Multispecies Catch at Age Model (MSCAGEAN): incorporating predation interactions and statistical assumptions for a predator-prey system in the eastern Bering Sea

1
Multispecies Catch at Age Model (MSCAGEAN)
incorporating predation interactions and
statistical assumptions for a predator-prey
system in the eastern Bering Sea

• Jesus Jurado-Molina
• University of Washington
• Patricia A. Livingston
• Alaska Fisheries Science Center

2
Fisheries models

Assumption Population Isolated Constant Natural
Mortality
Age Structured Models
Statistical Assumptions
Statistical Catch at Age Models
3
Fisheries models

Predation Equations MM1M2
No statistical assumptions on error structure
included
Virtual Population Analysis
Multispecies Virtual Population Analysis (MSVPA)
4
Multispecies Catch at age Analysis?

Predation interactions
Age structured model
Statistical assumptions on error structure
Multispecies Catch at age Analysis
5
Objectives

• To add the predation equations to a CAGEAN model
  (MSCAGEAN)
• Comparison of the MSCAGEAN results to the ones
  estimated with the multispecies VPA, the
  Multispecies Forecasting Model (MSFOR) and the
  single species CAGEAN.

6
Input and outputs of MSVPA-MSFOR
7
Statistical models
Error assumption
Model equations
Prior information
Data
8
Predation equations

S - suitability coefficient of prey p for
predator i BS - suitable prey biomass R - annual
ration of the predator i W - weight at age of
prey p M2 - predation mortality
9
Multispecies CAGEAN
Error assumption
Model equations
Predation equations
Prior information
Data
10
Equations
11
Advantages

• Multispecies approach
• We can use the tools used in single species stock
  assessments
• Likelihood profile
• Bayesian statistics (probability distributions)
• Model selection (Akaikes information
  criterion,likelihood ratio )

12
Assumptions

• Stomach content measured without error
• Suitabilities constant (estimated as the average
  of the annual suitabilities)
• Recruitment for the simulation is log-normal
  distributed
• Recruitment of age-0 individuals for the
  simulation takes place in the third quarter

13
Walleye pollock and Pacific cod interactions

Fishery
Walleye pollock
Pacific cod
14
Methods

• Initial run of the MSVPA updated to 1998.
• Run of Multispecies forecasting (F40).
• Spawning Biomass in 2015 as indicator of
  performance.
• Multispecies Catch at Age Analysis updated to
  1998
• Single species CAGEAN updated to 1998

15
MSVPA and MSCAGEAN results Age-0 walleye
pollock Natural mortality (1990)
MSCAGEAN
MSVPA

1.55
M2 1.70 57
16
MSVPA and MSCAGEAN results suitability
coefficients

MSCAGEAN
MSVPA
0.303
0.683 0.140
17
MSVPA AND MSCAGEAN results Spawning biomass in
2015
MSCAGEAN
MSFOR

SUMSUML
SUMSUM1
SSB 5.52E6 2.36E6
SSB 1.33E7 5.25E6
18
MSCAGEAN and CAGEAN results Spawning biomass in
2015
CAGEAN
MSCAGEAN

SSB 1.19E07 5.06E06
SSB 1.33E07 5.25E06
19
Future tasks

• To implement the predation equations in the stock
  assessments methods used in the AFSC assessments