the future? the future? the future? the future?

Assessing the potential impacts of

global warming

From the Seattle Post-Intelligencer, October 20, 2005

Courtesy of Nate Mantua

Climate Change: Assessing its

Implications for Marine

Ecosystems

Three Types of ApproachesModel Evaluation and SelectionA Pair of Examples from the Bering Sea

Potential Approaches

• Empirical downscaling: Ecosystem indicators for stock projection models are projected from IPCC global climate model simulations.

• Dynamical downscaling: IPCC simulations form the boundary conditions for regional bio-physical numerical models with higher trophic level feedbacks.

• Fully coupled bio-physical models that operate at time and space scales relevant to regional domains.

Comparing empirical versus dynamical models for ecosystems projections

• What are the strong points of each technique?

• What are the pitfalls of each technique?

• What are the likely sources of discrepancies between projections made by these two techniques?

Framework for projecting impacts of climate change on marine ecosystems through

empirical downscaling

• Identify mechanisms underlying productivity • Evaluate individual climate model hindcasts; select

models that appear to be valid for a specific region and purpose

• Compile climate model projections and develop time series of environmental indices

• Incorporate environmental time series in forecasting models for fish

• Evaluate harvest strategies under a changing ecosystem

IPCC I.D. Country Atmosphere Resolution

Ocean Resolution

# of Control runs

# of 20c3m runs

# of A1B runs

1 BCCR-BCM2.0 Norway T63L31 (0.5 -1.5°) x 1.5°L35 2 1 1

2 CCSM3 USA T85L26 (0.3 -1.0°) x 1.0°L40 1 1

3 CGCM3.1(T47) Canada T47 L3 1 1.9° x 1.9°L29 2 5 5

4 CGCM3.1 (T 63 ) Canada T63L31 1.4 ° x 0.9 °L29 1 1 1

5 CNRM-CM3 France T42L45 182x152L31 3 1* 1

6 CSIRO-Mk3.0 Australia T63L18 1.875° x 0.925° L31 3 3 1

7 ECHAM5/ MPI -OM Germany T63L31 1.5°x1.5°L40

8 FGOALS -g1 .0 (IAP) China T42L26 1°x1°xL30 9 3 3

9 GFDL-CM2.0 USA 2.5°x2. 0° L24 1°x1°L50 5 3 1

10 GFDL-CM2.1 USA 2.5°x2. 0° L24 1°x1°L50 5 5 1

11 GISS -AOM USA T42L20 1.4°x1.4°L43 2 2 2

12 GISS -EH USA 5°x4°L20 2°x2° *cos(lat) L16 4 5 3

13 GIS S-ER USA 5°x4°L13 5°x4°L33 1 9 5

14 INM-CM3.0 Russia 5°x5°L21 2°x2.5°L33 2 1 1

15 IPSL-CM4 France 3.75°x2.5° L19 2°x1 °L31 3 1 1

16 MIROC3.2(hires) Japan T106 L56 0.28°x0.188° L47 1 1 1

17 MIROC3.2(medres) Japan T42 L20 (0.5° -1.4°)x 1.4° L44 3 3 3

18 ECHO-G (MIUB) Germany/Korea T30L19 T42L20 1 3 3

19 MRI-CGCM2.3.2 Japan T42 L30 (0.5° -2. 5°) x 2° L23 3 5 5

20 PCM USA T42L18 (0.5 -0.7°) x 0.7° L32 1

21 UKMO -HadCM3 UK 3.7°5x2.5° L15 1.25°x1.25° L20 2+1* 1 1

22 UKMO -HadGem1 UK 1.25°x1.875°L38 (0.33 -1.0°) x 1.0° L40 1+2* 2 1*

Sum 55 40

Models Contributed to IPCC AR4

Predation

SpawningEarly larvae

(spring)Late larvae

(fall) Age-1 recruits

Spatialdistribution

BiomassConsumption ratePrey composition

Spring conditions (Late) summer conditions

Prey

Timing of ice retreat

SpringSST

Prey

Summer SST

Wind mixingStability

Estimated effects of summer SST & predation on log-recruitment

8.0 8.5 9.0 9.5 10.0 10.5

89

1011

Summer SST, age-0

1000 1500 2000 2500

8.5

9.0

9.5

10.0

11.0

Predation, age-0

log-

recr

uitm

ent

R2 =0.44P = 0.001

Prediction interval

Simulate effect of increase in average SST on recruitment at three levels of predation

Low Med High

Bering Sea SST (JAS) - A1B Scenario

6

7

8

9

10

11

12

13

14

2000 2010 2020 2030 2040 2050

Mean SST (JAS)

Pollock Recruitment (A1B Scenario)

0

2

4

6

8

10

12

2000 2010 2020 2030 2040 2050

log(R)

Parameter Rationale Reliability Large-scale mean pressure/wind patterns

Upper ocean advection; Surface forcing for ROMS

Very Good

Large-scale upper ocean T/S and currents

Direct estimates; Lateral BCs for ROMS

Good

Sea ice (Winter/Spring) Cold pool extent; Nature of spring bloom

Good

Spring bloom timing LTL Community; Pollock recruitment?

Fair/Poor

Summer SST Stratification; Mixed layer depth

Fair

Summer wind mixing Stratification; Nutrient re-supply

Good

Quasi-quantitative Assessment of Global Climate Model Capabilities

Summer Cold Pool Extent

Parametric Term p-value----------------------------------------SLP (Winter) 0.0235SLP (Spring) 0.0052Ice Cover 0.0002 Ice Retreat 0.0756Wind Mixing (JJ) 0.6294

Total variance explained ~ 76%

(Using just Spring SLP & Ice Cover,total variance explained ~ 69%)

Cold Pool vs. SLP

Cold Pool vs. Ice

Cold Pool Projections

0

20

40

60

80

2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050

Extent of Survey Area (%)

CCSM3_1

CCSM3_2

CCSM3_3

CCSM3_4

CCSM3_5

CCSM3_6

CCSM3_7

CNRM_1

ECHOG_1

ECHOG_2

ECHOG_3

MIROCM_1

MIROCM_2

MIROCM_3

UKHAD_1

MEAN

Dynamical Modeling

Physical Forcing(Wind, temp, sun)

NutrientsNO3, NH4…

Primary Producers(Phytoplankton)

Secondary Producers(Zooplankton)

Higher trophic levels(Pollock etc.)

Model Grid for NPZ and FEAST

3D grid has 10km resolutionGrid size = 180*256*60Grid size = 180*256*60Grid size = 180*256*60Grid size = 180*256*60

DATA MODEL

T at M2

EUPHAUSIIDS

LARGE COPEPODS

MICROZOOPLANKTON

SMALL

PHYTOPLANKTON

LARGE

PHYTOPLANKTON

NITRATE AMMONIUM

Slow sinking

DETRITUS

BENTHIC

FAUNA

BENTHIC DETRITUS

IRON

ICE ALGAE

NITRATE AMMONIUM

SMALL COPEPODS

Excretion +

Respiration

ICE

OCEAN

BENTHOS

BESTBESTBESTBEST----NPZNPZNPZNPZ

modelmodelmodelmodel

Mortality

Predation

Egestion

Molting

JELLYFISH

Inexplicit

food source

FEAST

Fast sinking

DETRITUS

1999 2004

Zooplankon Biomass

Day 220 1999 2004

Microzooplankton 5.9 16

Small Copepods 2.8 4.5

Large Copepods 8.9 1.29

Euphausiids 0.57 3.9E-5

Compares ‘reasonably’ well to Coyle data … – but will the fish have enough to eat ?

Zooplankton group Model Component 1999 2004 Reference

Microzooplankton Small. and large microzooplankton

27.8 mg_m -3 <20 mg_m -3 Strom and Fredrickson

2007 Olson and Strom, 2002

Pseudocalanus sp. Small copepods 1.8 mg_m-3 4 .63 mg_m-3 Coyle et. al., 2007 Calanus marshallae Large copepods 57mg_m-3 0.01 mg_m-3 Coyle et. al., 2007 Thysanoessa spp. Euphausiids 0.7 mg_m-3 0. 047 mg_m-3 Coyle et al., 2007

C. Melanaster Jellyfish 348 mg_m-3 7. mg_m-3 Coyle et al ., 2007

Brodeur et al ., 2002 Phytoplankton Phytoplankton 2 ug Chl -a_l-1 1.2 ug Chl -a_l -1 Whitledge F OCI data

FEAST model for forage species and predators

• Bioenergetics of feeding, growth, spawning• Focus on data-driven functional response between

predator and prey• Use allometric relationships for rates• Diet preferences based on stomach data• Movement (towards prey concentrations, away from poor

conditions, migration for spawning)• Currently includes pollock, cod, and arrowtooth flounder

FEASTwww.bsierp.nprb.org

Infauna

Epifauna

Other fish

Small/Large Copepods

NPZ

FEASTMarine mammals

Marine mammals

Marine mammals Marine mammals

Economic and spatial fishery predictions

Fish growth over timeF

ish

size

cla

ss

Time

1999 2004

Final Remarks• From present to mid-21st century, climate change

liable to be dominated by thermodynamic effects as opposed to dynamic effects (e.g., winds).

• Open questions: (1) Are the ocean components of global climate models sufficient for climate/ecosystem studies? (2) What is the best way to use existing climate model simulations for regional applications?

• The output from global climate models (perhaps subject to statistical downscaling) can complement that from vertically-integrated numerical models with full dynamics.

Comparing empirical versus dynamical models for ecosystems projections

• What are the strong points of each technique?

• What are the pitfalls of each technique?

• What are the likely sources of discrepancies between projections made by these two techniques?

iPdiPtnPlPsPtP

KtnrToptntrrespNCktbmCrespCTrefJ

RgR0i

ksowthinhibbetaI

alphaIQ10JelTt

Q10JelrfpEupJelfpNCJel

mpredJelgammaJel

eJelprefPSprefPL

eexDeex

rmortRresQres

RupPKupPkupDRup

Q1Bremin

wNCrisewDetwPLpPS

mpredEmpredC

mMZmaxmPLmaxmPSgammaE

Q10ETQ10EeEup

fpMZEupgammaNC

fNCeNC

gammaCQ10C

eCgammaMZ

mMZeMZ

ktbmMZTmaxMZ

respMZktbmPLTmaxPLktbmPSTmaxPS

respPLrespPS

k2PLpsiPL

DiLalphaPL

k_chlKext

ccrPLXi

PARfrac

NO

3_m

ean1

22N

O3_

surf

122

NO

3_m

eanT

may

N

O3_

mea

nBm

ay

NH

4_m

eanT

may

N

H4_

mea

nBm

ayP

hS_m

eanT

PhL

_mea

nT

MZ

_mea

nT

NC

a_m

eanT

C

al_m

eanT

E

up_m

eanT

Je

l_m

eanT

N

O3_

mea

nT

NO

3_m

eanB

N

H4_

mea

nT

NH

4_m

eanB

N

O3_

mea

nAll

PhS

_mea

nAll

PhL

_mea

nAll

NH

4_m

eanA

llM

ZS

_mea

nAll

MZ

L_m

eanA

llN

Ca_

mea

nAll

Cal

_mea

nAll

Eup

_mea

nAll

Det

_mea

nAll

PLi

_mea

nAll

NO

i_m

eanA

llN

Hi_

mea

nAll

Jel_

mea

nAll

Fe_

mea

nAll

Parameter Ranking

Model Diagnostics

Par

amet

ers

± 60%