Development and Application of Cayuga Lake Model (CLM2D) · 2014. 5. 19. · &MPV 727$/ 4416 hrs. &$/ J,1'6 0.02% '$7 3(5,2' Start Date: 5/1/2013 - 00:00 End Date: 10/31/2013 - 23:00

Development and Application of Cayuga Lake Model (CLM2D)

Rakesh K. Gelda

Upstate Freshwater Institute, Syracuse

1

Technical Briefing (CLTAC) Cayuga Lake Modeling Project May 19, 2014 Albany, NY

2

segments la

yers

inflow

outflow

Two-Dimensional Model

CE-QUAL-W2 (W2; US Army Corps of Engineers): dynamic, laterally averaged, two dimensional (longitudinal-vertical) model

applied successfully to 100s of waterbodies worldwide hydrodynamic submodel predicts water surface elevations, velocities,

and temperatures provides transport framework for a water quality model

Datasets used for model development and application

Model Run Year

Calibration 2013

Validation – primary 2012

Validation – secondary 1998–2006

Hindcast (extended application); probabilistic water quality simulations, if needed

1987–1997, 2007–2011

Model Setup

3

Longitudinal Grid

4 0 m 13 27 40 53 67 80 93 107 120 133 m

Depth X UTM

3.73e+5 3.74e+5 3.75e+5 3.76e+5 3.77e+5 3.78e+5

Y U

TM

4.701e+6

4.702e+6

4.703e+6

4.704e+6

4.705e+6

4.706e+6

L62

1

IL

LSC intake

McKinney's Point

pile clusterLSC discharge

CHWWTP

IAWWTP

48 segments

* Selected locations

Distance from downstream boundary (railroad bridge; km)

051015202530354045505560

Ele

vatio

n (m

)

-20

-10

0

10

20

30

40

50

60

70

80

90

100

110

120

LSC

dis

char

ge

LSC intake

Cayuga-AES power plantintake

Longitudinal-Vertical Grid - Cayuga Lake

48 segments 132 layers 3720 cells

5

Cayuga-AES power plant intake

LSC intake

LSC

dis

char

ge

6

Fall Creek

Cayuga Inlet Six Mile Creek

Taughannock Creek Salmon Creek

LSC discharge

IAWWTP discharge CHWWTP discharge

LSC intake

Cayuga-AES power plant intake and discharge

outflow Inflows and Outflows

5 Tributary Inputs 4 Point Source Inputs 2 Withdrawals 1 Outflow distributed inflows for minor tributaries

Meteorological Data

Meteorological Data Source Availability Notes

Piling Cluster (Cornell University) 10/27/2011–12/31/2013

10 min frequency; missing air T and dew T for 1/3/2013– 5/13/2013 were filled-in with the data from Ithaca Airport

Game Farm Road (Cornell University)*

1987–2013 hourly frequency; missing data (0.8% days) were filled-in with the data from Ithaca Airport

Ithaca Airport (NOAA)* 1996–2013 hourly frequency; missing data (0.2% days)

*Wind speed data from Game Farm Road and Ithaca Airport were adjusted to the height of wind measurement (8 m) at the piling cluster location, according to logarithmic boundary layer.

7

• air temperature, dewpoint temperature, wind speed, wind direction, solar radiation

Wind Distribution (Wind Rose Plots) May−October, 2013

8

WRPLOT View - Lakes Environmental Software

WIND ROSE PLOT:

COMMENTS: COMPANY NAME:

MODELER:

DATE:

1/22/2014

PROJECT NO.:

NORTH

SOUTH

WEST EAST6%

12%

18%

24%

30%

WIND SPEED (m/s)

>= 11.1

8.8 - 11.1

5.7 - 8.8

3.6 - 5.7

2.1 - 3.6

0.5 - 2.1

Calms: 0.02%

TOTAL COUNT:

4416 hrs.

CALM WINDS:

0.02%

DATA PERIOD:

Start Date: 5/1/2013 - 00:00End Date: 10/31/2013 - 23:00

AVG. WIND SPEED:

3.74 m/s

DISPLAY:

Wind SpeedDirection (blowing from)


WIND ROSE PLOT:


MODELER:

DATE:

3/31/2014

PROJECT NO.:

NORTH

SOUTH

WEST EAST4%

8%

12%

16%

20%

WIND SPEED (m/s)

>= 11.1

8.8 - 11.1

5.7 - 8.8

3.6 - 5.7

2.1 - 3.6

0.5 - 2.1

Calms: 8.88%

TOTAL COUNT:

4412 hrs.

CALM WINDS:

8.88%

DATA PERIOD:


AVG. WIND SPEED:

2.68 m/s

DISPLAY:



WIND ROSE PLOT:


MODELER:

DATE:

1/29/2014

PROJECT NO.:

NORTH

SOUTH

WEST EAST3%

6%

9%

12%

15%

WIND SPEED (m/s)

>= 11.1

8.8 - 11.1

5.7 - 8.8

3.6 - 5.7

2.1 - 3.6

0.5 - 2.1

Calms: 20.77%

TOTAL COUNT:

4333 hrs.

CALM WINDS:

20.77%

DATA PERIOD:


AVG. WIND SPEED:

3.10 m/s

DISPLAY:


Piling Cluster Game Farm Road Ithaca Airport


WIND ROSE PLOT:


MODELER:

DATE:

1/22/2014

PROJECT NO.:

NORTH

SOUTH

WEST EAST6%

12%

18%

24%

30%

WIND SPEED (m/s)

>= 11.1

8.8 - 11.1

5.7 - 8.8

3.6 - 5.7

2.1 - 3.6

0.5 - 2.1

Calms: 0.02%

TOTAL COUNT:

4416 hrs.

CALM WINDS:

0.02%

DATA PERIOD:


AVG. WIND SPEED:

3.74 m/s

DISPLAY:


Point Source Inflows

9

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Flow

(m3 s

-1)

0

1

2

32000-2013average

LSC discharge


Flow

(m3 s

-1)

0.0

0.5

1.0

1.51995-2013average

IAWWTP


Flow

(m3 s

-1)

0.0

0.1

0.2

0.31999-2002average

CHWWTP

2012 2013

Flow

(m3 s

-1)

0

5

10

152000-2013Cayuga-AES

Point Source Inflow Temperatures

10


Tem

pera

ture

(°C

)

5

10

152005-2013average

LSC


Tem

pera

ture

(°C

)

5

10

15

20

25

30

1998-2010, 2013average

IAWWTP

2012 2013

Tem

pera

ture

(°C

)

0

5

10

15

20

25

30Cayuga-AESall years:

CHWWTP discharge temperature = IAWWTP discharge temperature Years < 2012: Cayuga-AES discharge temperature = Model-predicted Cayuga-AES withdrawal temperature + 8.67 °C

Tributary Data

11

Tributary Flow Records Temperature Records

Fall Creek 1925-present 9/21/2011−11/18/2011;

5/13/2013−11/1/2013

Cayuga Inlet 1937-9/30/2011;

6/1/2012-present 5/9/2013−11/13/2013

Salmon Creek 2006-present limited during 2013

Six Mile Creek 1995-present limited during 2013

- unmonitored flow: pro-rated according to Fall Creek flow/watershed area - unmonitored temperatures: adopted from USGS monitored temperature at

Allen Creek near Rochester

12

2013


Flow

(m3 s

-1)

0

25

50

75

100(a) Fall Creekavg Q = 6.02 m3 s-1

2013


Flow

(m3 s

-1)

0

25

50

75

100(b) Cayuga Inletavg Q = 2.6 m3 s-1

2013


Flow

(m3 s

-1)

0

25

50

75

100(c) Salmon Creekavg Q = 3.48 m3 s-1

2013


Flow

(m3 s

-1)

0

25

50

75

100(d) Six Mile Creekavg Q = 2.05 m3 s-1

Tributary Inflow

Model Performance

13

2013 • depth-profiles of temperature at selected sites and days [observations from

Seabird Casts] • color-contours of temperature at the LSC intake location [observations from

thermistor string] • timeseries plots of temperature at different depths at the LSC intake location

[observations from thermistor string] • timeseries plot of temperature at site 2 [observations from a single thermistor

near surface] • spectral analysis of thermistor string measurements

2012 • color-contours of temperature at the LSC intake location [observations from

thermistor string]

1998−2006 • timeseries plot of temperature at site near piling cluster [observations from a

single thermistor at ~ 1 m deep]

Model Performance – Profiles 2013

14

07May13 18Jun13 25Jul13 20Aug13 17Sep13 15Oct13

(1) s 5 (2) s 5 (3) s 5 (4) s 5 (5) s 5 (6) s 5

RMSE=0.82 RMSE=0.32 RMSE=0.43 RMSE=0.81 RMSE=0.51 RMSE=0.76MAE=0.45 MAE=0.24 MAE=0.32 MAE=0.54 MAE=0.28 MAE=0.44

10:36A 10:39A 11:11A 10:33A 12:10P 10:10A

0 10 20

Elev

atio

n (m

)

-150

153045607590

105120

0 10 20 0 10 20 0 10 20

Temperature (°C)

0 10 20 0 10 20 30

07May13 18Jun13 25Jul13 20Aug13 17Sep13 15Oct13

(1) s 1 (2) s 1 (3) s 1 (4) s 1 (5) s 1 (6) s 1

RMSE=1.39 RMSE=0.32 RMSE=1.24 RMSE=0.44 RMSE=2.57 RMSE=1.15MAE=1.10 MAE=0.32 MAE=1.08 MAE=0.36 MAE=2.57 MAE=1.14

4:48P 3:06P 2:00P 4:22P 10:39A 4:49P

0 10 20

Elev

atio

n (m

)

105

110

115

1200 10 20 0 10 20 0 10 20

Temperature (°C)

0 10 20 0 10 20 30

site 5

site 1

Observations 2013 Observed temperatures at LSC intake location

2013

May Jun Jul Aug Sep Oct

Elev

atio

n (m

)

40

60

80

100

120

3 °C 7 11 16 20 24

15

measurements were made every 30 sec at 5 m depth interval data shown are at every 6 hours at 0.5 m depth interval

thermistor string temperature data at the LSC intake site

Predicted temperatures at LSC intake location

2013


Elev

atio

n (m

)

40

60

80

100

120

3 °C 7 11 16 20 24

16

Model Performance – Predictions 2013

predictions shown are at every 6 hours at 1 m depth interval

temperature predictions at the LSC intake site

17

Model Performance – Timeseries

Jun Jul Aug Sep

Tem

pera

ture

(°C

)

0

5

10

15

20

25

30observedpredicted

(b) 100 m el. (depth 17 m)RMSE = 1.8

Jun Jul Aug Sep

Tem

pera

ture

(°C

)

4

6

8

10

12

14

16

18

20

observedpredicted

(c) 90 m el. (depth 27 m)RMSE = 1.0

thermistor string at LSC intake location measurements are shown as 15-min averages and predictions are shown 1/hour

Model Performance – Timeseries

2013


Tem

pera

ture

(°C

)

0

5

10

15

20

25

30

observedpredicted

site 2, 2 mRMSE (5/1-10/31) = 1.78 °CRMSE (6/1-10/31) = 1.26 °C

18

single thermistor at site 2, 2 m deep measurements are shown at every 15-min interval and predictions are shown once every hour

19

Model Performance – Spectral Analysis

• comparison of internal wave spectrum generated from thermistor string data and model predictions

• June 1−July 13, 2013; data from 100 m elevation (17 m deep) thermistor

Frequency (d-1)

0.01 0.1 1 10 100

Spe

ctra

l Den

sity

1e-5

1e-4

1e-3

1e-2

1e-1

1e+0

1e+1

Period (d)

0.010.1110100

Spe

ctra

l Den

sity

0.01

0.1

1

10

Period (d)110100

Frequency (d-1)0.01 0.1 1

Spe

ctra

l Den

sity

0.01

0.1

1

(a) observed

(b) predicted

5.3 d 3.3 d

5.1 d3.3 d

Observed temperatures at LSC intake location

2012

Jun Jul Aug Sep Oct Nov Dec

Elev

atio

n (m

)

40

60

80

100

120

3 °C 7 11 16 20 24

20

Observations 2012

measurements were made every 30 sec at 5 m depth interval data shown are at every 6 hours at 0.5 m depth interval

thermistor string temperature data at the LSC intake site

Predicted temperatures at LSC intake location

2012

Jun Jul Aug Sep Oct Nov Dec

Elev

atio

n (m

)

40

60

80

100

120

3 °C 7 11 16 20 24

21

Model Performance – Predictions 2012

data shown are at every 6 hours at 1 m depth interval

temperature predictions at the LSC intake site; model validation

22

Model Performance – 1998-2006

• Temperature observations were made with a single thermistor at a frequency of 1 per hour deployed near piling cluster ~ 1 m deep

• extended validation


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(a) 1998RMSE = 0.87 °C

• RMSE is for April−October period


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(b) 1999RMSE = 1.84 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(c) 2000RMSE = 1.28 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(d) 2001RMSE = 1.62 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(a) 2002RMSE = 2.17 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(b) 2003RMSE = 2.10 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(c) 2004RMSE = 2.54 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(d) 2005RMSE = 2.14 °C


Tem

pera

ture

(°C

)

0

10

20

30

observedpredicted

(e) 2006RMSE = 1.72 °C

23

9/11 9/12 9/13 9/14 9/15 9/16

Win

d Sp

eed

(m s

-1)

-10

-5

0

5

10

N36W

S36E

Visualization of Hydrodynamic Features

First horizontal mode linear internal wave T1 = 80 hours T1/4 = 20 hours upwelling

southern end

24

Visualization of Hydrodynamic Features

• a 58-seconds movie (10 hours of simulation per second of animation); captures internal wave field

• simulation of hydrodynamics in Cayuga Lake - September, 2013 • depiction of mixing mechanisms*

• standing linear internal waves • progressive internal waves • internal wave breaking near shelf; localized intense turbulent

mixing • internal surges and hydraulic jumps (interaction of waves with

topographic features) • formation and propagation of nonlinear internal wave packets

• watch from the point of view of mixing, transport and redistribution of

plankton and nutrients; transport and resuspension of sediments

* details can be found in Fischer et al. (1979); Imberger (1985); Imboden and Wüest (1995); Boegman (2009)

25

southern end

southern end

Cayuga Lake: 8/31/2013−9/26/2013

Model Application

2D Graph 1

X Data

0 10000 20000 30000 40000 50000 60000

Y D

ata

-20

0

20

40

60

80

100

120

140

Col 1 vs Col 2

SRP

Mussels Excretion

Outflow Uptake

At steady-state:

• using the transport framework as an analytical tool • goal: evaluate implications of hypolimnetic mussel excretion for internal SRP

cycling and the consistency of measured SRP profiles • adopt an over simplified, epilimnion-only SRP model

26

conservative behavior for SRP in hypolimnion assumed

Example of a Complex

Water Quality Model

27

Zhang, Culver, Boegman (2008); Lake Erie

Estimation of Excretion Rate

28

• in-situ measurement difficult • measure biomass density and apply a literature value of mass-specific excretion rate

range: 0.084 − 0.66 µmolP/gDW/hr (n=10) (Nalepa and Schloesser 2014)

SRP Simulation Runs

29

Run ID Excretion Rate (mg/m2/d) Specification

Uptake Rate (1/m)

Run 11 estimated depth-specific 0.0


Run 1 50% of estimated depth-specific

0.1

Run 14 estimated depth-specific

0.5

effects of environmental variables on excretion not considered

SRP Simulation Results

• site 3 • selected depth-profiles of SRP; May-September 2013

30

0 25 50

Elev

atio

n (m

)

-150

153045607590

105

SRP (µg/L)

0 25 50 0 25 50 0 25 50 75

21 May, 2013 18 June, 2013 15 July, 2013 17 September, 2013


Uptake Rate (1/m)


SRP Simulation Results • site 3 • selected depth-profiles of SRP; May-September 2013

31

0 7 14

Elev

atio

n (m

)

-150

153045607590

105

Run 0

SRP (µg/L)

0 7 14 0 7 14 21


0 7 14


Uptake Rate (1/m)



32

0 7 14

Elev

atio

n (m

)

-150

153045607590

105

Run 0Run 1

SRP (µg/L)

0 7 14 0 7 14 0 7 14 21



Uptake Rate (1/m)


Run 1 50% of estimated depth-specific 0.1


33

0 7 14

Elev

atio

n (m

)

-150

153045607590

105

Run 0Run 1Run 14

SRP (µg/L)

0 7 14 0 7 14 0 7 14 21



Uptake Rate (1/m)


Run 1 50% of estimated depth-specific 0.1


SRP Simulation Results

• phytoplankton uptake of SRP in the epilimnion is important

• SRP concentration gradient across the thermocline allows diffusion and causes the characteristic shape of the SRP depth-profile as observed in the lake

• in the deep hypolimnion, SRP appears to be in quasi-steady state

• low levels of SRP in the metalimnion (z > 1% light level) suggest action of a sink process

34

Mussels: Some Thoughts, Issues

• uncertainty in biomass specific excretion rate (0.33 µmolP/gDW/hr)

• effects of environmental variables (temperature, turbidity, phytoplankton, pH, salinity, Ca), substrate, and currents on excretion rate – largely unknown

• longitudinal variations in excretion rate

• full P cycle will be necessary to explain SRP in the lake

• processes to consider: adsorption/desorption; zooplankton excretion; phytoplankton respiration; losses from settling of detritus; bio-deposits (pseudo feces)

35

algal P non algal P SRP/SUP

excretion of SRP

bio-deposits

Δ tissue • other features of metabolism will need to be considered

Summary

• A 2-D hydrodynamic/transport model, CE-QUAL-W2, has been setup and successfully tested for Cayuga Lake

• The model has been calibrated to data collected in 2013 and validated for 2012, 1998-2006; additional testing in future

• This model will be ready to serve as the transport submodel for a forthcoming (Phase 2) water quality model for the lake

• A preliminary application of the model demonstrates the effects, and importance of mussels excretion of soluble reactive phosphorus

36

Further Model Applications

• additional tracer analysis – mussel fluxes

– fate of interflows

– travel times for events

– south tributary transport • base flow

• runoff events

• material budgets

• kinetic insights

37

Documents

Development and Application of Cayuga Lake Model (CLM2D) · 2014. 5. 19. · &MPV 727$/ 4416 hrs. &$/ J,1'6 0.02% '$7 3(5,2' Start Date: 5/1/2013 - 00:00 End Date: 10/31/2013 - 23:00