Climate to Fish to Fishers:Climate to Fish to Fishers:Coupled Model DevelopmentCoupled Model Development

Kate Hedstrom, ARSC/UAFKate Hedstrom, ARSC/UAFSeptember, 2010September, 2010

I am not alone!I am not alone!

• Enrique Curchitser, RutgersEnrique Curchitser, Rutgers• Kenny Rose, LSUKenny Rose, LSU• Jerome Fiechter, UCSCJerome Fiechter, UCSC• Alan Haynie, NOAA-AFSCAlan Haynie, NOAA-AFSC• Jon Wolfe, Brian Kaufmann, NCARJon Wolfe, Brian Kaufmann, NCAR• Seth Danielson, UAFSeth Danielson, UAF• NSF funding via CAMEO, Multi-scale NSF funding via CAMEO, Multi-scale

modelingmodeling• ARSCARSC

OutlineOutline

• Overall goalsOverall goals• Coupled modelsCoupled models

– ROMSROMS– NEMURONEMURO– ClimateClimate– FishFish– Fishing fleetFishing fleet

• How are we doing?How are we doing?

Early 1970’s

Mid1970’s

1980’s

Changes in species compositionin small mesh bottom trawls inPavlof Bay. Rev. Aquat. Sci. (1992)

Pacific Decadal OscillationPacific Decadal Oscillation

“warm phase” “cool phase”

From Nate Mantua

Regime shift and PDORegime shift and PDO

From Nate Mantua

The state of the fisheriesThe state of the fisheries

Another way of looking at itAnother way of looking at it

It is not a jokeIt is not a joke

Changes in Fish CompositionChanges in Fish Composition

• When they happen, is it overfishing or When they happen, is it overfishing or is it climate change?is it climate change?

• Goal is to model the complete range of Goal is to model the complete range of processes from climate to fisheries to processes from climate to fisheries to get both top-down and bottom-up get both top-down and bottom-up effectseffects

• Can we do it?Can we do it?

BEST-BEST-BSIERPBSIERP

End-to-End ModelsEnd-to-End Models

• They are coming! BEST-BSIERP is one They are coming! BEST-BSIERP is one example of how to do itexample of how to do it

• They are primarily built from existing They are primarily built from existing stand-alone componentsstand-alone components

• NPZ and fish models meet at NPZ and fish models meet at zooplanktonzooplankton– Closure for getting phytoplankton rightClosure for getting phytoplankton right– Fish foodFish food

Some ProblemsSome Problems

• Zooplankton biomass vs. stagesZooplankton biomass vs. stages• Functional groupingsFunctional groupings

– Diet shifts in fishDiet shifts in fish– Prey selection by zooplanktonPrey selection by zooplankton

• New organismsNew organisms– JellyfishJellyfish– PeoplePeople

• ValidationValidation

A Simpler Problem?A Simpler Problem?

• Sardine – Anchovy Sardine – Anchovy cyclescycles– Well-studied species with Well-studied species with

population cycles observed population cycles observed in many systemsin many systems

– Teleconnections across Teleconnections across basinsbasins

• Good case studyGood case study– Forage fish tightly coupled Forage fish tightly coupled

to NPZto NPZ– Important ecologically and Important ecologically and

widely distributedwidely distributed– Low frequency variabilityLow frequency variability Provided by: Salvador E. Lluch-Cota

Source: Schwartzlose et al., 1999

Source: MacCall, 1990

Californian Anchovy Larval Abundance

April 1952

Low Anchovy Abundance

April 1965

High Anchovy Abundance

Our Model ComponentsOur Model Components

• Community Earth System Model Community Earth System Model (CESM)(CESM)

• Regional Ocean Modeling System Regional Ocean Modeling System (ROMS) with NEMURO NPZ(ROMS) with NEMURO NPZ

• Superindividual model of sardines Superindividual model of sardines and anchovies (Rose)and anchovies (Rose)

• California and Mexican sardine California and Mexican sardine fishery (Haynie)fishery (Haynie)

Regional Ocean Modeling Regional Ocean Modeling System (ROMS)System (ROMS)

• Community coastal ocean modelCommunity coastal ocean model– Over 750 svn downloadsOver 750 svn downloads– In active development with:In active development with:

• online discussion forumonline discussion forum• documentation wikidocumentation wiki• bug tracking systembug tracking system

– Annual meetings, outside the US in Annual meetings, outside the US in alternate yearsalternate years

More ROMSMore ROMS• Finite volumeFinite volume• HydrostaticHydrostatic• Terrain-following Terrain-following

vertical coordinatevertical coordinate• Structured Structured

horizontal gridshorizontal grids• MPI parallelMPI parallel• Several NPZ Several NPZ

optionsoptions

ROMS DevelopersROMS Developers

Some Regional DomainsSome Regional Domains

Eastern Eastern Bering SeaBering Sea

• Goal is to model:Goal is to model:– Ocean physicsOcean physics– Phytoplankton bloomPhytoplankton bloom

• Spring, fall, ice algaeSpring, fall, ice algae

– ZooplanktonZooplankton– FishFish– Fishing fleetFishing fleet

• Timing of spring Timing of spring bloom depends on bloom depends on sea ice meltsea ice melt

Model ValidationModel Validation

From SethDanielson

Model ValidationModel Validation

From SethDanielson

Summary of Bering SeaSummary of Bering Sea

• More model comparisons to come More model comparisons to come (Seth’s thesis)(Seth’s thesis)

• Hindcast first, future scenarios to Hindcast first, future scenarios to comecome

• Proof of concept that model will run Proof of concept that model will run with forcing and boundary conditions with forcing and boundary conditions from global climate modelfrom global climate model– One-way nestingOne-way nesting

ROMS-CESM CouplingROMS-CESM Coupling

• With:With:– Enrique Curchitser (Rutgers)Enrique Curchitser (Rutgers)– Bill Large (NCAR)Bill Large (NCAR)– Jim Hurrell (NCAR)Jim Hurrell (NCAR)– Jon Wolfe (was at NCAR)Jon Wolfe (was at NCAR)– Brian Kaufmann (NCAR)Brian Kaufmann (NCAR)– Jerome Fiechter (UCSC)Jerome Fiechter (UCSC)– Fei Chai (UMaine)Fei Chai (UMaine)– Justin Small (NCAR)Justin Small (NCAR)

The multi-scale problemThe multi-scale problem

Dickey, 1991, 2003

Downscalin

gUpscalin

g

SST, August 18, 2000SST, August 18, 2000

Background and motivationBackground and motivation

WCRP strategic framework

•Improved predictions of changes in statistics of regional climate, especially extreme events, are required to assess impacts and adaptation

•Recognize the need to improve representation of weather and climate link

•Working hypothesis is that the internal dynamics of the system are more accurately represented at higher resolution

Nested Regional Climate Models (NRCM) Nested Regional Climate Models (NRCM) at NCARat NCAR

Seamlessly integrate weather (WRF), high resolution ocean (ROMS) and climate models (CESM) to:

•Better capture and investigate important space/time scale interactions

•Develop approaches for reducing biases

•Inform the development of next-generation Earth System Models

•Apply the model to challenging science and important societal questions

•Assist decision-makers to plan for regional changes

Re-designing Re-designing the CESMthe CESM

POP

CESM StrategyCESM Strategy

• The coupling interval is one day, with the The coupling interval is one day, with the atmosphere running aheadatmosphere running ahead

• The composite ocean receives hourly The composite ocean receives hourly winds, etc. and interpolates to ROMS winds, etc. and interpolates to ROMS grid(s)grid(s)

• POP runs for that day, sends “curtains” out POP runs for that day, sends “curtains” out for ROMS nestsfor ROMS nests

• ROMS runs for that dayROMS runs for that day• Composite ocean merges SST from both Composite ocean merges SST from both

ocean models for CAMocean models for CAM

The new global SST (NCAR-The new global SST (NCAR-CESM)CESM)

A closer look at the down-scaled region A closer look at the down-scaled region (temperature and wind anomalies)(temperature and wind anomalies)

Summary of NRCMSummary of NRCM• Mechanics of coupling Mechanics of coupling

is workingis working– Three-year old code baseThree-year old code base– Currently updating to new Currently updating to new

codescodes

• Funded project to study Funded project to study air-sea carbon fluxes in air-sea carbon fluxes in three upwelling three upwelling systemssystems

Fish (and Fishers) ProjectFish (and Fishers) Project• With:With:

– Jerome Fiechter (UCSC)Jerome Fiechter (UCSC)– Kenny Rose (LSU)Kenny Rose (LSU)– Enrique Curchitser (Rutgers)Enrique Curchitser (Rutgers)– Bern Megrey (NOAA-AFSC)Bern Megrey (NOAA-AFSC)– Alan Haynie (NOAA-AFSC)Alan Haynie (NOAA-AFSC)– Miguel Bernal (IEO, Spain)Miguel Bernal (IEO, Spain)– Salvador Lluch-Cota (CIBNOR, Mexico)Salvador Lluch-Cota (CIBNOR, Mexico)– Others…Others…

Full Life CycleFull Life Cycle• SuperindividualsSuperindividuals• ReproductionReproduction• GrowthGrowth• MovementMovement

– Feeding and spawningFeeding and spawning

• MortalityMortality– FishingFishing– PredatorsPredators– StarvationStarvation

Many ChallengesMany Challenges

• Behavior should include spawning and Behavior should include spawning and feeding migrations, predator feeding migrations, predator avoidance, etc.avoidance, etc.

• Growth requires knowledge of Growth requires knowledge of bioenergetics – grow or make eggs?bioenergetics – grow or make eggs?

• Mortality from starvation – don’t all Mortality from starvation – don’t all starve at oncestarve at once

• Spawning new superindividuals in Spawning new superindividuals in bounded memory spacebounded memory space

MethodsMethods

• ““Fish” as modified floatsFish” as modified floats• Fixed number of fish per species per Fixed number of fish per species per

yearclassyearclass• Limit number of yearclasses, killing off Limit number of yearclasses, killing off

too old fishtoo old fish• Feedback to NPZ-type model, NEMURO Feedback to NPZ-type model, NEMURO

for now (PICES)for now (PICES)• Fish-eat-fish and fishing fleets require Fish-eat-fish and fishing fleets require

knowledge of fish in i,j spaceknowledge of fish in i,j space

fish_listfish_list

Fish GrowthFish Growth

• Compute change in weightCompute change in weight• Bioenergetics-basedBioenergetics-based• Consumption determined by:Consumption determined by:

– Zooplankton in cell (NEMURO)Zooplankton in cell (NEMURO)– Other individual fish in the neighborhoodOther individual fish in the neighborhood

• Once mature, allocate energy to Once mature, allocate energy to growth or reproductiongrowth or reproduction

MovementMovement

• Eggs, yolk-sac and larvae are moved Eggs, yolk-sac and larvae are moved by the currentsby the currents

• Juveniles, sub-adults and adults move Juveniles, sub-adults and adults move by behaviorby behavior

• Two choices (for now):Two choices (for now):– Kinesis (Humston et al., 2004), sum of Kinesis (Humston et al., 2004), sum of

random plus moving to better temperaturerandom plus moving to better temperature– Railsback, look for foodRailsback, look for food

Using Humston BehaviorUsing Humston Behavior

Using Railsback BehaviorUsing Railsback Behavior

Create new SuperindividualsCreate new Superindividuals• Fixed number per day of spawning – Fixed number per day of spawning –

spawning happens in a fixed time spawning happens in a fixed time windowwindow

• Find out how many adults spawn that Find out how many adults spawn that day, how many eggsday, how many eggs

• Could have:Could have:– No eggsNo eggs– Fewer cells with eggs than new SIsFewer cells with eggs than new SIs– More cells with eggs than new SIsMore cells with eggs than new SIs

BisectionBisection• Master node builds an egg array with Master node builds an egg array with

all of the eggs in each cellall of the eggs in each cell• Array is for the entire gridArray is for the entire grid• Successively divide up domain in i,j Successively divide up domain in i,j

directions until available directions until available superindividuals are filledsuperindividuals are filled

• Toss out empty partitionsToss out empty partitions• Keep a sorted (by egg count) linked Keep a sorted (by egg count) linked

list of partitionslist of partitions

BisectionBisection

Another ExampleAnother Example

Test of SpawningTest of Spawning

MortalityMortality

• Starvation – bioenergetics are Starvation – bioenergetics are comingcoming

• Predator – not full lifecycle:Predator – not full lifecycle:– Modeled on albacoreModeled on albacore– Migrate in to feed, spawn elsewhereMigrate in to feed, spawn elsewhere

• Fishing fleetsFishing fleets

Fishing FleetFishing Fleet

• Targeting sardines only in the Targeting sardines only in the California CurrentCalifornia Current

• Movement based on engineering, Movement based on engineering, economics, behavior (evaluated economics, behavior (evaluated once daily in model)once daily in model)

• Maximize revenue based on Maximize revenue based on expected CPUEexpected CPUE

Fishing FleetFishing Fleet

- Six ports: (CA(3), OR(1), WA(2))

- Num of boats per ports: 10-30

- Average catch per boat: 40-60 tons

- Boat motoring speed: 20 km/h

-Time to fish/process catch: 2 hours

- Average price for catch: 0.1-0.15 $/kg

NEMUROMS.SAN (1994-2004)Fully coupled ROMS + NEMURO + Fish + Predators + Fleet

(20,000 individuals; 1,000 predators; 100 fishing boats)

Future Fish PlansFuture Fish Plans

• Tuning of bioenergetics and behaviorTuning of bioenergetics and behavior– Warm/cold behavior switchWarm/cold behavior switch– Make them look like sardines and Make them look like sardines and

anchoviesanchovies

• Name this thing!Name this thing!• Write some papersWrite some papers

SummarizingSummarizing

• We have developed a new approach to We have developed a new approach to regional downscaling of climate models regional downscaling of climate models that maintains the important feedbacks that maintains the important feedbacks in the system.in the system.

• The new resolved physical scales are The new resolved physical scales are closer to what is needed for biology.closer to what is needed for biology.

• We are developing an individual based We are developing an individual based model for fish and fishers which is model for fish and fishers which is tightly coupled to the physics and lower tightly coupled to the physics and lower trophic level models.trophic level models.

• Feedbacks in the biological models may Feedbacks in the biological models may also be important.also be important.

Some further thoughtsSome further thoughts• End-to-end models are coming. The multi-End-to-end models are coming. The multi-

scale problem exists in biology as well, and scale problem exists in biology as well, and needs to be addressed.needs to be addressed.

• Due to the complexity, different approaches Due to the complexity, different approaches are likely to be better suited for certain are likely to be better suited for certain problems. problems.

• Models will not be too useful in a data Models will not be too useful in a data vacuum. We need the same big-picture vacuum. We need the same big-picture approach to observations as is being used in approach to observations as is being used in the development of complex models.the development of complex models.

Why ROMS FrameworkWhy ROMS Framework

Fish IBM

Sardines

Anchovies

Predators

Regional Ocean Circulation Model

NPZ Component (multiple)

Floats Component

Data Assimilation

Climate Coupling

Fishing Fleet