Introduction

The AGEPRO (Age-Structured Projection) model is a stochastic simulation framework developed to support fishery management by evaluating the uncertainty of future stock tracjectries under alternative harvest strategies. Initially created in 1994 to determine optimal rebuilding strategies for depleted fish stocks, AGEPRO has been applied in multiple New England groundfish assessments (e.g Northeast Fisheries Science Center [NEFSC] 1994; New England Fishery Management Council [NEFMC] 1994; NEFMC 1996) and has undergone several updates since its original documentation (Brodziak and Rago. Manuscript. 1994; Brodziak, Rago, and Conser 1998). Written in the C programming language, AGEPRO efficienyly peforms stochastic projections of age-structured populations, generating outputs such as landings, spawning stock biomass, and fishing mortality across userspecified time horizons. The model is available in the [NOAA Fisheries Intergrated Toolbox]

AGEPRO is designed to generate stochastic projections that inform fishery management decisions by quantifying the uncertainty in population and fishery outcomes over time. It simulates the dynamics of an exploited age-structured population, projecting variables such as landings, spawning stock biomass, fishing mortality, and population age composition Figure 1. The acronym “AGEPRO” highlights its focus on age-structured projections, distinguishing it from biomass-based approaches. Users can evaluate alternative harvest strategies by specifying annual quotas or fishing mortality targets.

AGEPRO incorporates three main sources of uncertainty in population projections: (1) recruitment, (2) initial population size, and (3) process error in biological and fishery parameters.

Recruitment

Recruitment is the primary stochastic component of AGEPRO projections, typically representing the number of age-1 fish entering the population each year. The model includes fifteen recruitment options, ranging from empirical bootstrap sampling to parametric stochastic models. A deterministic recruitment trajectory can also be specified see Recruitment Model 3

Initial population size

The second source of uncertainty concerns the initial population size at the start of the projection period Figure 1. This can be represented as a probability distribution estimated using methods such as bootstrapping or Markov chain Monte Carlo sampling, or as a fixed point estimate if uncertainty is not considered.

Process error

The third uncertainty source reflects random variation (process error) in population and fishery processes over time. AGEPRO models these as multiplicative lognormal deviations with a mean of one and a specified coefficient of variation. The user can choose to apply process errors to parameters such as:

1. Natural mortality at age
2. Maturation fraction at age
3. Stock and spawning stock weight at age
4. Mean population weight at age
5. Fishery selectivity, discard fractions, and weights at age.

These simulated time series can be saved in auxiliary files to document the stochastic realizations underlying each projection.

Brodziak, Jon, P. J. Rago, and Ramon Conser. 1998. “A General Approach for Making Short-Term Stochastic Projections from an Age-Structured Fisheries Assessment Model,” January. https://doi.org/10.4027/fsam.1998.52.

Brodziak, Jon, and P. Rago. Manuscript. 1994. “A General Approach for Short-Term Stochastic Projections in Age-Structured Fisheries Assessment Models.” Manuscript. 1994.

NEFMC. 1996. “Amendment 7 to the Northeast Multispecies Fishery Management Plan.” Newburyport, MA.

New England Fishery Management Council [NEFMC]. 1994. “Amendment 5 to the Northeast Multispecies Fishery Management Plan.” Newburyport, MA.

Northeast Fisheries Science Center [NEFSC]. 1994. “Report of the 18th Northeast Regional Stock Assessment Workshop. Stock Assessment Review Committee (SARC) Consensus Summary of Assessments.” Northeast Fisheries Science Center Reference Document ; 94-22. https://repository.library.noaa.gov/view/noaa/8551.