DWQ-2024-004517

1
Utah Lake Water Quality Study (ULWQS) Science Panel March 23, 9:00 AM to 12:00 PM Virtual Meeting Meeting Summary - FINAL
ATTENDANCE:
Science Panel Members: Zach Aanderud, Mitch Hogsett, Theron Miller, Hans Paerl, and Tim Wool Steering Committee Members and Alternates: Scott Bird, Eric Ellis, Chris Keleher, and John Mackey Members of the Public: Dave Epstein and David Richards Utah Division of Water Quality (DWQ) staff: Scott Daly and Nicholas von Stackelberg Technical Consultants: Kateri Salk
Facilitation Team: Heather Bergman and Samuel Wallace
ACTION ITEMS
Who Action Item Due Date Date Completed
David Richards Send his report on no detrital rain to Samuel Wallace and Scott Daly to distribute to the Science Panel. May 1
Tetra Tech Develop options for aggregating the Utah Lake data for the stressor-response analyses for discussion by the Nutrient Criteria Development Subgroup.
April 19 April 19
Samuel Wallace and Scott Daly Reach out to absent Science Panel members to see if they are interested in joining the Criteria Development Subgroup.
March 27 March 27
DECISIONS AND APPROVALS No formal decisions or approvals were made at this meeting. CURRENT STATUS OF THE ULWQS PROCESS Scott Daly, DWQ, presented a brief overview of the current status of the ULWQS process. His comments are summarized below.
• The Science Panel is transitioning from the research phase to the application of findings phase.
• The Science Panel is being asked to review the studies and consider how the findings can be applied to the development of the numeric nutrient criteria (NNC), in-lake nutrient and watershed models, and responses to the Steering Committee charge questions.
OVERVIEW OF AVAILABLE SEDIMENT NUTRIENT RECYCLING STUDIES Dr. Kateri Salk, Tetra Tech, provided an overview of the available sediment nutrient recycling studies. Her presentation is summarized below.
2
Overview
• The purpose of today’s discussion is to:
o Review studies conducted on Utah Lake sediment nutrient cycling
o Contextualize the importance of sediment characterization in the NNC development
• The Science Panel can use the existing sediment data and studies to respond to charge questions posed by the Steering Committee, inform the Utah Lake Nutrient Model, inform the Technical Support Document development, and inform implementation planning. More specifically, the Science Panel can use the sediment research to respond to sediment-related charge questions, inform the sediment-relevant parts of the Utah Lake Nutrient Model (e.g., sediment diagenesis, settling), and inform potential timescales for implementation planning.
• There have been several key studies to date, including: o Sediment Fluxes and Equilibrium Phosphorus Concentration Study o Littoral Sediment Study
o Phosphorus-Binding (P-Binding) Study
o Timpanogos Special Service District (TSSD) Limnocorral Study
o Carbon, Nitrogen, and Phosphorus (CNP) Study
o Dr. Brett’s mass balance analysis
o Other studies from the literature CNP Study and SedFlux Model
• Tetra Tech conducted the CNP Study with oversight from the Science Panel. Tetra Tech began the study by reviewing the literature on nutrient-relevant processes and pools in Utah Lake. Using the relevant literature, Tetra Tech developed a conceptual model that included sediment pools and fluxes across the sediment-water interface. They also developed an external mass balance that did not involve sediments and an internal mass balance that primarily incorporated sediments through the SedFlux model.
• The conceptual model provides an overview of the relevant phosphorus processes and chemical forms between the water column and sediments. The conceptual model provides a value for a specific phosphorus pool or process and assigns a level of confidence to that value. Some pools and processes have not been measured directly and are filled in with literature-derived values. One note is that the dissolved forms of phosphorus tended to be released from the sediment, while total phosphorus had net input to the sediments. The rates were also scaled up to tons/day within the lake, creating high confidence bands due to spatial and temporal variability.
• A similar conceptual model was created for nitrogen. Like phosphorus, dissolved nitrogen forms tended to be released from the sediment, while total nitrogen had a net input to the sediments. Nitrogen is unique to phosphorus in that it has an atmospheric input and output, which is documented in the conceptual model. Ultimately, the conceptual model quantifies the processes and integrates multiple studies into a chart.
• The purpose of the SedFlux model is to simulate nutrient fluxes and sediment oxygen demand (SOD) across the sediment-water interface. The model was adapted from the original work for QUAL2K and WASP (DiToro 2001). The model setup involved inputting data for Utah Lake and reaction network parameters from Su and von Stakceholer (2020). The model was not calibrated, so the Science Panel should defer to the EFDC/WASP model results and field observations when available, as these will be more accurate measurements.
• The SedFlux model simulates fluxes across the sediment-water interface for ammonium, nitrate, and soluble reactive phosphorus (SRP). Based on the given inputs, the SedFlux
3
model projected a positive flux of ammonium and SRP from the sediment to the water column. The model also projected a positive flux of nitrate (from the sediment to the water column in the summer) and a negative flux (from the water column to the sediment) in the spring and fall. The model also predicted high flux rates under high organic matter supply and more variable rates when the water column was shallow.
• The SedFlux model also simulates SOD. The model predicted higher flux rates under high organic matter supply and deeper water column. The modeled SOD was higher than the measured SOD by order of magnitude, indicating that the modeled rates are likely unrealistic. Tetra Tech investigated why the modeled rates were much higher than the measured rates. They found that SOD was not particularly sensitive to the reaction network parameters. SOD was sensitive to water column dissolved oxygen concentrations and the settling rate of particulate organic carbon. Several hypotheses as to why this dynamic was occurring include:
o The sediment dilutes incoming particulate organic carbon
o Frequent resuspension may lead to SOD becoming biochemical oxygen demand
o SedFLux may not capture important factors driving SOD.
• The SedFlux outputs for SRP, ammonium, and nitrate were comparable to other field-based studies (Hogsett et al. 2019 and Goel et al. 2020). The SedFlux outputs for SOD were substantially higher than the other field-based studies.
Sediment Charge Questions
• The Steering Committee had three charge questions related to sediments. The questions were: 1. What are current sediment equilibrium phosphorus concentrations (EPC) throughout the lake? What effect will reducing inputs have on water column concentrations? If so, what is the expected lag time for lake recovery after nutrient inputs have been reduced? 2. What is the sediment oxygen demand of, and nutrient releases from, sediments in Utah Lake under current conditions? 3. Does lake stratification [weather patterns] play a result in anoxia and phosphorus release into the water column? Can this be tied to harmful algal bloom formation?
• In response to the first question, Dr. Ramesh Goel quantified the sediment EPC in the main body of the lake and Provo Bay in the Sediment Fluxes and EPC Study. Those results were based on a few observations, and there was noise around some of the relationships in that study. The P-Binding Study will be able to provide a more precise EPC value using controlled batch experiments. The high-level overview of the Sediment Fluxes and EPC Study is that one should expect some degree of enhanced sediment loading following water column reductions until equilibrium is reached. Dr. Brett’s mass balance analysis will help the Science Panel and Steering Committee understand the lag time better.
• The second charge question asks what the SOD and the nutrient releases from sediments in Utah Lake are under current conditions. Hogsett et al. (2019), Randall et al. (2019), Goel et al. (2020), and Tetra Tech (2021) came to similar conclusions. Those conclusions are that the sediments are a net sink for total nutrients, so more nutrients are coming into Utah Lake than out. Additionally, bioavailable forms of nitrogen and phosphorus are released from the sediments. The nutrient release rates are spatially variable as well. The SOD tends to be positive, depending on the organic matter content of the sediment.
• The third question asked how lake stratification plays into anoxia. Utah Lake is a large system with a large fetch. There is evidence of transient thermal stratification in calm periods but no persistent seasonal stratification. Thus, there is no pattern of hypolimnetic
4
dissolved oxygen depletion and nutrient accumulation. It is possible that local zones of anoxia form. Some sediment phosphorus is bound to redox-sensitive compounds, so when there are local zones of anoxia, phosphorus bound to iron and manganese may be released. Lastly, frequent wind-driven mixing brings surface sediments into contact with the water column, which could be tied to harmful algal bloom formation.
Next Steps
• The P-Binding Study Subgroup will meet in April to finalize the P-Binding Study. The results of the Littoral Sediment Study will also provide additional insight into the questions around sediment nutrient recycling. Science Panel Clarifying Questions Science Panel members asked clarifying questions about the overview of available sediment nutrient cycling studies. Their questions are indicated below in italics, with the corresponding responses in plain text. How will the EFDC/WASP model integrate with episodic wind-driven resuspension events? The model can incorporate resuspension as part of the shear stress modeling. It is unknown what the exact plan is to model wind-driven resuspension at this time. The modeling team will present at the Science Panel meeting on March 24. Public Clarifying Questions Members of the public asked clarifying questions about the overview of available sediment nutrient cycling studies. Their questions are indicated below in italics, with the corresponding responses in plain text. Can the EFDC/WASP model simulate sediment-water fluxes if carp are removed and mollusks are added to the system? The modeling team will be able to respond to this question at the March 24 meeting. As far as current meeting participants know, the scenario management plans are related more to nutrient management than specific habitat management strategies.
Does the nutrient cycling flowchart account for the export of carp and midges leaving the lake? This part of the nutrient budget relies on the mass and stoichiometry of the organisms. The researchers did their best to fill in the specific pools and processes with available data. PRESENTATION ON THE ROLE OF INTERNAL NUTRIENT CYCLING IN A FRESHWATER SHALLOW ALKALINE LAKE Dr. Mitch Hogsett, ULWQS Science Panel, presented research on the role of internal nutrient cycling in Utah Lake. His presentation is summarized below.
Overview
• Dr. Hogsett’s research on SOD, nutrient fluxes, and sediment characteristics occurred from 2010 to 2012. The study was commissioned during discussions to establish a total maximum daily load (TMDL) for dissolved oxygen in the Jordan River. Those discussions prompted interest in the chemical characteristics of Utah Lake (the source of water for the Jordan River).
• The study employed a measurement device that has two chambers. One chamber has an opening at the bottom to measure SOD and other water quality parameters. The other chamber is closed at the bottom to measure oxygen demand in the water column. The
5
measurement device allowed researchers to compare SOD to the oxygen demand in the water column and capture the interaction between the sediments and the water column. The measuring device also captured the interaction between dissolved oxygen and nutrients in the sediment.
• Researchers used scuba gear to deploy the measurement device into the sediments. The installation process took about two to three hours. The installation also required connecting the device to a float tube to provide a power source.
• The device measured the oxygen demand in the water column under dark conditions, so photosynthesis was not a factor. It also measured the oxygen demand of the sediments and water column, dissolved oxygen saturation, and ambient conditions.
• The initial sediment and water column oxygen demand values were less than the ambient conditions because the measurement device was installed at the sediment-water interface. The device measured oxygen demand at a lower depth than the data collected under ambient conditions.
SOD Results
• The researchers collected SOD fluxes in August, September, and October. They selected these sampling times to measure SOD under favorable conditions for harmful algal blooms (e.g., warm water temperatures, which will accelerate biological processes). The highest SOD flux measured was -4.61 g/m2*day in Provo Bay. The next highest SOD flux measured was -2.04 g/m2*day in the northeast area of the main basin. On the east side of the lake, there were lower SOD values.
• The research team also measured surface sediment volatile solids representing organic matter and found a strong relationship between sediment organic matter and SOD. The presence of organic matter in the sediments increases the decay potential. When organic material decays, it turns oxygen into carbon dioxide. The highest SOD was in Provo Bay and the west side of Utah Lake, closer to urban centers.
• The research team measured water column dissolved oxygen depletion rates at each site. They could then calculate the ambient oxygen demand using the SOD and water column oxygen demand.
• The research team also calculated the percentage of oxygen demand associated with the sediments. The percent of oxygen demand associated with the sediment ranges from 10% to 40%. These values indicate that the sediment plays an important role in oxygen dynamics and that organic matter comes from the water column.
• The temperature of the sampling events ranged from 17 degrees Celsius to 23 degrees Celsius. Biological processes accelerate as temperature increases. The research team used known relationships between temperature and the speed of biological processes to standardize the SOD values at 20 degrees Celsius and 13 degrees Celsius. The SOD values were standardized to 13 degrees Celsius because it was the annual average temperature of Utah Lake based on grab samples from 2015 to 2020.
• There was no change in SOD in the downstream Jordan River from summer to winter. In the winter, biological oxygen demand decreased dramatically, but the sediments remained equally active in the winter than in summer. The hypothesis for this observed dynamic in the Jordan River was that there was so much organic matter in the sediments that in the summertime, there was an inhibition of dissolved oxygen diffusing into the sediments. In more organic sediments, the SOD would remain more constant throughout the year than in less organic sediments.
6
Nutrient Flux Results
• Utah Lake has many pathways for nutrient cycling (e.g., organic matter decay, periphyton growth, nitrification, and denitrification). The research team wanted to measure nutrient levels to understand better what processes were occurring at the sediment-water interface and water column.
• The sediment nutrient flux data indicates a positive ammonium flux, which suggests organic matter decay. There was also a positive flux of nitrate or non-detect values at all sites except one, which suggests nitrification occurs at the sediment-water interface. There was also a positive flux of orthophosphate or non-detect values at all sites except one, which suggests organic matter decay.
• The largest fluxes of ammonium and orthophosphate occurred at Provo Bay and on the east side of Utah Lake. These sampling locations are also where the sediments had the highest phosphorus concentration in other studies.
• A SOD flux of 1.5 grams of dissolved oxygen/m2*day converts to 0.014 grams of phosphorus/m2*day. The 0.014 grams of phosphorus/m2*day released from the sediments aligns with some of the orthophosphate flux measurements collected. These results beg the question of why there is not as much phosphorus flowing from Utah Lake into the Jordan River.
• The research team measured the nutrient rates for the water column. They found that phosphorus and nitrogen are removed from the water column. The research team could not identify the exact mechanism for nutrient removal because biological and abiotic processes (e.g., calcium-phosphorus interactions) could affect the measured outcome.
• Based on the sediment and water column nutrient fluxes, the average annual load of phosphorus and nitrogen across all sites would be 1,500 tons of phosphorus/year and 7,500 tons of nitrogen/year. These values suggest that the cycle of nutrient uptake by algae, the algae death, the decay, and the re-release of the nutrients in the sediments could occur up to seven times a year.
• The data collected on the main body of Utah Lake (i.e., Provo Bay excluded) shows that the average annual load to the main body of Utah Lake is 950 tons of phosphorus/year and 4,750 tons of nitrogen/year. The average annual load for the Utah Lake main body based on the 13 degrees Celsius standardized data would be 520 tons of phosphorus/year and 2,612 tons of nitrogen/year. Sediment Composition and Mineralogy Results
• The total and volatile solid data indicate more organic matter in the surface sediments in Provo Bay than in the main body of Utah Lake. The data also indicates that organic matter decreases deeper into the sediments.
• Roughly 60% of Utah Lake sediments are associated with carbonates, particularly calcite. There are also clays and silica oxides present in the sediments. The sediments at the entrance of Provo Bay notably had a high percent composition of silica oxides.
• The phosphorus-speciation data indicates that the majority of the phosphorus in the sediment was associated with calcium (40% to 83%). The phosphorus was also bound to iron and manganese. Science Panel Discussion and Comments Science Panel members provided comments on Dr. Hogsett’s internal nutrient cycling study. Their comments are summarized below.
• The dataset from this study will be useful in developing the in-lake water quality model.
7
• A high percentage of the phosphorus in the sediments is either iron or manganese bound. The calculation that standardized the nutrient loading to Utah Lake for the sediments based on the average annual temperature (13 degrees Celsius) may not be an accurate loading estimate since some of the phosphorus is bound to iron or manganese, not in an organic form.
• One difference between the Jordan River and the main body of Utah Lake is that according to this study, the sediments in Utah Lake were driven by organic matter decay. Organic matter decay was the source of nutrients and SOD in the Jordan River. There is less organic matter in Utah Lake than in the Jordan River. Downstream in the Jordan River, a reduction in temperature did not decrease nutrient loading. The internal nutrient cycling study did not collect data in the winter, so it is uncertain how temperature impacts nutrient loading in Utah Lake. There is more potential for temperature effects in Utah Lake than in the Jordan River.
• The methane oxidizers do not function as a result of temperature. They are active at low temperatures. This dynamic may be important for Provo Bay, where high methane release rates were observed. These methane levels were not observed in the main body of Utah Lake.
• In Dr. Aanderud’s Bioassay Study, he observed that the phosphorus being released from the sediments was largely tied up in the organic phosphorus fraction.
• In Dr. Ramesh Goel’s, University of Utah, 2020 report, he noted that denitrification and nitrogen fixation are data gaps that could be addressed in future research. Denitrification and nitrogen-fixation data at the sediment-water interface would help build an understanding of the nitrogen fluxes. The presence of nitrogen-fixation bacteria in the sediments does not necessarily mean nitrogen fixation is occurring. Additionally, nitrogen fixation is likely less important than denitrification regarding net flux.
Science Panel Clarifying Questions Science Panel members asked clarifying questions about Dr. Hogsett’s internal nutrient cycling study. Their questions are indicated below in italics, with the corresponding responses in plain text. What is known about nitrogen fixation from this study? Other studies published after this one have indicated a high potential for nitrogen fixation. There is the potential for bacterial nitrogen fixation at the sediment-water interface, even with ammonium fluxing out. Dr. Zach Aanderud’s Bioassay Study measured nitrogen fixation but not at the sediment-water interface.
Any thoughts on the cycling of nitrogen based on the results of this study? The study focuses on phosphorus cycling because nitrogen fixation and denitrification impact nitrogen cycling. Nitrogen fixation and denitrification were not measured in this study. Will the model be able to account for the rapid cycling of nutrients? Studies have indicated high internal cycling of nutrients in the lake. The rates presented in this study are not of concern to the modelers. Ensuring the model is accurate will be a function of correctly estimating the supply of nutrients from the water column so that the in-lake nutrient model accurately feeds to the sediment diagenesis model to simulate nutrient recycling.
Public Clarifying Questions Members of the public asked clarifying questions about Dr. Hogsett’s internal nutrient cycling study. Their questions are indicated below in italics, with the corresponding responses in plain text.
8
Are the SOD rates typical of other temperate eutrophic lakes? Yes, they are. The Provo Bay SOD rates may be comparatively higher. Another caveat for the study is that the measurement device’s chambers have a circulating velocity of 10 centimeters/second. Since Utah Lake is shallow, there is expected to be more sediment movement due to external factors (e.g., wind), which would increase the SOD rates by accelerating the interactions between the water column and oxygen. The SOD rates in Utah Lake may be higher due to that mixing within the lake.
Is the SOD rate in Provo Bay closer to the SOD rate found in wetlands? Would that imply that Provo
Bay should be managed more as a wetland? The Provo Bay SOD rates are related to organic matter. If Provo Bay were to dry, the SOD rates and organic matter build-up would reflect a wetland. It is uncertain what this conclusion means for future management options. Public Discussion and Comments Members of the public provided comments on Dr. Hogsett’s internal nutrient cycling study. Their comments are summarized below.
• The detrital rain (i.e., the organic material falling from the water column) may not only be algae. It could also be composed of zooplankton and decaying fish.
• Dr. David Richards, OreoHelix Consulting, did a study on detrital rain. This report will help create an understanding of nutrient cycling. Dr. Richards will send the report to Samuel Wallace and Scott Daly to distribute to the Science Panel. STRESSOR-RESPONSE ANALYSIS UPDATE AND NEXT STEPS Dr. Kateri Salk, Tetra Tech, presented an update on the stressor-response analysis and the next steps for the analysis. Her presentation is summarized below. Overview
• The NNC Technical Support Document (TSD) provides the technical basis for developing NNC to protect the designated uses. The designated uses of Utah Lake are recreation, aquatic life, agriculture, and downstream. The Science Panel will guide Tetra Tech in conducting analyses to develop multiple lines of evidence in the NNC framework.
• There are three primary lines of evidence: a) reference-based, b) stressor-response analysis, and c) scientific literature. The stressor-response analyses will rely on output from the Utah Lake nutrient model and statistical models. Additionally, the stressor-response analyses will incorporate in-lake monitoring data on water quality parameters and the application of the Environmental Protection Agency’s (EPA) ambient water quality criteria nutrient models.
• The primary datasets for the stressor-response analyses will include water chemistry, continuous buoy data, phytoplankton, and zooplankton.
o The water chemistry datasets will include surface and integrated grab data on multiple parameters from many sites around the lake. DWQ and Wasatch Front Water Quality Council (WFWQC) will primarily provide the water chemistry data. o The continuous buoy data will include surface data for dissolved oxygen, pH, temperature, turbidity, chlorophyll, and phycocyanin fluorescence parameters. The data will come from four sites. DWQ is the primary provider of this data.
o The phytoplankton dataset will capture surface composite and surface “scum” samples. The key parameters of this data are phytoplankton taxa abundance and toxins. The data will come from many sites around the lake. DWQ and WFWQC are the primary providers of this data.
9
o The zooplankton dataset will come from zooplankton tows, presumably capturing surface data. The key parameter will be zooplankton taxa abundance. The data will come from many sites around the lake. WFWQC, Brigham Young University, and Utah State University will be the primary providers of this data. This dataset requires harmonizing the data because each dataset uses different methodologies and identifies different taxa.
• The general approach for the stressor-response analysis is to tie ecosystem stressors to responses of interest through statistical models (e.g., linear regression, quantile regression, and logistic regression). The stress-response analysis involves assigning a threshold for the response and associating that with a threshold in the stressor. This analysis provides support for the threshold of the stressor that is protective of designated uses. The stressor-response analysis accounts for uncertainty and protectiveness and can incorporate confidence/credible intervals and prediction intervals/quantiles.
• The purpose of today’s discussion is to: o Highlight data availability for stressor-response relationships of interest o Make decisions about the aggregation approach o Set the stage for future feedback on the statistical approach and management-relevant decisions
Stressor-Response Relationships: Microcystin versus Cyanobacteria and Chlorophyll
• The first stressor-response relationship analyzed was the relationship between microcystin concentration and cyanobacteria counts. The analysis indicates a positive correlation between cyanobacteria cell count and microcystin. Linear and logistic regression may be appropriate to assess the stressor-response relationship. Linear regression creates a line of best fit. The logistic regression provides the probability that the microcystin concentration will cross the eight micrograms/liter threshold (the EPA 2019 recommended recreational criteria) at a certain cyanobacterial cell count. Additional analyses were conducted on the relationship between Dolichospermum, Microcystis, and Plankothrix cell count and microcystin concentrations.
• Several factors explain variability in cyanobacteria cell counts and microcystin concentrations. Whether a sample came from surface “scum” or whether it came from a composite sample affected cyanobacteria cell counts and microcystin concentrations. Additionally, the location of the sampling site within the lake affected cyanobacteria cell count and microcystin concentrations.
• EPA’s national model links chlorophyll, cyanobacteria, and microcystin from the National Lake Assessment (NLA). Due to its morphometry, Utah Lake is at the edge of the distribution of NLA lakes. The model allows users to input location-specific settings based on ecoregion and maximum lake depth. Due to Utah Lake’s unique system, a site-specific analysis may be more informative than the national model. Stressor-Response Relationships: Cyanobacteria versus Chlorophyll
• The relationship between cyanobacteria and chlorophyll could be used to relate cyanobacterial abundance thresholds to potential chlorophyll targets. EPA identifies potential cell count thresholds.
• Biovolume and relative abundance could also be used to relate cyanobacteria abundance thresholds to a potential chlorophyll target. The EPA has a review of state and international guidelines/action levels. The World Health Organization (WHO) also provides guidance on targets and thresholds.
10
Stressor-Response Relationships: pH versus chlorophyll
• The next stressor-response relationship is between pH and chlorophyll. The expectation is to see diel pH cycles align with photosynthesis and respiration cycles; photosynthesis increases pH, while respiration decreases it. There may be seasonal cycles that affect pH as well. The pH criteria is a “not to exceed” criteria, which makes it easier to link pH back to primary production. Utah Lake also has high alkalinity, so it is typical to see pH exceedances above 9 rather than exceedances below 6.5.
• The best pH data to link pH to chlorophyll comes from the continuous sondes because pH follows a diel cycle. The best chlorophyll data comes from the grab data, which measures chlorophyll as a concentration. Since there is more data for pH than chlorophyll, there is a need to consider an aggregation approach to link parameters across space and time. Stressor-Response Relationships: Dissolved Oxygen versus Chlorophyll
• The dissolved oxygen data is collected continuously from sondes. There are three nutrient criteria measures for dissolved oxygen: daily minimum (five and 3 milligrams/liter), seven-day mean (four and six milligrams/liter), and 30-day mean (5.5 milligrams/liter). There are also different criteria thresholds for early and late life stages. Some exceedances occur at sites across Utah Lake, but most occur in Provo Bay.
• Similar to pH, the dissolved oxygen data comes from continuous sondes, and the chlorophyll data comes from grab samples, so there is a need to consider an aggregation approach to link parameters across space and time. Stressor-Response Relationships: Zooplankton versus Chlorophyll
• The next stressor-response relationship analyzed is the relationship between zooplankton and chlorophyll. The EPA national model analyzes the trophic relationship between phytoplankton and zooplankton. The analysis indicates a tight linkage between chlorophyll and zooplankton observed at lower chlorophyll levels, but the relationship becomes decoupled at higher chlorophyll levels. As chlorophyll levels increase, the growth of zooplankton per unit increase of chlorophyll becomes slower. Due to this relationship, the Science Panel will need to select a slope threshold that protects against trophic decoupling to identify a chlorophyll target.
• Utah State University, Brigham Young University, and the WFWQC have zooplankton data for Utah Lake. Since each dataset uses different methodologies and taxa output, there is a need to determine how to combine the different datasets. The Science Panel will need to consider the metrics of interest (e.g., total taxa, certain taxa, etc.) and how those metrics can link to the protection of designated uses. The Science Panel will also need to consider whether to update the national model with Utah Lake-specific data or use the national model conceptual set up to inform the site-specific model.
Stressor-Response Relationships: Public Perception/Annual Visitation versus Chlorophyll,
Cyanobacteria, and Clarity
• A user perception is underway, led by Dr. Jordan Smith from Utah State University. The survey will quantify user perceptions to incorporate into the stressor-response analysis.
• The Utah Department of Natural Resources maintains visitation records of Utah Lake State Park. These records are an imperfect measurement to track annual visitation because it only captures annual visitation to the State Park, not total visitation to Utah Lake. Additional data from the user perception survey may be able to help.
11
• To assess the relationship between annual visitation and cyanobacteria, chlorophyll, and clarity, the stressor-response analysis can link monthly visitation totals to chlorophyll, cyanobacteria, and clarity measurements for a subset of the growing season.
Stressor-Response Relationships: Clarity versus Chlorophyll Secchi depth is negatively related to chlorophyll. Many variables affect Secchi depth, considering that two-thirds to three-quarters of turbidity in Utah Lake is non-algal. There is a need to establish a threshold for Secchi depth value protective of designated recreation uses. Stressor-Response Relationships: Chlorophyll versus Total Nitrogen and Total Phosphorus
• Once chlorophyll targets that are protective of uses are identified, the next step is to link chlorophyll to nutrients. The Science Panel will need to consider the specific mechanistic linkages.
• There is a positive relationship between nutrient concentrations and chlorophyll concentrations. There is some variability in the relationship due to nutrient limitation and non-bioavailable nutrient pools. The unequal variance may be suited for quantile regression.
• The EPA national models include chlorophyll-nutrient models. The chlorophyll-nutrient model considers other pools of phosphorus and nitrogen that are not phytoplankton in origin (e.g., sediment-bound phosphorus, inorganic nitrogen). There is an option to input lake-specific settings into the model (e.g., ecoregion, maximum lake depth, dissolved organic carbon, and turbidity). The results from the EPA national model can be used as another line of evidence to pair with the site-specific analysis. Stressor-Response Relationships: Cyanobacteria versus Total Nitrogen and Total Phosphorus Previous analyses related cyanobacteria cell count with total nitrogen and total phosphorus values. The previous analysis used quantile and logistic regression with a threshold of 100,000 cells/milliliter. A similar analysis could be conducted for this stressor-response analysis. Stressor-Response Relationships: Clarity versus Total Nitrogen and Total Phosphorus Secchi depth is negatively related to total phosphorus and total nitrogen concentrations. There is variability in this relationship due to the high proportion of non-algal turbidity in Utah Lake. There is a need to establish a threshold for Secchi depth value protective of designated recreation uses. Science Panel Clarifying Questions Science Panel members asked clarifying questions about the stressor-response analysis update. Their questions are indicated below in italics, with the corresponding responses in plain text. Is there photosynthetically available radiation (PAR) data available? There is a good amount of PAR data available. It does not go back as far as other datasets, but several sites have monthly data. The PAR data could be used to calculate light attenuation coefficients. Science Panel Discussion and Comments Science Panel members provided comments on the stressor-response analysis update. Their comments are summarized below.
• The number of carp in Utah Lake affects the zooplankton-algae relationship in Utah Lake by altering the zooplankton population. Because of the carp, most zooplankton is less than one-tenth of a millimeter. In the absence of carp, larger zooplankton would serve as top-down
12
control on algae. There are mitigating and altering factors involved with zooplankton populations that the Science Panel should account for in the stressor-response analysis.
• Secchi depth can be related to algae in a stratified lake, but in Utah Lake, high turbidity could be from a wind event. Secchi depth may not be a relevant measurement in Utah Lake, given the high level of non-algal turbidity. Thinking of another direct metric with fewer confounding variables, such as PAR data, may be useful. Public Comments Members of the public provided comments on the stressor-response analysis update. Their comments are summarized below.
• Given the alkalinity of Utah Lake, the pH criteria of nine may not be attainable for Utah Lake.
• Utah Lake has naturally elevated pH levels. However, pH is easier to measure than chlorophyll. Spikes in pH may indicate algal blooms, so it is worth exploring the relationship between pH and chlorophyll to determine those relationships.
• A pH of nine is the standard to determine whether a designated use is protected. A target can be created if pH can be reliably linked to primary production. There is a stipulation for background conditions if high pH levels occur regardless of primary production. The stressor-response analysis will help determine the difference between pH levels driven by primary production and naturally occurring variability.
• Dr. David Richards is working on a multi-metric index for zooplankton and phytoplankton. The researchers working on that analysis have identified over two dozen zooplankton metrics. Having the Science Panel review that data and index might be helpful. STRESSOR-RESPONSE ANALYSIS DECISION POINT DISCUSSION Dr. Kateri Salk, Tetra Tech, shared the key decision points for the stressor-response analysis the Science Panel will need to consider. The decision points and the Science Panel discussion are summarized below.
Decision Point Overview
• All the stressor-response analyses shared in the previous presentation were a spatial and temporal aggregation of all available data. As part of the stressor-response analysis, the Science Panel will need to consider if there are different approaches to aggregating the data. Some of the aggregation approaches the Science Panel will need to consider include the following:
o Seasonal aggregation
o Depth considerations
o Period of interest
o Spatial aggregation
• The Science Panel will need to define the growing/recreation season. Tetra Tech suggests defining the growing season from April to September. Once the Science Panel has defined the growing season, they will need to decide the statistical metric for aggregating the data to the growing season (e.g., mean, geometric mean, median).
• The Science Panel will need to define the depth that represents the surface. Tetra Tech suggests defining the surface as less than one meter and any data labeled as an integrated or composite surface sample.
• The next decision point for the Science Panel is whether to use all the available data or set a period of interest for relevant data. The greatest sample density has occurred in the past five to ten years, but there are collected samples from the late 1980s.
13
• The final decision is to assess Utah Lake as one assessment unit or break it into regions. One approach could be to break Utah Lake into two assessment units: the main basin and Provo Bay. Goshen Bay could also as act as a separate zone. Breaking the lake into different assessment units may lead to different targets for each region, which would have management implications.
• Once the group has decided how to break or not break the lake into different regions, the next consideration is how to aggregate data spatially and temporally. There are several potential sequences to aggregate data, such as:
o Aggregate all sites for a given date, then aggregate to the growing season
o Aggregate each site to the growing season, then aggregate across sites
o Aggregate all samples across sites and dates for a growing season Science Panel Clarifying Questions Science Panel members asked clarifying questions about the stressor-response analysis decision points. Their questions are indicated below in italics, with the corresponding responses in plain text. How will the Science Panel deal with having more data in recent times than in the past? More samples
mean that more blooms will be documented. This question is a common issue in stressor-response analyses. The question may justify using more data than less to capture variability. What percentage of the data samples are being collected regularly versus sporadically? The exact number of samples being collected regularly versus sporadically is not known. The majority of the data in the watershed chemistry comes from DWQ and WFWQC’s regular sampling programs. Why have sampling events steeply increased over the past several years? More resources and funding have led to an increase in sampling events. The end of the Clean Lake Program in the 1990s resulted in a decrease in sampling events. The WFWQC started collecting samples monthly on Utah Lake around 2015. DWQ’s monitoring program began in 2016. The variability in the frequency of sampling events since 2015 may be due to accessibility issues (i.e., sometimes the lake dries out or the wind makes collecting samples prohibitive). Are there distinct zones in Utah Lake that justify breaking the assessment into multiple areas?
• From the assessment perspective, the rationale for evaluating Provo Bay separately from the rest of the main basin is that Provo Bay is hydrologically distinct and acts as a different system. Considering the hydrologic distinctions, there might be a good argument for separating Goshen Bay as a distinct zone. Dividing the main basin into separate zones may not make sense since there are no observed hydrologic distinctions within the main basin.
• The more distinct zones there are in the analysis, the more difficult it will be to monitor, track, and report progress. Science Panel Discussion and Comments Science Panel members provided comments on the stressor-response analysis decision points. Their comments are summarized below.
• There is an argument that conditions of Utah Lake have not gotten worse; there is just more data capturing harmful algal bloom events. One benefit of highlighting the bivariate relationship is that it does not matter when the data was collected. Through the bivariate
14
analyses, the goal is to capture a gradient that has low and high concentrations. Still, there may be an extrapolation issue if the data only capture a specific part of the gradient.
• Researchers have only begun sampling cyanotoxins in the past five or six years. Most cyanotoxins samples are taken from the beach, so they are not representative of the open lake. An aggregation of these metrics would be misleading, so the Science Panel should spend time with each to see if they apply to the stressor-response analysis.
• Tetra Tech is coordinating with assessments and standards staff to ensure that their plans for assessing these conditions align with the stressor-response analysis. For example, if an assessment only focuses on open-water sites, then the stressor-response analysis should focus on open-water data. The stressor-response analysis will return to how the designated uses are being assessed so that the Science Panel can relate that situation to the statistical relationship of interest.
• Over the past year, DWQ has identified closures to specific Utah Lake areas. For example, they may close a certain beach or marina but allow open water recreation. The analysis will have more promise if the Science Panel assesses the data considering these nuances.
• It would be helpful to analyze the stressor-response using two datasets. The first dataset would be the whole dataset, including all samples. The second dataset would be a subset of the whole dataset that comes only from the routine monitoring starting in 2015.
• Besides Goshen Bay and Provo Bay, there is also an east-west bifurcation in Utah Lake. Additionally, there is expected to be more development and growth on the west lake. It may be worth considering the east and west sides of the lake as separate zones, given the differences in conditions.
• There may not be value in dividing the main basin of Utah Lake into an east and west side. The strategies for addressing anthropogenic nutrient loading on the east side of the lake will also improve conditions on the west side.
• The wind is also a factor in Utah Lake. The wind can drive surface materials to different parts of the lake. The Science Panel should consider multiple options for how to divide the lake.
• The Science Panel will need to consider how the development of criteria for Utah Lake will impact future management actions. Public Comments Members of the public provided comments on the stressor-response analysis decision points. Their comments are summarized below.
• From an ecological perspective, it would be helpful to break the lake into profundal and littoral zones since littoral dynamics have more impact on the productivity of Utah Lake. Next Steps
• Today’s presentation aimed to prime the Science Panel for future discussions. The Science Panel will continue to work on stressor-response analyses throughout the year.
• Tetra Tech cannot proceed on any of the analyses until there are decisions on the aggregation. The goal is to present the first or second draft of the analyses at the May Science Panel meeting.
• The Criteria Development Subgroup will re-form to guide Tetra Tech on how to aggregate data. Mitch Hogsett, Theron Miller, Zach Aanderud, and Han Paerl volunteered to join the Criteria Development Subgroup. Thad Scott may have an interest in participating in the subgroup. Samuel Wallace and Scott Dally will contact the Science Panel members absent from the meeting to see if they are interested in joining the subgroup. To prepare for that
15
meeting, Tetra Tech will develop options for aggregating the Utah Lake data for discussion by the Nutrient Criteria Development Subgroup