San Luis Drain Shutoff Simulations and Link-Node Recalibration

1
San Luis Drain Shutoff Simulations and Link-Node
Recalibration

• Joel Herr and Katie van Werkhoven
• Systech Water Resources, Inc.
• San Joaquin River DO TMDL PI Meeting
• May 15, 2008

2
Outline

• San Luis Drain Shutoff Forecast / Hindcast
• Model Validation
• Link-Node Recalibration
• Conclusion

3
Simulation Runs

• Forecast
• Load shifting
• Add flow and loading from off week to week before
  shutoff
• No net load reduction, only shift
• Do nothing forecast
• Assume flow and loading trends continue
• Hindcast
• Load shifting
• Actual event use measured data
• Do nothing
• Use measured data except for continuous flow,
  load from Mud Slough

4
Forecast Results EC
Load shifting SL Drain shutoff forecast vs Do
nothing forecast
5
Forecast Results Phytoplankton
Load shifting SL Drain shutoff forecast vs Do
nothing forecast
6
Relative Error of Input Projections
Mud Slough Salt Slough SJR _at_ Lander Merced River Tuolumne River Stanislaus River
Flow 0 55 30 -21 12 -22
EC -15 -14 -13 2 29 -10
Phyto-plankton 250 -5 8 601 57 525
7
Hindcast Results EC
Actual SL Drain shutoff event vs Do nothing
hindcast
8
Hindcast Results Phytoplankton
Actual SL Drain shutoff event vs Do nothing
hindcast
9
Prediction of Flow at Vernalis
Flow at Vernalis
10
Prediction of EC at Vernalis
EC at Vernalis
11
Phytoplankton at Vernalis
Phytoplankton at Vernalis
12
Response at Crows Landing
13
Response at Vernalis
14
Response at Brandt Bridge
15
D.O. at Rough Ready Island

• Effect on DO at RRI
• Was minimal 0.01 mg/l
• Signal of decreased phytoplankton mostly
  dispersed by Brandt Bridge
• Effect spread out over 2-3 weeks
• What does this mean?
• Load shifting signal does not transmit well to
  DWSC
• Little effect for short-term loading reduction

16
What did we learn?

• Reducing Mud Slough phytoplankton load may be
  effective at reducing phytoplankton at Vernalis
  Old River
• DWSC DO response is small, delayed, and diffused
  over longer time period

17
Outline

• San Luis Drain Shutoff Forecast / Hindcast
• Model Validation
• Link-Node Recalibration
• Conclusion

18
Validation Flow at Vernalis
19
Validation EC at Vernalis
20
Validation Phytoplankton at Vernalis
21
Phytoplankton Sampling
22
Whats up with 2007?

• Flow simulation very accurate
• Simulated EC 20 less than observed
• Simulated Phytoplankton much less than observed
• Model error or data error?
• Where are errors originating?

23
Data status for 2007

• Meteorology complete
• Tributary inflows mostly complete
• Merced River has missing time periods
• Los Banos Creek flow estimated
• Less data for agricultural drains
• Tributary water quality complete
• Diversions 2007 flow data included for 1 of 4
  IDs estimated for the rest

24
Possible Sources of Model Error

• Newman Wasteway
• DMC Recirculation occurred 8/15-9/12/2007
• Salt Slough / Mud Slough / Los Banos Ck
• Monitoring stations upstream of SJR
• Phytoplankton growth may occur before entering
  the SJR
• Unknown inputs not included in model

25
Location of EC Error

• EC is conservative
• Calculate mass error (simulated observed load)
  at Crows Landing
• Calculate mass error at Vernalis
• EC error at Crows Landing accounts for 48 of
  error at Vernalis

26
Location of Phytoplankton Error

• Fixed phytoplankton at Crows Landing Simulation
  shows Vernalis is OK

27
Phytoplankton Growth
28
Model changes

• Mud Slough, Salt Slough, Los Banos creek
  downstream of monitoring stations added to model
  domain
• 20-25 phytoplankton growth before reaching SJR
• Newman Wasteway added to model domain, DMC
  recirculation of 2004 and 2007 simulated
• Operational for limited time, dilution effect

29
Model Changes
30
Outline

• San Luis Drain Shutoff Forecast / Hindcast
• Model Validation
• Link-Node Recalibration
• Conclusion

31
Link-Node Recalibration Goals

• Improve D.O. simulation from Old River to Channel
  Point, maintain calibration at Rough Ready
  Island
• Improve phytoplankton simulation, especially from
  Old River to Channel Point
• Increase ammonia from Old River to Channel Point

32
Link-Node Recalibration Method

• Recalibration attempted with global reaction
  rates (unsuccessful)
• Link-Node code upgraded for reaction rates which
  vary by node
• Reaction rates recalibrated in two sections DWSC
  and elsewhere

33
Link-Node Recalibration

• (Live demonstration of simulation results)

34
Brandt Bridge Phytoplankton

• (live demonstration of Link-Node output)

35
Brandt Bridge Dissolved Oxygen
36
Garwood Bridge Phytoplankton
37
Garwood Bridge Dissolved Oxygen
38
Rough Ready Dissolved Oxygen
39
What happens now?

• San Joaquin River Model Interface (WARMF
  Link-Node) will continue to be available for
  download
• Updates will be available with new projects
• Users can get technical support from Systech
  joel_at_systechwater.com925-355-1780