Loading...
HomeMy WebLinkAboutDWQ-2024-004516 1 Utah Lake Water Quality Study (ULWQS) Science Panel March 24, 10:00 AM to 12:00 PM Virtual Meeting Meeting Summary - FINAL ATTENDANCE: Science Panel Members: Zach Aanderud, Greg Carling, Mitch Hogsett, Theron Miller, Hans Paerl, and Tim Wool Steering Committee Members and Alternates: Eric Ellis and Chris Keleher Members of the Public: Dave Epstein, Dan Potts, and David Richards Utah Division of Water Quality (DWQ) staff: Scott Daly and Nicholas von Stackelberg Technical Consultants: Rene Camacho, Kevin Kratt, and Kateri Salk Facilitation Team: Heather Bergman and Samuel Wallace DECISIONS AND APPROVALS No formal decisions or approvals were made at this meeting. OVERVIEW OF THE WATER QUALITY ANALYSIS PROGRAM (WASP) MODEL SEDIMENT DIAGENESIS ROUTINE Tim Wool, ULWQS Science Panel, presented an overview of the WASP model sediment diagenesis routine. His presentation is summarized below. Sediment Diagenesis Routine Background and History • Dr. James Martin was the expert on modeling sediment diagenesis and made major contributions to this module in his career. • Before the sediment diagnosis model, there was no way to simulate the impacts of sediment oxygen demand (SOD) on nutrient releases in water quality models. Modelers used to calibrate SOD to observed levels. The challenge with only calibrating the model to observed SOD values is that the model could not then simulate the impact of management practices on SOD and nutrient fluxes. The sediment diagenesis routine was developed so that the model could simulate how changes in management practices impacted SOD and nutrient fluxes. • The sediment diagenesis routine all comes from the same source. The source of the routine is modeler Dr. Dominic DiToro, currently at the University of Delaware. All model programs (WASP, QUAL2K, etc.) use the same routine. Dr. James Martin worked with model developers to ensure the code was unified and working appropriately across models. Sediment Diagenesis Routine Overview • The sediment diagenesis is fixed in two layers. The top sediment layer is the active layer that interacts with the overlying surface water. This layer can be oxic or anoxic depending on the conditions of the overlying water column and underlying sediment layer. The second, lower layer is the underlying sediment and is assumed to be anoxic. 2 • Materials pass between the two sediment layers through a diffusive process. The top layer rates are controlled by molecular diffusion or bioturbation (e.g., worms moving the sediment). The bioturbation is a function of productivity going above the top layer. • The sediment diagenesis model is not a sediment transport model. • There are several inputs into the model to control how the sediment is mixed. There are many inputs into the model, but the modelers have found that only a few parameters need to be adjusted when put into this module. • The WASP model will provide downward organic carbon, nitrogen, and phosphorus fluxes from the water column into the top sediment layer. When the organic material settles into the sediments, that is when the sediment diagenesis model begins to process the material. The WASP model can also simulate downward fluxes of silica if it is relevant. • The model breaks the material into key classes. The key classes organize the material by their reactivity. The classes are labile, refractory, and inert. For example, dead algal material will be classified as labile. The modeler has to program what fraction of the setline material is labile, refractory, and inert. • The model calculates the mass balance for each particulate organic matter for each of the reactivity classes. The model accounts for bioturbation, diffusion (into the water column and the sediments), diagenesis, burial, and influx. • There is a large list of kinetic constants put into the model. Many of the kinetic constants do not have to be changed. Sediment Diagenesis – Nitrogen and Methane • To simulate ammonia cycling, the sediment diagenesis routine has to solve for oxygen before it begins simulating nutrient cycling. The routine will calculate the SOD first. Then, if there is oxygen in the top layer of the sediment, it will incorporate nitrification to simulate ammonia fluxes. • The ammonia simulation has partitioning coefficients so that ammonia sorbs and desorbs as a function of concentration. • The ammonia inputs are simple. The routine uses a half-saturation constant for oxygen to simulate nitrification. • The nitrite cycling simulation functions similarly to the ammonia cycling simulation. The sediment diagenesis routine does not have a state variable for nitrite, so everything is assumed to be in the nitrate form in the water column. The routine simulates nitrification from nitrite to nitrate. • In the top layer of the sediment, the route accounts for the nitrification (from nitrite to nitrate) and denitrification (from nitrate to nitrite) occurring depending on oxygen concentrations. The routine simulates the denitrification occurring in the anoxic, lower sediment layer. The routine then models nitrate diffusion between all the layers using diffusion coefficients. • The routine calculates methane flux from the top layer to the water column. The WASP model does not incorporate methane flux. Other models, like QUAL2K, can simulate methane fluxes. Sediment Diagenesis – Phosphorus and Silica • The sediment diagenesis routine is a separate program that WASP talks to during runtime. There are multiple ways to link multiple segments to a single SOD calculation. Connecting multiple cells to a single SOD segment is possible. 3 • The sediment diagenesis routine calculates SOD first. Once the module solves for dissolved oxygen in the water column, it will calculate the dissolved oxygen concentration in the top layer as a function of what is settling out and diffusing in. • The routine will simulate phosphate cycling. The phosphate cycling uses a partitioning coefficient and inputs similar to the ammonia simulation. Sediment Diagenesis Routine Summary • The inputs to the sediment include: o Carbon, nitrogen, and phosphorus fluxes o Rates and constants o Characteristics of the water column as a function of time and space:  Ammonia  Nitrite  Nitrate  Phosphate  Oxygen  Salinity  Available silica  Methane  Temperature  Salinity • The WASP model will feed nitrate, ammonia, orthophosphate, oxygen, and organic material from the water column into the top layer of the sediments. The organic material will then be broken into different classes. • The processes are fairly slow, so the model can be run for a long period. It can also be run repeatedly until it reaches equilibrium. • The important outputs from the sediment diagenesis are: o Ammonia flux to the water column o Nitrite flux to the water column o Nitrate flux to the water column o Phosphate flux to the water column o Methane flux to the water column o SOD rates o Sulfide flux to the water column o Dissolved silica flux to the water column • Some of the limitations of the sediment diagenesis routine include the following: o It does not account for iron and manganese. o It also has a fixed two-layer system, so it is not possible to integrate multiple layers into the model. o It cannot simulate the impacts of sediment transport (e.g., scouring and sedimentation). o It does not simulate the impact of benthic algae. o It does not simulate the impacts of rooted macrophytes. Science Panel Clarifying Questions Science Panel members asked clarifying questions about the sediment diagenesis routine. Their questions are indicated below in italics, with the corresponding responses in plain text. 4 How does the sediment diagenesis routine handle sediment resuspension? The Environmental Fluid Dynamics Model (EFDC) and the Simulating WAves Nearshores (SWAN) wind-wave model simulate sediment resuspension. However, the sediment transport dynamic cannot be linked to the sediment diagenesis module. Is there any way to use free oxygen to estimate the release of nutrients from iron and manganese? There is a lot of iron and manganese-bound phosphorus in the sediments. The relationship between iron, manganese, and oxygen is difficult to simulate. The redox chemistry of iron and manganese is very difficult to model. Iron and manganese are not a state variable in the model. UTAH LAKE WASP MODEL ENHANCEMENTS AND PROGRESS UPDATE Rene Camacho, Tetra Tech, presented an update on the Utah Lake WASP model enhancements and progress. His presentation is summarized below. Model Enhancement Overview • The primary objective of the model enhancement is to develop a predictive model of hydrodynamics and water quality (organic matter, nutrient cycling, and phytoplankton dynamics) in Utah Lake. • The intermediary goals are to: o Improve representation of physical processes by implementing a wind-wave model coupled to EFDC o Incorporate sediment diagenesis in all bottom cells of the model domain o Incorporate pH and alkalinity as simulated state variables of the water quality model o Improve overall model performance, stability, and runtime efficiency • Today’s presentation will provide a background on individual models, how the model works together, the model hydrodynamic performance, and model enhancements. Background on Individual Models and How They Work Together • The EFDC is a hydrodynamic model. The EFDC model solves the equations of mass and momentum transport. It can be used in a 2-D/3-D orthogonal curvilinear grid to simulate any shape of the environment. It also provides solutions for salinity, temperature, and conservative tracers with full-density feedback to handle stratified conditions. • The SWAN model will be coupled with the EFDC to simulate wind-induced waves and their impact on shear stress. Wind on the surface of a water body causes orbital circulation, which impacts the bottom of the water body. The shear stress applied to the bottom will cause sediment to move. The SWAN model, coupled with the EFDC model, can reproduce this process, which is particularly important in shallow-lake ecosystems. • The WASP model is the water quality model. The model accounts for the different species of nutrients, including dissolved organic and inorganic forms of nutrients, and uptake and excretion of phytoplankton biomass. It also accounts for photosynthesis and respiration impacts on oxygen. The WASP model will simulate the biochemistry dynamics in Utah Lake. • The hydrodynamic models (EFDC and SWAN) operate separately from the water quality model (WASP). The outputs from the hydrodynamic models are sent to the water quality model via a hydro linkage file. The hydro linkage file contains information on grid cell volumes, velocities, temperatures, and inorganic suspended solids. The WASP model incorporates the information from the file to calculate the mass balance related to nutrient cycling and eutrophication. 5 • The modelers will run EFDC to simulate the bottom elevation, wind, water surface elevation, and currents. The EFDC sends this information to SWAN, which simulates significant wave height, wave length, wave period, and wave direction. The SWAN uses these simulations to calculate energy, dissipation, and radiational stresses; all variables are then sent back to EFDC. • The model uses a computational grid to simulate nutrient cycling in the lake. The modeling team uses two vertical layers to represent nutrient cycling in the water column. The first vertical layer is for the littoral areas subject to wetting and drying, and the other is for the main water body. • One of the output variables of the sediment transport model is water surface elevation. The modeling team is tracking water surface elevation to ensure volumes are well represented in the model. Other output variables include temperature and significant wave height. The modelers compare the simulated data with observed data to track how representative the simulated conditions are of observed conditions. • The modelers also tracked the shear stress simulated by EFDC only and EFDC and SWAN linked. The version of the model with EFDC and SWAN linked incorporates wind-induced impacts to calculate shear stress, resulting in a significant increase in shear stress in the model. Water Quality Model Enhancements – Sediment Diagenesis • The modeling team has made a significant effort to improve different aspects of the model. • The first update to the model was the incorporation of sediment diagenesis in all computational cells. There are two approaches to integrating sediment diagenesis into the model: descriptive and predictive. o The descriptive approach requires the user to define SOD values at a specific location and time. The descriptive approach is computationally efficient and simple to handle in the model if the data is available. One downside of the descriptive approach is that it does not account for changes based on the nutrient fluxes from the water column to the sediments. o The predictive approach incorporates a mechanistic representation of nutrients in the sediments based on sediment organic matter content and the settling fluxes of organic matter. An advantage of the predictive approach is that it accounts for the nutrients settling from the water column, the nutrient content of the sediments, and the different processes related to the breakdown of organic matter. A downside of this approach is that it requires inputting constraints, which require a lot of data. It is also computationally extensive, particularly when modeling wetting and drying in the littoral zones. • In coordination with Tim Wool, the modeling team has incorporated the predictive approach for sediment diagenesis into the model. The modeling team is using the Littoral Sediment Study data and the nutrient fluxes from Hogsett to constrain the model. • The model can now reproduce the organic matter settling processes into the sediments. The model can take changes in load and feed them into the sediment diagenesis routines to simulate the impacts of management changes on SOD and nutrient fluxes. One caveat is that changing these processes occurs over long periods, so the model has to be run over a fairly long period under new conditions to reach a new equilibrated sediment value. • The sediment diagenesis module will provide SOD values based on specific locations and periods. It will also simulate pH values at the bottom of the lake. 6 Water Quality Model Enhancements – Sediment Transport • The modeling team is also working on sediment transport model enhancements. The coupling of EFDC and SWAN alone did not impact the water quality model. They considered two options to improve the model to leverage the EFDC-SWAN linkage. o The first option was to send the shear stress value from EFDC directly to WASP via the hydrodynamic file; the WASP model would simulate inorganic suspended sediments (ISS) based on the shear stress values from EFDC. One limitation of this option is that the model was not gaining much efficiency. o The second option was to have EFDC simulate the impact of shear stress on ISS and send the ISS values directly from EFDC to WASP. The WASP model now receives the ISS concentrations from EFDC and uses them to calculate light extinction and water clarity. Water Quality Model Enhancements – Phosphorus-Binding (P-Binding) • Tetra Tech, with the Science Panel, evaluated two approaches to simulate p-binding processes in the model. o The first approach was to develop a mechanistic representation of the inorganic carbon cycle to estimate the carbonate budget. The advantage of this approach is that it would be a complete representation of the inorganic carbon cycle/buffer where phosphorus would bind to carbonate and precipitate. It would also track process impacts on pH. The disadvantage of this approach was that this option is not available in WASP, so the modelers would need to develop and incorporate this process into the model. It would be computationally intensive and require extensive data to constrain the model. o The other approach was to use a partition coefficient to represent the precipitation of calcite and phosphorus into the sediments. The modelers would calculate the concentration of the particulate form of inorganic phosphorus by multiplying the dissolved phosphorus concentration by the partitioning coefficient. The model assumes thermal and chemical equilibrium and that the process is instantaneous. The advantage of this approach is that it is simple and computationally inexpensive. It also captures the net impact of the precipitation of phosphorous. The disadvantage of this approach is that it is an approximation. This approach also makes it so that the p-binding process is independent of changes in pH. Dr. Josh LeMonte’s, Brigham Young University (BYU), P-Binding Study indicates that the partition coefficient is not constant and varies with pH and other biogeochemical processes. • The modeling team is using the partition coefficient approach to simulate p-binding processes and calibrating the model using the data from Dr. Josh LeMonte’s P-Binding Study. Science Panel Discussion and Comments Science Panel members discussed and commented on the Utah Lake WASP model enhancements and progress. Their comments are summarized below. • The model will provide useful insights into the initial mass balances. It will be exciting to talk about the model in the context of other Utah Lake processes, such as calcium-phosphorus precipitation and iron and phosphorus interactions. 7 Science Panel Clarifying Questions Science Panel members asked clarifying questions about the Utah Lake WASP model enhancements and progress. Their questions are indicated below in italics, with the corresponding responses in plain text. Is it possible to adjust the CNP molar ratios of detritus if it is widely different than the Redfield ratio based on its composition of zooplankton and algae? • That is an option. The model has a variable stoichiometric relationship for phytoplankton groups. A similar thing could be implemented for zooplankton. • Anything incorporated into the mechanistic model needs to be calibrated. Without a zooplankton setline rate to calibrate, it would be difficult to know whether the model is accurately simulating conditions. Other lines of evidence in the NNC framework will be important to incorporate zooplankton considerations. Dr. Greg Carling, BYU, and Dr. James Martin discussed linking the WASP model to PHREEQC to simulate phosphorus-calcite binding. How would that work? Tim Wool and Dr. James Martin have discussed this possibility more generally and not specifically for Utah Lake. The information available for Utah Lake is likely not enough to link WASP with PHREEQC. Public Clarifying Questions Members of the public asked clarifying questions about the Utah Lake WASP model enhancements and progress. Their questions are indicated below in italics, with the corresponding responses in plain text. At the bottom of Utah Lake, there is a very loose layer above the sediment that is thick and deep but mobile. This layer forms anytime the wind blows and resuspends the sediment. Some refer to this layer as ooze or goo. How does the model capture this layer in its simulations? • This layer is referred to as the nepheloid layer in limnology. Others refer to it as flock. • The short answer is that the model does not explicitly represent the layer in the sediment transport simulation. The SWAN wind-wave model and portrayal of sediment shear stress is the first step to integrating this layer into the model. The model captures how much sediment is stirred up from the lake bottom and incorporated into the water column. • Utah Lake is unique because it does not stratify, and wind plays an important role in resuspending sediment. Even if the model does not directly address this layer of ooze, the Science Panel continues to talk about and consider these types of dynamic processes in the lake. Will the modelers field test the model with photosynthetically available radiation (PAR) readings? The modeling team plans on comparing the model outputs with the observed PAR data. The model team uses about 45% of the total active variation as PAR. Based on the observed data, the 45% value is a pretty constant number. Is the modeling team building the WASP model or another group? Tetra Tech is building the WASP model with support from Tim Wool. The sediment diagenesis routine incorporates algal biomass settling. Does it incorporate zooplankton? The model incorporates zooplankton descriptively, but it is not a state variable in the model. 8 Public Discussion and Comments Members of the public discussed and provided comments on the Utah Lake WASP model enhancements and progress. Their comments are summarized below. • The ooze at the bottom of the lake may be composed of ultra-processed biological material (e.g., algae, zooplankton, and decaying fish). Someone should sample this ooze to determine what it is made of. Previous studies have tracked total and volatile solids in sediment cores, and the data has been consistent across studies. • Some researchers have collected freeze cores, showing a fairly clear divide between the sediments and the water column. While cores may not be able to capture this layer of ooze, freeze cores should be able to because the freeze core is a thin sheet of aluminum that freezes everything around it. It might be good to look more closely at those freeze cores to determine what is occurring at the bottom of the lake. Another method could be to pump the layer of ooze for sampling. NEXT STEPS • Tetra Tech will continue to make progress on the model. Their next task is to calibrate the lake model against observed conditions to see how well it performs. After the calibration, they will conduct a sensitivity and uncertainty analysis. The Science Panel will potentially discuss the model updates at the May Science Panel meetings. • The next Science Panel meetings will be in person on May 18 and 19. More information will be forthcoming.