Loading...
HomeMy WebLinkAboutDWQ-2024-004546 1 Utah Lake Water Quality Study (ULWQS) Science Panel May 19, 9:00 AM to 5:00 PM Provo Airport – Skyview Lounge Meeting Summary - FINAL ATTENDANCE: Science Panel Members: Janice Brahney, Mike Brett, Mitch Hogsett, Theron Miller, Hans Paerl, Thad Scott, Tim Wool Steering Committee Members and Alternates: Eric Ellis Members of the Public: Jeff DenBleyker, Bob Marshall, Dan Potts, and David Richards Utah Division of Water Quality (DWQ) staff: Scott Daly Technical Consultants: Jon Butcher, Rene Camacho-Rincon, Maddie Keefer, Kevin Kratt, and Kateri Salk Guest Presenters: Josh LeMonte Facilitation Team: Heather Bergman and Samuel Wallace ACTION ITEMS Who Action Item Due Date Date Completed Tetra Tech and Janice Brahney Identify for what species researchers screened using eDNA in King (2019). June 28 Assess the Paleolimnology Study sediment cores to see how the dam's construction may have impacted lake levels and, in turn, impacted benthic and epiphytic diatom levels, particularly in open water. June 28 DECISIONS AND APPROVALS No formal decisions or approvals were made at this meeting. SCIENCE PANEL DIRECTION The Science Panel supported having Tetra Tech proceed with the following listed elements to run as the reference scenario through the Utah Lake Watershed Model: • Changed elements: o Change land cover based on pre-EuroAmerican settlement vegetation from the LANDFIRE database to account for natural wildfires o Remove water withdrawals (i.e., irrigation, public water supply) and releases o Remove irrigation (no agricultural land) o Remove point source discharges (no developed land) o Remove septic systems (no developed land) • Maintained elements: o Dams (e.g., Deer Creek Reservoir, Jordan River outflow) and stream hydraulics (i.e., channel geometry) 2 o Subbasin boundaries, drainage divides, and stream routing o Weather conditions o Calibrated parameter values for natural land covers (e.g., forest) REFERENCE CONDITIONS ANALYSIS Dr. Kateri Salk, Tetra Tech, presented the reference conditions analysis. The presentation, the subsequent Science Panel discussion, and public comments are summarized below. Reference Conditions Analysis Presentation Below is a summary of the reference conditions analysis presentation. Tetra Tech Reference Condition Analysis Overview • The reference conditions analysis is a component of the Technical Support Document (TSD). The TSD provides the technical basis for developing numeric nutrient criteria (NNC) to protect designated uses. • The reference conditions analysis will include results from the paleolimnological studies and the Utah Lake Nutrient Model prediction/extrapolation of reference conditions. • The typical approach for a reference conditions analysis is to identify comparable systems in a location with minimal human impact. Herlihy and Sobota (2013) outlined the approach for identifying reference lakes in the National Lakes Assessment (NLA), which involves a three-stage process for identifying reference lakes based on three criteria. Since Utah Lake is so shallow and alkaline, it has few systems to which it can be compared. As a result, the Science Panel will need to look at other evaluations to identify applicable reference conditions. • The questions for the Science Panel are: o What did Utah Lake look like pre-Euro-American settlement? o What would the system look like if it was minimally impacted by humans? • The reference conditions analysis is intended to set a floor for Utah Lake conditions and add context for how the lake has changed over time. It is not intended to set a goal for future lake conditions. • The reference conditions analysis will incorporate the paleolimnological reconstruction of past conditions. The paleolimnological reconstruction will quantify pre-settlement nutrient conditions and how they have changed over time. The reconstruction will be based on multiple studies and Steering Committee charge question responses. • The reference conditions analysis will also use model-based predictions. The approach involves running the watershed model under a reference conditions scenario (to be discussed by the Science Panel) to generate a watershed nutrient loading scenario. The watershed conditions generated by the model can then be used as boundary conditions for the Utah Lake Nutrient Model. • The technical consultants will use several paleolimnological studies in the analysis. The ULWQS funded a paleolimnological study, with Dr. Janice Brahney as the principal investigator, which has led to theses from King (2019) and Devey (2021). There are other recent theses/dissertations, including Macharia (2012), Tate (2019), and Williams (2021). Other historical studies related to the paleolimnological reconstruction of Utah Lake include Brimhall (1972), Bolland (1974), Sonerholm (1974), Brotherson (1981), and Javakul et al. (1983). If Science Panel members know of other relevant studies, they should contact Tetra Tech. • Several lines of paleolimnological evidence in the TSD include qualitative information on community composition and semi-quantitative information on oligo/meso/eutrophic 3 conditions over time. The quantitative information includes results from analysis conducted on sediment nutrient concentrations and isotope values. It is important to note that some cores have reliable dating, while others do not. Combining the different lines of evidence will help form a holistic picture of reference conditions. Tetra Tech Paleolimnological Lines of Evidence • Three studies have explored benthic diatoms over time: Bolland (1974), Jakuval et al. (1983), and Brahney et al. (2021). The studies suggested that benthic and epiphytic diatoms dominated in preindustrial conditions, with an increasing prevalence of planktonic diatoms approaching the present day. These studies provide evidence for clearer water conditions pre-Euro-American settlement. • In the planktonic diatom community, there is a shift from oligo-mesotrophic to eutrophic diatom taxa from deep to shallow core (Bolland 1974). Additionally, pollution-tolerant species increase moving up-core (Brahney et al. 2021). This evidence supports a shift from oligo-mesotrophic to eutrophic conditions. • King (2019) used dating techniques to analyze the phytopigments in the cores from 1720 to the present day. King's thesis (2019) provides evidence of increased diatom production around 1890 and a transition from diatoms to cyanobacteria and green algae production in recent decades. King (2019) also showed a greater chlorophyll degradation rate post-1890, with the highest chlorophyll concentration near the core surface. The chlorophyll a, pheophytin a, and b-carotene core data suggest a decrease in production around 1950, consistent with the installation of wastewater treatment plants (WWTPs) in the 1950s. This evidence supports a shift from oligo-mesotrophic conditions to eutrophic conditions. This study does not capture the impacts of recent upgrades to the WWTPs on water quality. • King (2019) assessed the Goshen Bay sediment core using the eDNA. The eDNA records show an increasing abundance of cyanobacteria post-1900 and representation of hardstem Bulrush pre-1900. These results are consistent with other lines of evidence that capture a transition of Utah Lake from a macrophyte-dominated system to one with more pelagic- type growth. For example, the original survey of Utah Lake in 1872 noted extensive Bulrush in Goshen Bay. • There is evidence in the paleolimnological record of other taxa. The paleolimnological evidence points to a historical macrophyte-dominated, clear-water condition. Present-day macrophyte restoration depends on several factors, not just the nutrient regime, such as carp, non-algal turbidity, algal production, and hysteresis. • King et al. (2023) created a model to understand the level of clarity needed to establish a self-stabilizing submerged macrophyte community. The study found that water clarity would need to be consistent with chlorophyll concentrations of less than 18 micrograms/L and Secchi depth of around one meter: much greater water clarity than currently exists in Utah Lake. • Paleolimnological evidence shows an increase in cladocerans and a decrease in ostracods post-1890. • Several studies have assessed nutrient concentrations in sediment cores. The studies have assessed concentrations of exchangeable phosphorus, iron/manganese-bound phosphorus, organic phosphorus, calcium carbonate-bound phosphorus, and refractory phosphorus. The Brimhall (1972) and Bolland (1974) studies found increased phosphorus concentrations in their cores from 20 centimeters to the present. Devey (2021) found that exchangeable and calcium-bound phosphorus concentrations increase over time, while iron and aluminum- bound phosphorus are relatively stable. This evidence supports a shift to more eutrophic conditions from preindustrial times. 4 • Macharia (2012) and King (2019) analyzed nutrient-stable isotopes (δ15N and δ13C). They found an increase in δ15N values post-1900. Macharia (2012) also found that δ15N was positively correlated with percent nitrogen in sediment bulk organic matter. These results suggest a greater prevalence of wastewater-derived nitrogen, consistent with population growth. The increase in δ15N values may be aligned with wastewater delivery and treatment upgrades. In particular, Provo sewers were built in 1908, and the Provo treatment plant was completed in 1956. The Orem sewers were built in 1945, and the treatment plant was completed in 1959. • Sediments store integrated information about system conditions. Sediment nutrient concentrations are proportional to water column concentrations but do not have a direct 1:1 relationship. Converting sediment concentrations to water column concentrations could be difficult in Utah Lake due to the influence of alkalinity, phosphorus speciation, equilibration, and water levels, which make it difficult to determine the proportional relationship between sediment and water column nutrient concentrations. Many equilibration efforts would be needed to develop a reliable sediment-to-water column nutrient concentration conversion. An approach more likely to be successful would be to calculate relative differences to build a relative understanding of changes in water column nutrients over time. Additionally, another approach could be to pursue a mass balance approach. • One limitation of the paleolimnology lines of evidence is that additional changes have co- occurred alongside nutrient regime changes (e.g., carp introduction, hydrologic changes, climate change). There is a need to evaluate lines of evidence holistically to see where they agree in direction and magnitude. The reference scenario for the mechanistic model will mimic present-day, minimally disturbed conditions. • The next step for the paleolimnology lines of evidence is for the ULWQS Paleolimnology Study Subgroup to discuss and finalize the study. Science Panel Clarifying Questions Science Panel members asked clarifying questions about the reference conditions analysis. Their questions are indicated in italics below, with the corresponding responses in plain text. Are the Utah Lake sediment cores banded by snowmelt? No, it is rare to have the cores banded by snowmelt. How did the construction of the dam impact lake levels and, in turn, impact benthic and epiphytic diatom levels, particularly in open water? Tetra Tech and Dr. Brahney can examine the available evidence to respond to this question. How are the cores dated? • Investigators in Brahney et al. (2021) used lead-210 and cesium-137. Different methods were effective based on the core. They used Bayesian techniques to produce age-depth models with calculated uncertainty. • Investigators who dated cores in other studies used a variety of techniques. Lead-210 and carbon-14 were common. What is the oldest date the dating method can capture? It depends on the core. 5 What is meant by "pollution-tolerant" species? What type of pollution? The term pollution-tolerant is related to nutrient pollution and other stressors as well. What is the resolution of the King (2019) core dating, particularly for recent decades? According to the graphs, there is an increase in green algae pigments over time but a decrease in more recent years. What can that recent decrease be attributed to? • The concentration of green algae pigments in the King (2019) cores indicate a significant increase in green algae pigments after 1870. There has been an even more significant increase in green algae pigment concentrations close to the surface of the sediment cores within the past two decades. • However, there is a decrease in green algae pigments at the surface of the sediment in the most recent years. The decrease in green algae pigments at the surface of the sediment could be an artifact of the sampling methods, but it could also be a real trend. Some recent work out of Brigham Young university (BYU) has shown a decrease in blooms over the past decade. It is difficult to tell whether the decrease in pigments is an artifact or real trend without collecting more samples that show the same or different trend. Could resuspension near the surface of the sediments explain the decrease in green algae pigments? Several potential variables could impact the surface of the sediment cores (e.g., organisms growing in the sediment, greater exposure to light and oxygen, resuspension), so it is difficult to assign a specific reason that explains the decrease in green algae pigments. It would be easier to assess this trend if more samples were available. Even with the decrease, the current concentration of green algae pigment is statistically higher than historical green algae pigment concentrations (i.e., 1870). What is the length of the King (2019) cores? The cores varied in size. The cores in King (2019) ranged from 60 centimeters to one meter. Are there other cores from King (2019) that may provide additional evidence on whether a decrease in green algae pigments is occurring near the surface? King (2019) assessed multiple cores, but pigment data was collected for two cores from Bird Island and Goshen Bay. There was insufficient funding to measure pigments in more than two cores. Given the particle size, what is the sediment thickness at which resuspension occurs? It is unclear how to measure the sediment thickness at which resuspension occurs. The view of researchers is that resuspension blurs the signal in Utha Lake more than in banded lakes, but the law of superposition still holds. The impacts of resuspension likely impact the signal in the more recent sediments than historical ones. Is there evidence that shows how sediment resuspension rates have changed over time? Dr. Brahney could assess the sediment cores to understand better how sediment resuspension rates change over time. Did researchers in King (2019) use eDNA to assess fish or animal abundance? The researchers could only attain reliable results for plants and algae. Not enough fish remains were left in the sediment to assess fish populations using eDNA reliably. They had criteria on which data was appropriate and not for eDNA analysis. What is the percentage increase in total phosphorus concentrations from preindustrial times? • Because of variability, the Science Panel would need to set bounds to calculate the percent increase in total phosphorus concentrations over time. 6 • The exchangeable phosphorus concentrations are not a good indicator for total phosphorus concentrations in the water column because exchangeable phosphorus can be mobile in the sediment. The calcium-bound phosphorus is the only reliable indicator for phosphorus concentrations in the water column. There has been a 100% increase in calcium-bound phosphorus from pre-settlement conditions to modern levels. • Using sequential extractions, Devey (2021) measured different phosphorus constituent concentrations over time. Sequential extractions are not the best approach because the method involves using harder acids to target a specific fraction. Since sequential extraction has limitations in its precisions, the researchers used electron microscopy to identify calcium carbonate particles and the associated phosphorus content. They then used physical separation methods to refine the particles and measure the calcium-bound fraction. Using multiple techniques allowed researchers to identify the amount of calcium- bound phosphorus precipitating in the water column. In the isotope analysis, in what forms can carbon be found? The isotope study identified what percentage of the carbon is organic carbon. The majority of the other carbon is bound in mineral form. Some carbon may be from plastic materials. The isotope analysis shows a large increase in δ15N post-1850 (approximately an increase of 18 milligrams/liter). Are there any potential explanations for the sudden increase? There was a large restructuring of Utah Lake post-1850. Carp were introduced to Utah Lake around this time. Additionally, there was a loss of macrophytes. The Paleolimnology Study researchers have found a dramatic increase in the degradation rate of pigments after 1850. Historically, sediments were not mixing in the water column, especially when macrophytes were present. Post- 1850, there was a massive decay of vegetation and resuspension of sediments. After a sudden increase in δ15N after 1850, the percentage of δ15N decreased slightly over time. Are there any potential explanations as to why the δ15N percentage decreased? It is important not to over-interpret the record. Potential explanations could include carp or other animal sources (e.g., tanning factories, cattle grazing). Has an analysis been completed to compare changes in sediment cores to population density/growth? Macharia (2012), King (2019), and Devey (2021) compared pigment and phosphorus concentration data to population density. Are there plans to publish the theses that came from the recent paleolimnology data? Yes, King has submitted her 2019 paper. Devey is in the process of writing two papers, which he plans to submit. Public Clarifying Questions Members of the public asked clarifying questions about the reference conditions analysis. Their questions are indicated in italics below, with the corresponding responses in plain text. What does it mean that the reference conditions analysis will help determine what Utah Lake "looked like?" The reference conditions analysis will provide evidence of the past conditions of the lake and how it functioned. 7 All lakes transition from oligo-mesotrophic to eutrophic conditions. Is it possible to parse if and how European settlement accelerated that transition? The dating combined with other human population markers can help provide evidence on whether a shift in trophic conditions is human-driven or from natural variability. When did the WWTP facilities come online? The records indicate that the Orem and Provo treatment facilities came online in 1954 and 1959. The WWTP technology has changed over time. Does the eDNA evidence from King (2019) track phragmites? The eDNA research from King (2019) was not intended to capture a comprehensive list of taxa. Tetra Tech can revisit the King (2019) thesis to determine what species they screened for in their eDNA samples. Will the reference conditions analysis account for first-hand historical records? Trappers and hunters who passed through in the 18th and 19th centuries have documented conditions on the lake. The reference conditions analysis will incorporate historical records of Utah Lake conditions. If Science Panel members or members of the public have relevant historical records of Utah Lake, they should send those to Tetra Tech. Did researchers find shells within the sediment cores? Researchers in Devey (2021) did not find fragments of shells in their sediment cores. When was the pump house installed, and how could the installation of the pump house impact Utah Lake conditions? • The pump house was installed at the turn of the 20th century. • In the 1930s, water was pumped out of Utah Lake to the extent that most of the lakebed was dry, except for a small pool. This level of dryness impacted fish movement. • There is a signal in the sediment cores of a shift in the 1930s. The investigators associated that signal with severe drought. It will be difficult to distinguish the impact of the pump station compared to natural drought. Science Panel Discussion on the Reference Conditions Analysis • Not all lakes transition to eutrophic conditions. For example, macrophytes and bottom flora can stay the same as a lake matures and becomes a meadow. The evidence for a lake becoming eutrophic is whether planktonic communities become the dominant feature of productivity. • Understanding the sediment rate and impacts of sediment resuspension in Utah Lake will be important. Public Comment on the Reference Conditions Analysis • Because of sediment disturbance, it may be difficult to rely on sediment core data for the last 50 to 70 years. For more recent years, the Science Panel should consider relying on direct lake samples from 1970 onward, including the algae samples collected by Dr. Sam Rushforth. • The installation of the pump house on Utah Lake has changed water levels and functionally scrubbed the lake shorelines of Bulrush, opening up the opportunity for the expansion of phragmites. 8 • Sediment resuspension could be correlated to lake depth; as the wind blows, a shallower lake would be more vulnerable to sediment resuspension than a deeper one. • Utah Lake once contained 14 mollusk species; today, only two taxa remain. The mollusk shells were composed of calcium carbonate. The presence of native mollusks affected the water column, periphyton, and calcium in the sediment. • In addition to the introduction of carp, the loss of native species, such as Bonneville Cutthroat Trout, also impacted the lake's physical and biological conditions. WATERSHED MODEL REFERENCE SCENARIO DISCUSSION The Science Panel discussed building the reference condition scenario to run through the mechanistic models. Members of the public also provide comments on the watershed model reference scenario. Their comments are summarized below. Tetra Tech Background and Context of the Reference Scenario • The term "reference" can be interpreted in different ways. It could mean removing human- influenced land cover, human-influenced hydrologic changes, or all traces of human impact. An extreme sample of a "reference" scenario would be a simulation of Lake Bonneville conditions. • Utah Lake does not have a comparable system to consider as a reference. Therefore, there is a need to develop a simulation of a comparable system with minimal anthropogenic nutrient inputs. The question for the Science Panel is, "What would Utah Lake look like under minimal human influence?" • The purpose of the reference scenario is to serve as a "floor" for Utah Lake conditions, not set a goal for future conditions. • Once the Science Panel identifies reference conditions, the Tetra Tech modeling team will run the Utah Lake watershed model using those inputs. The watershed model will generate outputs for watershed nutrient loading and hydrology. The Tetra Tech modeling team will use the watershed model outputs as conditions in the Utah Lake Nutrient Model. The Utah Lake Nutrient Model will then provide data on water column nutrients and primary production under the reference scenarios. • The proposed watershed model reference scenario includes the following changed elements and maintained elements: o Changed elements: ▪ Change land cover based on pre-EuroAmerican settlement vegetation from the LANDFIRE database to account for natural wildfires ▪ Remove water withdrawals (i.e., irrigation, public water supply) and releases ▪ Remove irrigation (no agricultural land) ▪ Remove point source discharges (no developed land) ▪ Remove septic systems (no developed land) o Maintained elements: ▪ Dams (e.g., Deer Creek Reservoir, Jordan River outflow) and stream hydraulics (i.e., channel geometry) ▪ Subbasin boundaries, drainage divides, and stream routing ▪ Weather conditions ▪ Calibrated parameter values for natural land covers (e.g., forest) 9 Science Panel Discussion The Clean Water Act assigns designated uses based on the condition of water bodies in 1975, so it may be reasonable to use 1975 conditions as the reference scenario. Public Clarifying Questions Members of the public asked clarifying questions about the watershed reference conditions scenario. Their questions are indicated in italics below, with the corresponding responses in plain text. The maintained elements list includes stream routing. Would this include maintaining the Strawberry Reservoirs diversions, which transports water from the Colorado River Basin to the Bonneville Basin, in the reference conditions scenario? The watershed model currently captures the diversions from the Colorado River Basin to Strawberry Reservoir, data that the Central Utah Water Conservancy District provided. Removing the Deer Creek Dam and Strawberry Reservoir from the watershed model as part of the reference condition scenario would be extremely difficult. Public Comment The diversion system from the Colorado River Basin to the Bonneville Basin impacted the geomorphology of the Provo River. Science Panel Direction The Science Panel supported having Tetra Tech proceed with the listed reference conditions to run through the watershed model. WATERSHED MODEL DEVELOPMENT UPDATE Maddie Keefer, Tetra Tech, provide an update on the watershed model development. The presentation, the subsequent Science Panel discussion, and public comments are summarized below. Watershed Model Development Update Presentation Below is a summary of the ULWQS watershed model update presentation. Watershed Model Overview • The watershed model will quantify nutrient load contributions to Utah Lake by source, simulate the impact of management actions (e.g., permit limits, best management practices) on nutrient loading to Utah Lake, and evaluate alternative watershed conditions (e.g., climate, land use). • The watershed model will not be able to evaluate other pollutants (e.g., bacteria, metals), simulate individual or field-scale best management practices (i.e., best management practices must be simulated in aggregate), or identify certain pollution issues (e.g., locations of failing septic systems). • The Science Panel worked with Tetra Tech to select a model for watershed simulation. With Science Panel input, Tetra Tech defined specific criteria related to watershed characteristics, simulation capabilities, source representation, usability, and general platform functionality. These criteria were used to rank 11 modeling platforms quantitatively. The assessment identified Hydrologic Simulation Program – FORTRAN (HSPF) as the top-ranked model. 10 • The process for developing the model can be characterized into five steps: 1) gather data, 2) build model, 3) calibrate model, 4) assess current conditions, and 5) run scenarios. At this time, Tetra Tech is working on calibrating the model. • Tetra Tech follows the Watershed Modeling Quality Assurance Project Plan (QAPP). The QAPP is the project plan to ensure quality objectives are met for measuring and modeling data. The QAPP provides a framework to ensure the model will support project goals and objectives. It also sets guidelines to collect data and specifies quality assurance/quality control (QA/QC) activities to assess model performance. Lastly, the QAPP outlines a methodology for assessing the model's usability. • The geographic extent of the watershed model is the entire Utah Lake drainage. The watershed model geography spans three different catchments. The watershed model does not simulate the lake itself; it will be connected to the Utah Lake Nutrient Model, which will simulate the internal Utah Lake water quality dynamics. • Two moderately sized reservoirs in the watershed model geography are Deer Creek Reservoir and Jordanelle Reservoir. The watershed model is limited because it cannot model the complex water quality processes occurring in these reservoirs. As a result, the watershed model uses the Deer Creek Reservoir as a boundary for simulating water dynamics and balances. • The watershed modeling team divided the entire geographic area into HUC12 watersheds and modified the boundaries to account for gaging stations and water diversions. They then used climate, geology, topography, and land use/cover influence on runoff and stream water quality to identify discrete hydrologic response units (HRUs). Each HRU has unique characteristics that delineate it from other areas in the drainage. Lastly, because the watershed modeling team used 2016 National Land Cover Data (NLCD), the dataset did not account for fires after 2016. The modeling team integrated post-2016 fire perimeters to identify local land use changes. • The watershed model considers impervious surfaces as different from other surfaces. The modeling team used a high-resolution geospatial dataset to capture impervious surfaces, including buildings and roads. • The modeling team used the US Department of Agriculture's Soil Survey Geographic Database (SSURGO) to account for different soil types through the drainage. They divided the soil types into two hydrologic soil groups: higher infiltration and lower infiltration. They then incorporated slope into the model using the US Geological Survey (USGS) 10-meter digital elevation model and categorizing slope into low (<10 degrees), medium (10-30 degrees), and high (>30 degrees). • The watershed model captures the irrigation of agricultural lands and lawns/landscapes. The modeling had to estimate irrigation rates using irrigation demand. They did so by estimating evapotranspiration using data from the Utah Climate Center. They then used crop coefficients to estimate water demand by crop type. They then calculated irrigation demand, equal to crop water demand minus precipitation. • There are many diversions and releases in the watershed model geographic area. The modeling team obtained diversion and release data from the Central Utah Water Conservancy District, Provo River Water Users Association, Utah Division of Water Rights, and annual distribution system reports. The model then represents diversion time series as withdrawals and external water imports as releases. • The watershed model represents weather using an hourly time series. The model splits the drainage into 13 unique weather zones based on 30 years of Parameter elevation Regression on Independent Slopes Model (PRISM) data. The full QAPP will document how 11 the modeling team conducted QA/QC analysis on the data and used it to derive a gridded weather dataset. • Lastly, the modeling team incorporated permitted point source discharges based on discharge monitoring report data and monthly operating reports. The watershed model does not include discharges from Timpanogos Special Services District (TSSD) or Orem WWTP because those facilities discharge directly into Utah Lake. Those sources are accounted for in the Utah Lake Nutrient Model. Watershed Model – Hydrology Calibration and Water Balance Analysis Overview • Tetra Tech completed the hydrology calibration in early 2023 and presented the results to the Science Panel Modeling Subgroup. • The QAPP outlines the hydrology calibration methodology. As part of the methodology, the modeling team: o Calibrated the model to multiple endpoints (e.g., remotely sensed snow depth and water storage, actual evapotranspiration, and daily, monthly, and cumulative gaged flow) o Used multiple visuals for total flow, seasonal/monthly flows, and high/low flow distribution and calculated statistical metrics (e.g., Nash Sutcliffe efficiency (NSE) coefficients) o Obtained daily streamflow records from USGS monitoring sites o Sought to obtain the best overall fit at multiple locations, with the priority on larger tributaries to Utah Lake (Spanish Fork and Provo River) • The calibration compared observed and simulated flow at ten sites. The modeling team calculated the monthly NSE to measure observed and simulated flow differences statistically. The results of the calibration indicated good agreement between simulated and observed flows. • Tetra Tech then conducted a water balance analysis as the final step in the hydrology calibration process. Streams in the Utah Lake watershed have significant interactions with deep groundwater. The deep groundwater recharge occurs at the foot of the mountains and discharges back into the streams near the lake or directly into the lake. It is difficult to simulate this interaction in the watershed model, which only simulates local shallow groundwater pathways. • The modeling team compared the groundwater discharge outputs from the watershed model with the groundwater discharge outputs from the USGS Groundwater Modeling Reports. The USGS Groundwater Modeling Reports estimates the annual average of groundwater discharge in the Utah Lake Watershed to be between 205,000 and 340,000 acre-feet/year; the ULWQS watershed model estimates the groundwater discharge to be 323,957 acre-feet/year. The analysis concludes that the watershed model outputs closely match the USGS groundwater studies even though the physical mechanisms represented in the watershed model do not fully align with the USGS conceptual model. • The modeling team compared the Utah Lake inflows based on reported lake levels and water balance to the flows predicted by the watershed model at a monthly time scale. The overall water balance of Utah Lake can be characterized as flow-in - flow-out +groundwater input – evaporation + precipitation P wastewater discharges. The modeling team rearranged the formula to calculate the flow-in value using the expression: difference in Utah Lake volume + flow-out - groundwater input + evaporation - precipitation - wastewater discharges. Using the rearranged formula, the modeling team ran the Utah Lake Nutrient Model to calculate the approximate annual flow-in value as 515,480 acre-feet/year. The watershed 12 model has almost an identical flow output to the lake of 515,751 acre-feet/year. Each of the variables in the water balance equation came from a variety of data sources. • The modeling team compared monthly inflows between the observed and simulated data in a time series. The time series shows that there are some seasonal discrepancies. Those discrepancies are likely due to the lack of seasonal specification of groundwater inflow and wastewater discharges, coupled with the uncertainty in estimating evaporative losses. • The modeling team compared cumulative inflow between the observed and simulated data. The cumulative inflow comparison suggests that the current version of the watershed model is an unbiased estimate of flow to Utah Lake. Watershed Model – Water Quality Calibration Process Update • The modeling team is currently conducting the sediment and nutrient calibration. • One objective of the watershed model is to represent the sediment behavior of the watershed based on the available information on the morphological characteristics of the rivers and streams. • The sediment calibration involves calibrating the model to multiple endpoints, including: • Daily, monthly, and cumulative monitored sediment concentrations • Sediment source assessments • Reach/sediment balance • The calibration is guided by multiple visuals and statistical metrics related to sediment loading rates from the landscape, seasonal/monthly total suspended solid (TSS) concentrations, high/low flow distribution, and average and median concentration and load error. The modeling team obtained TSS records from the Utah stream monitoring sites. The calibration seeks the best overall fit at multiple locations, prioritizing larger tributaries to Utah Lake (Spanish Fork, Provo River). • There are nine calibration sites. Most of the sites have a high number of data points. Two sites with less data available include the Starvation Creek/Upper Soldier Creek site (11 samples) and the Thistle Creek site (21 samples). • Following the sediment calibration, the modeling team will conduct a nutrient calibration. The main objective of the nutrient calibration is to obtain an acceptable agreement of observed and simulated concentrations while maintaining the instream water quality parameters within physically realistic bounds and the point loading rates with the expected ranges from the literature. • The steps to conduct the nutrient calibration are in order: • Estimate all mode parameters, including land use-specific accumulation and depletion/removal rates, washoff rates, and subsurface concentrations • Compare simulated nonpoint loading rates with the expected range of nonpoint loadings from each land use and adjust loading parameters when necessary to improve agreement and consistency • Calibrate instream water temperature • Compare simulated and observed instream concentrations at each of the calibration stations • Analyzing the results of comparisons in steps 3 and 4 to determine appropriate instream and nonpoint parameter adjustments, and repeat those steps as needed until calibration targets are achieved • The timeline is to complete the sediment calibration in May 2023. The modeling team will then work on the water quality calibration, with an expected completion in June 2023. They will then begin the sensitivity analysis and model documentation. 13 Science Panel Clarifying Questions Science Panel members asked clarifying questions about the watershed model update. Their questions are indicated in italics below, with the corresponding responses in plain text. According to the water balance analysis, the total evaporation loss is 588k acre-feet/year, and the flow-in is 515k acre-feet/year. How could the evaporation loss be greater than the flow-in value? The modeling team calculated the flow-in value using the expression: difference in Utah Lake volume + flow-out - groundwater input + evaporation - precipitation - wastewater discharges. In this case, the modeling team is solving for the flow-in value. Will the nutrient calibration differentiate between particulate and dissolved forms of phosphorus and nitrogen? Yes, the modeling team will calibrate the model to various constituents of nitrogen and phosphorus in addition to total phosphorus (TP) and total nitrogen (TN). They will calibrate the model to organic nitrogen, nitrate, nitrite, organic phosphorus, and orthophosphate concentrations. However, there is less observed data that differentiates between dissolved and particulate forms, so the calibration will most likely rely on TN and TP values. How does the loading contribution of the 2023 inflow compare to past years? There was a similar runoff event in the 1980s in terms of magnitude but not timing. The high amount of precipitation that has fallen in a relatively short time frame in 2023 is unique compared to recent years. Public Clarifying Questions Members of the public asked clarifying questions about the watershed model update. Their questions are indicated in italics below, with the corresponding responses in plain text. A significant sediment load is coming into Utah Lake from the Spanish Fork. The amount of precipitation and sediment loading flowing to Utah Lake is historical. Can the sediment loading from this year (2023) be used to calibrate the model? The modeling team is calibrating the model with data between 2015-2020. It is unsurprising that people are observing high levels of sediment coming from the Spanish Fork, as the Spanish Fork reaches have the highest concentration of TSS compared to other reaches. Will calibrating the model using data from 2015 to 2020 result in a missed opportunity to calibrate the model to a significant event, like the one occurring this year? • The amount of precipitation falling this year will likely result in high nutrient loads but lower concentrations. • Some events occurred between 2015 and 2020 that represent large sediment contributions to the lake, such as large debris flows from the fires in the Spanish Fork drainage. These types of events will be accounted for in the calibration timeframe. Could the watershed model be used to simulate whether sediment and nutrient loading from tributaries may lead to observed algal blooms in Utah Lake? The watershed model alone cannot simulate the likelihood that sediment and nutrient loading will result in observed algal blooms. The Utah Lake Nutrient Model can use the outputs from the watershed model at various locations and translate how those contributions may impact algal blooms. 14 Science Panel Discussion • By calibrating the model to data from 2015 to 2020, there is a missed opportunity to capture an episodic event like the one occurring this year. The sediment and nutrient contributions from the 2023 inflow will spread across the lake. It would be advantageous to capture these types of sediment and nutrient loads in the model calibration. • It will be difficult to gather the meteorological data to run the watershed model for this year to compare the observed and simulated outputs. There may be a future opportunity to conduct a comparative budget analysis based on the meteorological year. Public Comment • Evaporation losses are often difficult to characterize in marginally freshwater lakes due to changes in evaporation rates caused by total dissolved solids. Measuring and characterizing these evaporation rates is a currently active area of research. • The sediment contributions from the Spanish Fork are tremendous this year. Some of the sediment comes from an area of the Spanish Fork Canyon that burned a few years ago. The delta is beginning to form at the mouth of the Spanish Fork. • Tributaries, like the Spanish Fork and Hobble Creek, contribute sediment to the lake, and in years of high flows, temporary deltas can form. However, the lake eventually distributes the sediment to make the lakebed smooth and flat. • Some sediment input from the Spanish Fork is composed of mica schist. The mica schist acts like little mirrors that reflect light out of the lake, lowering evaporation rates. • The Provo River Delta will impact the amount of sediment entering Utah Lake because the sediment will settle in the delta rather than the lake. It would be advantageous for the model to capture this new dynamic. UTAH LAKE NUTRIENT MODEL UPDATE Rene Camacho, Tetra Tech, provided an update on the development of the Utah Lake Nutrient Model. The presentation, the subsequent Science Panel discussion, and public comments are summarized below. Utah Lake Nutrient Model Update Presentation Below is a summary of the ULWQS nutrient model update presentation. Utah Lake Nutrient Model Overview • The Utah Lake Nutrient Model simulates internal hydrodynamics using the Environmental Fluid Dynamics Model (EFDC). The EFDC model uses a 2-D/3-D orthogonal curvilinear grid to solve mass and momentum transport equations. The EFDC model provides solutions for salinity, temperature, and conservative tracers. The EFDC is the backbone of the entire Utah Lake Nutrient Model. • The Utah Lake Nutrient Model also uses the Simulative WAves Nearshores (SWAN) model to simulate wind-induced waves. The SWAN model is coupled with the EFDC model to simulate the orbital circulation between the lake surface and bottom due to wind. This simulation ultimately calculates the excess shear stress that can induce sediment transport. • The modeling team has made improvements in linking the EFDC and SWAN models. The linkage between the two programs begins by running the EFDC model. The EFDC model produces information on the grid, bottom elevation, wind, water surface elevation, and currents and sends it to the SWAN model. The SWAN model receives these inputs and simulates significant wave height, wave length, wave period, and wave direction. The SWAN 15 model then calculates energy, dissipation, and radiational stresses and sends that information to the EFDC model. • The EFDC and SWAN models are then linked to the Water Quality Analysis Simulation Program (WASP) models. The WASP model is the most comprehensive model available. The WASP model can simulate multiple dynamic processes, including nutrient cycling, phytoplankton succession, and organic matter decomposition. For example, the WASP model can simulate phytoplankton biomass using the stoichiometry of different phytoplankton groups. When phytoplankton dies and settles as particulate detrital matter, the model can simulate the dissolution of organic matter into dissolved organic matter and the transformation from dissolved organic matter to inorganic nutrients. • The sediment diagenesis module is integrated into the WASP model to simulate the nutrient processes occurring in the sediments, including how organic matter breaks down into different forms of nitrogen and phosphorus. The outputs from the sediment diagenesis module are then exchanged between the sediment and water column. The sediment diagenesis module does not communicate with the sediment transport model. • The EFDC and SWAN models communicate to generate hydrodynamic outputs (e.g., grid cell volumes, velocities, temperatures, inorganic suspended sediments, and shear stress), which are sent to the WASP model. The WASP model calculates the mass balance for different nutrient and water quality constituents and produces outputs like nutrient concentrations, algae mass, biochemical oxygen demand, and dissolved oxygen. • The computational grid for Utah Lake is divided into one square kilometer cells. Based on bathymetry data, the cells are then delineated by "deep" and "shallow" cells. • The model integrates meteorological data. Air temperature, relative humidity, and altimeter pressure data come from the Provo Municipal Airport weather station. Data from buoy stations are used as supplemental data. Precipitation, solar radiation, and cloud cover came from different data sources. • The wind speed and direction data came from the Provo Airport and Utah Lake weather stations near Mosida. Buoy data from 2019 and 2020 supplement this data; however, the Provo Airport weather station has the longest and most comprehensive time series. • Other inputs into the model include nutrient loading from tributaries and the two direct point sources on the lake from TSSD and Orem WWTP. Model Hydrodynamic Calibration Performance • Before the Utah Lake Nutrient Model is calibrated to water quality metrics, it must first be calibrated to hydrodynamic metrics. • The modeling team compared observed and simulated data for water surface elevation and current velocity. They calibrated the model using water surface elevation because it is a good indicator of how well the model represents volumes in the system. • When calibrating the Utah Lake Nutrient Model, it is important to note that it generates outputs at a one-square-kilometer resolution. Essentially, the model estimates what occurs within the one square kilometer cell. Because the calibration involves comparing the broader outputs from the model to specific observed data, the modeling team is looking for reasonable agreement between the two data sources. • The water surface elevation and current velocity calibration results indicate reasonable agreement between the observed and simulated data. • The modeling team also calibrated the Utah Lake Nutrient Model using significant wave height. The results of the calibration indicate that the model is well-constrained. • The modeling team uses statistical methods to compare observed and simulated results in their calibration. The statistical values they generate include the mean absolute error, root- 16 mean-square (RMS) error, Norm RMS error, and index of agreement. The statistical values indicate that the model outputs reasonably agree with the observed data for the hydrodynamic calibration. • The modeling team also calibrated the model using temperature. The calibration results indicate good agreement between the observed and simulated data. • The modeling team used data from the DWQ hydrometer tests to break the sediment into different size classes for the inorganic sediment transport model. The modeling team then calibrated the model shear stress outputs with critical shear stress estimates. DWQ estimates the critical shear stress to be between 0.38 and 2.81 Pa; they also estimate the wave shear estimate to be 0.17 Pa. • The modeling team simulated shear stress using only the EFDC model and the EFDC-SWAN linked model and compared the results. The linked EFDC-SWAN model produces much higher shear stress values than the EFDC model alone, emphasizing the importance of wind- induced shear stress in the system. • The modeling team plotted the simulated sheer stress and simulated TSS values against the observed TSS values. By coupling the EFDC-SWAN programs, the model links TSS increases to wind-induced shear stress better. • The modeling team compared the observed and simulated TSS values. The calibration is only in draft form, so there are no statistical conclusions to draw from the comparison. Draft Model Water Quality Calibration Performance • The modeling team is in the process of comparing the model outputs with observed data for chlorophyll a, dissolved oxygen, total nitrogen, and total phosphorus. • The sediment diagenesis module uses carbon, nitrogen, and phosphorus content from the ULWQS Littoral Sediment Study (2022) and fluxes from the Littoral Sediment Study (2022) and Hogsett et al. (2018). • The Utah Lake Nutrient Model simulates phosphorus-binding (P-binding) by using partition coefficients from the ULWQS P-Binding Study (Kd [20-130] liters/kilogram). • The calibration results include the most recent estimates of atmospheric deposition from the Science Panel. • The calibration results are in their draft form. The modeling team has not yet drawn conclusions from the calibration. The modeling team's next step is to refine and finalize the water quality calibration. Science Panel Clarifying Questions Science Panel members asked clarifying questions about the Utah Lake Nutrient Model update. Their questions are indicated in italics below, with the corresponding responses in plain text. In some cases, the simulated results for DO are zero. Is that because the site went dry? Yes. Public Clarifying Questions Members of the public asked clarifying questions about the Utah Lake Nutrient Model update. Their questions are indicated in italics below, with the corresponding responses in plain text. 17 The R2 value comparing simulated and observed water surface elevation was close to 100%, while the R2 value comparing simulated and observed significant wave height was 75%. Is this difference notable to the modeling team? The significant wave height fit between observed and simulated values is good. The statistical metrics are for two different variables, so they do not have a direct comparison against each other. The WASP model conceptual framework includes periphyton. Where is the data on periphyton coming from? The WASP model is not simulating periphyton specifically. Although modelers can put specific data on periphyton into the WASP model, it is not necessary. Science Panel Discussion It was interesting to see the model results related to the current velocities. When sediment oxygen demand (SOD) was measured in the lake in Hogsett et al. (2018), the device used to measure SOD mixed the sediment at 10 centimeters/second. Utah Lake Nutrient Model Development Next Steps The modeling team will be moving from the hydrodynamic calibration to the water quality calibration. The results of the water quality calibration will be shared at the next Science Panel meeting in June. RESEARCH PRESENTATION – PHOSPHORUS-BINDING (P-BINDING STUDY) Dr. Josh LeMonte, Brigham Young University (BYU), presented the preliminary results from the ULWQS P-Binding Study. The presentation, the subsequent Science Panel discussion, and public comments are summarized below. P-Binding Study Research Presentation Below is a summary of the ULWQS P-Binding Study presentation. P-Binding Study Field Data Overview • The key findings from the P-Binding Study are that sediment and water geochemical trends change with depth and that the sorption and desorption from sediments are site, pH, and phosphorus loading dependent. • The P-Binding Study addresses the following ULWQS charge questions: o What is the current state of the lake with respect to nutrients and ecology? o What is the role of calcite "scavenging" (i.e., binding) in the phosphorus cycle? o How do sediments affect nutrient cycling in Utah Lake? o What are current sediment equilibrium phosphorus concentrations (EPC) throughout the lake? • The sediment geochemistry across the lake is not uniform, and it changes. The primary study used to characterize sediment phosphorus content across the lake are Randall et al. (2019), and the primary study used to characterize sediment mineralogy is Sonerhelm (1974). • The P-Binding Study researchers sampled seven sites across Utah Lake using bulk top sediment and freeze cores. The sampling sites included Saratoga Springs, West of Vineyard, Pelican Point, Provo Marina, Provo Bay, East of Bird Island, and Goshen Bay. • The research team assessed the sediment mineralogy of the cores collected at each site. The dominant mineral throughout Utah Lake is calcite, except Provo Bay, which is primarily composed of quartz. 18 • The phosphorus concentrations of the sediment cores ranged from 674 mg/kg to 957 kg/mg. The West of Vineyard and Saratoga Springs sampling sites had the lowest sediment phosphorus concentrations, while Provo Bay had the highest. • The research team conducted sequential extractions of the sediment cores. They also sent samples to the Brookhaven National Laboratory to process at its electron microscopy facility. The electron microscopy results show whether phosphorus molecules are bound to iron, calcium, or magnesium. It can also show the oxidation state of iron. The research team found mixed iron oxidation states in the surface sediment west of Provo Marina and Provo Bay. The electron microscopy analysis for all analyzed phosphorus "hotspots" indicated a high amount of calcium bonding in Provo Bay and West of Provo Marina. • The research team installed redox probes to collect measurements at each site. The probes were connected to buoys with batteries and solar panels to act as an energy source. The probes show that the redox potential varies across time and depth. Some of the oxidizing conditions may be due to storm events, which may have stirred the sediments. Additionally, the attachment between the buoys and the probes may have dislodged the probe, exposing it to well-mixed water. The research team is looking for a wind speed dataset to compare to the redox results. • The average redox condition for the Bird Island dataset is +50 millivolts, while the average redox condition for Provo Bay is -200 millivolts. All the sites where the researchers installed redox probes (Bird Island, Provo Bay, Provo Marina, and West of Vineyard) had average redox values indicative of reducing conditions, suggesting that it would be more prevalent for iron-2 to be present in the sediment than iron-3. • The research team also collected porewater samples. They installed peepers in the sediments, which collected samples at the sediment-surface interface and deeper into the sediments. They left the peepers in the sediment for two weeks to equilibrate. Because the lake changes and the sediment moves across the two-week equilibration period, they also put Velcro on the side of the peepers to track where the sediment-water interface was at the extraction time. • The results from the peepers suggest that the average porewater redox potential decreases rapidly near the sediment-water interface. At 20 centimeters above the sediment-water interface, the redox value is approximately 200 millivolts. The redox values rapidly decrease from zero to eight centimeters deep into the sediment, and at eight centimeters deep, the redox value is approximately -150 millivolts. • The average porewater pH stays fairly consistent across depth. There is a slight observed increase in pH over time, but statistically, the pH stays the same. • The average porewater conductivity decreases with depth. • Overall, total phosphorus concentrations are higher in porewater than in the water column. The general trend is that phosphorus concentration increases as the depth of the sample increases, but the phosphorus concentration trends are largely site dependent. For example, the Provo Bay sample has a sharp increase in phosphorus concentrations at 20 centimeters deep, a unique characteristic of this sample. • The research team sampled the water quality at the surface, middle, and bottom waters. They found that Provo Bay had the highest total phosphorus concentrations compared to the other sites, and Goshen Bay had slightly higher total phosphorus concentrations compared to the other sites. They also found that bottom waters had the highest total phosphorus concentrations compared to surface and middle waters. The researchers captured differences in surface and middle waters at Pelican Point, Goshen Bay, and Provo Bay samples. 19 • The research team separated the colloidal phases and extracted the colloids to assess how much phosphorus each colloidal phase contained. They separated the colloidal phases into four size classes (one micron, 0.45 microns, 0.1 microns, and 0.05 microns). They found that all the colloidal phases contained phosphorus. This line of research may be something for other investigators to explore in the future. P-Binding Study Experimental Results • The research team conducted two types of experiments to determine the partitioning coefficients and rate of kinetics (i.e., how quickly partitioning is occurring): sorption isotherm experiments and time-resolved stirred flow experiments. • The time-resolved stirred flow experiments determine the phosphorus sorption reaction rates. The results from the stirred flow experiments indicate that most phosphorus sorption occurs rapidly (less than thirty minutes) in Utah Lake sediment. The research team determined these rates by conducting the stirred-flow experiments at the natural pH 8.5. The sorption/desorption hysteresis (reversibility) experimental results are forthcoming. • The researchers ran the sorption isotherm experiments using sediments from all seven sampling sites and pure calcite. These experiments provide sorption maxima and partitioning coefficient values. • The research team modeled sorption isotherm data using linear, Freundlich, and Langmuir approaches. The linear approach is preferred if there is a good fit because it is simpler. One benefit of the Langmuir approach is that it allows the research team to generate the sorption maxima (i.e., the maximum amount of phosphorus the sediments could bind). • The research team conducted the sorption isotherm experiments using different phosphorus concentration values to identify the sorption maxima. • The results of the sorption isotherm experiments indicate that the partitioning coefficient increases as pH increases across all sites. The average partitioning coefficient is 40 across all samples at all pH levels. At a pH of 8, the average partitioning coefficient is around 20, and at a pH of 9, the average partitioning coefficient is closer to 80. These results indicate that more sorption is occurring at higher pH levels. • The research team conducted one set of sorption isotherm experiments without microbes. Under these abiotic conditions, there is a decrease in sorption, which suggests that some of the sorption is microbially mediated. • The results of the sorption isotherm experiments also indicate the sorption maxima, on average, is around 2,000 milligrams/kilogram of wet sediment. This value indicates that the sediments can hold onto more phosphorus than it is currently holding onto. • The Utah Lake system is complex; the sorption maxima value represents the greatest amount of phosphorus the sediment could hold onto if conditions are ideal. There are equilibrium controls that may prevent the sediment from reaching the sorption maxima values. These results beg the question: Does the sediment's affinity to bind phosphorus change as the phosphorus concentration changes? The research team calculated the sorption origin values (i.e., where the sediment shifts from sorption to desorption conditions) to try to answer this question. • For the Provo Bay samples at pH 8.5, the linear sorption origin is 0.74 milligrams/liter, the Freundlich sorption origin is 0.58 milligrams/liter, and the Langmuir sorption origin is 0 milligrams/liter. These results indicate the sediment changes from source to sink at 0.58 to 0.74 milligrams/liter of dissolved phosphorus in the water column. • The experiment conducted on Provo Marina samples at pH 8.5 indicates that the sediments will turn from a source to a sink at 0.21 to 0.80 milligrams/liter of dissolved phosphorus in the water column. 20 Science Panel Clarifying Questions Science Panel members asked clarifying questions about the P-Binding Study. Their questions are indicated in italics below, with the corresponding responses in plain text. Are the peepers measuring total phosphorus concentrations in the porewater? The peepers are measuring dissolved phosphorus concentrations in the porewater. Did the research team digest the porewater samples? No. The peeper data indicates that there are sometimes pH values lower than the pH values conducted for the sorption isotherm experiments (7.5 to 9). Why are the pH values in the lake so low? The pH levels in the peeper data may be an artifact of the sampling method. The research team added deionized water to the peepers to let it equilibrate. Dr. LeMonte does not have confidence that the pH values from the peeper porewater data are as low as it is showing. Is the information from the P-Binding Study being used in the Utah Lake Nutrient Model sediment diagenesis module? No, the results from the P-Binding Study are not incorporated into the sediment diagenesis module. The partitioning coefficient values from this study are used in the WASP model. What conditions might change the sorption origin values? A change in pH would affect the sorption origin values. The sorption origin will differ at a pH of 9 versus 7. How would one know whether to use the sorption origin based on the linear, Freundlich, or Langmuir approach? The research team statistically analyzed the fit of each approach to the data. The approach that produced the best fit statistically would be the more appropriate approach to select. Are there any issues with people forcing a fit of the data? The question of how to select the line of best fit depends on whether one is assessing inorganic or organic phosphorus. The literature suggests that if one is assessing organic forms of phosphorus, it is better to use the line of best fit that passes through zero. The research team is still working on how to determine the line of best fit for determining the sorption origin. Science Panel Discussion • DWQ and Tetra Tech have compiled wind speed datasets, which may be useful to the P- Binding Study research team. • The sorption origin values would suggest that the sediments almost always act as a phosphorus source to the water column. Some phosphorus entering Utah Lake is being bound into mineral forms that are not readily available to biological processes. • Many factors will influence the pH in Utah Lake. Cyanobacterial blooms will raise pH. For example, following an algal bloom in the summer, the pH in Utah Lake can get as high as 9 in the main lake and 10 in the marinas. Photosynthesis will drive pH up during the day, and respiration will drive it down at night. • The sediments will act as a sink when the pH is high, potentially making phosphorus the limiting factor during high pH conditions. 21 Public Comment Utah Lake is a self-regulating body of water. Over the past 30 years, phosphorus loading has at least doubled, if not tripled, but the phosphorus concentration in the water column has stayed the same. There would have to be a 95% to 98% reduction of the inflowing phosphorus to get to a situation where the phosphorus concentration in the water column would decrease. Returning Utah Lake to mesotrophic or oligotrophic conditions is a nearly impossible task. P-Binding Study Next Steps Dr. LeMonte is working on the final report from the P-Binding Study. Once the report is final, the P- Binding Study Subgroup will reconvene to provide any feedback. They will have more time to discuss the study's final results at that time. RESEARCH PRESENTATION – UTAH LAKE MASS BALANCE Dr. Mike Brett, University of Washington, presented his mass balance analysis. The presentation, the subsequent Science Panel discussion, and public comments are summarized below. Utah Lake Mass Balance Research Presentation Below is a summary of the ULWQS mass balance research presentation. Mass Balance Analysis Overview • The mass balance analysis considers the lake as one giant system rather than hundreds of smaller systems. The analysis quantifies nutrient inputs, sediment storage and release, and long-term recovery in Utah Lake. • Quantifying nutrient fluxes in Utah Lake is important because the lake frequently has toxic cyanobacteria blooms. Cyanobacteria blooms are strongly associated with high phosphorus and nitrogen concentrations. The nutrients come from natural watershed weathering and biological processes. However, high nutrient concentrations are usually associated with anthropogenic inputs. • The nutrient mass balance in lakes follows a simple formula: change in concentration = inputs - outputs - removal. • It is conventional to assume that Utah Lake is in a steady-state condition to characterize long-term average conditions. Under steady-state conditions, the change in concentration is assumed to be zero. This assumption means that: 0 = inputs – outputs – removal. • The formula can be rearranged to the expression: inputs = outputs + removal. • The input side of the equation is the phosphorus entering Utah Lake, and the outputs represent the phosphorus leaving Utah Lake. The phosphorus removal from the system occurs when phosphorus is lost to the sediment. • The mass balance model quantifies net removal. This means the model represents the cumulative outcome of several dynamic processes, not the processes themselves. The loss of phosphorus to the sediments is assumed to be a removal because only phosphorus in the water column can support algal growth. The mass balance model assumes that phosphorus is either in the water column or in the sediments at any given time; it cannot be in both places at once. • This equation represents the general mass balance model: lake volume = water flow into the lake * phosphorus concentration of inflow – water flow out of the lake * phosphorus concentration of outflow – first-order loss rate of nutrients * phosphorus concentration of the first-order loss. • The steady-state assumption allows the researcher to assume lake volume does not change. The general mass balance model equation can be rearranged to: water flow into the lake * 22 phosphorus concentration of inflow = water flow out of the lake * phosphorus concentration of outflow + first-order loss rate of nutrients * phosphorus concentration of the first-order loss. • The mass balance equation is further rearranged to account for the impact of water residence time and evaporative losses on phosphorus concentrations. Lake total phosphorus concentrations are determined by the flow-weighted input concentration, the first-order net loss rate to the sediments, and the amount of time that phosphorus can be lost to the sediments. • The mass balance dataset is based on data originally calculated by Dr. LaVere Merritt and subsequently by Dr. Kateri Salk, Scott Daly, and various Science Panel members and subcommittees. The values used for the mass balance analysis are as follows: o Total phosphorus concentration of the lake = 68 micrograms/liter o Total phosphorus concentration of the inflow – 343 micrograms/liter o Outflow:inflow ratio = 0.29 o Water residence time = 1.09 years o First-order net loss rate to the sediments = 4.31/year o Total phosphorus removal = 1 – (outflow * outflow total phosphorus concentration/inflow * inflow total phosphorus concentration) = 0.937 • The mass balance values are based on the five-year average of data from 2015-2020. Residence time varies because volume and inflow vary. • The following sources are assumed to provide nutrients to Utah Lake: tributaries, drains, precipitation, WWTPs, and atmospheric deposition. Each source is estimated and accounted for in the mass balance analysis. Mass Balance Analysis Results • The results of the mass balance analysis indicate that Utah Lake is an exceptional lake in that it has a high removal rate of phosphorus, which means it is good at storing phosphorus in the sediment. One potential explanation is that the phosphorus is binding to calcite. • The mass balance analysis can assess how changes in phosphorus concentrations of WWTP effluent will impact total phosphorus concentrations in the lake. As WWTP effluent concentrations decrease, capital and operations and maintenance costs, energy use, and greenhouse gas emissions increase. In another mass balance analysis conducted in the Spokane Basin in Washington, the results indicated a threshold at which point the costs of decreasing WWTP effluent concentrations outweighed the benefits produced. The mass balance analysis could identify a similar threshold for Utah Lake. • One question that the mass balance analysis can help answer is how long it would take for the Utah Lake system to achieve a new steady state. The results of the mass balance model predict that the transition to a new steady state for Utah Lake will primarily be governed by the removal of phosphorus to the sediments. According to the mass balance analysis, achieving a new steady state would take less than a year. • One question often asked about the mass balance model is whether internal loading is included in this model. Internal loading is included in the model because the model estimates the net retention rate. The Upper Klamath Lake is a good example of how to estimate internal loading at a lake-wide scale. Upper Klamath Lake has a large amount of internal phosphorus loading and large algal bloom events. During the summer, phosphorus internal loading increases dramatically; the mass balance for Upper Klamath Lake indicates that the internal loading is much higher than the flow input concentrations. In Upper Klamath Lake, the phosphorus is stored in the sediments, and the active blooms further draw phosphorus into the water column. 23 • Utah Lake has a much higher total phosphorus input concentration than Upper Klamath Lake. However, the total phosphorus concentration in the Utah Lake water column does not come near the total phosphorus input concentration. If internal loading in Utah Lake were 1,500 tons of phosphorus/year, that would be equivalent to adding ~2,200 micrograms of phosphorus/liter to the lake water each year, above and beyond the contributions from all external inputs. • The long-term monthly average concentrations of the Utah Lake water column can help quantify summer phosphorus releases from the sediment back to the water column. The long-term dataset indicates a 24 microgram/liter increase in the total phosphorus water column concentration during summer. This increase is equivalent to about 30 tonnes of phosphorus/year, which can be attributed to internal loading. If a more flexible approach is used to estimate the internal loading during the summer months, then the estimated internal loading is about 45 tonnes of phosphorus/year. The internal loading in Utah Lake tends to follow the summer phytoplankton bloom and not vice-versa. This pattern may be because the phytoplankton can access the phosphorus due to vertical migration. • The total phosphorus loading estimates to Utah Lake include 133.4 tonnes of TP/year from WWTPs, 49.6 tonnes of TP/year from tributaries, 32 tonnes of TP/year from atmospheric deposition, and 45 tonnes of TP/year from internal loading. • The conclusions from the mass balance model are that: o Utah Lake has high phosphorus inputs, mostly from WWTPs. o Conversely, the lake effectively removes phosphorus from the water column and sequesters it in the sediments. o However, lake TP concentrations are still directly linearly related to input concentrations. o Once external inputs are reduced, recovery is predicted to be very rapid. o Internal loading increases summer TP concentrations by about 33 micrograms/liter. Science Panel Clarifying Questions Science Panel members asked clarifying questions about the mass balance analysis. Their questions are indicated in italics below, with the corresponding responses in plain text. What are the current phosphorus concentrations of WWTP effluent? It varies by WWTP. The TSSD effluent has a concentration of 0.6 to 0.7 milligrams of phosphorus/liter. How does the mass balance model account for sediment resuspension and settling? The mass balance model says a phosphorus molecule is either in the water column or sediments; it cannot be in both. The model accounts for sediment resuspension and settling by estimating the net retention rate. How deep is Upper Klamath Lake? Upper Klamath Lake is, on average, three meters deep. Does Upper Klamath Lake experience hypoxia above the sediment? Yes, that is a difference between Upper Klamath and Utah Lake. Upper Klamath Lake is also very nitrogen limited. 24 Does Upper Klamath Lake remain nitrogen limited through the summer, even with the presence of nitrogen-fixing microbes? Aphanizomenon dominates Upper Klamath Lake. Aphanizomenon can fix nitrogen, which results in large pulses of nitrogen, stimulating microcystis blooms. Aphanizomenon remains the dominant genus in Upper Klamath Lake. Public Clarifying Questions Members of the public asked clarifying questions about the mass balance analysis. Their questions are indicated in italics below, with the corresponding responses in plain text. What is meant by the term "recover?" The term "recover" refers to how long it would take for the lake to reach a new steady state condition. It is not intended to allude to reference conditions. Science Panel Discussion • If phosphorus is stored in the sediment, lowering the phosphorus concentration in the water column would induce more phosphorus coming from the sediment, extending the time it takes to reach a new equilibrium. • The estimated phosphorus contribution of WWTP to Utah Lake (133.4 tonnes/year) is outdated. Once the upgrades are completed to the WWTPs, the new estimate will be closer to 40 tonnes/year. • The atmospheric deposition estimate of 32 tonnes of phosphorus/year is far below other estimates of 130 tonnes of phosphorus/year. • The mass balance analysis is based on total phosphorus and does not distinguish between bioavailable and biologically unavailable forms of phosphorus. • It is important that the Science Panel carefully consider how decreases in WWTP effluents will impact the lake. Upgrades to WWTPs are costly, and a high level of confidence that decreases in phosphorus inputs will impact the lake is needed to justify these costly upgrades. • The mass balance model expects a rapid change in the lake's steady state as concentrations decrease. As WWTPs improve their facilities and release less nutrients into the lake, this model suggests that there will be a rapid change to the lake over the next few years. These next few years will be a way to examine how quickly the steady state of Utah Lake changes in response to decreasing nutrient inputs. • Decreases in external phosphorus loading may result in a simultaneous decline in phosphorus sedimentation rates. If this dynamic were to unfold, Utah Lake's recovery period would lengthen, too. • The mass balance model is intended to be a helpful planning tool for all sources. The vision for the tool is to transform it into an interactive model through which Science Panel members could change different variables and see how the lake responds. • The mass balance model uses data from 2015 to 2020, so it does not account for recent WWTP upgrades. An interactive model would allow Science Panel members to input the new WWTP phosphorus concentration values to see how the lake would respond. Similarly, the Science Panel members could change other loading variables, not only WWTP inputs. • The mass balance model would complement the Utah Lake mechanistic models and provide multiple lines of evidence. • Nitrogen is an important limiting factor to algal growth, particularly in the summer. The Science Panel should continue considering the relationship between nitrogen and algal growth. The mass balance analysis focused on phosphorus primarily because it is more 25 difficult to build a mass balance model for nitrogen, given the influence of nitrification and denitrification on nitrogen loading. Public Comment • Over the past thirty years, the loading contributions from treatment plants doubled, while at the same time, there was no change in observed chlorophyll levels in the lake. • Utah Lake is surrounded by the playa that formed as Lake Bonneville contracted. This playa contributes a high amount of atmospheric deposition to Utah Lake, estimated to be two to three times higher than 32 tonnes of phosphorus/year. • There are three to four inches of "ooze" at the bottom of Utah Lake that has accumulated over the past 50 years. This ooze is a tremendous reserve of phosphorus that contributes to algal growth. These phosphorus reserves would make it unlikely that Utah Lake would recover within a year. • Carp removal has not resulted in positive impacts on Utah Lake. Ultimately, carp removal reduced intraspecies competition, replacing starving carp with stronger, healthier carp. The Provo River Delta will help strengthen June sucker populations, with the goal of displacing common carp in the lake. If possible, in the future, it would be beneficial to account for the Provo River Delta in the mass balance model. Mass Balance Analysis Next Steps Dr. Mike Brett will continue to work on the mass balance analysis, with the goal of creating a unified mass balance for Utah Lake and publishing the results. NEXT STEPS • Tetra Tech will incorporate Science Panel feedback and continue to make progress on the S- R analyses. Tetra Tech will also continue to make progress on the reference conditions analysis. • The Science Panel will be meeting in person in June. Potential topics for the June agenda include: o An update on the Paleolimnological Study o An overview of results from a recent study conducted by Dr. Steve Nelson at BYU o An overview of ecological monitoring efforts o A boat tour of Utah Lake • Once the P-Binding Study is completed, the P-Binding Study Science Panel Subgroup will meet to review the study and provide final feedback.