Statistical analysis for marine resources (completed)

Project description

Statisticians at Department of Mathematics are involved in a range of different projects within marine resources, in close cooperation with the Norwegian Computing Center, the Institute of Marine Reasearch and Centre for Ecological and Evolutionary Synthesis.  We also have close relationships to

Many of these projects are related to estimation of abundance for different species. Such estimates are typically based on population dynamic models. Assuming N_y,a is the total number of a species in year y at age a, a simple dynamic model is

N_y,a=N_y-1,a-1exp(-M_y,a)-C_y,a

where M_y,a is the natural mortality rate and C_y,a is the number of fish caught by the fishery. The model is typically made stochastic by distributional assumptions about the mortality rate. The fishery industry report total weight of hauls, while a sample of individual fish are  taken and measurements of weight, length and (sometimes) age are provided. In addition relative abundance indices based on catch per unit effort from separate research vessels provide information about abundance.  Many of the  subprojects involve sophisticated statistical modelling and computational tools for improving existing technology.

Objectives

The aim in all projects is to provide the fishery managers with better tools for specification of relevant information for performing decisions, e.g. setting quotas, opening/closing areas for fisheries,  constructing sampling design for data collection.

Outcomes

In addition to software useable for fishery management, important outcomes have been publications in high-ranked journals. See publication list on sub-projects.

Cooperation

Statisticians at Department of Mathematics are involved in a range of different projects within marine resources, in close cooperation with the Norwegian Computing Center (NR), the Institute of Marine Reasearch (IMR) and Centre for Ecological and Evolutionary Synthesis (CEES). Many of these projects are related to estimation of abundance for different species. Such estimates are typically based on population dynamic models. Assuming

Sub-projects

• Dynamic abundance estimation: PhD project for Peter Maisha (finances by Statistics_for_Innovation) where posterior output from catch-at-age estimation is combined with abundance indices for a full Bayesian estimation procedure. Inference is based on importance sampling/sequential Monte Carlo methods.
• Catch-at-age: In this NR based and IMR financed project, a Bayesian hierarchical model is built up for estimating catch at age based on landings data from fishery catch. Markov Chain Monte Carlo methods are used for estimation.
  • Publications:
    • Hirst, Aanes, Storvik, Huseby, Tvete (2004): Estimating catch at age from market sampling data ny using a Bayesian hierarchical model
    • Hirst, Storvik, Aldrin, Aanes, Huseby (2005): Estimating catch-at-age by combining data from different sources
    • Hirst, Storvik, Rognebakke, Aldrin, Aanes, Vølstad (2013): A Bayesian modelling frame for the estimation of catch-at-age of commercially harvested fish species
•  In Bycatch: In this NR based and IMR financed project, geostatistical and GAM modelling are combined to perform prediction of bycatch.
  • Publications
    • Aldrin, Mortensen, Storvik, Nedreaas, Agglen, Aanes (2011): Improving management decisions by predicting fish bycatch in the Barents Sea shrimp fishery

• Infectious diseases in aquaculture: NR project that develop stochastic simulation models to describe how infectious diseases are transmitted between fish farms.

Financing

The Norwegian Research Council,Statistics_for_Innovation, the Institute of Marine Reasearch (IMR)