Position Descriptions

2020 Great Lakes Summer Fellows Program

  1. Improving water level forecasts in Lake Champlain basin: skill assessment of operational atmospheric models
    Mentors: Dima Beletsky (CIGLR, beletsky@umich.edu), Daniel Titze (CIGLR), James Kessler (NOAA GLERL)

Project 1 Information:

In recent years, severe floods caused by intense rain events and spring runoff caused significant destruction of property and infrastructure in the Lake Champlain basin. In addition, high lake water levels provided conditions for more shoreline destruction by wind waves and storm surges. To predict water level changes due to runoff, wind and waves, a real-time flood forecast modeling system is being developed for the basin. The modeling framework consists of a hydrodynamic lake model based on the Finite Volume Community Ocean Model (FVCOM) driven by hydrologic runoff from the Weather Research and Forecasting-Hydrologic distributed model (WRF-Hydro) and wind wave model based on the WAVEWATCH III model. Meteorological forcing in hydrologic and lake models comes from operational National Weather Service models and assessment of its accuracy is critical for improving water level predictions in the basin.

Research Questions:

  • What is the accuracy of overland and overlake forcings generated by operational atmospheric models that drive hydrologic, hydrodynamic, and wave models in Lake Champlain basin?
  • How does water level forecast accuracy compare when generated by different forcings?

Project Activities: Meteorological forcings that drive WRF-Hydro, WAVEWATCH III and FVCOM models are produced by the following operational atmospheric models: GFS (Global Forecast System) model, HRRR (High Resolution Rapid Refresh) model and its latest update HRRRx. Modeled meteorological variables will be extracted and compared with archived observations. In addition to comparing the skill of the forcing itself, FVCOM and WRF-Hydro model output and skill from runs with the various forcings will be compared.

  • Assess the skill of overland and overlake forcings by comparing atmospheric model output with observations of wind, precipitation, air temperature, humidity, etc.
  • Compare hydrologic and lake model outputs produced with different meteorological forcings.
  • Compare hydrologic and lake model skill produced with different forcings using observations at stream gages and lake level gages.

Required & Desired Skills: Candidates should have strong scientific and computing skills, and experience with data analysis (e.g., using Python, R, IDL). Knowledge of the UNIX/LINUX environment and background in meteorology, hydrology or oceanography is desired.

    1. Identifying patterns in Great Lakes environmental data collected by an underwater glider
      Mentors: Michael Fraker (CIGLR, mfraker@umich.edu), Russ Miller (CIGLR)

    Project 2 Information:

    CIGLR and NOAA GLERL have operated an underwater glider since 2012 in Lakes Michigan, Ontario, and Huron that has collected physical, chemical, and biological data along transects in each lake. These data provide a unique horizontal viewpoint of several key ecosystem variables, but have not been comprehensively analyzed as of yet. Exploring the spatial, seasonal, and between-lake patterns in the glider data may help clarify ecosystem processes and drivers that other monitoring data cannot.                                                      

    Research Questions: How much variation in the spatial patterns of key ecosystem variables do we see over a season, between years, and between lakes? Are patterns correlated with variation in key environmental drivers (e.g., precipitation, wind direction, spring warming rate) or other local or regional drivers?

    Project Activities: The fellow will become familiar with the data being collected and how it may be useful to the study of Great Lakes ecosystem dynamics. S/he will identify patterns in the data using approaches for analyzing large spatial and temporal datasets (e.g., random forest and gradient forest analysis, regression). 

    • Become familiar with the types of data available and how they may be used
    • Use various spatial statistics to identify patterns

    Required & Desired Skills: Ideally, the fellow will have an interest in analyzing large datasets and experience using the statistical package R. Knowledge of specific statistical approaches is useful, but not required. A familiarity with aquatic ecology also would be useful, but is not required.

    1. Environmental mechanisms influencing seasonal progression of phytoplankton in the western basin of Lake Erie
      Mentors: Reagan M. Errera (NOAA GLERL, reagan.errara@noaa.gov), Jim Hood (Ohio State University)

    Project 3 Information:

    The 2019 cyanobacterial harmful algal bloom (cHAB) in the western basin of Lake Erie was considered relatively severe compared to blooms over the last two decades. While bloom size has been related to discharge from the Maumee River watershed, seasonal succession and the competitive factors influencing species succession pre-, during, and post-bloom remain elusive. Identifying the key environmental factors that influence the shift in phytoplankton succession and dominance leading to a cHAB can improve cHAB forecasting and diagnose bloom and post-bloom drivers. In 2019, the GLERL-CIGLR monitoring program collected phytoplankton community composition via phytoplankton counts and FluoroProbe (bbe moldaenke©), providing the opportunity to examine bloom dynamics using multiple methods.

    Research Questions:

    • What is the seasonal progression of phytoplankton in the western basin of Lake Erie from April – Oct 2019; Do phytoplankton counts relate to FluoroProbe readings and phytoplankton group distinctions?
    • Which environmental factors are correlated to shifts in phytoplankton species dominance? Do shifts in the community composition of primary producers translate into community shifts in higher trophic levels (i.e., zooplankton)?
    • Can fluctuations and factors identified in 2019 phytoplankton succession provide additional understanding of bloom dynamics since 2015?

    Project Activities:

    • Identify key time periods within the phytoplankton assemblage composition when population demographics deviate based on phytoplankton composition and biomass.
        • Analyze phytoplankton counts provided by BSA Environmental Services, Inc. from Lake Erie in 2019.
        • Compare results to FluoroProbe readings and determine calibration
    • Identify environmental mechanisms related to phytoplankton population shifts and response strength.
        • Correlate environmental factors to phytoplankton community changes through both univariate and multivariate statistical
        • Perform a retrospective analysis based on mechanistic drivers identified in 2019 and to Fluoroprobe data collected between 2015 -2018 (based on findings in Activity 1).

    Required & Desired Skills: An understanding of aquatic ecology, with a basic understanding of phytoplankton groups and factors influencing growth characteristics and population dynamics. General understanding of statistics (experience with multivariate statistics desired but not required) and experience with statistical programs (such as R, Python, or Matlab).

      1. Estimating Bythotrephes abundance and potential impacts by three sampling methods:  MOCNESS, laser optical counter, and standard zooplankton net
        Mentors: Henry Vanderploeg (NOAA GLERL, henry.vanderploeg@noaa.gov) and Edward Rutherford (NOAA GLERL)

      Project 4 Information:

      Bythotrephes is an invasive, visually-preying cladoceran zooplankton that has wreaked havoc on the food web of the Great Lakes by preying on other zooplankton that serve as food for fish early life stages. Ongoing research by NOAA GLERL and CIGLR aims to characterize the diel vertical migration of Bythotrephes, larval fishes, and zooplankton to better understand their interactions. We have analyzed our zooplankton samples from standard vertical opening and closing nets and from tows of a laser optical counter (LOPC) that allow us to estimate abundance by plankton size. Both methods give wildly different estimates of Bythotrephes abundance, and we have reason to suspect that the standard net tows underestimate Bythotrephes abundance.  We have also collected samples for larval fishes, the shrimp-like crustacean Mysis, and other zooplankton with a large Multiple Opening-Closing Net Environmental Sampling System (MOCNESS), but have not examined the samples for Bythotrephes.  We believe the MOCNESS method will give better results, and with the aid of the LOPC, allow us to determine fine-scale distribution of Bythotrephes and better estimate its consumptive impact on other zooplankton.

      Research Questions:

      • What is the vertical distribution, abundance, size (biomass), and reproductive status of Bythotrephes determined by the MOCNESS?
      • How does estimated abundance and biomass of Bythotrephes sampled by MOCNESS compare with estimates using other sampling methods, and what would these varying estimates imply for our estimates of consumptive impact of zooplankton by Bythotrephes?

      Project Activities: We hypothesize that the MOCNESS provides estimates of Bythotrephes abundance and biomass that are much higher than what was collected by standard net tows, and that using MOCNESS has the further advantage of allowing direct comparison with other zooplankton predators (larval fish and Mysis) collected in the same samples.

      For this project, the fellow will:

      • Examine archived samples to determine concentration, biomass, and fecundity of Bythotrephes in different depth zones sampled by MOCNESS.
      • Compare estimates determined by MOCNESS with other sampling methods.
      • Participate on at least one 5-day research cruise to sample zooplankton, larval fishes, Mysis, and Bythotrephes using different sampling technologies, and as time allows examine samples collected this field season.
      • Work with mentors to determine the consequences of the new estimates to evaluate consumptive impact of Bythotrephes by application of a bioenergetics model.

      Required & Desired Skills: Ideal candidates should be familiar with zooplankton and limnology or biological oceanography. A solid background in statistical analysis is necessary and working in R is desirable. Experience on large lake or oceanographic cruises is useful but not required.

        1. Evaluation of next generation runoff risk tools combining process-based and statistical models
          Mentors: Brent Lofgren (NOAA GLERL, brent.lofgren@noaa.gov), Lacey Mason (NOAA GLERL), Yao Hu (University of Delaware), Lindsay Fitzpatrick (CIGLR)

        Project 5 Information:

        Agricultural fertilizers applied just before a heavy rainfall can wash away into local waterways, costing farmers money and causing serious downstream water quality problems. A collaborative project led by the NOAA National Weather Service (NWS) has produced Runoff Risk decision support tools to help farmers determine the best time to apply fertilizers based on weather forecasts and soil moisture conditions. A research team from NOAA GLERL, CIGLR, and the University of Delaware is working with the NWS to develop the next generation of Runoff Risk tools using newer technology including the NOAA National Water Model. This will enable finer-scale forecasts and continuous improvements.

        Research Questions:

        • How well does the NOAA National Water Model (NWM) directly simulate runoff at the edge-of-field scale?
        • How much is this improved by using a statistical procedure to use specific NWM outputs to predict edge-of-field runoff?

        Project Activities:

        • Interact with other project members to acquire modeled output and “ground truth” data at edge of field locations.
        • Conduct an analysis to evaluate and compare the skill of the two modeling frameworks (National Water Model fluxes vs. statistical postprocessing of National Water Model output) at predicting the risk of runoff at edge of field locations.
        • Provide recommendations to the team in the form of a final presentation, including graphics summarizing model skill and the intercomparison results.
        • Attend bi-weekly project meetings and present progress.

        Required & desired skills: Ability to conduct data analysis and, in cooperation with team members, interpret the results to assess the suitability of different methods. A background in hydrologic modeling and ability to code in scripting languages such as Python, Matlab, or R is desired.

          1. Defining bottlenecks to fish larvae growth, survival, and potential recruitment in Lake Michigan
            Mentors: Ed Rutherford (NOAA GLERL, ed.rutherford@noaa.gov), Doran Mason (NOAA GLERL), and Henry Vanderploeg (NOAA GLERL)

          Project 6 Information:

          The declining productivity of Great Lakes food webs is thought to affect reproductive success of prey fish that support valuable fisheries. Survival and potential recruitment of fish (i.e., numbers of young adults) often is determined during the egg and larval stages. Recent studies in Lake Michigan indicate larval fish growth rates, condition, and their zooplankton prey densities have declined since invasive quagga mussels irrupted in the early 2000s. Analysis of fish larvae diets suggests that Dreissena mussel veligers have replaced native zooplankton as prey, and may have caused the decline in fish larvae growth rate. Energy density of veligers and its relative importance to fish larvae growth and survival are unknown.

          Research Questions:

          • What is the relative importance of Dreissena mussel veligers in the diets of larval fishes?
          • Can a diet of veligers support the historic growth and production of larval fishes in the Great Lakes?

          Project Activities:

          • Help conduct surveys and laboratory analysis to learn how density, availability, and energetic content of Dreissena veligers affect diet and growth of larval fish (yellow perch, alewife, bloater).
          • Supplement data analysis of new samples collected during diel surveys in Lake Michigan in 2020 with analysis of prior plankton collections in Lake Michigan in 2019.
          • Use data from the field surveys and lab analysis to configure and run bioenergetics models of larval fish growth to understand whether consumption of invasive mussel veligers or changes in lake temperature have affected larval growth rates.

          Required & Desired Skills: A strong background in freshwater or marine science and completion of an introductory statistics course are required. Experience working on boats in variable weather conditions is desired.

            1. Connection between Great Lakes and Arctic Ice Cover in Response to Teleconnection Patterns
              Mentors: Jia Wang (NOAA GLERL, jia.wang@noaa.gov), Ayumi Fujisaki-Manome (CIGLR), Philip Chu (NOAA GLERL), Yoyo Lin (CIGLR)

            Project 7 Information:

            Great Lakes ice cover and thickness are controlled by local meteorology (e.g. air temperature), which is heavily impacted by external, remote forcing from teleconnection patterns.  Both Great Lakes and Arctic sea ice variability is driven by a combination of these teleconnection patterns, such as Arctic Oscillation/North Atlantic Oscillation, El Niño-Southern Oscillation, Pacific Decadal Oscillation, and Atlantic Multidecadal Oscillation, whose thermodynamic impacts are difficult to separate. CIGLR and GLERL intend to conduct in-depth research linking climate teleconnection patterns to the Great Lakes and Arctic climate and ice cover/thickness, leading to development of hindcast models: multi-variable non-linear regression models. The project is part of the prediction of ice cover/thickness in the Great Lakes and the Arctic in response to a changing climate on seasonal, interannual, and decadal time scales, which enables us to provide information to broader users in search and rescue operations, navigation (commercial shipping), and recreational ice fishing during winter season. These forecasts provide decision makers with tools to aid in protecting the Great Lakes and the Arctic community.

            Research Questions:

            • In addition to local meteorology, is Great Lakes ice cover correlated to Arctic sea ice cover in response to the same teleconnection patterns? 
            • If so, how strong are these correlations and are they linear or non-linear?
            • Can we hindcast and predict Great Lakes and Arctic ice cover using climate indices?

            Project Activities: The fellow will participate in research to investigate the relationship among atmospheric teleconnection patterns and lake/sea ice cover in the Great lakes and the Arctic Ocean (or in their subareas). The predominant relationship identified in this project will be used to develop a tool to predict Great Lakes and Arctic ice cover. Specific activities include:

            • Investigate the relationships among climate indices, Great Lakes ice cover, and Arctic sea ice cover using data analyses, including time series analysis, correlation, and/or empirical orthogonal function analysis, and regression analysis.
            • Identify predominant relationships between ice cover in a Great Lake (or a certain subarea within a lake) and in a region in the Arctic Ocean.
            • Identify climate indices that control the above predominant relationship in lake/sea ice cover.
            • Develop a tool to forecast ice cover in the Great Lakes and the Arctic Ocean using the relationships identified from the above activities.

            Required & Desired Skills: Experience with at least one programming language (e.g., R-software, Fortran, Python) is required. Data analyses and statistics background are desired.

              1. Toward flood resilience in Great lakes coastal regions: numerical modeling of flood risks in different zones under climate change
                Mentors: Yi Hong (CIGLR, yhon@umich.edu), Eric J. Anderson (NOAA GLERL)

              Project 8 Information:

              In May 2019, new high water level records were set on Lakes Erie and Superior, and there has been widespread flooding across Lake Ontario for the second time in three years. These conditions have had wide spread impacts on many coastal communities. Coastal inundation can be caused by a variety of phenomena such as intense rainfall and extreme storm surges. In order to support resilience in coastal communities to extreme weather and climate change, a multi-scale hydrologic and hydrodynamic approach is necessary.

              Research Questions:

              • Can improved understanding of the frequency and intensity of extreme weather events improve our ability to predict coastal flooding?
              • How do we couple hydrological and coastal hydrodynamic models to improve flood forecasting?

              Project Activities: This project aims to use hydrologic and hydrodynamic models (FVCOM, WRF-Hydro) to evaluate coastal inundation in the Great Lakes. The summer fellow will use existing models to explore a series of hydro-meteorological scenarios for assessing the flood risks of different coastal regions. Project tasks may include:

              • Collecting and analyzing data for model configuration;
              • Implementation of the model with different simulation scenarios;
              • Extracting and analyzing the model outputs.

              Required & Desired Skills: We are looking for a highly motivated graduate/undergraduate student with interest in hydrology, hydrodynamics, environmental science, and numerical modeling. The ideal candidate will have or be working toward a degree in an appropriate discipline (such as Environmental Science, Engineering, etc.) with strong skills in data analysis and computer science.

                1. Climatic influences on antipredator phenotypic plasticity in larval amphibians
                  Mentor: Michael Fraker (CIGLR, mfraker@umich.edu)

                Project 9 Information:

                Climate change can impact ecosystems by changing the conditions under which species interact (e.g., by changing resource availability or the patterns of key abiotic factors such as spring warming rates). Predator-induced phenotypic plasticity is common across many taxa (including zooplankton, macroinvertebrates, and fish), and nonlethal predator-prey interactions mediated by induced traits can strongly influence the structure and functioning of the ecosystems in which they occur. Identifying how climatic factors influence phenotypically-plastic traits will clarify how a key component of many ecosystems may react to future conditions. Tadpoles have become a model system for research on phenotypic plasticity because both the physiological mechanisms regulating their plastic responses and the community- and ecosystem-level effects have been well-studied.

                Research Questions:

                • Using existing samples of tadpoles and predatory invertebrates from a 20-year biannual survey of 39 pond communities, how do the morphological phenotypes of 6 species of tadpoles vary between relatively wet years (e.g., 2002) and drought years (e.g., 2007)?
                • Which abiotic and biotic variables are common drivers among species?

                Project Activities: All individuals (tadpoles and invertebrates) from each sample year will be digitally photographed, then morphological characteristics will be measured. The morphological measurements will be combined with data on the abundance of all pond species, pond characteristics (e.g., hydroperiod, canopy cover), and meteorological data, then analyzed using multivariate approaches. The project will focus on the wettest and driest years first, then expand to other years as time permits.

                • Digitize samples
                • Measure morphology of all individuals using ImageJ
                • Use multivariate statistical approaches (e.g., CCA, random forest) to identify patterns and drivers of tadpole phenotypes

                Required & Desired Skills: Ideally, the fellow will have an interest and background in aquatic community ecology and climate change. Experience using the statistical package R also is useful, but not required.