Presentation is loading. Please wait.

Presentation is loading. Please wait.

Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks CWEMF Annual Meeting April 16, 2012 Paul Hutton, Ph.D., P.E.

Published byBryce Small Modified over 7 years ago

Similar presentations

Presentation on theme: "Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks CWEMF Annual Meeting April 16, 2012 Paul Hutton, Ph.D., P.E."— Presentation transcript:

1 Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks CWEMF Annual Meeting April 16, 2012 Paul Hutton, Ph.D., P.E.

2 Acknowledgements Dr. Sujoy Roy, Tetra Tech Dr. Limin Chen, Tetra Tech Sanjaya Seneviratne, DWR

3 Summary Findings Additional review is needed before firm conclusions can be reached. ANNs appear to faithfully emulate DSM2 turbidity fate and transport during the season of interest (i.e. Dec-Feb). ANNs appear to provide a promising foundation for representing turbidity- based regulations in CalSim.

4 Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks Results Model Development Background Next Steps

5 RPA Component 1: Protection of Adult Delta Smelt Life Stage … delta smelt have historically been entrained when first flush conditions occur in late December. In order to prevent or minimize such entrainment, Action 1 shall be initiated on or after December 20 if the 3 day average turbidity at Prisoner’s Point, Holland Cut, and Victoria Canal exceeds 12 NTU… Action 1 shall require the Projects to maintain OMR flows no more negative than -2000 cfs… Source: Remanded USFWS 2008 Biological Opinion

6 DSM2 Turbidity Fate & Transport

7 Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks Results Model Development Background Next Steps

8 ANN Training Data DSM2 Simulations Run Hydrology & Operations Turbidity Boundary Conditions FreeportVernalisYoloCosumnesMokelumneCalaveras 1HistoricalLow 2HistoricalMidLowMid 3HistoricalHighLowHigh 4HistoricalLowHighLow 5HistoricalMidHighMid 6HistoricalHigh 7Historical w/o ExportsLow 8Historical w/o ExportsMidLowMid 9Historical w/o ExportsHighLowHigh 10Historical w/o ExportsLowHighLow 11Historical w/o ExportsMidHighMid 12Historical w/o ExportsHigh Notes: (1) DCC gates closed; (2) South Delta barriers not installed; (3) Constant Martinez & agricultural return turbidity boundary conditions

9 ANN Model Structure Matlab Feed Forward y(t) = f(x(t-1), …., x(t-d)) Inputs = 6 boundaries (3 flow & 3 turbidity) Hidden Neurons = 10 Time delay = 1-2 days Outputs: turbidity at 6 locations

10 ANN Model Structure Boundary Input (Daily Averages) Flow –North Delta (Freeport + Yolo) –East Side Streams –Calculated Old & Middle Rivers Turbidity (Flow-weighted) –North Delta (Freeport + Yolo) –East Side Streams –Vernalis

11 11 11 22 33 44 55 66 66 11 22 33 44 55 Sacramento River at Rio Vista San Joaquin River at Prisoners’ Point Old River at Quimby Island Middle River at Holt Old River at Bacon Island Clifton Court Forebay Entrance ANN Model Structure Output Locations

12 ANN Model Structure Training Process DSM2 data points are randomly assigned: –Training 60% –Validation 20% –Testing 20% Training data are used to compute network parameters. Intermediate results are iteratively compared with validation data until residual error is minimized. Testing data are independent of training and validation data and are used to evaluate network predictive power.

13 Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks Results Model Development Background Next Steps

14 Model Results Old River @ Quimby Island (Dec-Feb)

15 Model Results: Summary Statistics ANN Turbidity (ntu) = Ф 1 + Ф 2 * DSM2 Turbidity (ntu) LocationDailyMonthly Ф1Ф1 Ф2Ф2 R2R2 Ф1Ф1 Ф2Ф2 R2R2 Sacramento River @ Rio Vista 3.50.970.941.11.010.99 Old River @ Quimby Island 2.00.890.831.70.910.96 Old River @ Bacon Island 1.80.820.781.50.850.93 San Joaquin River @ Prisoner’s Point 3.70.810.763.00.870.92 Middle River @ Holt 2.00.760.691.70.820.89 Clifton Court Forebay Entrance 3.10.750.731.30.900.91

16 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity Old River @ Quimby Island Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 30 ntu East Side Turbidity = 30 ntu

17 Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks Results Model Development Background Next Steps

18 Evaluate auto-regressive networks Explore tidal input variable Implement methodology in CalSim

19 Next Steps (cont’d) CalSim Implementation Decision statement: Reduce pumping as needed to increase OMR flows, thereby controlling turbidity levels as defined by existing or alternative regulations. Develop 82-year turbidity time series for Delta inflows Integrate information into monthly time step Refine ANN training (and associated data) as needed

20 Paul Hutton, Ph.D., P.E. phutton@mwdh2o.com

21 EXTRA SLIDES

22 Turbidity Boundary Conditions Freeport Flow Range (cfs) Low (50%)Mid (75%)High (90%) < 10,000101520 12,500203040 17,500304070 22,5004060100 27,50060100160 32,50070140280 37,50090160320 45,000100170350 55,000100175300 65,000100140240 >70,000100140180

23 Turbidity Boundary Conditions (cont’d) Vernalis Flow Range (cfs) Low (50%)High <2,00015100 2,75020100 4,25025100 7,5002590 15,0002060 >20,0001560

24 Turbidity Boundary Conditions (cont’d) Flow Range (cfs) LowMidHigh <10020 1,000304060 5,00060120200 10,000100200300 >30,000100150200 Calaveras Yolo Bypass Flow Range (cfs) LowMidHigh <5020 10030 40 >1,0004070100

25 Turbidity Boundary Conditions (cont’d) Flow Range (cfs) LowMidHigh <10020 500305080 >1,0004070100 Cosumnes River Mokelumne River Flow Range (cfs) LowMidHigh <10010 500305080 1,00050100180 2,00080200280 >3,000100300

26 Model Results Sacramento River @ Rio Vista (Dec-Feb)

27 Model Results Old River @ Prisoner’s Point (Dec-Feb)

28 Model Results Middle River @ Holt (Dec-Feb)

29 Model Results Old River @ Bacon Island (Dec-Feb)

30 Model Results Clifton Court Forebay Entrance (Dec-Feb)

31 Model Results 2009-10 Historical Conditions Old River @ Quimby Island Sacramento River @ Rio Vista

32 2009-10 Historical Conditions

33 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity Sacramento River @ Rio Vista Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 30 ntu East Side Turbidity = 30 ntu

34 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity San Joaquin River @ Prisoner’s Point Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 30 ntu East Side Turbidity = 30 ntu

35 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity Middle River @ Holt Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 30 ntu East Side Turbidity = 30 ntu

36 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity Old River @ Bacon Island Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 30 ntu East Side Turbidity = 30 ntu

37 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity Clifton Court Forebay Entrance Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 30 ntu East Side Turbidity = 30 ntu

38 Steady State Flow-Turbidity Relationship as a Function of North Delta Turbidity Clifton Court Forebay Entrance Steady State Assumptions North Delta Flow = 30,000 cfs East Side Flow = 1500 cfs Vernalis Turbidity = 100 ntu East Side Turbidity = 30 ntu

39 ANN Model Structure Matlab Autoregressive y(t) = f(x(t-1), …., x(t-d)) Boundary Inputs = 6 (3 flow & 3 turbidity) Recursive Input = 6 (turbidity) Hidden Neurons = 10 Time delay = 1-4 days Outputs: turbidity at 6 locations

40 Spring-Neap Effect on Turbidity Clifton Court Forebay Entrance 1994-95 Study 4 Study 10

Download ppt "Modeling Delta Flow-Turbidity Relationships with Artificial Neural Networks CWEMF Annual Meeting April 16, 2012 Paul Hutton, Ph.D., P.E."

Similar presentations

Pilot Project Alternatives FRANKS TRACT San Joaquin River False River Old R i v e r Sacramento River Operable Gates in Three Mile Slough Alt #1 Operable.

Pilot Project Alternatives FRANKS TRACT San Joaquin River False River Old R i v e r Sacramento River Operable Gates in Three Mile Slough Alt #1 Operable.
Upstream Boundary (Section V.1). 1. Delta boundary, daily inflow 2. Extend upstream channel by 45 miles, daily inflow 3. Hourly inflow at Freeport boundary.

Upstream Boundary (Section V.1). 1. Delta boundary, daily inflow 2. Extend upstream channel by 45 miles, daily inflow 3. Hourly inflow at Freeport boundary.
Climate Change and Water Resources Challenge in California Francis I Chung, Ph.D., P.E. Department of Water Resources.

Climate Change and Water Resources Challenge in California Francis I Chung, Ph.D., P.E. Department of Water Resources.
SWRCB Modeling I Historical Modeling Tara Smith, Delta Modeling Section Department of Water Resources Tara Smith, Delta Modeling Section Department of.

SWRCB Modeling I Historical Modeling Tara Smith, Delta Modeling Section Department of Water Resources Tara Smith, Delta Modeling Section Department of.
Overview of TM6: Simulation of Selenium Fate and Transport in North San Francisco Bay Limin Chen, Sujoy Roy, and Tom Grieb Tetra Tech, Inc., Lafayette,

Overview of TM6: Simulation of Selenium Fate and Transport in North San Francisco Bay Limin Chen, Sujoy Roy, and Tom Grieb Tetra Tech, Inc., Lafayette,
Neural Networks. Background - Neural Networks can be : Biological - Biological models Artificial - Artificial models - Desire to produce artificial systems.

Neural Networks. Background - Neural Networks can be : Biological - Biological models Artificial - Artificial models - Desire to produce artificial systems.
Reclamation Mid-Term Operational Modeling Seasonal to Year-Two Colorado River Streamflow Prediction Workshop CBRFC March 21-22, 2011 Katrina Grantz, PhD.

Reclamation Mid-Term Operational Modeling Seasonal to Year-Two Colorado River Streamflow Prediction Workshop CBRFC March 21-22, 2011 Katrina Grantz, PhD.
The Upper Rio Grande. Multi-objective River and Reservoir System Modeling Flood Control Water Supply Navigation Aquatic/Riparian Habitat Recreational.

The Upper Rio Grande. Multi-objective River and Reservoir System Modeling Flood Control Water Supply Navigation Aquatic/Riparian Habitat Recreational.
In-Delta Storage Process OverviewProcess Overview Program BenefitsProgram Benefits Project CostsProject Costs IssuesIssues Proposed Work Plan for FY 2003Proposed.

In-Delta Storage Process OverviewProcess Overview Program BenefitsProgram Benefits Project CostsProject Costs IssuesIssues Proposed Work Plan for FY 2003Proposed.
A look at how MWD & other water agencies can make an extra $1.5 Billion Dollars by counting Yolo Bypass flows as if the same water flowed past the I Street.

A look at how MWD & other water agencies can make an extra $1.5 Billion Dollars by counting Yolo Bypass flows as if the same water flowed past the I Street.
City of Surrey Fire Service: Reducing Uncertainty for Annual Staffing Costs.

City of Surrey Fire Service: Reducing Uncertainty for Annual Staffing Costs.
Periodic Review of the 1995 Water Quality Control Plan for the San Francisco Bay/Sacramento-San Joaquin Delta Estuary Comments of The Bay Institute on.

Periodic Review of the 1995 Water Quality Control Plan for the San Francisco Bay/Sacramento-San Joaquin Delta Estuary Comments of The Bay Institute on.
Using Neural Networks in Database Mining Tino Jimenez CS157B MW 9-10:15 February 19, 2009.

Using Neural Networks in Database Mining Tino Jimenez CS157B MW 9-10:15 February 19, 2009.
Outline What Neural Networks are and why they are desirable Historical background Applications Strengths neural networks and advantages Status N.N and.

Outline What Neural Networks are and why they are desirable Historical background Applications Strengths neural networks and advantages Status N.N and.
Delta Flows: Data and Calibrations Cathy Ruhl US Geological Survey California Water Science Center, Sacramento, CA.

Delta Flows: Data and Calibrations Cathy Ruhl US Geological Survey California Water Science Center, Sacramento, CA.
Position Reconstruction in Miniature Detector Using a Multilayer Perceptron By Adam Levine.

Position Reconstruction in Miniature Detector Using a Multilayer Perceptron By Adam Levine.
Hanoi, January 28 th 2015 Quang Dinh DEIB – Politecnico di Milano IMRR Project Emulators of the Delta model 7 – Emulators of the Delta model INTEGRATED.

Hanoi, January 28 th 2015 Quang Dinh DEIB – Politecnico di Milano IMRR Project Emulators of the Delta model 7 – Emulators of the Delta model INTEGRATED.
Study for Fish Screens at Existing Delta Facilities Dr. Leah Orloff Water Resources Manager CUWA Board MeetingJuly 24, 2009.

Study for Fish Screens at Existing Delta Facilities Dr. Leah Orloff Water Resources Manager CUWA Board MeetingJuly 24, 2009.
1 Development of Common Assumptions Common Model Package for the CALFED Surface Storage Investigations Presentation To: California Water and Environmental.

1 Development of Common Assumptions Common Model Package for the CALFED Surface Storage Investigations Presentation To: California Water and Environmental.
Time-Series Forecasting Overview Moving Averages Exponential Smoothing Seasonality.

Time-Series Forecasting Overview Moving Averages Exponential Smoothing Seasonality.

Similar presentations

Ads by Google