• Climate Change Site
  • Drivers of Climate Change and Variability on the Northeast Shelf
  • Projected Change
  • Historical Analyses
  • Projected Impacts
  • Implications
  • Protected Species
  • Literature Cited
• Current Conditions
  • Current Update
  • Sea Surface Temperature
  • Bloom Development
  • Bloom Start/Magnitude
  • SST Distribution
  • Chlorophyll
  • Survey Temperature
  • Satellite SST
  • Extended SST
  • Thermal Transition
  • Species Distribution
  • Kernel Density
  • SST Forecast
  • Previous Update
• Ecology of the Ecosystem
  • NEUS Shelf Ecology
  • Climate Influences
  • Physical Habitat
  • Oceanography
  • Phytoplankton
  • Zooplankton
  • Fish and Squid
  • Bottom Life
  • Deep Corals
  • Invasive Species
  • Protected Species
  • Human Dimensions
  • Site Map
• Ecosystem Status
  • Main Findings
  • Executive Summary
  • 1. Introduction
  • 2. Climate Forcing
  • 3. Physical Pressures
  • 4. Production
  • 5. Benthic Invertebrates
  • 6. Fish Communities
  • 7. Protected Species
  • 8. Human Dimensions
  • 9. Ecosystem Services
  • 10. Stressors and Impacts
  • 11. Status Determinations
  • 12. Synthesis
  • Contact/Contributors
  • Literature Cited
  • Glossary
  • Site Map
• Modeling Approaches
  • Ecosystem Modeling
  • Modeling Review
  • Overview Document
• Spatial Analyses
  • Home page
  • Gadoid Fish
  • Other Demersal Fish
  • Flatfish
  • Invertebrates
  • Elasmobranchs
  • Pelagic Fish
  • Other Fish
• Ecosystems Home

Modeling Approaches

Ecosystem Modeling

Development of multispecies and ecosystem models has an extensive history at the Northeast Fisheries Science Center (NEFSC).  In particular, NEFSC has been exploring options for development of a flexible analytical framework for Ecosystem-based Fishery Management EBFM) in consultation with the New England and Mid-Atlantic Fishery Management Councils and the Atlantic States Marine Fishery Commission. in the Northeast.  Principal elements of the approach include:

1. establishment of a transparent connection between single species and ecosystem-based advice using multispecies assessment models as a natural bridge, 
2. development of multiple operating models to test assessment models and candidate management procedures,
3. application of assessment models spanning a spectrum of complexity to evaluate the issue of model uncertainty,
4. application of formal strategies of multimodel inference in applying results from the multispecies assessment models,
5. use of these results to assess uncertainty and risk, and
6. evaluation of tradeoffs in a bioeconomic context. 

The models under development are designed to accommodate spatial structure and to incorporate consideration of climate variability and change. 

We are currently developing a prototype multispecies analysis for Georges Bank (Figure 1). Data streams feeding into this process encompass fishery-dependent (both ecological and social-economic) sources, fishery-independent surveys, food habits data to identify and quantify biotic interactions among species, and oceanographic and climate data to track external forcing mechanisms. To further enhance communication with stakeholders we are developing options for data and model visualization to aid in the interpretation of multispecies model outputs.

The core analytical elements of the process involve development and testing of a set of indicators, multispecies assessment models, social-economic modules linked to the assessment models, and forecast models developed outside the assessment model framework to complement predictions made using these assessments.  The interplay between the operating models and the other analytical elements of the approach is envisioned as an iterative process (Figure 1).  The analysis culminates in a risk analysis accounting for key uncertainties and in the context of multiple candidate management procedures.   The process is designed to provide management advice in the form of annual catch limits to match existing requirements under current management approaches on Georges Bank.  The results will be provided as an interactive web-based product (Figure 1).

Our overall approach entails the use of four different assessment models encompassing simple multispecies production models applied both to individual species and to defined functional groups, multispecies delay-difference models that implicitly accommodate simple demographic structure (again for individual species and functional groups), and a complex multispecies statistical catch-at-age model applied to individual species (Figure 2).  Single species analogues of these models are familiar to resource managers in the region and we have deliberately attempted to frame our approach in a way that trades on this familiarity. Multiple estimation techniques including maximum likelihood, state-space, genetic algorithm and Bayesian methods will be applied to the production models in both aggregated and disaggregated forms to assess aspects of estimation uncertainty.

 A key issue in assessment and management of the Georges Bank system is the centrality of the mixed-species nature of the fishery.  We define our functional groups as species that are caught together and share basic ecological characteristics (similarity in life history attributes, body size, etc.).  Our interest in testing the performance of assessment models based on functional groups defined in this way centers both on their importance as key structural elements of the system and recognition that we cannot fully control the fishing mortality rates on the individual species comprising these mixed-species assemblages.  These species, inter alia, share similar histories of exploitation and environmental forcing. Tests will be made to assess the performance of the functional group models against models in which the full species identity of all components is retained to see if they offer any advantage in assessing mixed-species fisheries.

Economic modules link to the assessment models to produce revenue streams and measures of profitability. They are being developed for direct use in tradeoff analysis.   For the economic module we are also employing an empirical multispecies portfolio model approach to assess risk.  We are developing forecast models using new methods in nonlinear time series analysis to complement the assessment models. We are using two operating models to serve as a virtual test beds to examine the performance of the assessment models and to evaluate the efficacy of alternative management procedures.  These models, Hydra and EcoSim, are currently in different stages of development. Hydra has been developed at NEFSC. It is spatially structured, deals with length structured populations, and allows for multiple fleet sectors.  The model is designed to accommodate climate/environment forcing on biological and ecological processes. The operating model will be used to test the performance characteristics of several simpler assessment models that can be used to provide reference points for management action.

NEFSC is also working collaboratively with the Northeast and Mid-Atlantic Regional Planning Bodies and their associated regional ocean councils to develop analytical methods to support marine Ecosystem-Based Management (mEBM) in the region.  The focus of mEBM (which encompasses EBFM) is on broad multisectoral concerns and human use patterns in the ocean.  The elements of our Ecosystem-Status Report are critical components of this overall effort.  A central management tool for mEBM is the application of spatially explicit strategies for ocean management.

In the accompanying document, we provide an overview of the spectrum of NEFSC ecosystem models that have been developed in support of marine Ecosystem-Based Management and Ecosystem-Based Fishery Management in this region.  We also provide a link to an external evaluation of our modeling efforts by an external peer review panel conducted by the Center for Independent Experts.