Greg Ruiz, Kate Murphy, Jen Boehme, Ari Balsom, George Smith, Chris Brown and Monaca Noble

Smithsonian Environmental Research Center

Ballast Water Exchange Verification: Discrimination Using Chemical Characteristics of Oceanic Water

Identify characteristics that reliably discriminate between ballast water of coastal vs. oceanic origin.

Inform the development of an in-situ (field) instruments for BWE verification.

Aims

Tracer Qualities:

• Unambiguous discrimination between coastal and oceanic water

• Robust to potential contaminants from ships and tanks

• Easily and reliably measured in ships

• Stable over time in ballast tanks

• Consistent results for varying source and exchange regions

Potential tracers of mid-ocean exchange

Biological tracers (plankton)BUT –difficult to identifydifficult to quantifypopulations vary greatly in space and time

Chemical tracers easier to quantifydistribution easier to predict, both

* within ballast tanks* globally

0%

50%

100%

0 100 200

miles offshore

[ Tra

cer ]

HypothesisNot

exchanged(non-complying)

exchanged(complying)

The chemical signature of exchanged ballast

water is more similar to water from the mid-

ocean than water from its original coastal

source.

80° W 60° W 40° W 20° W 0°

0°

20° N

40° N

60° N

Fos

BN

Cruise Tracks: N. Pacific & N. Atlantic

150oE 175oE 160oW 135oW 110oW 10oN

20oN

30oN

40oN

50oN

60oN

1. CONTROL = not exchanged

2. TREATMENT = exchanged

Experimental Design

cargo cargo cargo cargo5 4 3 2 1

5 4 3 2 1

3. SHIPSIDE = Ocean Reference Set

• Salinity

• Radium (223Ra, 226Ra, 228Ra)

• Trace elements (Ag, Ba, Mn, Mo, P, U, V)

• CDOM fluorescence signatures (chromophoric dissolved organic matter)

Objective

Identify tracers that, alone or in combination, reliably discriminate between ballast water of coastal vs. oceanic origin.

Tracers:

Results: Salinity

MV Lily Fortune

initial final

Sal

inity

[ppt

]

0

5

10

15

20

25

30

35

Unexchanged (Control) Exchanged

MV Asahi Sunrise

initial final

Shipside

} +15%

Tank Pair 1

initial final

223 R

a [d

pm l-1

]

0

1

2

3

4

5

6

Control Treatment

Tank Pair 3

initial final

Tank Pair 4

initial final

Results: 223Ra(Data from “MV Asahi Sunrise”)

|| + +| ..................+.......O| . O . +| . + O O.| . O O .+| . .| . + +.| . O O O .| . + O + .| . O O .| . O + .| . + O .| . O + .| . O .| +. O + .| .O O .| . + .| + + .| O..+....................... | + +||-----------------------------------------

tracer 1

tracer 2

Compare with ballast water samplesDm

2 = (xi - x)' S-1(xi - x)(Mahalanobis Distance Statistic)

BDm

Define an “ocean reference set”x = tracer mean in open oceanS = tracer covariance matrix

Multivariate Model: Trace Metals

0

50

100

150

T0T1

T2

T3 Sal+ Ba

+ Mn+ P

+ Mo+ V

Dm

TimeTracers

Unexchanged Ballast Water

0

50

100

150

Sal+ Ba

+ Mn+ P

+ Mo

+ VT0 (0 %)

T1 (55 %)T2 (83 %)

T3 (93 %)

Dm

Time,% Exchange Tracers

Exchanged Ballast Water

Percent Exchange0 55 83 93

Dm

Number of Tracers1 2 3 4 5 6

Dm

Results: Trace Metals

Results: CDOM(Data from “MV Asahi Sunrise”)

Fluo

resc

ence

[QSE

]

3

6

9

12

15

18

250300350400450

300350

400 450 500550

600650

3

6

9

12

15

18

300 350400 450

500550

600 650

Emission wavelength [nm]

3 6 9 12 15 18

Excitation wavelength [nm]

A. Before BWE B. After BWE

CDOM Analysis BWE Experiments on trans-oceanic cruises on commercial ships (N. Pacific & N. Atlantic).

2 treatments (Control, Exchanged) per cruise3 – 5 ballast tanks per treatment3 – 5 time-points per ballast tank2 – 4 replicates per depth

Ocean dataset: 12-30 / cruise

> 690 samples in database

Samples collected in amber glass DOM-clean bottles and analyzed at one of 3 analytical labs (i.e. 3 different instruments).

Excitation-Emission Fluorescence Spectroscopy used to generate ‘EEMs’.

Data Analysis via Parallel factor analysis (PARAFAC)

300 400 500 6000

0.2

0.4

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

0.3

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

0.3

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.2

0.4

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

0.3

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.2

0.4

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.2

0.4

0.6

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

1 2 43 5

6 87 9

PARAFAC – 9 component model

240 260 280 300 320 340 360 380 400 420 440300

350

400

450

500

550

600

1

2

3

4

5

6

7

240 260 280 300 320 340 360 380 400 420 440300

350

400

450

500

550

600

-1.5

-1

-0.5

0

0.5

1

1.5

240 260 280 300 320 340 360 380 400 420 440300

350

400

450

500

550

600

1

2

3

4

5

6

7

Raw Data Model

+

Residuals

=

PARAFAC – Parallel Factor Analysis

9 Components

0

2

4

6

8

1 2 3 4 5 6 7 8 9

300 400 500 6000

0.2

0.4

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

0.3

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

0.3

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.2

0.4

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.1

0.2

0.3

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.2

0.4

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

300 400 500 6000

0.2

0.4

0.6

Wavelength(nm)

Load

ing

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

1 2 43 5

6 87 9

PARAFAC – 9 component model

Effect of Ballast Water Exchange

salinity: 36 33 30 27

Ocean water (>100 miles offshore)

Fos LF BN AS0

2

4

6

8

salinity: 36 33 36 33

• BWE greatly reduces concentrations of fluorescent components

• Exchanged CDOM > Ocean CDOM

• Signatures from all high salinity sources are similar

Exchanged Ballast Water

Fos LF BN ASMax

imum

Inte

nsity

(QS

E +

/- S

E)

0

2

4

6

8

salinity: 36 33 36 33

Unexchanged Ballast Water

Fos LF BN ASMax

imum

Inte

nsity

(QS

E +

/- S

E)

0

2

4

6

8 C2 C3 C4 C8

Is fluorescence predictable in the ocean?

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Ocean Fluorescence: 99% Upper Confidence Limits

LF AS BN Fos all

Max

imum

Inte

nsity

(QS

E +

/- S

E)

0.0

0.5

1.0

1.5

2.0

2.5

C2 C3 C4 C8

C2

C3

C4

C8

Exchanged: 99% Upper Confidence Limits

LF AS BN SF PS Fos

Max

imum

Inte

nsity

(QS

E +

/- S

E)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

C2 C3

Is fluorescence predictable following BWE?Unexchanged: 99% Lower Confidence Limits

LF AS BN SF PS Fos0

2

4

6

8

10

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

C2 C3

Design parameters for handheld fluorometers

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Ex. (nm)

Em

. (nm

)

250 300 350 400 450300

400

500

600

Component 2 Component 3

Future Directions

What are the best tracer sets?• Maximize discrimination & minimize noise• Robust to many sources of variation:

(e.g. ships, source regions, ocean, residence time..)

How robust is approach to different global regions and ports?• Collaborative expansion project with Singapore & New Zealand

How does approach perform for Overseas and Coastwise Traffic to the U.S. Pacific coast, as a Model System?

• Distance from shore (decay function) – seasonal measures• Within bay (port) variation – seasonal / spatial measures

Pacific Basin as a Model System

90° E 135° E 180° E 135° W 90° W

45° S

0°

45° N

Pacific Ocean

Port SurveyCruise80nmi transect

Acknowledgements

Host Shipping Companies:•NYK Bulkship (USA) LTD., Gateway Maritime Corp. / Sincere Industrial Corp., Matson Navigation Company, Bergesen DY ASA., Sea River Maritime, the Alaska Tanker Company, BP Amoco PLC and Krupp Seeschiffahrt GmbH

CDOM Analysis:• University of S. Florida (P. Coble, R. Conmy)• University of Maine (J. Boehme)• Denmark National Environmental Research Institute (W. Martinsen)

PARAFAC:• Denmark National Environmental Research Institute (C. Stedmon)• Denmark Royal Veterinary and Agricultural University (R. Bro)

Sponsors:

• California State Lands Commission• National Sea Grant Program• New Zealand Government (MAF)• U.S. Coast Guard

KAS

LF

140° E 160° E 180° E 160° W 140° W 120° W

20° N

40° N

60° N

80° W 60° W 40° W 20° W 0°

0°

20° N

40° N

60° N

Fos

BN

Cruise Tracks: N. Pacific & N. Atlantic