Results From Comparative Ecosystem Studies within GLOBEC: Examples from SPACC, CCC and ESSAS

Results From Comparative Ecosystem Studies within GLOBEC: Examples from

SPACC, CCC and ESSAS

Ken DrinkwaterInstitute of Marine Research and Bjerknes Center

for Climate Research, Bergen, Norway

US GLOBEC MeetingBoulder, Colorado Feb. 17, 2009

Outline

•Why comparative studies?•Types of comparisons•Examples

Species –Cod, LobsterEcosystems – Upwelling, Subarctic Seas

•Problems•Concluding Remarks

Why Comparative Studies of Ecosystems?

3. Ecosystems are complex – can help determine what is a fundamental process and what is unique.

2. Provides insights that one cannot obtain by looking at a single ecosystem

4. Increases statistical degrees of freedom

5. Sharing approaches and methodologies

1. Cannot run controlled experiments on ecosystems

Types of Comparisons

3. Same species but different geographic (hydrographic) regions, e.g. Atlantic cod, American lobster.

1. Same ecosystem type but different geographic regions (upwelling regions, subarctic seas)

There are many types of comparative studies.

4. Same ecosystem (geographic area) but at different times (forcing), e.g. cold period vs. warm period.

2. Different ecosystems, e.g. tropics vs. polar ecosystems.

Atlantic Cod (Gadus morhua)

CCC (Cod and Climate Change)

Cod Distribution

Temperature plays a large role in determining the Growth Rate of cod

The relative size of a 4-year old as a function of mean bottom temperature.

Brander 1994

The same information shown graphically, i.e. the relative size of the age 4 cod at different temperatures.

Cod Recruitment and TemperatureCod Recruitment and Temperature

Mean Annual Bottom Temperature11

10

9

8

7

6

4

3

2

Temperature Anomaly

Warm Temperatures

decrease Recruitment

Warm Temperatures

increase Recruitment

Log2 Recruitment Anomaly

Faroes

Iceland

Newfoundland

North Sea

Irish Sea

-2 -1 0 1 2

-2 -1 0 1 2

-2 -1 0 1 2

-2 -1 0 1

-2 -1 0 1 2 -2 -1 0 1 2

-2 -1 0 1 2

-2 -1 0 1 2

Georges Bank

4

W. Greenland

4

2

0

-2

-

4

2

0

-2 -1 0 1 2

4

2

0

-2

Celtic Sea

-2

4

2

0

-2

4

2

0

-2

4

2

0

-2

4

2

0

-2

4

2

0

-2

4

2

0

-2

Barents Sea

American Lobster (Homarus americanus)

Lobster Landings Magdalen Island

In the late 1980s landings rose dramatically with suggestions that it was due to good management.

Again a dramatic rise in landings in the late 1980s-early 1990s and claims that management was working well.

US and Canadian Lobster LandingsUS and Canadian Lobster Landings

The US and Canadian landings for each of the lobster management regions showed similar trends (except for 4 out of 34) with different management strategies. Could not have been management strategies. It was not increased effort.

SPACC (Small Pelagics and Climate Change)

Kawasaki, 1983;Bakun, 1997

Bakun, 1989; 1997

Synchrony in Upwelling Areas

Lead to much research on ecological teleconnections.

RIS

0

150

300

450

0

350

700

1050

HumboldtA

nchovy x10 4 (tons)

Sar

dine

x 1

0 4 (to

ns) 0

20

40

60

80

0

10

20

30

40

California

0

150

300

450

0

15

30

45Japan

-2.5

0

2.5

1920 1940 1960 1980

Sardine

Anchovy

Also found tendency for anchovy and sardines to be out of phase (but not always or everywhere).

Lluch-Belda et al., 1989

ESSAS (Ecosystem Studies of Sub-Arctic Seas)

Coccolithophore Bloom at the eastern entrance to the Barents Sea

SSTs in the Labrador Sea

NORCAN (Norway-Canada Comparison of Marine Ecosystems)

NORCAN

• Workshop Funded by NRC and DFO• Held in Bergen in December 2005•Meeting in St. John’s May 2006 and writing groups met in Bergen 2007•Decided to write papers along discipline lines – physical oceanography, phytoplankton, zooplankton, fish (3) and marine mammals.•Drafts nearing completion and hope to submit mid-2009

-1.5

-1.0

-0.5

0.0

0.5

1.0

1975 1980 1985 1990 1995 2000 2005

YEAR

TE

MP

AN

OM

AL

Y (O

C)

-20

-15

-10

-5

0

5

10

15

20

25

CIL

AN

OM

AL

Y (

KM

2)

FUGLØYA-BJØRNØYA NEWFOUNDLAND CIL LABRADOR CIL

-1.5

-1.0

-0.5

0.0

0.5

1.0

1975 1980 1985 1990 1995 2000 2005

YEAR

TE

MP

AN

OM

AL

Y (O

C)

-20

-15

-10

-5

0

5

10

15

20

25

CIL

AN

OM

AL

Y (

KM2

)

VARDØ-N NEWFOUNDLAND CIL LABRADOR CIL

65 62 59 56 53 50 47 44 41

LONG ITUDE

42

44

46

48

50

52

54

56

58

LATITUDE

NEW FOUNDLAND

LABRADOR SEA

LABRADOR

STANDARD SECTIONS200 M500 M1000 M2000 M3000 M4000 M

DEPTH

STATION 27

FUGLØYA - BJØRNØYAVARDØ-N

KOLASEAL ISLAND

BONAVISTA

•Ocean Temperature Trends between two regions Out-Of-Phase prior to mid-1990s

•Similar Trends Recently

-1.0

-0.5

0.0

0.5

1.0

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

YEAR

TE

MP

AN

OM

AL

Y(°

C)

-25

-20

-15

-10

-5

0

5

10

15

20

25

CIL

AN

OM

AL

Y (

MB

)

KOLA SECTION TEMP TRENDS NEWFOUNDLAND CIL LABRADOR CIL

INDEX 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960

NAO -0.25 -0.09 -0.56 0.60 -0.40 0.98 1.25 -0.48 0.64 0.94 1.29

NUUK 0.80 0.51 1.09 0.43 0.57 1.32 0.18 0.98 1.48 0.72 1.60

IQUALUIT 0.36 0.16 0.78 0.18 0.17 1.92 -0.03 -0.06 1.02 -0.04 1.10

CARTWRIGHT 0.24 1.27 1.72 0.81 0.42 1.50 0.22 -0.05 1.48 0.62 1.50

ST JOHN'S -0.42 1.88 1.44 1.09 0.74 -0.26 0.04 -0.72 1.20 -1.01 0.87

NL SEA ICE                      

ICEBERGS 0.44 1.09 1.08 1.02 0.65 1.02 0.99 -0.25 1.11 0.10 0.73

S27 SURFACE T -0.15 1.13 1.16 0.37 -0.92 -1.73 -0.14 -1.57 0.52 -0.84 0.40

S27 BOTTOM T -1.11 1.01 1.06 1.05 0.58 0.75 0.77 -0.03 1.10 -0.60 1.19

S27 AGERAGED T 0.31 1.92 0.90 2.83 0.14 1.27 1.28 -1.10 2.47 0.59 0.52

HAMILTON BANK SURFACE T 0.50 1.65 0.74 -0.15 1.45 -1.67 -0.86 -1.16 0.12 -0.21 -0.06

HAMILTON BANK BOTTOM T -0.63 1.14 -0.87 0.34 -1.23 -0.09 -0.13 -0.73 0.47 -0.47 -0.10

FLEMISH CAP SURFACE T -1.27 1.06 1.42 0.13 0.19 0.08 -0.30 -0.62 0.50 -0.16 0.36

FLEMISH CAP BOTTOM T -0.44 0.60 0.32 0.33 0.92 1.22 0.11 -0.54 0.90 -0.45 0.69

SEAL ISLAND AVG T 1.13 1.13 -0.17 0.56 -0.86 0.39 -0.07 0.00 0.02 0.83 1.61

BONAVISTA AVG T -0.57 -0.05 -0.52 -0.07 0.09 0.09 0.71 0.82 1.06 0.40 0.73

FLEMISH CAP AVG T   1.41 2.01 1.24 0.10 0.03 0.32 0.03 1.29 0.45 1.14

ST. PIERRE BANK BT -0.46 0.64 1.84 -0.84 1.30 -1.16 -1.99 -0.54 1.29 -1.51 -0.51

SEAL ISLAND CIL 0.23 0.85 0.06 0.21 -0.68 0.95 -0.19 -0.68 0.28 -0.32 -0.20

BONAVISTA CIL -0.02 0.52 -0.39 0.74 0.05 -0.03 1.06 1.37 0.91 0.38 0.64

FLEMISH CAP CIL   0.90 0.94 1.86 -0.08 -0.28 0.58 -0.52 1.78 1.47 0.94

S27 SURFACE S 0.50 -0.25 -0.69 0.32 1.09 0.48 0.84 -0.33 -0.05 0.77 0.13

S27 AVERAGED S 0.64 0.74 -0.58 0.12 1.03 -0.54 1.63 -0.43 -1.08 -0.45 -0.14

SEAL ISLAND AVG S 1.05 1.05 -0.06 0.60 -0.12 0.92 0.60 0.66 -0.77 0.14 0.47

BONAVISTA AVG S 1.39 0.17 0.04 0.53 1.63 -0.08 1.63 2.24 0.65 -0.08 0.53

FLEMISH CAP AVG S   -0.83 0.15 0.34 1.02 0.44 1.61 1.71 -0.73 0.83 0.34

2.26 19.63 12.91 14.66 7.86 7.53 10.10 -1.98 17.66 2.09 15.77

Normalized indices (blue-cold, fresh; red-warm, saline) used to estimate overall index for both Labrador and Norwegian regions.

These standardized anomalies also show change from out of phase prior to the mid-1990s and in phase since then.

NORTH ATLANTIC WINTER SLP FIELDS

MEAN ANOMALY

1991

2000

2003

HISTORICAL PATTERN- COLD IN WEST WARM IN EAST

EASTWARD DISPLACEMENT

WESTWARD DISPLACEMENT

(Marine Ecosystem Comparisons of Norway and the United States)

MENU

NOAA Fisheries

MENU• Workshop Funded by NRC• Held outside Bergen in March 2007• Brought data to the table•Divided into 2 groups: (1) response to recent changes and (2) structure and function of ecosystem. •Five papers have been accepted for publication in PiO.

Eastern Bering Sea (EBS)

Gulf of Alaska (GOA)

(NOR/BAR)

Alaska

Russia

Canada

USA

Greenland

GB

Norwegian Sea

Barents Sea

GOM0

250,000

500,000

750,000

1,000,000

1,250,000

1,500,000

EBS GOA GOM GB NOR BAR

are

a (

sq

km

)

35

45

55

65

75

85

EBS GOA GOM GB NOR BAR

lati

tud

e (

de

gre

es

N)

Area

LatitudeGulf of Maine / Georges Bank (GOM/GB)

Menu Regions

Highly Advective Systems

Strong Tidal Currents and Mixing in subregions

Mean Mean CirculationCirculation

Monthly meanMonthly meansea-surfacesea-surface

temperaturetemperatureanomaliesanomalies1900-20061900-2006

Correlations between annual heat fluxes and SST temperatures

Pacific Ecosystems: Significant correlations, 25-30% of SST variance accounted for.

Atlantic Ecosystems: Weak and non-significant correlations.

Suggests that warming due to advection in the Atlantic while in Pacific air-sea fluxes play a significant role.

Gulf of Alaska

Gulf of Maine

Barents Sea

In GoA surface freshening due to local runoff

In GoM freshening due to advection from the North (Arctic?)

In Barents Sea increasing salinity due to higher salinity in Atlantic Water.

Salinity

0

100

200

300

400

500

600

Bering GOA GOM / GB Norwegian Barents

Prim

ary

prod

uctio

n (g

C m

-2y-1

Total annual net PP 1998-2006 average (± 2 SD)

SeaWiFS climatology – Chl. a (Apr-Jun)

Bering Sea / Gulf of Alaska Norwegian Sea / Barents Sea

Gulf of Maine /Georges Bank

Source: http://oceancolor.gsfc.nasa.gov/cgi/level3.pl Mueter et al., in press

EBS GOA GOM/GB NOR BAR

10

20

30

40

Nitr

ate

co

nce

ntr

atio

n (

µM

)

Productivity increases with nitrate content of deep source waters

Approximate rangeNitrate in source waters and

total annual primary production

GOM/GB

Mueter et al., in press

Effect of SST on primary production, 1998-2006

2 4 6 8 10

200

300

400

500

Annual mean SST (°C)

Total annual net

primary production(gC m-2)

Barents Sea(P = 0.093)

Norwegian Sea(n.s.)

Bering Sea(P = 0.039)

Gulf of Maine/Georges Bank(P < 0.001)

Gulf of Alaska(n.s.)

Mueter et al., in press

MENUII

NOAA Fisheries

MENUII• With the success of MENU, NOAA and

IMR administrators encouraged MENU participants to submit full proposals

• Decided that emphasis would be model comparisons and ecosystem indicators

• 4 types of models: ECOPATH, production models, biophysical models (3-D hydrodynamic models up to zooplankton) and system models (includes fish and fisheries (ATLANTIS)

Ecosystems are created in Atlantis three-dimensionally, using linked polygons that represent major geographical features. Information is added on local oceanography, chemistry and biology such as currents, nutrients, plankton, invertebrates and fish.

The model then simulates ecological processes such as:consumption and production, waste production, migration,Predation, habitat dependency, mortality.

The Atlantis framework used for management strategy evaluation incorporates a range of sub-models for each major step in the management cycle. They simulate the marine environment, the behaviour of industry, fishery monitoring and assessment processes, and management actions and implementation.

ATLANTIS

MENUII• Same model different regions• Different models for same region• Determine what we learned from each

of the models

A good forecaster (modeller) is not smarter than everyone else, he merely has his ignorance better organised. -Anonymous

MENUII•Norwegian component funded by RCN (2009-2011)

•US component submitted to CAMEO but not funded in first round. Hoping to obtain funds to carry out work from other sources.

Prediction is difficult, especially if it involves the future.

Nils Bohr

Prediction is easy, getting it right is the difficult part!

Prediction is easy, getting it right is the difficult part!

What does ”right” imply?

-Some quantifiable measure of how well the model fits the observations

-For future projections where we won’t have observations need some quantifiable measure of the uncertainty.

Observationalists and Modellers need to work closer together

- Modeller’s to help determine what, where and how often observationalists should measure.

- Observationalists should provide more feedback on model results (requires available model results, positive criticisms)

- All motherhood statements but not generally done

Some Problems

• For data comparisons, the data or datasets should be similar. Not always possible.•Forcing is based on large model and data based datasets (e.g. NCEP) that usually have some problems • When using different models there is the difficulty of knowing if one is comparing ecosystems or models

Concluding Remarks

• Comparative studies are a useful way to gain insights into marine ecosystems• They often lead to shifts in our thinking about what is important and what is not.• Bring comparative datasets to the table•For models need to develop new and better measures of uncertainty• Need to make sure that observationalists and modelers do not work independently.

Thank You!

SST Anomalies in the North Atlantic during 1990-1994

HISTORICAL PATTERN- COLD IN

WEST WARM IN EAST

SST Anomalies in the North Atlantic

during 2004

BROAD-SCALE WARMING

NOAA Optimum Interpolation SST, NOAA-CIRES Climate Diagnostics Center

Maximum SST 1997-2006

Sea-Ice Cover Sea-Ice Cover AnomaliesAnomalies

Bering Sea

Barents Sea

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

-0.4

0

0.4

-1

-0.5

0

0.5

1

1.5

Zooplankton anomalies: Evidence of top-down and bottom-up control

Barents Sea

Norwegian Sea

Gulf of Maine /Georges Bank

Bering Sea

Nor

mal

ized

ano

mal

y(B

iovo

lum

e)N

orm

aliz

ed a

nom

aly

(Bio

mas

s )

Napp & Shiga (unpublished)

Based on Valdés et al. (2006)

r = 0.60P = 0.002

Mueter et al., in press

-1.0 -0.5 0.0 0.5

-1.5

0.0

1.5

-1.0 0.0 1.0

-20

1

Fish: SST & cod recruitment

Barents Sea(Atlantic cod)

Georges Bank (Atlantic cod)

Bering Sea(Pacific cod)

Gulf of Alaska(Pacific cod)

SST anomaly

Log(

Recr

uitm

ent)

ano

mal

y 1977-2005 only!

Georges Bank and Barents Sea figures from:Planque & Frédou (1999)

-1.0 0.0 1.0

-20

1

-1.0 -0.5 0.0 0.5

-1.5

0.0

1.5

Mueter et al., in press