Platte West Well Field A Protective Modeling Approach

Transcription

1 Platte West Well Field A Protective Modeling Approach Presented to: 2007 Eastern South Dakota Water Conference / 52 nd Annual Midwest Ground Water Conference June 25, 2007 Presented by: Luca DeAngelis, P.E., P.G. Layne Christensen Co-authors: Glenn Dostal, P.E. (HDR), Kevin Tobin, P.E. (MUD)

2 Overview Project Location Saunders and Douglas Counties, Nebraska Project Objectives Primary - Provide additional water supply to Omaha Metro Area Meet demand estimates for year 2030 Secondary - Increase system reliability Omaha currently gets 70% of drinking water from Missouri River

3 Omaha - Water Supply Sources mgd mgd Florence Florence Surface Surface Water Water mgd mgd Platte Platte West West Wells Wells mgd mgd Peak Peak shaving shaving Bedrock Bedrock Wells Wells Study Study Area Area 60 mgd Platte South Wells

4 Major Project Components Well Field 42 vertical wells with groundwater designation Direct recharge from the Platte River Combination of river and storage wells 2 48-inch diameter pipes crossing Platte River Compliance with Clean Water Act Section 404 Permit 100 million gallon per day (mgd) water treatment plant

5 EW-8 EW-11 EW-10 EW-9 EW-7 EW-6 EW-5 EW-4 EW-3 EW-2 EW-1 Unique Project Setting Ithaca Mead Superfund Superfund Site Site Yutan Venice Wann Platte Platte West West Well Well Field Field Platte West well field located: Three miles from a Superfund Site Five miles from the City of Lincoln, NE well field Memphis Lincoln Lincoln Well Well Field Field Ashland

6 Role of Groundwater Modeling Purpose Ensure compliance with Clean Water Act Section 404 Permit Evaluate impact of well field operation on other users Impact on Superfund Site Impact on Lincoln well field Develop contingencies for Platte River Low flow conditions River contamination scenarios

7 404 Permit Overview Clean Water Act Section 404 Permit Responsible agencies United States Army Corps of Engineers (USACE) United States Environmental Protection Agency (EPA) Restricts well field pumping Annual average of 52 mgd Maximum daily rate 104 mgd Specifies no impact on Superfund Site No impact on existing containment system No impact on contaminant transport

8 Protective Modeling Concept All groundwater modeling involves professional judgment Protective modeling: Develop a model framed with real world data Use model input parameters that will overestimate impact from operation of Platte West well field Provide realistic (probable) estimates for system response Ensure that any errors are conservative or protective to desired outcome Be protective of Superfund site cleanup efforts and Lincoln well field water rights Integrate groundwater and surface water Review historical streamflow record Establish low flow quantiles Review historical groundwater data Review groundwater/surface water interaction

9 Protective Model Scenarios Flow (cfs) Platte River Flow Frequency Curve Leshara Gauge Ave. Ave. flow flow = 4,600 4,600 cubic cubic feet feet per per second second (cfs) (cfs) 100 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Exceedance Frequency (%) Streamflow Average October 2003 Mar-05 Period of Record 1990 to % 99% flow flow = cfs cfs Develop reasonable worst case scenarios Groundwater systems near rivers are most stressed in the summer High water demand Low surface water flow Protective simulations should reproduce summer conditions Platte River simulate summer flows at 99% exceedance (100 year return period) Simulate highest well field flow 75 mgd for July and August Include pumping from other well fields Include crop irrigation (over 500 wells 150 mgd!)

10 Hydrogeologic Setting GW Flow Shaded Relief Map of Study Area GW Flow Platte River Alluvial Aquifer Todd Valley: Ancestral stream valley of Platte River. Not connected to present day Platte River.

11 Modeling Procedure Desktop Study Develop conceptual model Extend computer model to physical boundaries!!! Collect Field Measurements Head targets October 2003 and March 2005 Flux targets Seven United States Geological Survey (USGS) stream Construct model and calibrate to steady-state state Model constructed using MODFLOW (flow model) and MODPATH (particle tracking model) codes Conduct a sensitivity analysis of calibration Use ASTM standard as guidance (279 model simulations) Predictive simulations Steady State simulations Transient simulations Six year flow field Predictive Sensitivity Analysis Approximately 60 model simulations!

12 Model Construction Colon Wahoo Cedar Bluff Sweadeburg Ithaca Mead Mercer Leshara Memphis Yutan Ashland 31 6 Valley Venice Wann Elk City Waterloo Model grid smallest cells 100 ft x 100 ft Model Area 650 Square Miles Model Boundaries River Cell No Flow Cell General Head Boundary Cell Agricultural Drain Tile Cell

13 NWann SWann Model Calibration Cedar Bluff Wahoo Colon Sweadeburg Ithaca Mead Mercer Leshara Yutan Valley Venice Elk City Waterloo Memphis Ashland Wann Calibration or History Matching A demonstration that the model is capable of reproducing measured heads, flows, and response to stress as observed in the real world Monitoring Well/ Calibration Target 107 monitoring well locations

14 Goals for Calibration - History Matching Key to transition of model from conceptual to predictive 1. Exhibit a calibration error between simulated and observed water levels of less than 5% 2. Visual match Observed groundwater flow direction and gradients Superfund Site plume movement 3. Coherence with expected values for streamflow Baseflow values in creeks, less than or equal to 10 percentile of measured flow in creek 4. Exhibit less than a 1% difference in water mass balance

15 Calibration Results Numerical Calibration NRMS = 1.4% ARM = 2.1 feet Scattergram: Shows random distribution of error Conclusion Excellent statistical calibration with random error

16 Calibration - History Match to Superfund Plume MW63B year MW40B MW41B10 year MW55B MW53B MW56B 20 year MW30B MW43B MW31B MW33B MW64B 15 year MW52BMW54B 25 year year MW42B MW44B Particles Particles released released in in and and tracked tracked forwards forwards under under no-pumping no-pumping scenario scenario Model Model predicts predicts TCE TCE in in Johnson Johnson Creek Creek between between USACE USACE reports reports TCE TCE detections detections in in Johnson Johnson Creek Creek in in (maximum (maximum concentration concentration detected detected at at location location indicated). indicated). MW29B MW32B MW34B MW45B MW46B MW-39A Wann MW36B MW35B MW-38A Platte River MW62B MW37B

17 Predicative Simulations Steady State Annual average conditions Pumping rate of 52 MGD per permit Transient Pumping rate changes every 2 months (41 to 75 MGD) Seasonal river flow/stage Annual average (50% exceedance) - all months except 99% exceedance for months of July August and December- January Includes spatially distributed irrigation wells

18 Seasonal River Flows 2002 Drought Conditions Photos from Ashland, NE Normal Flow Drought < 1,000 cfs July to August

19 Impact of Irrigation Pseudo Steady State Monitoring Well August 2006 to August 2007 Hydrograph Water Level Elevation (ft msl) Irrigation Impact /1/2006 9/1/ /2/ /2/ /3/2006 1/3/2007 2/3/2007 3/6/2007 4/6/2007 5/7/2007 6/7/2007 7/8/2007 8/8/2007 Date

20 Predictive Steady State Model Predicted Drawdown Wahoo Colon Sweadeburg Ithaca Leshara Mead Yutan Memphis PLATTE Ashland Valley RIVER Venice Wann Waterloo Steady State Conditions Drawdown Contour (feet) Platte West Former Nebraska Ordnance Plant (Superfund Site) Ashland Well Field (City of Lincoln)

21 Predictive - Transient Cedar Bluff Wahoo Colon Sweadeburg Ithaca Mead Mercer Leshara Memphis Yutan Ashland Valley Venice Wann Elk City Waterloo Platte River Platte River Transient Irrigation Wells 550 wells 150 mgd Irrigators typically consist of center pivot systems (1,000 gpm plus)

22 Comparison of Transient vs. Steady State Drawdown (feet) Red Steady State Green Transient Transient Cone Larger Lower river flows Concentrated irrigation pumping

23 Identify Source of Water to Wells Summary: 71% - Direct Depletion 84 % - Direct or Indirect Depletion

24 Reverse Particle Tracking Waterloo Mead Yutan Venice Particles Tracked Backwards 20 yrs Identifies Source Water to Wells Wann Memphis

25 Impact on Superfund Site Yutan Venice Wann Procedure: Particles start outside of known contaminant extents Particles tracked under no Platte West pumping conditions Particles are tracked with Platte West pumping at 52 mgd Particle travel path is measured Result is predicted movement under Protective model scenario

26 Model Review Detailed Review Provided by Several Agencies: USACE Omaha District Kansas City District HTRW Center of Expertise EPA Region 7 Center for Subsurface Modeling Support (CSMoS) Nebraska Department of Environmental Quality (NDEQ) Lower Plate North Natural Resource District (LPNNRD) Public Comments

27 Model Uses Develop hydraulic and chemical monitoring programs Develop hydraulic monitoring well priority structure Develop predicted operational water level range for 17 hydraulic monitoring wells Hydraulic wells equipped with transducers

28 Hydraulic Monitoring Network Priority One Wells Priority Two Wells Priority Three Wells

29 Example Monitoring Well Operational Range (Priority One Well)

30 Future Work Model Post Audit Evaluate accuracy of model predictions Compare to real world data Annual 404 Permit Report Simulate previous 6 months pumping perform Post Audit Predict impact of next 6 months pumping Hydraulic and Chemical Monitoring Field Test to Estimate Conductance Better quantify connection with Platte River

31 Conclusions MODEL REPRODUCES KNOWN REAL WORLD CONDITIONS Platte West Well Field Obtains Approximately 85% of Flow from Depletion (Direct or Indirect) of Platte River Well Field Reaches Steady State Approximately 1.5 Years Steady State Simulations are a Reasonable Approximation of Transient Simulations Platte West operations will not impact Superfund Containment/City of Lincoln System Protective Modeling Approach Help develop monitoring programs Planned model updates should improve predictive capabilities

32 THANK YOU ARE THERE ANY QUESTIONS?