Measures of larval lake whitefish length and abundance as early predictors of year-class strength in Lake Michigan

Many lake whitefish stocks in Lake Michigan have experienced substantial declines in growth and condition since the 1990s. Reduced growth and condition could result in reduced quality or quantity of eggs produced by spawning females, which in turn could negatively impact recruitment. We evaluated the potential for reduced recruitment by measuring early life stage density and length, and we discuss the utility of these measures as early indicators of lake whitefish year-class strength. Overall, mean larval density (number per 1000 m³ ± SE) in Lake Michigan was greater in 2006 (373.7 ± 28.3) than in 2005 (16.6 ± 24.8); whereas, mean length (mm ± SE) of larval lake whitefish was smaller in 2006 (12.87 ± 0.07) than in 2005 (14.38 ± 0.13). The ratio of zooplankton to fish density did not show an expected relationship with larval fish density or length. Rather, variation in larval density was best explained by a multiple-regression model that included larval length, spring wind intensity, and adult stock density as predictor variables. Our results suggest that the density of larval lake whitefish is not directly regulated by temperature or zooplankton density at the time of emergence, but that a potential for density-dependent regulation exists when larval emergence rates are high. We conclude that the observed declines in growth and condition of adult lake whitefish are not resulting in substantial reductions in recruitment.     

Energy Density of Three Fishes from Minnesota Waters of Lake Superior

Energy densities of three salmonine prey species in Lake Superior were determined by bomb calorimetry to examine seasonal trends, differences due to gender and to establish a relation with percent dry weight. These results from April through October for 1992 and 1993 were compared with other studies from the Great Lakes region. Across the 2-year sampling period rainbow smelt Osmerus mordax energy density decreased from April to June and July then significantly increased into September. Lake herring Coregonus artedii energy density was highest in June, decreased significantly by August, and remained stable through October. Bloater C. hoyi energy density was lowest in July, then increased in August and early September before decreasing in late September. There were no significant differences in energy density between males and females for rainbow smelt and lake herring collected on the same date across sampling dates. However, bloater energy density was significantly higher for females than for males on two dates. Energy density estimates for bloater in Lake Superior were lower than those reported from Lake Michigan. Energy density values of rainbow smelt and lake herring fell within the range reported in Lake Michigan and Lake Ontario (rainbow smelt), and from two inland lakes in Minnesota and Lake Mendota, Wisconsin (lake herring). Percent dry weight was a good predictor of energy density (P < 0.001, r² = 0.92–0.99) for all three species for both years. Percent dry weight vs energy density was significantly different between years for bloater and lake herring, and was significantly different for all three species compared with populations in other lakes in the Great Lakes Region.     

Reproductive Potential and Fecundity of Lake Trout Strains in Southern and Eastern Waters of Lake Ontario, 1977–1994

We assessed the reproductive potential of various genetic strains of hatchery lake trout (Salvelinus namaycush) in southern and eastern Lake Ontario from indices of fecundity and indices of male abundance. Indices were constructed from catches of mature lake trout in gill nets during September 1980 to 1994 after correcting for mortality from sea lampreys (Petromyzon marinus) which occurred between September sampling and late fall spawning. Strain and age were assigned to individual lake trout based on clipped fins and maxillary bones or coded wire tags. Fecundity-length relationships for fish of the same age, determined from mature females collected in 1977 to 1981 and 1994, were not different (P > 0.05) among genetic strains. For all strains combined, fecundity-length relationships in 1977 to 1981 were not different among fish of various ages but in 1994, age-5 and -6 fish had fewer eggs (P < 0.003) than age-7 fish, and age-7 fish had fewer eggs (P < 0.003) than fish of age 8, 9, or 10. Annual indices of fecundity varied 19 fold and indices of mature males varied 11 fold; both indices were low in the early 1980s, increased sharply in the mid 1980s, and peaked in 1993. The strain which dominated fecundity and mature male indices shifted during the study from Seneca Lake strain to Lake Superior strain and then back to Seneca Lake strain. However, changes in either reproductive potential or genotypes do not appear responsible for the abrupt appearance of naturally-produced yearling lake trout throughout southern and eastern Lake Ontario in 1994–1995, the first widespread occurrence of juveniles produced by hatchery lake trout in Lake Ontario.     

Determining the efficacy of microsatellite DNA-based mixed stock analysis of Lake Michigan's Lake Whitefish commercial Fishery

Management of commercially exploited fish should be conducted at the stock level. If a mixed stock fishery exists, a comprehensive mixed stock analysis is required for stock-based management. The lake whitefish Coregonus clupeaformis comprises the primary commercial fishery across the Great Lakes. Recent research resolved that six genetic stocks of lake whitefish were present in Lake Michigan, and long-term tagging data indicate that Lake Michigan's lake whitefish commercial fishery is a mixed stock fishery. The objective of this research was to determine the usefulness of microsatellite data for conducting comprehensive mixed stock analyses of the Lake Michigan lake whitefish commercial fishery. We used the individual assignment method as implemented in the program ONCOR to determine the accuracy level at which microsatellite data can reliably identify component populations or stocks. Self-assignment of lake whitefish to their population and stock of origin ranged from > 96% to 100%. Evaluation of genetic stock discreteness indicated a moderately high degree of correct assignment (average = 75%); simulations indicated supplementing baseline data by ～ 50 to 100 individuals could increase accuracy by up to 4.5%. Simulated mixed stock commercial harvests with known stock composition showed a high degree of correct proportional assignment between observed and predicted harvest values. These data suggest that a comprehensive mixed stock analysis of Lake Michigan's lake whitefish commercial fishery is viable and would provide valuable information for improving management.     

Precision of Ages Determined from Scales and Opercles for Yellow Perch Perca flavescens

Scales and opercles were used to age yellow perch Perca flavescens collected in 1989 from Lake Madison (South Dakota), Dauphin Lake (Manitoba), and southern Lake Michigan (Indiana). Three readers aged fish from Lake Madison and Dauphin Lake once and two readers aged fish from Lake Michigan twice. The coefficient of variation (CV) was calculated to compare precision. Ages determined from opercles were as precise as those from scales for fish from Lake Madison (CV = 0 for both structures), and more precise than ages from scales for fish from Dauphin Lake (CV_opercle = 14.0, CV_scale = 27.4, p < 0.001) and Lake Michigan (CV _opercle = 10.6, CV_scale = 13.9, p < 0.001). The high precision of scale and opercle ages for yellow perch from Lake Madison can be attributed to the fast growth rate of fish from that lake and also that only age 1 and 2 fish were aged. The greater precision of opercle ages in comparison to scale ages for Dauphin Lake and Lake Michigan yellow perch can be attributed to ease of recognition of false annuli on opercles as well as to difficulty in distinguishing between false and true annuli crowded on the edge of scales from mature, slower growing fish. Because true annuli are more easily recognized on opercles, ages determined from opercles may be more accurate than ages determined from scales for yellow perch growing at slow or moderate rates.     

Space-for-time proxy for climate change in deep lakes in the Canadian cordillera: Seasonality along a latitudinal climate gradient

A comparison of the annual thermal histories in seven deep, intermontane lakes in western Canada along a north–south transect from 49.67° to 61.25° N reveals latitudinal trends in thermal structure; specifically: 1) summer maximum and winter minimum mean temperatures decrease ~  0.2°C per degree latitude, 2) the dates of maximum and minimum heat content lag ~ 1 and ~  6 days per degree latitude, respectively, 3) the lengths of summer and winter stratification lag ~  8 days per degree latitude, 4) the timing of the spring and autumn turnovers lags ~  4 days per degree latitude, 5) the timing of the onset and break-up of ice cover lags ~ 6 days per degree latitude, and 6) the rate of response to atmospheric heating and cooling is ~  1MJ m^− 2 day^− 1 per degree latitude. We propose that insight into the response of a lake at a specific latitude to climate change can be acquired by examining the present conditions of a morphologically and dynamically similar lake at an appropriate latitude and altitude further south. Of interest is the potential transition of a lake from one mictic state to another, with consequences to the resupply of nutrients to the euphotic zone and to the phenology of spring and fall phytoplankton blooms. The absence or presence of ice cover and timing of freeze-up and break-up also affect a lake's heat budget and thermal history. The full interpretation of climate signals in lake systems requires an understanding of lake-specific physical forcing and response mechanisms.     

Diets of Siscowet Lake Trout from the Apostle Islands Region of Lake Superior, 1993

Stomach contents of 384 siscowet lake trout (Salvelinus namaycush siscowet) from the Apostle Islands region of Lake Superior were examined to determine if diets exhibited seasonal trends or shifted with predator size. The information was also used to qualitatively compare siscowet and lean lake trout (S. namaycush) diets. Stomachs from siscowet less than 500-mm total length (TL) contained primarily macroinvertebrates (clams, Mysis relicta, and insects), sculpins (Cottidae; mean TL = 45 mm), and coregonines (Coregonus spp.; mean TL = 120 mm). Stomachs from siscowet between 500- and 699-mm TL contained items similar to smaller siscowet, but burbot (Lota lota; mean TL = 162 mm and a 254-mm siscowet were also present. Stomachs from siscowet longer than 699-mm TL contained fishes such as burbot (mean TL = 246 mm) and coregonines (mean TL = 240 mm). Dietary proportions of coregonines increased significantly with increasing siscowet length groups. The proportion of the total number of identified taxa and frequency of occurrence of terrestrial insects and sculpins declined significantly across all length groups between June and December. Siscowet and lean lake trout diets were not found to be substantially different and a potential for competition exists.     

Dreissenid mussels are not a “dead end” in Great Lakes food webs

Dreissenid mussels have been regarded as a “dead end” in Great Lakes food webs because the degree of predation on dreissenid mussels, on a lakewide basis, is believed to be low. Waterfowl predation on dreissenid mussels in the Great Lakes has primarily been confined to bays, and therefore its effects on the dreissenid mussel population have been localized rather than operating on a lakewide level. Based on results from a previous study, annual consumption of dreissenid mussels by the round goby (Neogobius melanostomus) population in central Lake Erie averaged only 6 kilotonnes (kt; 1 kt = one thousand metric tons) during 1995–2002. In contrast, our coupling of lake whitefish (Coregonus clupeaformis) population models with a lake whitefish bioenergetics model revealed that lake whitefish populations in Lakes Michigan and Huron consumed 109 and 820 kt, respectively, of dreissenid mussels each year. Our results indicated that lake whitefish can be an important predator on dreissenid mussels in the Great Lakes, and that dreissenid mussels do not represent a “dead end” in Great Lakes food webs. The Lake Michigan dreissenid mussel population has been estimated to be growing more than three times faster than the Lake Huron dreissenid mussel population during the 2000s. One plausible explanation for the higher population growth rate in Lake Michigan would be the substantially higher predation rate by lake whitefish on dreissenid mussels in Lake Huron.     

Trends of PCB concentrations in Lake Michigan coho and chinook salmon, 1975–2010

The manufacture and use of polychlorinated biphenyls (PCBs) was banned in the United States in 1977 after it was determined that these compounds adversely affect animals and humans. The Wisconsin Department of Natural Resources has quantified total PCB concentrations in Lake Michigan chinook (n = 765) and coho (n = 393) salmon (Oncorhynchus tshawytscha and Oncorhynchus kisutch, respectively) filets since 1975. We analyzed these data to estimate trends in PCB concentrations in these fish (1975–2010). We used generalized linear models with a gamma error distribution and log link fit to the untransformed concentrations. Trend patterns were examined using graphical smoothing and generalized additive models. We identified a candidate set of models that included time trend and other predictor variables. Using the Akaike Information Criterion to select among models we found the best models for both species included piecewise linear time trends, total body length, % lipid, and collection season as predictor variables. The intersection of the two trends was 1985 for chinook salmon and 1984 for coho salmon. PCB concentrations in both species increased with body length and % lipid, and were higher for individuals caught in the fall. Our data reveals a dramatic decline in PCB concentrations of − 16.7% and − 23.9% per year for chinook and coho, respectively, up until the intersection year likely reflecting implementation of restrictions on Aroclor-based PCBs. After the intersection year to 2010, PCB concentrations declined at an annual rate of − 4.0% (95% CI: − 4.4% to − 3.6%) and − 2.6% (95% CI: − 3.3% to − 1.9%) for chinook and coho, respectively.     

A Comparison of Photosynthate Allocation in Lakes

We compared the relationships between photosynthate allocation to protein, carbohydrate, lipid and low molecular weight (LMW) fractions and the variables daylength and water temperature in Lakes Huron, Michigan, and Ontario as well as three smaller headwater lakes in the Lake Ontario drainage. In all lakes investigated the allocation of recently produced photosynthate to carbohydrate was strongly related to daylength (% carbohydrate = −3.5 * daylength (hr) + 72.8; n = 59, r² = 0.56). The percentage of photosynthate allocated to protein was a function of water temperature in all lakes although the y-intercept for the protein-temperature relationship was much lower in the three headwater lakes and Lake Ontario (% protein = 0.50 * temperature (°C) + 6.1; n = 37, r² = 0.52) than in Lake Huron and Lake Michigan (% protein = 0.68 * temperature (°C) + 24.2; n = 23, r² = 0.49). The increase in allocation to protein was related to a decrease in allocation to low molecular weight material (% LMW = −1.1 * % protein + 57.13; n = 60, r² = 0.72). The percentages of photosynthate in lipid and LMW material were not related to any of the environmental variables measured. Assuming that photosynthate allocation is related to biochemical composition, the phytoplankton in Lakes Huron and Michigan were more protein rich for a given temperature than those in Lake Ontario and in the smaller inland lakes. The protein deficit was due to an increase in allocation to LMW material.     

Using LiDAR to reconstruct the history of a coastal environment influenced by legacy mining

LiDAR (light detection and ranging) data can be used to create fine digital elevation and bathymetric models (DEMs). Here we examine natural coastal erosion in Grand Traverse Bay, Michigan, a part of Keweenaw Bay in Lake Superior, and discuss how a variety of geological features (submersed river bed and channels associated with the Houghton Low; Nipissing dunes) interact with long-term sediment accumulation patterns. The geological features also modify migrating tailings from a legacy mining site. The combination of LiDAR derived images and aerial photographs allowed us to reconstruct the historical movement of tailings along the coastline. A total of 22.8 million metric tonnes (Mt) of stamp sand were discharged into the coastal environment off Gay, MI. Over a span of 80 years, beaches to the southwest of Gay have progressively received 7.0 Mt (30.7%) of the mass eroded from the original pile, whereas 11.1 Mt (48.7%) have moved into the bay. The total amount accumulated along the beaches now greatly exceeds the mass remaining on the original tailings pile (3.7 Mt; 16.2%). Bathymetric differences between two LiDAR surveys (2008 and 2010) were also used to estimate the mass, and to track the movement of migrating underwater stamp sand bars. These bars are moving southwesterly towards Buffalo Reef, creating a threat to the lake trout and lake whitefish breeding ground.     

Age and growth of round gobies in Lake Michigan,with preliminary mortality estimation

The round goby (Neogobius melanostomus) is a prevalent invasive species throughout Lake Michigan, as well as other Laurentian Great Lakes, yet little information is available on spatial variation in round goby growth within one body of water. Age and growth of round goby at three areas of Lake Michigan were studied by otolith analysis from a sample of 659 specimens collected from 2008 to 2012. Total length (TL) ranged from 48 to 131 mm for Sturgeon Bay, from 50 to 125 mm for Waukegan, and from 54 to 129 mm for Sleeping Bear Dunes. Ages ranged from 2 to 7 years for Sturgeon Bay, from 2 to 5 years for Waukegan, and from 2 to 6 years for Sleeping Bear Dunes. Area-specific and sex-specific body–otolith relationships were used to back-calculate estimates of total length at age, which were fitted to von Bertalanffy models to estimate growth rates. For both sexes, round gobies at Sleeping Bear Dunes and Waukegan grew significantly faster than those at Sturgeon Bay. However, round goby growth did not significantly differ between Sleeping Bear Dunes and Waukegan for either sex. At all three areas of Lake Michigan, males grew significantly faster than females. Based on catch curve analysis, estimates of annual mortality rates ranged from 0.79 to 0.84. These relatively high mortality rates suggested that round gobies may be under predatory control in Lake Michigan.     

Feeding ecology of age-0 lake whitefish in Saginaw Bay,Lake Huron

Age-0 lake whitefish Coregonus clupeaformis (11–160 mm total length) were collected from Saginaw Bay, Lake Huron during April–November 2009 and 2010 for diet analysis and for the evaluation of ontogenetic changes in feeding ecology. Lake whitefish ≤ 50 mm ate mainly zooplankton, after which their diets switched mainly to benthic macroinvertebrates. Cyclopoida were the dominant prey consumed by very small lake whitefish (< 17 mm) and the most frequently selected zooplankton type for individual small fish. Once lake whitefish reached 18–19 mm, Cyclopoida in the diet declined and cladocerans emerged as an important diet item. Daphnia were the most common cladoceran in the diets, but for fish 31–50 mm Bosminidae were also relatively important. Although the shift to Daphnia could represent an ontogenetic point when lake whitefish were large enough to effectively handle this prey, it also took place when the relative availability of Daphnia increased. Once lake whitefish were > 50 mm, Chironomidae larvae became a dominant prey item and this shift to benthivory coincided with a 55% increase in length-adjusted energy content between June and July. However, as fish grew (around 110–120 mm), Sphaeriidae and the benthic zooplankton Chydoridae became increasingly important in the diet. As these less energetically rich prey were incorporated into the diet, there were corresponding 21 and 15% decreases in length-adjusted energy content from July to August and September, respectively.     

Carbon,nitrogen and phosphorus loadings from tilapia fish cages in Lake Malawi and factors influencing their magnitude

Lake Malawi is a great lake not only because of its large size (30,800 km²); but also because it hosts between 700 and 1000 fish species, mostly endemic cichlids. The fish are most abundant and more diverse in the littoral zone. Cage aquaculture in Lake Malawi started in 2004 in the nearshore waters of the southeast arm. Approximately 50 cages were operational by 2009 harvesting about 20 tons fresh fish/cage/year. In 2007, data were aggregated from cage feed and production records to produce a mass balance for carbon (C), nitrogen (N) and phosphorus (P) added to cages in the form of feed and juvenile fish for 22 production cycles. Nutrient losses from cages to the surrounding environment accounted for between 81 and 91% for C, 59 and 80% for N and 85 and 92% for P. Production strategies such as stocking density and feeding rates were also analysed to assess their effect on fish harvests and nutrient loads from the fish cages. The study showed that nutrient loads from the cages were significant and can be exacerbated by poor feed quality, overstocking, stocking of premature fish and use of lower than recommended feeding rates to grow the fish to target sizes. The inefficient production strategies necessitated longer production cycles (376 ± 42 days) and more feed usage to achieve market sizes of ≥ 300 g/fish than if generally recommended stocking and feeding rates were followed. If cage aquaculture continues to expand, it will become an important new source of nutrients, not only locally, but on a lake-wide basis.     

Temporal and spatial differences in deposition of organic matter and black carbon in Lake Michigan sediments over the period 1850–2010

Inorganic carbon (IC), total organic carbon (OC), and black carbon (BC) were analyzed in eight sediment cores obtained from deep water (>30 m) sediments in the Chippewa and south Chippewa basins, as well as Green Bay in Lake Michigan. These cores were segmented at high resolution and radio-dated to reconstruct a detailed history of deposition to the lake both spatially and temporally since ca. 1850 CE. To help interpret the depositional record, cores were also characterized for stable isotopes (¹³C and ¹⁵N), as well as particle size distribution, density, organic matter (OM), and other parameters. Fine (silt and clay) sediment particles contained OM of primarily lacustrine algal biomass origin. Sedimentation fluxes showed large increases in OM and OC fluxes through much of the lake during the onset of industrialization and the period of rapid industrialization to onset of Great Lakes environmental legislation. In contrast, fluxes and loading of BC increased dramatically in the southern basin until the 1930's, then decreased substantially after the 1940's. This observation was due largely to results from site M009 nearest the steel mills and industrial zones of Chicago and northern Indiana. Together, whole lake loadings of OM and BC provide evidence that changing industrial activity and legislation intended to curb air pollution in the Great Lakes region have had a fairly rapid and dramatic impact. In contrast, legislation intended to decrease eutrophication through reductions in nutrient loading to the lake have not had a similar impact on sedimentation of OM in the lake.     

Horizontal and Vertical Distributions of Mercury in 1983 Lake Superior Sediments with Estimates of Storage and Mass Flux

Persistent elevated mercury concentrations in some species of Lake Superior fish and the general lack of information on mercury storage in and fluxes to the lake's sediments prompted the analysis of samples that were collected in 1983. Results of the analyses support the conclusion that Lake Superior sediments have mercury concentrations above background levels at all sites sampled. For those cores which penetrated the sediment deeply enough (the majority of the cores), background mercury concentrations ranged between 0.016 and 0.048 mg/kg. Mercury concentrations in surficial sediments ranged between 0.027 and 0.96 mg/kg. The maximum mercury concentration found in sub-surface sediments (2 to 20 cm deep) was 6.5 mg/kg. The surficial 20 cm of sediment contained 342 metric tons of mercury of which 51% or 174 metric tons was anthropogenic. The surface 2 cm of sediment contained 29 metric tons of mercury of which 76% or 22 metric tons was anthropogenic. Estimated total mercury fluxes to surficial sediments ranged between 0.1 and 10 ng/cm²/y with a mean of 3.2 ng/cm²/y. Background total mercury fluxes to the lake ranged between 0.20 and 0.72 ng/cm²/y with a mean of 0.48 ng/cm²/y. Estimated fluxes of anthropogenic mercury to surficial sediments ranged between –0.42 and 10 ng/cm²/y with a mean of 2.7 ng/cm²/y. The fluxes reported here are only the second known reporting of mercury fluxes to Lake Superior sediments. The inventory of mercury in the sediments is the first reported.     

Comparison of Otolith and Scale Ages for Yellow Perch from Lake Michigan

The age composition of the yellow perch (Perca flavescens) population in Lake Michigan is assessed annually by five management agencies, but all agencies do not use the same structure to estimate ages. The reliability of the most commonly used structure, scales, has not been formally evaluated for this population. We compared ages estimated by three readers from scales and sagittal otoliths for 150 yellow perch from southwestern Lake Michigan. The maximum age of yellow perch determined from scales and otoliths was 12. Otoliths had better precision (reproducibility) and usually had more annuli than scales for all three readers. Scale ages were usually younger than otolith ages when otolith ages were ≥ 7. Chi-square tests revealed significant differences (P < 0.05) between the age distributions determined from scales and otoliths for two of the three readers. We recommend use of otoliths for aging Lake Michigan yellow perch greater than 150 mm in length because of greater precision, easier readability, and detection of more annuli.     

Sensitivity of Ancient Lake Ohrid to Local Anthropogenic Impacts and Global Warming

Human impacts on the few ancient lakes of the world must be assessed, as any change can lead to an irreversible loss of endemic communities. In such an assessment, the sensitivity of Lake Ohrid (Macedonia/Albania; surface area A = 358 km², volume V = 55 km³, > 200 endemic species) to three major human impacts—water abstraction, eutrophication, and global warming—is evaluated.It is shown that ongoing eutrophication presents the major threat to this unique lake system, even under the conservative assumption of an increase in phosphorus (P) concentration from the current 4.5 to a potential future 9 mg P m⁻³. Eutrophication would lead to a significant reduction in light penetration, which is a prerequisite for endemic, deep living plankton communities. Moreover, a P increase to 9 mg P m⁻³ would create deep water anoxia through elevated oxygen consumption and increase in the water column stability due to more mineralization of organic material. Such anoxic conditions would severely threaten the endemic bottom fauna. The trend toward anoxia is further amplified by the predicted global warming of 0.04°C yr^–1, which significantly reduces the frequency of complete seasonal deep convective mixing compared to the current warming of 0.006°C yr⁻¹. This reduction in deep water exchange is triggered by the warming process rather than by overall higher temperatures in the lake. In contrast, deep convective mixing would be even more frequent than today under a higher temperature equilibrium, as a result of the temperature dependence of the thermal expansivity of water. Although water abstraction may change local habitats, e.g., karst spring areas, its effects on overall lake properties was shown to be of minor importance.     

Systemic,early-season Microcystis blooms in western Lake Erie and two of its major agricultural tributaries (Maumee and Sandusky rivers)

Recurrent, massive cyanobacterial blooms composed mainly of the genus Microcystis indicate a broad-scale re-eutrophication of Lake Erie. In the past, ameliorating eutrophication relied on intentionally decreasing point-source tributary nutrient, especially phosphorus, loads to the lake. However, recent research has shown that tributaries load not only nutrients but also bloom-levels of phytoplankton, including Microcystis. We built on this previous work by sampling earlier in the year and in much smaller tributaries in both the Maumee and Sandusky systems. We found Microcystis wet biomasses in these tributaries averaged 3.16 mg/L (± 0.59 mg/L, one standard error of the mean) in 2009 and 3.42 mg/L (± 0.55 mg/L) in 2010. Importantly, we found Microcystis in small ditches in March, much earlier than previously observed. Microcystis biomass did not directly correspond to measured phosphorus, chlorophyll, or phycocyanin concentrations likely reflecting complexities associated with lagged physiological responses and/or non-linear growth relationships. Consequently, our findings emphasize that Microcystis blooms form a more broad-scale problem than previously documented, occurring far upstream much earlier in the year.     

Evidence for a Trophic Cascade Effect on North-shore Western Lake Erie Phytoplankton Prior to the Zebra Mussel Invasion

Increasing summer total phosphorus (TP) concentrations measured in samples from a municipal water intake off the north shore of western Lake Erie during 1976 to 1983 were inconsistent with TP loads to the western basin of Lake Erie and with phytoplankton densities in the intake samples, both of which declined over the same time interval. The long-term (1976 to 1988) summer TP data were inversely correlated (r = −0.858) with summer average maximum daily wind velocities, suggesting that low wind velocities contributed to anoxia at the sediment-water interface and high sediment TP release rates in summer. While TP loading reductions in the late 1960s and early 1970s likely contributed to phytoplankton declines, continued phytoplankton declines during the late 1970s to early 1980s could not have been caused by continued reductions in TP loadings while TP concentrations increased. The phytoplankton declines of the 1980s are more likely attributable to changes in the trophic cascade associated with dramatic declines in some species of zooplanktivorous fish during the 1970s and 1980s as a result of a restored walleye population. Long term phytoplankton densities were fit (R² = 0.902) to a multiple regression model with western Lake Erie TP loads and an index of zooplanktivore density as independent variables; the zooplanktivore component of this model was the most significant contributor to the prediction of phytoplankton density. The implications of these findings for maintenance of good lake water quality include the need to maintain strong piscivore populations as well as reduced phosphorus loads.