not to the mid-depth and shallow CSMI stations. Publications
to date summarize these data only through 2012 (e.g., Barbiero et al. 2012, Dove and Chapra 2015,
Mida et al. 2010), so for context to the 2015 CSMI we include plots here that extend the GLNPO station
timeseries through 2015. For TP, the plots (Fig. 4) continuance confirm patterns already noted in the
above publications, namely that concentrations remain low relative to the 1980s and early 1990s, that
summer concentrations remain slightly lower than spring concentrations, and that differences between
deep (~100m stations) and extra-deep stations (~150m) early in the time series are no longer evident.
For CHLA, the extended the time series (Fig. 5) confirms that concentrations have remained
approximately level over the last decade, that intra-annual variability remains lower now than during
the 1980s and early 1990s, that marked differences between spring and summer CHLA early in the time
series no longer exist, and that concentrations at deep stations remain slightly less than at extra-deep
stations.
In contrast to the strong data record for offshore station, very little nearshore data is archived in the
GLNPO long-term dataset. Spatially well-distributed data for regions of Lake Michigan shallower than
100 m are available only for the early 1980s. The 2015 CSMI thus provides a picture of nearshore spatial
patterns in Lake Michigan that has been unavailable for many years. For TP (Fig. 6 left-hand panels),
boxplots show no indication of levels being higher at shallower relative to deeper stations in either
spring or summer and in either 1983 (GLNPO monitoring data) or 2015 (CSMI data). In other words, the
TP data show no evidence for the expectation that nutrient levels would be highest closest to the
shoreline from which the loading presumably emanates. In contrast, patterns among nearshore depths
are quite different between 1983 and 2015 and between seasons for CHLA (Fig. 6 right-hand panels). In
1983 (GLNPO monitoring data), box-plots for both spring and summer show a steadily increasing level of
83
CSMI Lake Michigan 2015 Report
-------
CHLA from the deep to mid-depth to shallow stations, with the CHLA magnitude and the difference
among depths being largest in spring. In 2015 (CSMI data), box-plots do not show CHLA differences
among depths in spring and the pattern of differences among depths in summer is the opposite of the
pattern seen in 1983.
Taken together, the water quality spatial and temporal patterns presented here are consistent with
previous findings (e.g., Fahnenstiel et al. 2010, Rowe et al. 2017, Yousef et al. 2017) of long-term
declines in TP and CHLA and temporal (spring vs. summer) decoupling of nutrients to plankton
production coincident with the arrival of dreissenid mussels. The results also provide support for the
prediction that dreissenid mussels would substantially alter spatial patterns of water quality across the
nearshore-offshore gradient (Hecky et al. 2004, Nalepa et al. 2010, Yousef et al. 2017). However rather
than increasing nearshore CHLA as might be expected were dreissenids increasingly retaining and cycling
nutrients in the nearshore (the "nearshore shunt" hypothesis), our findings suggest that dreissenid
action is instead producing a homogenization of conditions horizontally (i.e., across depth gradients) in
Lake Michigan - which may also be responsible for the quite muted tributary loading response we
observed.
Spatial patterns and temporal trends in biota stable isotopes
We also examined spatial patterns of the carbon and nitrogen stable isotope composition in
zooplankton, quagga mussel (Dreissena bugensis), and other benthic invertebrates (oligochaetes). The
carbon stable isotope composition (i.e., its 513C value) is a diet tracer; the 513C value of an organism
reflects the isotopic composition of its prey ("you are what you eat"; Peterson and Fry 1987). If
organisms have different 513C values, this reveals they are eating different prey. In contrast, the nitrogen
stable isotope composition (515N value) traces both nitrogen source and trophic level. Previous studies
in the Great Lakes demonstrated that an enrichment of 1SN in the food web (as measured by tissue
samples of invertebrates, fish, or both) in either coastal wetlands or nearshore waters (<10 m depth)
was related to increased dissolved nutrients in coastal waters, as well as increased urban and
agricultural activity in the watershed (Peterson et al. 2007, Hoffman et al. 2012). If these human sources
of nitrogen strongly influence the food web, then we expect higher 51SN values in organisms captured in
the nearshore compared to the offshore, with the highest 51SN values associated with high N
concentrations near rivers with relatively large nutrient loads to Lake Michigan (e.g,. St. Joseph River).
Moreover, there is a consistent enrichment with 1SN with trophic level, such that consumers that feed
higher in a food web will have higher 51SN values than those feeding lower on the food web (Vander
Zanden and Rasmussen 2001).
Here, we focus on large zooplankton (those that were retained by a filter with a 153 |a,m nominal pore
size) and adult quagga (those with a shell length >15 mm), which were generally captured at most or all
stations sampled. Oligochaetes were also found in sediment dredge samples, though at only a few
stations. Among taxa, we observed substantial isotopic variation associated with depth, with a shift from
relatively low 51SN values and high 513C values in shallow, nearshore waters (18 m depth) to higher 51SN
values and lower 513C values in offshore waters (110 m depth; Fig. 7). For both zooplankton and quagga
mussel, there was a statistically significant different among depths in both 513C and 515N values (ANOVA;
513C zooplankton: p=0.003, 513C quagga: p=0.048, 51SN zooplankton: p=0.030, 51SN quagga: p<0.001;
84
CSMI Lake Michigan 2015 Report
-------
data generally met assumption of normality except quagga 513C values, for which a Kruskal-Wallis test
was used). The magnitude of the shift in 51SN values is much larger than that in 513C values. These 51SN
patterns are consistent with previous observations regarding nitrogen cycling in Lake Michigan, where
offshore enrichment in 1SN reflects is likely caused by relatively high denitrification rates in sediment
(Gardner et al. 1987). The shift in 513C is consistent with either a greater contribution of nearshore
carbon sources to the food web (benthic algae) or higher production compared to offshore waters (or
both; Sierszen et al. 2014). The relatively low 51SN values in nearshore water suggests that sampling at
this depth (18 m) was not sensitive to inputs of anthropogenic nitrogen. Among taxa, zooplankton and
quagga mussel were much more isotopically similar to each other than to the oligochaetes, which were
relatively 1SN- and 13C-enriched (Fig. 7). The oligochaete we sampled were embedded among the byssal
threads of the quagga mussels. The enriched isotopic composition implies they were feeding in a
microbial food web, consuming either particles colonized and processed by sediment bacteria, the
bacteria itself, or both. This is a distinct food web pathway from zooplankton and quagga mussel.
Comparing the zooplankton and quagga mussel, the question arises as to whether they are consuming
the same food items, as indicated by their isotopic composition. We found a seasonal shift in the
difference between zooplankton and quagga mussel (Fig. 8). In May, their isotopic composition was
well-differentiated, indicating they were not feeding on the same algae or particles during the winter
and spring. However, as summer progressed, they become increasingly isotopically similar, such that
they were not different from each other by September. The observation that zooplankton and quagga
mussel have relatively similar 513C and 51SN values indicates they are potentially in competition for the
exact same food source, presumably phytoplankton and sinking organic material largely composed of
phytoplankton. Based on paired net samples, the zooplankton samples were largely composed of
calanoid copepods (89.%, 75.5%, and 63.2% of biomass on average for the May, July, and September
sampling events). The calanoids were mostly large species, including the grazer Leptodiaptomus sicilis
and the omnivorous consumers Senecella calanoides and Limnocalanus macrurus. If they are in
competition (i.e., particle density is limiting to growth), this implies that quagga mussel could exert a
direct effect on the entire lake food web via competition with zooplankton. That is, quagga have the
potential to reduce the energy available to the pelagic food web, which includes zooplankton, alewives,
bloater, and Pacific salmon. The finding is notable because we would expect that quagga mussel and
zooplankton are most likely to compete for food in the winter and spring when particle resuspension
and mixing is occurring (Eadie et al. 1984), rather than in the summer when the lake is stratified.
References
Barbiero R.P., Lesht B.M. and Warren G.J. 2012. Convergence of trophic state and the lower food webs
of lakes Huron, Michigan, and Superior. J. Great Lakes Res. 38: 368-380.
Beletsky, D., and Schwab, D. 2008. Climatological circulation in Lake Michigan, Geophys. Res. Lett., 35,
L21604, doi:10.1029/2008GL035773.
Bolgrien, D.W. and Brooks, A.S. 1992. Analysis of thermal features of Lake Michigan from AVHRR
satellite images. J. Great Lakes Res. 18: 259-266.
85
CSMI Lake Michigan 2015 Report
-------
Dolan, D.N., and Chapra, S.C. 2012. Great Lakes total phosphorus revisited: 1. Loading analysis and
update (1994-2008). J. Great Lakes Res. 38: 730-740.
Dove, A., and Chapra, S. C. 2015. Long-term trends of nutrients and trophic response variables for the
Great Lakes. Limn. Ocean. 60(2): 696-721.
Eadie, B.J., Chambers, R.L., Gardner, W.S., and Bell, G.L. 1984. Sediment trap studies in Lake Michigan:
resuspension and chemical fluxes in the southern basin. J. Great Lakes Res. 10:307-321.
Fahnenstiel, G. L., S. Pothoven, H. Vanderploeg, D. Klarer, T. Nalepa, and D. Scavia. 2010. Recent changes
in primary production and phytoplankton in the offshore region of southeastern Lake Michigan. J. Great
Lakes Res. 36(s3): 20-29.
Gardner, W.S., Nalepa, T.F., and Malczyk, J.M. 1987. Nitrogen mineralization and denitrification in Lake
Michigan sediments. Limnol. Oceanogr. 32:1226-1238.
Han, H., Bosch, N. and Allan, J.D., 2011. Spatial and temporal variation in phosphorus budgets for 24
watersheds in the Lake Erie and Lake Michigan basins. Biogeochemistry, 102: 45-58.
Hecky, R. E., R. EH Smith, D. R. Barton, S. J. Guildford, W. D. Taylor, M. N. Charlton, and T. Howell. 2004.
The nearshore phosphorus shunt: a consequence of ecosystem engineering by dreissenids in the
Laurentian Great Lakes. Canadian J. Fish. Aquatic Sci. 61: 1285-1293.
Hoffman, J.C., Kelly, J.R., Peterson, G.S., Cotter, A.M., Starry, M., and Sierszen, M.E. 2012. Using 515N in
fish as an indicator of watershed sources of anthropogenic nitrogen: response at multiple spatial scales.
Estuaries and Coasts. 35:1453-1467.
Mida, J. L., Scavia, D., Fahnenstiel, G. L., Pothoven, S. A., Vanderploeg, H. A., & Dolan, D. M. 2010. Long-
term and recent changes in southern Lake Michigan water quality with implications for present trophic
status. J. Great Lakes Res. 36(s3): 42-49.
Nalepa, T. F., Fanslow, D. L., & Pothoven, S. A. 2010. Recent changes in density, biomass, recruitment,
size structure, and nutritional state of Dreissena populations in southern Lake Michigan. J. Great Lakes
Res. 36(s3): 5-19.
Peterson, B.J., and Fry, B. 1987. Stable isotopes in ecosystem studies. Ann. Rev. Ecol. Syst. 18: 293-320.
Peterson, G.S., Sierszen, M.E., Yurista, P.M., and Kelly, J.R. 2007. Stable nitrogen isotopes of plankton
and benthos reflect a landscape-level influences on Great Lakes coastal ecosystems. J. Great Lakes Res.
33(s3): 27-41.
Plattner S., Mason D.M., Leshkevich G.A., Schwab D.J., Rutherford E.S. (2006), Classifying and
forecasting coastal upwellings in Lake Michigan using satellite derived temperature images and buoy
data, J. Great Lakes Res. 32: 63-76.
Robertson, D.M. and D.A. Saad, 2011. Nutrient Inputs to the Laurentian Great Lakes by Source and
Watershed Estimated Using SPARROW Watershed Models. J. Am. Water Resources Assoc. 47: 1011-
1033.
86
CSMI Lake Michigan 2015 Report
-------
Rowe, M. D., Anderson, E. J., Vanderploeg, H. A., Pothoven, S. A., Elgin, A. K., Wang, J., and Yousef, F.
2017. Influence of invasive quagga mussels, phosphorus loads, and climate on spatial and temporal
patterns of productivity in Lake Michigan: A biophysical modeling study. Limn. Ocean. 62: 2629-2649.
Sierszen. M.E., Hrabik, T.R., Stockwell, J.D., Cotter, A.M., Hoffman, J.C., and Yule, D.L. 2014. Depth
gradients in food-web processes linking habitats in large lakes: Lake Superior as an exemplar ecosystem.
Freshwater Biol. 59:2122-2136.
Vander Zanden, M.J., and Rasmussen, J.B. 2001. Variation in 61SN and 613C trophic fractionation:
Implications for aquatic food web studies. Limnol. Oceanogr. 46: 2061-2066.
Yousef, F., Shuchman, R., Sayers, M., Fahnenstiel, G., and Henareh, A. 2017. Water clarity of the Upper
Great Lakes: Tracking changes between 1998-2012. J. Great Lakes Res. 43: 239-247.
Yurista, P. M., Kelly, J. R., Cotter, A. M., Miller, S. E., and Van Alstine, J. D. 2015. Lake Michigan:
Nearshore variability and a nearshore-offshore distinction in water quality. J. Great Lakes Res. 41: 111-
122.
CSMI Lake Michigan 2015 Report
87
-------
Figure 1. Transects sampled by CSMI in 2015 and associated rivers and loading categories, overlain on a
map of TP loading differences among shoreline locations. The map is excerpted from the Great Lakes
Environmental Assessment Mapping Project website (http://data.glos.us/gleam/lake-
stressors/nonpoint-pollution/phosphorus-loading.html), and is based on TP tributary loading data
compiled over 1994-2008. and propagated into the lake assuming distance-based decay.
Sturgeon ->
(ndriver, no load))
<- Frankfort
(no river, no load)
Manitowoc -> (Manitowoc
R, high load)
<- Ludington
(Pere Marquette R., med load)
Racine ->
(Root R., med load)
<- Saugatuck
(Kalamazoo R. - high load)
Waukegan ->
(no river, no load)
<- St Joseph
(St. Joseph R., high load)
88
CSMI Lake Michigan 2015 Report
-------
Figure 2. Graph showing relationship between surface-water TP measured on the May or July 2015 CSMI
cruises vs. 2002-2008 avg annual TP loading from the adjacent tributary (the three no-tributary transects
were assigned to zero load). The slopes of the lines in panel a) are not significantly different from zero
in a linear least-squares regression.
Tributary TP load (metric annual tons)
a) May
15
b)July
20 t
• deep
± mid
shallow
89
CSMI Lake Michigan 2015 Report
-------
Figure 3. Graph showing relationship between surface-water planktonic CHLA measured on the May or
July 2015 CSMI cruises vs. 2002-2008 avg annual TP loading from the adjacent tributary (the three no-
tributary transects were assigned to zero load). The slopes of the shallow-depth line in panel a) is
significantly different from zero in a linear least-squares regression (p=0.048) and the deep and shallow
lines in panel b) have slopes significantly different from zero (<0.001 and 0.018, respectively) whereas
the mid-depth line slope is not quite significant (p=0.071).
CD
3
C
O
03
C
<1)
o
c
o
o
a) May
2.0 -r
1.5
1.0
0.5
0.0
X
o
aj
"S
a)
o
4=
3
CO
b) July
4 -r
2 —
1 -h
0
0
100 200 300 400
Tributary TP load (metric annual tons)
90
• deep
± mid
shallow
CSMI Lake Michigan 2015 Report
-------
Figure 4. 1982-2015 time-series of upper-water column total phosphorous concentrations (mean ± 1
standard deviation) from a set of long-term monitoring stations (i.e., GLNPO monitoring data) measured
in either spring (top panel; April or May) or summer (bottom panel; August or September). Stations
were classified by lake depth as either deep (~100 m) or extra deep (<150 m). Plotted data meet the
criteria of being from stations sampled 10+ years, from the epilimnion or upper water column layers,
and from either late-spring (April or May) or late-summer (Aug/Sep).
Spring (Apr/May)
O)
3
12 T
8 --
4 --
0
deep (around 100m)
extra-deep (>150 m)
I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I
I 1 1 1 1 I
o5 Summer (Aug/Sep)
>. 15 -r
co lo '
CD
Q.
Q.
3 10 + T
5 "
0
iiM i5&
I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I
I 1 1 1 1 I
h
H
1982 1987 1992 1997 2002 2007 2012 2017
91
CSMI Lake Michigan 2015 Report
-------
Figure 5. 1982 to 2015 time-series of upper-water column chlorophyll a (CHLA) concentration (mean ± 1
standard deviation) from a set of long-term monitoring stations (i.e., GLNPO monitoring data) measured
in either spring (top panel; April or May) or summer (bottom panel; August or September). Stations
were classified by lake depth as either deep (~100 m) or extra deep (<150 m). Plotted data meet the
criteria of being from stations sampled 10+ years, from the epilimnion or upper water column layers,
and from either late-spring (April or May) or late-summer (Aug/Sep).
6 t Spring (Apr/May)
4 "
2 "
0
6
< >"
• deep
• extra-deep
f'ii
I'filW
0
Summer (Aug/
Sep)
- ¦ ¦ ¦ ¦ 1 ¦ ¦ ¦ ¦
;j jii'tiiiiiil'ijifii
1 ¦ ¦ ¦ ¦ 1 ¦ ¦ ¦ ¦ 1 ¦ ¦ ¦ ¦ 1 ¦ ¦ ¦ ¦ 1 ¦ ¦ ¦ ¦ 1
1982 1987 1992 1997 2002 2007 2012 2017
92
CSMI Lake Michigan 2015 Report
-------
Figure 6. Box plots comparing levels of TP (left panels) and CHLA (right panels) across depths in 1983
(GLNPO monitoring data) relative to 2015 CSMI stations. Stations were classified as shallow if bottom
depth was <30m, mid-depth if 30-60m, and deep if 70-120m. Top panels are for spring, bottom panels
for summer. Data for multiple sampling depths (e.g., upper water-column as well as near-bottom) are
included. Because the GLNPO database has historic data for mid-depth and shallow stations only in
1983, this comparison is not possible for intervening years.
25-1
20"
cn 15-
3
io-
5-
spring
4 -|
|-L.
J
11
*
o
40—i summer
o
30-
CT>
3 20-
10-
*
1983
.3-
CT>
3
<2^
o
1
0
5-
spring
CT>
33
<
_l
X2
1H
1—
summer
~ deep
n mid
shallow
2015
1983
* *
2015
CSMI Lake Michigan 2015 Report
93
-------
Figure 7. Carbon (A) and nitrogen (B) stable isotope ratios of quagga mussel, zooplankton (Zoop.), and
oligochaetes (Oligo.) at the 18 m (left), 46 m (center), and 110 m (right) stations. Sample sizes are given
in italics (same for 513C and 51SN values). Box plots show median, quartiles, and outliers.
94
CSMI Lake Michigan 2015 Report
-------
Figure 8. The difference between zooplankton and quagga mussel carbon (A) and nitrogen (B) stable
isotope ratios by month (May, July, September). Values below the line indicate zooplankton have a
lower 513C or 51SN value than quagga mussel, and vice versa for values above the line. Box plots show
median, quartiles, and outliers.
TT
O)
O)
Cl
-10
-Q-
May
Jul
Sep
95
CSMI Lake Michigan 2015 Report
-------
Application of a Nutrient Model to Address Nearshore Phosphorus Levels in Lake
Michigan
James Pauer, Terry Brown, Tom Hollenhorst,
U.S. EPA Office of Research and Development, Mid-Continent Ecology Division, Duluth MN.
Introduction
Concerns with the nearshore water quality of the Laurentian Great Lakes, such as excessive
eutrophication and harmful algal blooms, called for establishing a nearshore monitoring program
and an improved understanding of the watershed-nearshore link (Great Lakes Water Quality
Agreement, 2012). This is challenging, as sporadic runoff events and varying circulation patterns
cause the nearshore to be dynamic and exhibit large spatial and temporal gradients. The Grand
River is the largest discharger of phosphorus directly into Lake Michigan with the potential of
causing high levels of nutrient and phytoplankton at the discharge point which is close the Grand
Haven, Michigan (Figure 1). Mathematical models are powerful tools to understand nearshore
nutrient circulation, identify the main drivers of phosphorus in the nearshore, and to assist
stakeholders with management options to improve or maintain water quality. The objective of
this study is to investigate the impact of tributary phosphorus loadings and lake circulation on the
nearshore areas of southeastern Lake Michigan using a mathematical model and observational
data. This study will explore how these drivers influence the nearshore phosphorus levels
temporally and spatially. This work will be expanded in the future to address nearshore algal
dynamics. A second study objective is to develop a simple, transparent and transportable tool to
be easily applied in other ecological sensitive areas in Lake Michigan and the other Great Lakes.
Methods
A study area on the southeastern side of Lake Michigan was selected around the Grand and
Muskegon rivers, two tributaries that contribute substantial phosphorus loadings to the eastern
shore of the lake. The study area focuses on the nearshore within approximately 25 miles from
the coast. The model has a computational grid of 5476 small (1km x 1km) completely mixed
cells (Figure 1). The nearshore circulation (hydrodynamics) was provided by the US Naval
Research Lab (Stennis, MS). The model used simple phosphorus kinetic equations similar to the
lakewide phosphorus model developed by Chapra and Dolan (2012). The Cooperative Science
Monitoring Initiative (CSMI) results were used to estimate boundary and initial conditions, and
to ground-truth the model. No tributary loadings measurements were available beyond 2008, and
therefore loadings were estimated looking at historical loading trends (Dolan and Chapra, 2012,
Rossmann, 2006) and limited tributary measurements from the 2015 Lake Michigan CSMI. Due
to the uncertainty with these loading estimates, we also investigating the impact of alternative
tributary loads on model results.
Results and Discussion
96
CSMI Lake Michigan 2015 Report
-------
Using the estimated (base) total phosphorus (TP) loadings, this relatively simple model
somewhat over-predicts the 2015 TP observations. (Figure 2). The model prediction was higher
than the observations at a number of locations very close to the Grand River. Model prediction
that were much higher than the corresponding observations were also much more variable,
intimating how dynamic and variable the nearshore can be (top-left of Figure 2). A possible
reason for this discrepancy is a too high TP loading estimate for 2015, although the timing of the
loading and the accuracy of the circulations might also contribute to the high model predictions.
Running the model with lower TP loads (30% reduction of the base loads), the results has a
tighter fit to the observations, although it still over-predicts the same observations close to the
Grand River (Figure 3). We will further investigate the sensitivity of the loading estimate,
timing of the loads, and lake circulation on the model results using the 2015 CSMI dataset, as
well as data from other, previous sampling efforts.
Figure 4 shows spatial patterns of phosphorus in the nearshore of Lake Michigan at two dates in
summer of 2015. The model results show that the nearshore close to the river discharges and
within 1-3 km of the shore is often strongly impacted by the river's loads, but the nearshore
phosphorus pattern can also change significantly. Phosphorus at deeper areas of the nearshore
and further away from the discharge locations were much lower and approaching off-shore TP
concentrations. Future analyses will investigate how TP loadings and circulation patterns impact
the nearshore concentrations spatially and temporally.
This model, with limitations such as using simple phosphorus kinetic formulations and estimated
rather than measured phosphorus loadings, demonstrates that Lake Michigan tributaries
(watershed loading) can cause high phosphorus concentrations in the nearshore, but that it is
limited to zones of impact that can change relatively rapidly depending on the nearshore
circulation. We believe values from the model summarized spatially and temporally have
potential to help guide future nutrient criteria development efforts for the near shore. This work
also demonstrate that a simple model can be useful in guiding managers in making water quality
decisions, and such a model can easily be applied to other locations in Lake Michigan and the
Great Lakes. However, these models need to be thoroughly tested at these locations before any
model predictions should be made.
References
Chapra, S.C., Dolan, D.M., 2012. Great Lakes total phosphorus revisited: 2. Mass balance
modeling. J. Great Lakes Res. 38 (4), 741-754.
Dolan, D.M., Chapra, S.C. 2012. Great Lakes total phosphorus revisited: 1. Loading analysis and
update (1994-2008). J. Great Lakes Res. 38 (4), 730-740.
Great Lakes Water Quality Agreement. Protocol Amending the Agreement between Canada and
the United States of America on Great Lakes Water Quality, 1978, as Amended on October 16,
1983, and on November 18, 1987, Signed September 7, 2012, Entered into force February 12,
2013. http://ijc.org/files/tinymce/uploaded/GLWQA%202012.pdf (2012)
97
CSMI Lake Michigan 2015 Report
-------
Rossmann, R., 2006. Results of the Lake Michigan Mass Balance Project: Polychlorinated
Biphenyls Modeling Report. USEPA, Large Lakes Research Station, Grosse lie, MI (2006), p.
621
LMMB Project, 2006. Results of the Lake Michigan Mass Balance Project: Polychlorinated
Biphenyls. (Ed. Ronald Rossmann), pp579, Grosse lie, MI
CSMI Lake Michigan 2015 Report
98
-------
Figure 1: Computational grid (1km x 1km) of the study area. The area is southeast Lake
Michigan adjacent to the Muskegon and Grand Rivers.
CSMI Lake Michigan 2015 Report
99
-------
V/
i+f+
1:1
• < 20 m
* < 20 m, noon
• >=20 m
* > = 20 m, noon
10 20 30 40
Observed values
50 60
L
Obs. freq.
L
Model freq.
residuals
Figure 2: Model phosphorus results versus observational data (in |ig/L): Base TP load. The
circled area highlights areas where model prediction was much higher and variable than the
observations.
60-
50
40
30
20-
10-
0-
A
i>
v-.
¦ ii i /
••
10
1:1
< 20 m
< 20 m,noon
>= 20 m
>= 20 m, noon
L.
Obs. freq.
L
Model freq.
MIL
residuals
20 30 40
Observed values
50
60
Figure 3: Model phosphorus results versus observational data (in |ig/L): 30% reduction of the base
TP load. The circled area highlights areas where model prediction was much higher and variable
than the observations.
100
CSMI Lake Michigan 2015 Report
-------
Jun 15 *15 0.043 Jun 27'15 0.167
0 5 10 15 20 0 5 10 15 20
Figure 4: Nearshore spatial patterns of TP in the Lake Michigan for two dates in summer (2015).
Hot colors represent high TP concentrations and cold colors reflect low concentrations.
CSMI Lake Michigan 2015 Report
101
-------
"Data in Motion" - Continuous Water Sensor Data Collection for the 2015 Lake
Michigan CSMI
Tom Hollenhorst1, Laura Fiorentino2, Paul McKinney1, Terry Brown, Anett Trebitz1 and Joel
Hoffman1
U.S. EPA Office of Research and Development, Mid-Continent Ecology Division, Duluth MN.
2NOAA - Center for Operational Oceanographic Products and Services, Chesapeake, VA
Introduction
The 2015 Cooperative Science Monitoring Initiative (CSMI) coordinated investigations linking
Lake Michigan nearshore and offshore habitats and environments. Particularly, we were
interested in understanding the distribution and abundance of nutrients and other water
quality components at a high spatial resolution across the nearshore to offshore gradient.
These investigations will help inform the implementation of the nearshore framework as
envisioned by the Great Lakes Water Quality Agreement of 2012 (IJC 2016). To support this
effort, during the 2015 Lake Michigan CSMI we collected continuous, undulating tow and
autonomous glider data, including conductivity, temperature, pressure, fluorescence, and
optical backscatter across survey tracks crossing the northern half of Lake Michigan. During the
same period, we also collected discrete vertical water column profiles from our research vessel
the Lake Explorer II at locations spread along four spatially coincident transects. Although the
discrete vertical profiles are informative, we found the continuously collected data to be much
more so, particularly when identifying wind and weather-related upwelling and their effect on
lake thermal stratification. Researchers have long recognized the need for high resolution,
detailed information about the variability and patchiness of limnological phenomena like
upwellings (Megard et al. 1997). This is particularly true with transport processes along the
nearshore where conditions are especially variable and episodic due to the contributions of
tributaries after large rain events and the mixing that occurs in areas exposed to high amounts
of wind and wave energy. A better understanding of nearshore transport processes will support
a better understanding of several issues including nutrient loads, lower food web productivity,
linkages with harmful algal blooms, and movements of sediments. This is especially important
since stressors usually occur first and are felt the most strongly in the nearshore areas (Jetoo
and Krantsberge, 2014 and Bails et al. 2005). Because the nearshore is so dynamic we need
advanced technology to monitor and assess nearshore condition as well as long term trends
along the nearshore of the Great Lakes. The 2015 Lake Michigan CSMI provide an opportunity
to explore and compare some available technologies as applied to nearshore - offshore
gradients.
102
CSMI Lake Michigan 2015 Report
-------
Methods
Data Collection
We collected undulating tow data, as described by Yurista and Kelly (2007), combined with
discrete vertical water column profiles from the research vessel the Lake Explorer II along four
transects across northern Lake Michigan. The transects began in the nearshore at
approximately 18 meters depth and extended offshore to approximately 110 meters' depth.
The transects initiated near Sturgeon Bay, Wl, Frankfort, Ml, Ludington, Ml and Manitowoc, Wl
(Figure 1 - map of transects). After collecting the tow data, we sampled at point locations
starting at the deepest depth (110 m), then at 46 m and then at 18 m. At these stations we
collected a vertical conductivity, temperature and depth (CTD) profile and water samples from
at 2 m above the bottom, the mid-hypolimnion layer, at the depth of maximum fluorescence
and from the mid-epilimnion as interpreted from the downward CTD temperature profile. Data
were collected at each of three visits in June, July and September of 2015. During this time, we
also flew two extended Slocum glider across-lake transect missions partially coincident with the
ship transects, during July through August, and again from September through October. Dates
for the CTD casts, tow transects and glider missions are listed in Table 1. Sensors on the Slocum
glider, tow body and CTD Rosette (Figure 2 -photos of tow body, CTD Rosette glider) are listed
in Table 2.
Data processing and visualization
The glider data was imported and initial plots of the data were processed in MatLab. The tow
data was processed with a series of scripts with initial plots processed in ArcGIS 10.3. CTD casts
were processed with SeaBird Scientific SeaSoft-Win32 software (www.seabird.com), binned
into depths and trimmed to down cast only data and imported into a common Microsoft Access
database and plotted with R. We also used ESRI story maps (http://storymaps.arcgis.com) to
visualize the tow and glider data as an interactive map. A user can select points along the
transects and visualize interpolated vertical profiles for the different sensors on board the tow
body and glider at those locations. We also used Cesium, an open-source JavaScript library for
developing dynamic animated map displays on 3D globes and maps.
Results and Discussion
Data visualization and analysis
The tools and techniques available for data visualization and analysis have greatly increased in
just the last few years, as have the available platforms for sharing this type of data. For
example, Xu et al (2017) recently came out with a tool for near-real time visualization and
analysis of undulating tow data (which we have only begun to explore). In addition, the
manufacturers of the Slocum glider (www.teledynemarine.com/webb-research) have recently
released new mission control software for their gliders with more refined data visualization
tools. And more and more so, platforms like ESRI Story Maps, Google Earth, Qlik Sense,
Tableau, and other open source platforms like Cesium are being used to display and
103
CSMI Lake Michigan 2015 Report
-------
disseminate complex geo-spatial data sets like we've collected as part of the Lake Michigan
CSMI efforts.
We've visualized and analyzed our data using a variety of tools including MatLab, Cesium and
ESRI Story maps. Each application has advantages and strengths depending on the data and
visualization type. Figure 3 is a screen shot of a visualization of the glider-collected data
created using Cesium. Cesium can be used to either create interactive web interfaces to the
data or to record videos animating data through time, which helps to visually analyze the data.
Figure 4 is a screen shot of an ESRI story map for Lake Superior glider data to illustrate what we
hope to develop for the Lake Michigan 2015 glider and tow data. Figure 5 illustrates the dates
and locations of the segments associated with the two glider missions. Managing these large
complex data sets is extra challenging in terms of file sizes and complexity. Working with these
large continuous data sets has helped us increase our understanding of the capabilities and
usefulness of these processing and visualization tools, as we also develop workflows and
techniques for automating the process.
Comparisons across seasons and sensor platforms
Comparisons across seasons and sensor platforms revealed some interesting features and
differences in how each platform detected them. Although the discrete vertical profiles are
informative we found the continuously collected data to be much more informative especially
in terms of identifying wind and weather-related upwelling events, and their effect on lake
thermal stratification. In one case an upwelling noticeable in the glider data didn't seem to be
noticeable in the tow data (Figure 6), although the temperature stratification did seem weaker
at the times the tows were conducted. Unfortunately, except for the very first glider segment
of the first deployment, the tow and CTD data were collected days apart from when the glider
was in the same area making comparisons difficult (see Table 1 and Figure 5 for dates and
locations). That said, data collected by the glider did provide us with very high-resolution
observations of the conditions following a major upwelling event. The horizontal and vertical
resolution provided by the glider of the upwelling front are unprecedented in the Great Lakes.
The event occurred during the second deployment (segment 7, Oct. 4-7), and affected water
temperatures on the east side of the lake, near Ludington, Ml. The glider passed through this
area starting in the nearshore near Ludington and travelled west across the lake towards
Sturgeon Bay, Wl. The upwelling is apparent on the east side of the lake in the nearshore and
appears to extend about 25 kilometers off shore (Figure 6). We also considered wind data
leading up to this upwelling event as well as satellite imagery before and after the event. The
wind data from the National Data Buoy Center (NDBC) (Figure 7) shows strong winds blew from
the north and northeast during the four days before October 4th leading to the upwelling
observed along the eastern shore near Ludington. The surficial extent of the upwelling is
apparent in Advanced Very High Resolution Radiometry (AVHRR) satellite data collected on Oct.
10 (Figure 8). Although AVHRR satellite images are available daily, cloud cover over this part of
the lake didn't allow for a useful image until October 10th. The Great Lakes Aquatic Habitat
Framework (Wang et al 2015) developed a useful framework for mapping and classifying
ecosystems and their key driving variables for the Great Lakes (Riseng et al 2018). As part of
104
CSMI Lake Michigan 2015 Report
-------
that effort they modeled upwelling events from 1994 through 2013 using methods established
by Plattner et al. (2006). From that we see these events affect nearshore water temperatures
most frequently on the western side of Lake Michigan, but they do still frequently occur on the
east side near Ludington (Figure 9)
It's not surprising that discrete vertical CTD casts (Figure 10) are not effective for capturing
episodic, ephemeral phenomena like upwelling. The casts themselves represent a very discreet
temporal sample limited only to the time it takes to lower and raise the CTD rosette. Also, they
only represent a very limited spatial extent, represented only by a point location and the
vertical data associated with it. The tow data is clearly an improvement in that vertical
undulating tows are collected along a line, yet are still somewhat limited due to available ship
time and staffing. In this study our tows were integrated with the point sampling so that only
about half of each day was dedicated to the actual towing. We did detect a strong thermocline
in the September tow data, particularly on the eastern side of Lake Michigan (fig. 11).
Unfortunately, due to the short time available for towing we were not on site to capture the
upwelling we observed in early October with the glider.
The data collected by the glider, although still a relatively small sample when compared the
total volume and surface area of Lake Michigan across an entire season, has the best potential
for capturing ephemeral events like upwelling. This is especially true when glider missions are
integrated with near real-time satellite data and wind and weather data. That way glider
missions can be adapted and redirected repeatedly across the gradients (temperature,
fluorescence, etc.) associated with phenomena like upwelling, thermal bars, and tributary
inputs after storm events. In addition, the glider can remain deployed during storm events
when conditions are unfavorable for ship-based data collection.
Conclusions
Each of the Great Lake Cooperative Monitoring Initiatives collect exceedingly large amounts of
complex data spanning a wide range of geographies and time spans. This is necessary to gain
the knowledge and understanding of complex ecosystem processes in the Great Lakes. We've
demonstrated here the application of three different data collection systems that range from
discreet point orientated samples of the water column and relatively short continuous
undulating tow transects that were visited three times over the spring, summer and fall of
2015, compared with 2 relatively long continuous glider missions that collected continuous data
over a 28-day July-August mission, and a 27-day Sept.-Oct. mission. Each system has value due
to the specific capabilities of each (e.g. different sensors, sampling limitations etc.), but it seems
the successful integration of the data systems in time and space will prove the most useful. To
do that we will need well established data frames for storing these large complex data sets and
additional tools and algorithms for their analyses. That will further facilitate combining these
data collection systems with effective visualization and analysis tools along with available
remotely sensed imagery, weather data and modeled data like the upwelling index to provide
for adaptive sampling in near real time. We anticipate that these continuous autonomous
105
CSMI Lake Michigan 2015 Report
-------
sampling platforms combined with some or our more traditional techniques (CTD cast, plankton
nets etc.) and leveraged with innovative and adaptive sampling designs will be key factors in
recording and better understanding important physical and biological processes in the future.
106
CSMI Lake Michigan 2015 Report
-------
References
Bails, J., Beeton, A., Bulkley, J., DePhillip, M., Gannon, J., Murray, M., Regier, H., Scavia, D.,
2005. Prescription for Great Lakes Ecosystem Protection and Restoration: Avoiding the Tipping
Point of Irreversible Changes. Wege Foundation and Joyce Foundation.
Great Lakes Water Quality Agreement. Protocol Amending the Agreement between Canada and
the United States of America on Great Lakes Water Quality, 1978, as Amended on October 16,
1983, and on November 18, 1987, Signed September 7, 2012, Entered into force February 12,
2013. http://ijc.org/files/tinymce/uploaded/GLWQA%202012.pdf (2012)
IJC, 2016 The Great Lakes Nearshore Framework, https://binational.net/wp-
content/uploads/2016/09/Nea rshore-Framework-EN.pdf
Jetoo, S. and Krantzberg, G. 2014. Donning our thinking hats for the development of the Great
Lakes nearshore governance framework. Journal of Great Lakes Research 40 (2014) 463-465.
Megard R. O., Kuns M. M., Whiteside M. C., Downing J. A. 1997. Spatial distributions of
zooplankton during coastal upwelling in western Lake Superior, Limnology and Oceanography,
42, doi: 10.4319/lo.l997.42.5.0827.
Plattner, S., Mason, D. M., Leshkevich, G. A., Schwab, D. J., & Rutherford, E. S. 2006. Classifying
and forecasting coastal upwellings in Lake Michigan using satellite derived temperature images
and buoy data. Journal of Great Lakes Research, 32, 63-76.
Riseng, CM, KE Wehrly, L Wang, ES Rutherford, JE McKenna, Jr., LB Johnson, LA Mason, C
Castiglione, TP Hollenhorst, BL Sparks-Jackson, SP Sowa (2018) Ecosystem classification and
mapping of the Laurentian Great Lakes. Candian Journal of Fisheries and Aquatic
Sciences, https://doi.org/10.1139/cjfas-2017-0242
Yurista, P.M., and Kelly, J.R. 2007. Spatial patterns of water quality and plankton from high-
resolution continuous in situ sensing along a 537-km nearshore transect of western Lake
Superior, 2004. M. Munawar, I.F. Munawar (Eds.), State of Lake Superior, M. & R. Heath,
Ecovision World Monograph Series, Aquatic Ecosystem Health and Management Society,
Canada (2007), pp. 439-471
Wang, L, CM Riseng, LA Mason, KE Wehrly, ES Rutherford, JE McKenna, Jr., C Castiglione, LB
Johnson, DM Infante, S Sowa, M Robertson, J Schaeffer, M Khoury, J Gaiot, T Hollenhorst, C
Brooks, M Coscarelli (2015) A spatial classification and database for management, research, and
policy making: The Great Lakes aquatic habitat framework. Journal of Great Lakes
Research, 41(2): 584-596. http://dx.doi.Org/10.1016/i.iglr.2015.03.017
Xu, W., Collingsworth, P., Bailey, B., Mazur, M. C., Schaeffer, J., & Minsker, B. (2017). Detecting
spatial patterns of rivermouth processes using a geostatistical framework for near-real-time
analysis. Environmental Modelling & Software, 97, 72-85.
107
CSMI Lake Michigan 2015 Report
-------
Tables
Table 1. Data collection dates for 2015 Lake Michigan CSMI
Transects
Spring
Summer
Fall
Sturgeon Bay, Wl
May 30, 2015
July 14, 2015
Sept. 19, 2015
Frankfort, Ml
May 31, 2015
July 15, 2015
Sept. 20, 2015
Ludington, Ml
June 1, 2015
July 16, 2015
Sept. 21, 2015
Manitowoc, Wl
June 3, 2015
July 19, 2015
Sept. 22, 2015
Glider Missions
Start
Stop
Mission 1
July 14, 2015
Aug. 10, 2015
Mission 2
Sept. 22, 2015
Oct. 18, 2015
Table 2. Sensors on various sampling platforms used for CSMI
CTD Rosette
Tow Body
Slocum Glider
CTD
CTD
CTD
Transmissometer 1
Transmissometer 1
Backscatter
Transmissometer 2
Transmissometer 2
Fluorometer
Fluorometer
Fluorometer/Chlorophyll -A
Dissolved Oxygen
Dissolved Oxygen
Dissolved Oxygen
pH
pH
PARS
PARS
LOPC
CDOM
CSMI Lake Michigan 2015 Report
108
-------
Figures
Sturgeon Bay
Frankfort
cj>
Manitowoc
»»-« 89
Ludington
Saugatuck
Waukegon
110.94
K f
• St. Joes
0 25 50 100
Figure 1, Lake Michigan 2015 CSMI transects with dots indicating vertical CTD casts. Light gray
lines represent cross lake transects and distances.
Figure 2. Clockwise from top left: Photos of Slocum Glider, tow body, CTD Rosette glider,
Research Vessel the Lake Explorer II and tow body ready to be deployed.
109
CSMI Lake Michigan 2015 Report
-------
/ sci_flbbcd_chlor_unitg T .yd 283 sat
>ct 10 2Q1S 06 50.S8 GMT-0S00
-------
2015 Deployment 1
44°35'
44° 10'
2015 Deployment 2
43°45' -
43°20'
Dates:
Seg. 1: July 14-17
Seg. 2: July 17-19
Seg. 3: July 19-20
Seg. 4: July 21-23
Seg. 5: July 23-26
Seg. 6: July 26-29
Seg. 7: July 29-31
Seg. 8: July 31-Aug. 1
Seg. 9: Aug. 1-4
Seg. 10: Aug. 4-9
Seg. 11: Aug. 9-10
44°5T
44°28'
w 43°59'
43°30'
43°0r
Dates:
Seg. 1: Sept. 21-22
Seg. 2: Sept. 22-23
Seg. 3: Sept. 23-28
Seg. 4: Sept. 28-Oct. 1
Seg. 5: Oct. 1-2
Seg. 6: Oct. 2-4
Seg. 7: Oct. 4-7
Seg. 8: Oct. 7-10
Seg. 9: Oct. 10-14
Seg. 10: Oct. 14-14
Seg. 11: Oct. 14-15
-87°55'
-87°30' -87°05' -86°40'
Loneitude
-86° 15'
-S7°55' -S7°30' -87°05* -S6°40' -86°15'
Longitude
Figure 5. Dates of glider deployment segments.
CSMI Lake Michigan 2015 Report
111
-------
Temperature (*C)
0
50
•100
£150
m
w
1
41
f
1
r
200
250
G ider
0
14
50
12
100
10
150
8
200
6
250
4
Chlorophyll-a (pg I" )
abed
80
60 40 20
Distance (km)
Temperature ( C)
5 10 IS
20
60 40 20
Distance (km)
Chlorophyll-a (pg I"1)
0.5 1 1.5
T
i
J
i
y
4
r
1
i
i
i
—-profile a
i
- - profile b
i
profile c
i
i .
—• profile d
50
100
Q.
e;
Q
150
—profile a
- - profile b
profile c
—-profiled
50
100
Q.
Q
150
Figure 6, (top) Cross section of Lake Michigan temperature and chlorophyll-a from 2015
Deployment 2, segment 7 (for dates and location see Figure 5). The glider transited the lake
from east to west, completing 480 profiles. The maximum dive depth for the glider is 150 m,
and white area is no-data, (bottom) Representative profiles of temperature and chlorophyll-a
from different distances (locations are indicated in top images) along the segment. Low values
of Chlorophyll at the surface at 30 km, 50 km and 75 km are due to daytime quenching effects.
112
CSMI Lake Michigan 2015 Report
-------
Buoy 45002
Wind at NDBC station 45002
Sep 29-Oct 04 2015
NORTH
0,360
-------
CLOUDS
CLOUDS
Figure 8, Lake Michigan surface temperature on Sept, 26, 2015 (left) and Oct. 10, 2015 (right).
The left hand image, acquired prior to the wind event discussed in the text, shows warm
surface temperature across most of the lake, In the right hand image, acquired after the wind
event, cooler temperatures on the east side of the lake, near Ludington, Ml., are the result of
wind-driven upwelling of colder deep water to the surface.
Upwelling Index
High
0 12.5 25 50 75 100
— —— Miles
Figure 9. Lake Michigan Upwelling index summed from 1994 through 2013 using methods
developed by Plattner et al 2006. Data acquired from www.glahf.org,
114
CSMI Lake Michigan 2015 Report
-------
LuftngtonSepl-l10m
Temp (C)
5 10
Temp(C)
5 10
Ludington Sept-46m
Ludington Sept - 18m
Figure 10. Vertical CTD Cast for the Ludington, Ml transect at depths of 110, 46 and 18 meters
collected on Sept. 21, 2015 (before the upwelling event).
CSMI Lake Michigan 2015 Report
115
-------
S«*i21 LudlnfllKf.
CSMI
Lake Michigan
September 2015
I>epth (m)
G-20
20 - -iO
-413-50
- SO - SO
- SO -10 0
- 100-120
- 120 - UO
" 14C - 150
" 1SC - 1B3
- 1S0 -200
-200-220
- 220 -2-tn
- 240 - 253
- 2S0 - 290
Temperature ("Q
0.00 - 2.50 $ 10.00 - 12.50
2.50 - £,00 # 12jS3 - 15.00
¦EDO - 7.50 0 1500 -17.50
7 50 - 10.00 # 1750-20.00
SwfilWrr.'lJW/ !9 J02A IJtSA.O,H
9fc|ph*nvJMV SO JfT25 II;J/:?/
0 200 4DO 600 BOO ir0CD 1.200
OitUnc#
SvpftfrnlM -V Of:4S:24
Tw-nvfwraitur#'
Temiwiturij
Distant*
Umptmmbw jPJ JCT1S 01:31:03
700
Figure 11. Tow temperature data from all four transects during September 2015. Note the
strong thermal stratification occurring at this time.
116
CSMI Lake Michigan 2015 Report
-------
Report: Atrazine Concentrations in Lake Michigan:
Investigating Causes of the Recent Decline
Authors:
Kathryn A. Meyer, USEPA Great Lakes National Program Office (ORISE Fellow)
Russell G. Kreis, Jr., USEPA ORD/NHEERL/MED (Retired)
Kenneth R. Rygwelski, USEPA ORD/NHEERL/MED
Todd G. Nettesheim, USEPA Great Lakes National Program Office
Glenn J. Warren, USEPA Great Lakes National Program Office
Contact:
Glenn Warren
Email: Warren.glenn@epa.gov
Phone:
Address:
USEPA Great Lakes National Program Office
77 W. Jackson
Chicago, IL 60604
117
CSMI Lake Michigan 2015 Report
-------
Project Summary
The 1994-1995 Michigan Mass Balance Study (MMBS) observed and forecasted whole lake,
volume-weighted average atrazine concentrations for Lake Michigan. The atrazine
concentrations were well below the U.S. Environmental Protection Agency (EPA) biological
thresholds. But, with a decay estimated at less than 1% per year in the lake and knowing that
atrazine acts similarly to a conservative substance, concentrations were expected to increase
under current atrazine loadings. Ken Rygwelski and Russ Kreis developed a model to predict
future atrazine concentrations in Lake Michigan for a variety of different scenarios (Figure 1).
100% reduction of total load (scenario 4)
¦ Field data +/- one standard deviation
- • -Constant loads (scenario 3)
Zero vapor phase concentration (scenario 7)
90 r No wet deposition (scenario 6)
— 35% reduction of total load (scenario 8)
80 —¦ -100% reduction of tributary load (scenario 5)
1963 1993 2023 2053 2083 2113 2143 2173 2203 2233 2263
Year
Figure 1: Lake Michigan atrazine concentration prediction modeling (Kreiss, Rygwelski)
Lake Michigan atrazine sampling was done again in 2015. Water samples were taken at the EPA
Great Lakes National Program Office's Lake Michigan open water stations in August 2015 at
both the mid-epilimnion and mid-hypolimnion depths. These samples were analyzed using gas
chromatography - mass spectrometry by the U.S. Geological Survey National Water Quality
Laboratory in Lakewood, Colorado. The average atrazine concentration for Lake Michigan in
2015 was 36 ng/L. This concentration fell between the model predictions for the 100% reduction
of tributary loading and 100% reduction of total loading scenarios of the prediction model
(Figure 1).
Three hypotheses were investigated regarding why the Lake Michigan atrazine concentration in
2015 was less than forecasted.
1) Atrazine is degrading more rapidly than predicted in the model
2) A decline in atrazine usage has significantly reduced atrazine input to Lake Michigan
3) Sedimentation processes are removing atrazine from the water column
The first hypothesis was addressed through a literature review. Studies have found that atrazine
degradation may be occurring in soils, as atrazine degradation genes have evolved in many
bacteria species, so less atrazine may be reaching the lake. Faster biodegradation rates were
found in soils already treated with atrazine, too. The influence of photolysis in atrazine
118
CSMI Lake Michigan 2015 Report
-------
degradation, which could also be stimulated by deeper light penetration (supported by increasing
secchi depths) was also suggested. Atrazine-degrading bacteria have not yet specifically been
found in Lake Michigan.
To address the second hypothesis, atrazine use data in the Lake Michigan basin from the USGS
National Water-Quality Assessment (NAWQA) and the USDA National Agricultural Statistical
Service (NASS) were evaluated. There was an estimated 34% decline in annual atrazine use
from 1992-2015 in Lake Michigan counties (USGS NAWQA) and an estimated 18% decline in
annual atrazine application from 1990-2014 in Lake Michigan states (USDA NASS). Annual
atrazine use in the United States, which was assessed because atmospheric transport may also
affect atrazine concentrations, had an estimated 15% decline from 1992-2015 (USGS NAWQA).
Overall, the declines potentially contributed to lower atrazine concentrations.
The third hypothesis was address through another literature review. The prediction model
assumed sedimentation was negligible. A recent study by Guo et al. (2016) found annual atrazine
loadings to Lake Michigan sediments to be 44.4 +/- 25.6 kg/year; and that atrazine did not easily
desorb from particles. Although atrazine sedimentation in Lake Michigan has been recorded, the
amount of atrazine fluxing to the sediment is likely negligible compared to the large amount of
atrazine entering the lake through other pathways.
The overall general conclusions for this investigation were that decreased atrazine usage is
having an influence on the atrazine concentration in Lake Michigan, and that increased atrazine
degradation is also having an influence on atrazine concentrations in the lake.
Relevant Literature Sources
Cessna, A.J. Nonbiological degradation of triazine herbicides: photolysis and hydrolysis. "The
Triazine Herbicides." Ch. 23. 2008. 329.
Fenner, K., et al. Relating atrazine degradation rate in soil to environmental conditions:
Implications for global fate modeling. Environ. Sci. Technol. 2007, 41, 2840-2846.
Giardi, M.T., et al. Chemical and biological degradation of primary metabolism of atrazine by a
Nocardia strain. Agriculture Biology and Chemistry. 1985, 49, 1551-1558.
Guo, J., et al. Occurrence of atrazine and related compounds in sediments of Upper Great Lakes.
Environ. Sci. Technol. 2016, 50, 7335-7343.
Kurt-Karakus, P.B., et al. Metolachlor and atrazine in the Great Lakes. Environ. Sci. Technol.
2010. 44 (12), 4678-4684.
Mandelbaum, R.T., et al. Isolation and characterization of a Pseudomonas sp that mineralizes the
s-triazine herbicide atrazine. Appl. Environ. Microbiol. 1995, 61, 1451-1457.
Mueller, T.C., et al. Enhanced atrazine degradation is widespread across the United States. Pest.
Manag. 2017.
119
CSMI Lake Michigan 2015 Report
-------
Newton, R.J.; McLellan, S.L. A unique assemblage of cosmopolitan freshwater bacteria and
higher community diversity differentiate an urbanized estuary from oligotrophic Lake Michigan.
Frontiers in Microbiology. 2015, 6, 1028.
Rousseaux, S., et al. Isolation and characterization of new Gram-negative and Gram-positive
atrazine degrading bacteria from different French soils. FEMS Microbiology Ecology. 2001, 36,
211-222.
Rygwelski, K.R. (Ed.) Results of the Lake Michigan Mass Balance Project: atrazine modeling
report. U.S. Environmental Protection Agency, ORD/NHEERL/MED, Grosse lie, MI.
EPA/600/R-08/111, 140 pp.
Rygwelski, K.R., et al. Model forecasts of atrazine in Lake Michigan in response to various
sensitivity and potential management scenarios. J. Great Lakes Res. 2012, 38, 1-10.
Schottler, S.P.; Eisenreich, S.J. Herbicides in the Great Lakes. Environ. Sci. Technol. 1994. 28,
2228-2232.
Schottler, S.P.; Eisenreich, S.J. Mass balance model to quantify atrazine sources, transformation
rates, and trends in the Great Lakes. Environ. Sci. Technol. 1997. 31, 2616-2625.
Smalling, K.L.; Aelion, C.M. Distribution of atrazine into three chemical fractions: impact of
sediment depth and organic carbon content. Environ. Toxicol. Chem. 2004, 23 (5), 1164-1171.
Struger, J., et al. In-use pesticide concentrations in surface waters of the Laurentian Great Lakes,
1994-2000. J. Great Lakes Res. 2004, 30 (3), 435-450.
Tierney, D.P., et al. Predicted atrazine concentrations in the Great Lakes: Implications for
biological effects. J. Great Lakes Res. 1999. 25 (3), 455-467.
Topp, E. A comparison of three atrazine-degrading bacteria for soil remediation. Biol. Fertil.
Soils. 2001, 33, 529-534.
CSMI Lake Michigan 2015 Report
120
-------
Report: Examining Legacy and Emerging Contaminants
Lake Michigan Tributaries
Authors:
Marta Venier, Indiana University
Jiehong Guo, Indiana University
Kevin Romanak, Indiana University
Stephen Westenbroek, USGS Wisconsin Water Science Center
An Li, University of Illinois at Chicago
Russell G. Kreis, Jr., USEPA ORD/NHEERL/MED (Retired)
Ronald Hites, Indiana University
Contact:
Marta Venier
Email: mvenier@indiana.edu
Phone: 812-855-1005
Address:
School of Public and Environmental Affairs
702 N. Walnut Grove ave.
Bloomington, IN 47405
121
CSMI Lake Michigan 2015 Report
-------
Project Summary
High-volume water samples were collected every 3 weeks in 2015 from five tributaries to Lake
Michigan, and were subsequently analyzed for PCB congeners, mercury, and emerging chemical
flame retardants. The five tributaries were selected because they showed the highest loads of
PCBs in the 1994-5 Lake Michigan Mass Balance Project (LMMBP); the tributaries sampled
were the Grand, Kalamazoo, St. Joseph, and Lower Fox Rivers and from the Indiana Harbor and
Ship Canal (IHSC). A total of 59 samples were collected from these five tributaries by two
USGS field offices between April and December of 2015. The sampling procedure was similar
to that in the LMMBP, where 80-160 L of water were pumped through a 0.7 |im filter and
through XAD-2 packed resin columns.
The dissolved phase (XAD-2) and the particle phase (glass fiber filters) were analyzed separately
to study how the chemicals partition between the two phases. The XAD-2 resin or filter was
loaded in a Soxhlet extractor, spiked with the surrogate standards and extracted using a mixture
of hexane and acetone. Following additional concentration and extraction steps, the samples
were analyzed for PCBs and flame retardants, including organophosphate esters (OPEs),
brominated flame retardants (BFRs), and dechlorane-related compounds.
In the five tributaries sampled in this study, the geometric mean concentrations of EPCB (sum of
85 congeners) ranged from 1.52 to 22.4 ng/L. The highest concentrations of PCBs were
generally found in the Lower Fox River and in the Indiana Harbor and Ship Canal. The highest
BFR concentrations were measured in either the IHSC or the St. Joseph River. OPEs were the
most abundant among the targeted compounds with geometric mean concentrations ranging from
20 to 54 ng/L; OPE concentrations were comparable among the five tributaries. BFR
concentrations were about 1 ng/L, and the most abundant compounds were bis(2-ethylhexyl)
tetrabromophthalate, 2-ethylhexyl 2,3,4,5-tetrabromobenzoate, and decabromodiphenyl ether.
The dechlorane-related compounds were detected at low concentrations (< 1 pg/L). The fraction
of target compounds in the particulate phase relative to the dissolved phase varied by chemical
and tended to increase with their octanol-water partition coefficient.
During the work on flame retardants, a relatively new compound named Marbon was identified.
Marbon is isomeric with Dechlorane Plus (DP). Dechlorane Plus is commonly found in the
environment throughout the world, but Marbon has, so far, only been detected at low levels in
one sediment core collected near the mouth of the Niagara River in Lake Ontario. In addition to
the 59 Lake Michigan Tributary water samples, 10 surface sediment samples from the IHSC, and
2 surface sediment samples from the Chicago Sanitary and Ship Canal were analyzed for
Marbon. Three Marbon diastereomers were detected in the water and sediment samples from the
IHSC, which is far from the location of its previous detection in Lake Ontario. The sum of the
concentrations of the three Marbons was greater in the water from the IHSC (N = 11, median =
150 pg/L) compared to those in water from the other four tributaries (N = 11-13, medians = 0.9-
2.0 pg/L). Marbon concentrations in sediment samples from the IHSC were up to 450 ng/g dry
122
CSMI Lake Michigan 2015 Report
-------
weight. Anti-DP was also measured for comparison. Its concentrations were not significantly
different among the water samples, but its sediment concentrations in the IHSC were
significantly correlated with those of Marbon. The source of Marbon contamination in the IHSC
is not clear.
Loads for PCB, mercury, and flame retardant compounds were calculated for the five tributaries.
PCB data from this study were combined with PCB concentration data from other previous
studies involving open lake water, air, and sediment to calculate an updated mass budget. The
input flows of EPCBs from wet deposition, dry deposition, tributary loading, and air to water
exchange, and the output flows due to sediment burial, volatilization from water to air, and
transport to Lake Huron and through the Chicago Diversion were calculated as well as flows
related to the internal processes of settling, resuspension, and sediment-water diffusion. The net
transfer of PCBs was 1240 ±531 kg/yr out of the lake. This net transfer is 46% lower than that
estimated in 1994-5. PCB concentrations in most matrices in the lake are decreasing, which
drove the decline of all the individual input and output flows. Tributary loads at the Lower Fox
River and the Indiana Harbor and Ship Canal both decreased substantially relative to 1994-1995
loads. Atmospheric deposition to Lake Michigan has become negligible, while volatilization
from the water surface is still a major route of loss, releasing PCB from the lake into the air.
Large masses of PCB remain in the water column and surface sediments and are likely to
contribute to future efflux of PCBs from the lake to the air.
Mercury loads from all five tributaries to Lake Michigan were on the order of 50% to 75% lower
in 2015 relative to loads calculated for 1994-1995. The total mercury load in the Lower Fox
River was about 160 kg/yr in 1994-1995; the total mercury load in the Lower Fox River was
about 43 kg/yr in 2015. The total mercury load in the Grand River was about 36 kg/yr in 1994-
1995; the total mercury load in the Grand River was about 8 kg/yr in 2015. Similar decreases
were seen for the other three tributaries. Indiana Harbor and Ship Canal's total mercury load
decreased from about 4.3 kg/yr (1994-1995) to about 1.7 kg/yr. St. Joseph River's total mercury
load decreased from about 35 kg/yr (1994-1995) to about 12.6 kg/yr (2015). Kalamazoo River's
total mercury loads decreased from about 21 kg/yr (1994-1995) to about 9 kg/yr (2015).
Publications Related to this Work
Guo, J.; Venier, M.; Romanak, K.; Westenbroek, S.; Hites, R. A., Identification of Marbon in
the Indiana Harbor and Ship Canal. Environ. Sci. Technol. 2016, 50, (24), 13232-13238.
Guo, J.; Romanak, K.; Westenbroek, S.; Li., An, Kreis, R.G. Jr, Hites, R. A., Venier, M.;
Updated Polychlorinated Biphenyl Mass Budget for Lake Michigan. Environ. Sci. Technol.
2017, 51, (21), 12455-12465.
123
CSMI Lake Michigan 2015 Report
-------
Guo, J.; Romanak, K.; Westenbroek, S.; Hites, R. A.; Venier, M., Current-Use Flame
Retardants in the Water of Lake Michigan Tributaries. Environ. Sci. Technol. 2017, 51, (17),
9960-9969.
124
CSMI Lake Michigan 2015 Report
-------
bleeping
Bear Dunes
MO 3 2
Lower »
Fox River
MO 2 4
«64a
-|- Air (gas and precipitation?
• Open lake water
* Sediment core
Tributary water
M009
®
popu
Chicago~P
Indiana Harbor
Ship Canal
Grand Rive
aamazoo Rive
St. Joseph River
Kilometers
Figure 1. Sampling locations related to the 2015 CSMI Lake Michigan contaminant study
referenced in subsequent plots and the text.
CSMI Lake Michigan 2015 Report
125
-------
103
102 -
>*
D>
i. 101
"O
to
o
15 10° -
3
C
C
<
10
-1
10
¦2
IHSC
SJR
KR
GR
PCBs
1 1
OPEs
nFRs
1 1
PBDEs
LFR
Figure 2. Tributary loading of organophosphate esters (OPEs), non-BDE novel flame
retardants (nFRs), and polybrominated diphenyl ethers (PBDEs) to Lake Michigan.
Loadings for polychlorinated biphenyls (PCBs) from 2005-06 are included for reference.
Abbreviations: IHSC, Indiana Harbor and Ship Canal; SJR, St. Joseph River; KR,
Kalamazoo River; GR, Grand River; and LFR, Lower Fox River. The rivers are arranged
based on latitude from south to north. ANOVA results are to be read across the tributaries;
water from tributaries sharing the same letter do not have statistically different (p < 0.05)
concentrations.
CSMI Lake Michigan 2015 Report
126
-------
CD
Q.
(/)
C
o
-I—'
(0
I—
c
(D
O
c
o
O
104
103
102
101
105
103
101
10-1
-i 1 1 1 r
PCB-33 *
B D C D A
PCB-118
f
A C B C AB
1 1 1 1 r
^ PCB-52
A C B
-------
100
Lower
Fox River
H 1994-5
I—I 2005-6
r~! 2015
Grand
River
Kalamazoo
River
St. Joseph
River
Indiana Harbor
and Ship Canal
Figure 4. Tributary EPCB flows to Lake Michigan for 1994-5, 2005-6, and 2015.
Estimates for 1994-5 and 2005-6 were obtained from a previous study.
CSMI Lake Michigan 2015 Report
128
-------
Gross
volatilization
11101485
(539
Gas absorption
120 ±68,
Atmospheric (2310
deposition
Wet deposition
Dry deposjtidn
Chicago
Diversion
export
1 ±0 (1)
resuspension
iff us ion
240 ±317
(429)
'Sediment
burial
406 ±199
(984)
Strait of Mackinac
export
9 ± 2 (-2)
166 ±65 (379)
tributary loading
PCB inventory
Water column=2340 ± 543 (1510) kg
Sediment=11,000 ± 3110 (27,000) kg
(0-4cm)
Figure 5. Estimated total PCB mass budget flows (kg yr1) and inventories (kg) for 2010-2015 and
comparison to the 1994-5 mass balance results based on the MICHTOX model 16 (in parentheses)
in Lake Michigan. The blue and green layers represent water and sediment layers, respectively. The
thickness of the arrows indicates the magnitude of flows in 2010-2015.
CSMI Lake Michigan 2015 Report
129
-------
105
104 -
103 -
102
101
105 -
104 -
103 -
£
TCIPP
A A A A A
TNBP
0
B A A AB B
TPHP o
^ O Q i
A A A A B
102
TCEP
A A A A A
TDCIPP
E150PES
A BC CD AB D
A A A A B
Figure 6. Box and whisker plots of concentrations of organophosphate esters (OPEs) in
tributary water samples. Abbreviations: IHSC, Indiana Harbor and Ship Canal (N = 11);
SJR, St. Joseph River (N = 12); KR, Kalamazoo River (N = 12); GR, Grand River (N =
11); and LFR, Lower Fox River (N = 13); TCIPP, tris[(2R)-l-chloro-2-propyl] phosphate;
TNBP, tri-n-butyl phosphate; TPHP, triphenyl phosphate; TCEP, tris(2-chloroethyl)
phosphate; TDCIPP, tris(l,3-dichloro-2-propyl) phosphate; £15OPEs, sum of 15 OPEs.
Shown are the medians (black lines inside the box), the 25th to 75th percentiles (box), the
10th and 90th percentiles (whiskers), the minimum and maximum values (circles), and the
ANOVA results (letters at the bottom of each box). ANOVA results are to be read across
the tributaries; water from tributaries sharing the same letter do not have statistically
different (p < 0.05) concentrations. Tributaries are arranged based on latitude from south
to north.
CSMI Lake Michigan 2015 Report
130
-------
104
103
102
101
^ 10°
£ 104
"c 103
o
=S 102
(0
~ 101
O 10°
o 101
o
101
10° -
101
102
BDE-47
A A B A B
BDE-99
i
A AB C B C
BDE-209
BC AB C C A
S£
235BDEs
A A B B A
EHTBB
A B D AB C
BEHTBP
BTBPE
DBDPE
A B B AB C
AB A C BC ABC
0
C A B B B
Dec604-0
Mi rex
Ihi
A A A A A
Chlordane Plus
A A A A A
Dec602
555,
B B A B B
Figure 7. Box and whisker plots of concentrations of polybrominated diphenyl
ethers (PBDEs), non-BDE novel flame retardants (nFRs), and dechlorane related
compounds (Decs) in tributary water samples. Abbreviations: IHSC, Indiana
Harbor and Ship Canal (N = 11 for PBDEs and nFRs, N=8 for Decs); SJR, St.
Joseph River (N = 12 for PBDEs and nFRs, N=8 for Decs); KR, Kalamazoo River
(N = 12 for PBDEs and nFRs, N=8 for Decs); GR, Grand River (N = 11 for
PBDEs and nFRs, N=7 for Decs); and LFR, Lower Fox River (N = 13 for PBDEs
and nFRs, N=7 for Decs); £35BDEs, sum of 35 PBDEs; EHTBB, 2-ethylhexyl
2,3,4,5-tetrabromobenzoate; BEHTBP, di-(2-ethylhexyl)-tetrabromophthalate;
BTBPE, l,2-bis(2,4,6-tribromophenoxy)ethane; DBDPE,
decabromodiphenylethane; Dec604-0, hexachlorophenyl-norbornene; Dec602,
dechlorane 602. Shown are the medians (black lines inside the box), the 25th to
75th percentiles (box), the 10th and 90th percentiles (whiskers), the minimum and
maximum values (circles), and the ANOVA results (letters at the bottom of each
box). ANOVA results are to be read across the tributaries; water from tributaries
sharing the same letter do not have statistically different (p < 0.05) concentrations.
Rivers are arranged based on latitude from south to north.
CSMI Lake Michigan 2015 Report
131
------- |