NPDES Permits in New England
Bays Eutrophication Model Evaluation Group
Submitted to OMSAP April 29,2002
The Bays Eutrophication Model Evaluation Group (BEMEG) met on March 18, 2002 at U Mass Boston to review work conducted by HydroQual for the Massachusetts Water Resources Authority (MWRA) in response to questions and recommendations raised by BEMEG in 1999 and 2001. In particular, BEMEG was asked to review and comment on the following reports:
Calibration of the Massachusetts and Cape Cod Bays Hydrodynamic Model: 1998-1999. Report 2001-12. (Preliminary report reviewed March 2001.)
Addendum to "Bays Eutrophication Model (BEM): modeling analysis for the period of 1992-1994." Report 2001-13.
Boundary sensitivity analysis for the Bays Eutrophication Model (BEM). Report 2001-14.
Analysis of the addition of a third algal group to the Bays Eutrophication Model (BEM) kinetics. Report 2001-15.
BEMEG also heard a preliminary report on the efforts underway at U Mass Boston to transition the USGS/HydroQual Bays Hydrodynamic Model (BHM) and BEM to their new home at U Mass Boston, and an analysis of DO and chlorophyll variability found in the recent Battelle monitoring data. The meeting agenda and list of attendees are included in Appendices 1 and 2, and lists of previous HydroQual and MEG reports in Appendices 3 and 4 for completeness.
This report summarizes next the discussion, comments and concerns about the above HydroQual reports (with comments directed to the authors) and the U Mass Boston transition effort, and concludes with recommendations made by BEMEG for future work.
2. Review of HydroQual Reports
2.1 Calibration of the Hydrodynamic Model: 1998-1999 (Report 2001-12)
This report presents the Bays Hydrodynamic Model (BHM) results and comparisons with in-situ data for the period 1998-1999. The model is identical to that used for the earlier 1992-1994 model calibration period with the one modification that incident shortwave radiation is now allowed to penetrate into the water column rather than be absorbed entirely in the surface grid layer. This change improved the model's ability to predict near-surface temperatures. Overall the model results appear to be adequately "calibrated" and ready for use in the water quality modeling for this time period.
The BHM uses the Mellor-Yamada level 2.5 (MY2.5) turbulence closure model with the Galperin et al extensions to compute the vertical turbulent viscosity/diffusivities at each time step. Recent studies suggest that MY2.5 may underestimate vertical mixing in the presence of strong stratification and weak shear, i.e., those conditions typically found in the thermocline/pycnocline. The background minimum viscosity/diffusivities used here is Km = 0.05 cm(2)/s, consistent with the thermocline dye study of Geyer and Ledwell (1994). Hilton et al (1998) found that using Km = 0.5 cm(2)/s provided better simulations of near-surface mixing in Boston Inner Harbor. To help provide some insight into this issue, please add plot(s) showing the vertical viscosity and thermal diffusivity as a function of depth at two sites (one at/near the new outfall and the other in the center of Stellwagen Basin) over a tidal cycle during several representative periods (e.g., a calm summer (winter) day when vertical stratification is near its seasonal maximum (minimum)), plus a comparison of these results with relevant observations (preferably in Mass Bays or if none available, in the literature).
This report contains a lot of model documentation, but it is not clear if this is intended to be the definitive model documentation requested by BEMEG. If it is intended as the definitive documentation, then it is incomplete. Additional information about the following is needed:
Shortwave penetration (page 2-6). Many investigators use the two spectral component model to describe the vertical distribution of shortwave radiation : I(z) = I0 (a1 e(-k1z) + a2 e(-k2z)where Io is the total incident shortwave radiation corrected for reflection, a1 and a2 are the fraction of radiation carried by the shorter and longer wavelength components (with a1 + a2 = 1), and k1 and k2 are their extinction coefficients. Typical values for the southern flank of Georges Bank are a1 = 0.8, k1 = 1/1.4m, a2 = 0.2, and k2 = 1/6.3m. To help justify the one component model (equation 2-7) used here, please include model/data comparison plot(s) and describe the range of the extinction coefficient ke.
Surface heat flux (page 2-7). What input and model data and equations are used to compute the surface heat flux components? Since modeled heat flux depends on water surface temperature, whose prediction varies over a large number of surface grid cells and time steps, how does the model perform these calculations? The text suggests that surface evaporation (E) and precipitation (P) offset each other so that E+P is set to zero in the surface mass flux (which seems reasonable). Based on recent email from Rich Isleib, we understand that surface heat flux due to evaporation is included, but MEG would appreciate more details.
Boundary T/S relaxation times (page 2-11). The use of relaxation times for inflow conditions on the open boundaries is appropriate, but how were the values of 3 to 30 days chosen and were these calibrated?
GOMM (page 2-12). What is the Gulf of Maine Model (GOMM) referred to here?
2.2 Addendum to "Bays Eutrophication Model (BEM): modeling analysis for the period of 1992-1994" (Report 2001-13)
This report documents the boundary conditions used in the original and revised 1992 BEM simulations, plus the few changes in model coefficients made between these two simulations. The boundary conditions and water quality kinetics and model parameters used in the 1992-1994 calibration simulation are also documented.
This report also compares the BEM results for 1994 with the BEM solution computed using the same horizontal/vertical grid resolution used in the hydrodynamic model. As expected, the two solutions are similar overall, and the past BEM model projections made using grid aggregation should not be affected significantly. The higher spatial resolution model results exhibit reduced numerical diffusion, especially for properties like nitrite/nitrate that have relatively strong spatial gradients as compared with, say salinity, and provide a more detailed picture of spatial structure. For these reasons, all future BEM simulations should be run using the hydrodynamic model grid within the BEM domain.
2.3 Boundary sensitivity analysis for the BEM (Report 2001-14)
This report explores the sensitivity of BEM-predicted DO and DIN concentrations within the bays near the outfall (the near field (NF)) to changes in these variables imposed along the open boundary (i.e., the specified far field (FF) concentrations). The model results show a strong correlation of FF and NF DO, implying that oxygen is being imported to the NF more than generated within the model domain. However, because DO does not significantly affect other state variables, uncertainties in the boundary conditions for DO are not as critical as they might be. On the other hand, there is little correlation of NF and FF DIN concentrations, but a strong correlation of other NF variables to the FF DIN. That is, changes in the FF DIN concentration affect other variables within the model domain, more than changes in the FF DO concentration. These results point out the need for more data on the open boundary.
2.4 Analysis of the addition of a third algal group to BEM (Report 2001-15)
A third algal group was added to BEM to investigate if the fall phytoplankton bloom frequently observed in the bays could be accurately predicted for the period 1992-1994. With this fall diatom group included, the BEM did predict an annual fall bloom but the model-data agreement was only slightly better than without the fall group. The BEM could still not reproduce the chlorophyll maximum observed in October 1993, and model calibration had to be effected by changing the carbon/chlorophyll ratio and the boundary loadings, indicating that the fall algal group model was not capturing the actual internal algal dynamics. Despite the limited success, the effort of adding the third algal group was worth doing as it illustrates the limitations of our current understanding.
3. U Mass Boston Model Transition
The U Mass Boston (UMB) modeling group has successfully configured the latest version of the HydroQual hydrodynamic model code to run on the UMB computer system, and showed a comparison between their model simulation for 1994 and the existing HydroQual simulation. The two solutions were very similar, differing only in the near-surface temperature field due to the inclusion of vertical penetration of shortwave radiation in the latest model code. Further comparison cases will be conducted with identical physics and forcing to ensure that the HydroQual and UMB hydrodynamic models are producing identical solutions within machine error. After these tests, the UMB hydrodynamic model will be ready to use to produce the input fields required for the BEM. The UMB group has started working on the transition of the BEM, with the long-term objective being to simulate the 1998-2002 period for comparison with the field data during the pre- and post-start of the outfall.
Based on the presentations and discussions held at the meeting, the BEMEG makes the following recommendations for future work and consideration:
HydroQual be asked to run the BEM for 1998-1999 both with and without the third algal group, using the hydrodynamic model grid. These two simulations are of intrinsic interest for several reasons. The closing of the Nut Island treatment plant and start of secondary treatment at the Deer Island treatment plant in 1998 made a significant change in the distribution and forms of nutrient input to Boston Harbor, and it will interesting to determine if the higher spatial resolution BEM simulations will capture the effects of these input changes in and near Boston Harbor, especially the increased chlorophyll observed in Boston Harbor, and whether the addition of the fall algal group is more successful in these years. These simulations also allow more localized mass balance studies of different nutrients. The boundary forcing in 1998 and 1999 were considerably different, so these model simulations should provide additional insight into the bays behavior. The two-algal group simulation will also serve as the final test solution for the transition of BEM from HydroQual to UMB.
Additional documentation about the HydroQual hydrodynamic model following section 2.1 above should be added to the "Calibration of the Hydrodynamic Model: 1998-1999" report, so that this report when completed will be the definitive report on this model.
As shown effective by HydroQual, UMB should plan to run the BEM using the same spatial resolution used in the hydrodynamic model. In addition, UMB should consider if co-locating the BEM open boundary with the hydrodynamic model would improve BEM simulations.
The space and time scales resolved in the Bays hydrodynamic model are insufficient to simulate directly the physical processes that control the discharge of effluent from the outfall diffusers and immediate mixing with ambient fluid. To bridge this gap between "far-field" and "near-field", a mass- conservation approach is used to estimate the properties of the mixed water in the grid cells immediately over the outfall, which in turn is advected and mixed away from the outfall by the larger-scale physical processes represented in the hydrodynamic model. This approach has been examined by Blumberg et al. (1996) and Zhang and Adams (1998) through comparisons of the far-field model predictions with those from a near-field engineering model. Both groups find support for the approach, especially in looking at the far-field effects of the outfall effluent. The monitoring data recently collected after the outfall started operation provides an excellent opportunity to revisit the question of just how well the Bays hydrodynamic model does in predicting near-field mixing and plume behavior. In particular, the quality and quantity of in-situ near-field data now allow a detailed comparison with the hydrodynamic model predictions. BEMEG encourages UMB to conduct this study before conducting final post-outfall simulations, to determine the accuracy of this approach and give confidence that the hydrodynamic model is providing the correct near-field currents, water structure and effluent concentrations as input to the BEM sufficient for the environmental questions being asked by MWRA and EPA. Since the Bays model development effort was completed, Blumberg and co-investigators have developed a more direct approach to the scale mismatch problem, i.e., embeding a near-field model in the far-field model (P. Roberts, personal communication). Initial results reported by Connolly et al. (1999) and Roberts (1999) show that this approach predicts the outfall plume behavior well. The results of the detailed comparison between the present Bays hydrodynamic model and recent monitoring data requested above will help determine if this newer, more direct approach is needed in the Bays modeling effort.
These recent reports plus the Battelle presentation all point to the importance of the western Gulf of Maine on water properties and marine life in especially Massachusetts Bay. There is a clear need for continued and improved sampling of water properties along the upstream open boundary near Cape Ann. The addition of the GoMOOS buoys in this area is an important step forward, and MWRA should consider how to supplement the proposed measurements with nutrient sensors to improve our knowledge of the advective nutrient input into the bays, plus more frequent shipboard sampling to provide the spatial context for the moored measurements. The question of near-bottom advection of DO from the far field to the near field could be investigated with moored DO measurements at several sites, including at the Boston buoy near the outfall. Again, BEMEG encourages MWRA and UMB to think of ways to improve the in-situ sampling of the key variables that strongly influence the physics and biology in the bays, especially in the near-field around the outfall.
The HydroQual hydrodynamic model results for 1998-1999 exhibit several fascinating events (e.g., the wind-mixing event during May 7-14, 1998; Figure 3-30). If possible, MWRA should encourage detailed scientific study of some of these events where there is sufficient field data for model/data comparison. Such a study might be an excellent M.S. or Ph.D. thesis project that provides both deeper insight to how the bay works, and the strengths and weakness of the present models.
Blumberg, A.F., Z.-G. Ji, and C.K. Ziegler, 1996. Modeling Near-Field Plume Behavior using a Far-Field Circulation Model. Journal of Hydraulic Engineering, 122, 610-616.
Connolly, J.P., A.F. Blumberg, and J.D. Quadrini, 1999. Modeling fate of pathogenic organisms in coastal waters of Oahu, Hawaii. J. Environmental Engineering, 398-406.
Galperin, B., Kantha, L. H., Hassid, S., and Rosati, A., 1988. A quasi-equilibrium turbulent energy model for geophysical flows. J. Atmospheric Sci., 45, 55-62.
Geyer, W.R. and J. Ledwell, 1994. Final Report, Massachusetts Bay Dye Study. MWRA. Report ENQUAD 1994-17, pp 39.
Hilton, A., McGillivary, E., and Adams, E., 1998. Residence time of freshwater in Boston's Inner Harbor, J. Waterway, Port, Coastal and Ocean Engineering, 124(2): 82-89.
Roberts, P.J.W., 1999. Modeling Mamala Bay Outfall Plumes. I: Near field. J. Hydraulic Engineering, 565-573.
Zhang, X.-Y., and E.E. Adams, 1998. Simulating near field plumes with a far-field model. J. Hydraulic Engineering, 125,233-241.
Appendix A1. Meeting Agenda
Outfall Monitoring Science Advisory Panel
Bays Eutrophication Model Evaluation Group Meeting
March 18, 2002, 10:00 AM to 4:00 PM
U Mass Boston, Provost's Conference Room, 8th Floor Healey Library
10:00 - 10:15
Welcome, Purpose of Meeting, and Introductions
Bob Beardsley, WHOI, MEG Chair
10:15 - 10:30
Schedule for Future Modeling
Meng Zhou, U Mass Boston
10:30 - 11:15
Response to 1999 and 2001 MEG Recommendations
Jim Fitzpatrick and Richard Isleib, HydroQual
11:15 - 12:00
Questions Posed to the MEG:
- Is the Model Documentation Adequate
- Should We Use the Fine Grid Resolution?
- Is the Third Algal Group Justified?
- How Substantial is the Influence of the Boundary Conditions?
- Is the Hydrodynamic Run for 1998-1999 Adequate?
Jim Fitzpatrick and Richard Isleib, HydroQual
Mike Mickelson, MWRA
12:00 - 1:00
1:00 - 1:30
Insights from the Data into the DO and Chlorophyll Response
Scott Libby and Carlton Hunt, Battelle
1:30 - 3:30
MEG Discussion and Comments for the Final MEG Report
Appendix A2. Meeting Attendees
MEG Members: Bob Beardsley (chair), WHOI; Eric Adams, MIT; Jeff Cornwell, U. Maryland; Don Harleman, MIT; Jack Kelly, EPA Duluth MN Research Lab; Jay O'Reilly, NOAA Narragansett Lab; and John Paul, EPA Narragansett Lab.
Observers: Brad Butman, USGS; Cathy Coniaris, MADEP; Dave Dow, NMFS; Jim Fitzpatrick, HydroQual; Bernie Gardner, UMass Boston; Anne Giblin, MBL; Doug Hersh, MWRA; Carlton Hunt, Battelle; Russ Isaac, MADEP; Rich Isleib, HydroQual; Mingshun Jiang, UMass Boston; Ben Kelly, Save the Harbor/Save the Bay; Wendy Leo, MWRA; Pierre Lermusiaux, Harvard; Suh Yuen Liang, MWRA; Scott Libby, Battelle; Matt Liebman, EPA; Mike Mickelson, MWRA; Andrea Rex, MWRA; Dave Taylor, MWRA; Sal Testaverde, NMFS; John Warner, USGS; and Meng Zhou, UMass Boston.
Appendix A3. List of HydroQual Reports
(1) Water Quality Model: 1989-1991
"A water quality model for Massachusetts Bay and Cape Cod Bay: model design and initial calibration." Report 1993-05.
(2) Water Quality Model: 1989-1992
"A water quality model for Massachusetts and Cape Cod Bays: Calibration of the Bays Eutrophication Model (BEM)." Report 1995-08.
Final report reviewed by MEG in 1995.
(3) Water Quality Model: 1993-1994 with revised 1992
"Bays Eutrophication Model (BEM): modeling analysis for the period 1992-1994." Report 2000-02.
Final report reviewed by MEG in 1999.
(4) Hydrodynamic Model: 1998-1999
"Calibration of the Massachusetts and Cape Cod Bays Hydrodynamic Model: 1998-1999." Report 2001-12.
Preliminary report reviewed by MEG March 14,2001.
Final report reviewed by MEG March 18, 2002.
(5) Water Quality Model: 1992-1994 with revised 1992 --- Additional documentation and exploration of grid disaggregation for 1994
"Addendum to "Bays Eutrophication Model (BEM): modeling analysis for the period of 1992-1994." Report 2001-13.
Final report reviewed by MEG March 18, 2002.
(6) Water Quality Model: explore 1992 with reference to range of 1994-1999 boundary conditions
"Boundary sensitivity for the Bays Eutrophication Model (BEM)." Report 2001-14.
Final report reviewed by MEG March 18, 2002.
(7) Water Quality Model: explore 3rd algal group for 1992-1993
"Analysis of the addition of a third algal group to the Bays Eutrophication Model (BEM) kinetics." Report 2001-15.
Final report reviewed by MEG March 18, 2002.
Appendix A4. List of BEMEG Reports
Beardsley R, Adams EE, Harleman D, Giblin AE, Kelly JR, O'Reilly JE, Paul JF. 1995. Report of the MWRA hydrodynamic and water quality model evaluation group. Boston: Massachusetts Water Resources Authority. Report ENQUAD ms-037. 58 p.
Report of Bays Eutrophication Model Evaluation Group to OMSAP, June 13, 2000 http://www.epa.gov/region01/omsap/meg1299.html
Bays Eutrophication Model Evaluation Group comments on draft HydroQual report "Preliminary Calibration of the Massachusetts and Cape Cod Bays Hydrodynamic Model: 1998-1999", March 14, 2001
Will be posted at: http://www.epa.gov/region1/omsap/