Development of a climate-to-fish-to-fishers model: proof-of-principle using long-term population dynamics of anchovies and sardines in the California Current Jerome Fiechter, UC Santa Cruz (on Behalf of the NEMUROMS.SAN Group) ePOPf meeting, Portland, 21 Sep 2010
Chris Edwards University of California – Santa Cruz Dave Checkley Scripps Institution of Oceanography Alec MacCall National Marine Fisheries Service Tony Koslow SIO - CALCOFI Sam McClatchie Ken Denman Institute for Ocean Sciences (Canada) Francisco Werner Rutgers University Kenneth A. Rose Louisiana State University Enrique N. Curchitser Rutgers University Kate Hedstrom Arctic Region Supercomputing Center Jerome Fiechter University of California – Santa Cruz Miguel Bernal Instituto Español de Oceanografía (Spain) Shin-ichi Ito Fisheries Research Agency Salvador Lluch-Cota CIBNOR (Mexico) Bernard A. Megrey National Marine Fisheries Service
Motivation Much emphasis on climate to fish linkages Global change issues Bottom-up, middle-out, top-down controls Increasing pressure for ecosystem-based considerations in management Full array of interactions within an ecosystem, including humans (instead of single species) Development of end-to-end ecosystem models Multi-species, individual-based, physics to fish Proof of principle, continuation of NEMURO effort
Motivation The human component…
Proof of Principle Sardine – Anchovy cycles Well-studied species with population cycles observed in many systems Teleconnections across basins Good case study Forage fish tightly coupled to NPZ Important ecologically and widely distributed Low frequency variability Provided by: Salvador E. Lluch-Cota Source: Schwartzlose et al., 1999
Californian Anchovy Larval Abundance April 1952 Low Anchovy Abundance April 1965 High Anchovy Abundance Source: MacCall, 1990
End-to-End (E2E) Ecosystem Model (NEMUROMS.SAN) Sub-Model 1: 3-D ROMS for ocean circulation Sub-Model 2: NEMURO for NPZ Sub-Model 3: Multiple-species IBM for fish Sub-Model 4: Fishing fleet dynamics Today: progress to date Solved many numerical and bookkeeping issues Next is to add more realistic biology
Regional Ocean Circulation Model Why ROMS Framework Regional Ocean Circulation Model NPZ Component (multiple) Floats Component Climate Coupling Data Assimilation Fish IBM Sardines Anchovies Predators Fishing Fleet
ROMS-CCSM Coupling – Climate Forcing Focus on Upwelling Regions California Current System Humboldt Current System Benguela Current System
3-D ROMS for Ocean Circulation California Current Grid - 30 km horiz. resolution - 30 vertical levels Run duration - 40 years (1960-2000) - hourly time step BC/IC: SODA-POP - Monthly SSH, U, V, T, S (Carton et al., 2000) Surface forcing: CORE2 - 6-hrly wind stress, heat fluxes - Daily radiation, monthly E-P (Large and Yeager, 2008)
NEMURO for Lower Trophic Levels (Nutrients, Phytoplankton, and Zooplankton)
Fish IBM: Species Types Why individual-based model (IBM) Natural unit in nature Allow for complicated life history Plasticity and size-based interactions Conceptually easier movement Sardines and anchovy (full life-cycle) Reproduction, growth, mortality, movement Competitors for food (P,Z) Predator species (e.g., albacore) Consumption and movement only Relative biomass to scale mortality
Fish IBM: Growth and Mortality Bioenergetics for weight Consumption from multi-species functional response At maturity, allocate energy to growth or reproduction Mortality Natural (temperature and stage dependent) Predation (predator species dynamics, vary spatially) Fishing (fishing fleet dynamics, vary spatially) Maximum consumption: f(W, T) Holling Type II feeding curve
Fish IBM: Movement (Happiness-based behavior; Humston et al, 2004) Eggs, yolk-sac, and larvae moved by physics - Vertical position assumed at 50 m depth Juveniles and adults moved by behavior - Day-to-day vs. seasonal migrations Kinesis for sardines and anchovies (Happiness-based behavior; Humston et al, 2004) - Fitness for predator species (Maximized growth behavior; Railsback et al., 1999) Each fish has continous x, y, and z position - mapped to 3-D grid at each time step to determine cell location and local conditions
Fishing Fleet Targeting sardines in the California Current Movement based on engineering, economics, behavior (evaluated once daily in model) Maximize revenue based on expected CPUE - Num. of ports: 6 (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
CA Commercial Fish Landings (2009) Source: CA Department of Fish and Game Pacific Sardine (~38K tons) P. Mackerel (~5K tons) Fisheries Peak: ~700K tons (1936) N. Anchovy (~3K tons) Dover Sole (~3K tons) Sablefish (~2K tons)
Numerical Details Major numerical and bookkeeping challenges Solving everything simultaneously, eventually physics with an offline option However, it is the fish IBM that takes computing time Super-individual approach for IBM Sheffer et al., 1995 Allocation of super-individuals (array management) Super-individuals eating super-individuals Bisection to start new super-individuals as eggs each day based on spawner locations
NEMUROMS.SAN (1994-2004) Fully coupled ROMS + NEMURO + Fish + Predators + Fleet (20,000 individuals; 1,000 predators; 100 fishing boats)
Sample NEMUROMS.SAN output (1960-2000) Weight at Age Total Egg Production Diet Composition Mean Depth (m) Mean Temperature (C)
Next Steps It can be done – proof of principle 2 days to run 40 years with 20,000 individuals Target: 500,000 super-individuals Can we calibrate and validate this model? Very challenging: Physics, NPZ, Fish Is it useful? Need to increase biological realism Investigate the causes of low-frequency cycles in sardine and anchovy Parallel effort in Japan to provide a geographic and ecosystem contrast
Next Leaps Relate to other E2E ecosystem models Atlantis, ECOPATH w/ECOSIM , FEAST, … Assess use for ecosystem-based management Other species, fisheries in the California Current California Sea Lion foraging in 2004 and 2005 (Weise et al., 2006) Coho Salmon Survival (Peterson and Schwing, 2003)
ROMS NEMURO IBM
Biological Resolution 3D-NEMURO (1/12º) NEMUroms.SAN 3D-eNEMURO 3D-FeNEMURO 3D-NPZ 3D-PlankTOM5 3D-NEMURO(1º) SEAPODYM 2D-NEMURO.FISH Spatial Resolution 2D-NEMURO multi-box NEMURO.FISH 3-box NEMURO.FISH MFW- NEMURO 1D-NEMURO 0D-NPZ 1-box NEMURO.FISH Biological Resolution Prepared by Shin-ichi Ito
ROMS time-stepping (hourly) Age and clean old individuals Loop over cells in grid Loop over days in year Next year Growth NEMURO – NPZ Initial conditions If midnight then reproduction Hydrodynamics Determine fish in each cell ROMS time-stepping (hourly) Age and clean old individuals Next Loop over years in run Consumption from previous hour Small, large, and predatory zooplankton Loop over fish in cell If midnight then add new super-individuals Mortality Movement NEMUROMS.SAN FLOWCHART
Fish IBM: Kinesis Movement Inertial: f(Vt-1) = Vt-1·multiplier Random: g(ε) = ε·multiplier Velocity: f(Vt-1) + g(ε)
Fish IBM: Development Eggs, yolk-sac larvae, larvae Temperature-dependent stage durations Birthday on January 1 of each year Juvenile to subadult on first January 1 Subadult to adult with maturity at second birthday
Fish IBM: Reproduction Two strategies Capital and income Hybrid that allows switching Beginning and ending temperatures (or days) Check energy to initiate a batch Capital using condition income using yesterday’s weight change Batches develop based on temperature Accumulate eggs each day and locations of spawners
Fish IBM: Mortality Natural Starvation: too skinny Eggs: 0.1/day Yolk-sac: 0.05/day Larvae: 0.05/day Juveniles: 0.3/year Adults: 0.3/year Starvation: too skinny
Bisection
Fish IBM: Fitness Option for Movement (Railsback et al. 1999) Evaluate entire water column defined by 9 cells Project growth rate using interpolated fields at center of each cell Go towards center of cell with fastest growth Position updated hourly
Fish IBM: Kinesis Option for Movement (Humston et al. 2004) X and Y velocities of each individual is computed hourly based on kinesis behavior (response to temperature) Kinesis is the sum of random and inertial velocities (happiness) Horizontal done 24 times a day using first hour’s conditions; vertical is hourly
Density-Dependence Growth via feedback effects on prey Starvation Fecundity Other possibilities: Maturation (mediated through growth) Mortality via predation (movement only species) and fishing Movement Spreading out under high abundances Costs of inferior habitats?
Sample IBM output: early life history
ROMS and NEMURO Fields (1960-2000 means)
Sample IBM output: individuals and biomass
Sample IBM output: individuals and biomass
Sample IBM output: individuals and biomass
Sample IBM output: fish trajectories Locations every 5 days for one year Grey shading shows bottom topography
Sample IBM output: vertical position (every 5 days)