Assessing Biodiversity from Resource Trawl Surveys

The spatial distribution of species richness. Colors represent the expected number of species to be found in that area, with darker colors representing a greater number of species. A. Cooper and A. Solow
The spatial distribution of species richness. Colors represent the expected number of species to be found in that area, with darker colors representing a greater number of species. A. Cooper and A. Solow
Project Summary:

This project explores spatial and temporal patterns of fish biodiversity by performing a variety of statistical analyses on the Northeast Fisheries Science Center trawl survey data.

Location:

Gulf of Maine

Significance

By their very nature, surveys such as the annual Northeast Fisheries Science Center trawl survey are plagued with sampling error, especially when fish are in low abundance. Imperfect as they are, surveys are often the best data available and are extrapolated to provide biodiversity data to fisheries managers.

Advanced statistical methods hold promise as a means by which these data interpretations and extrapolations can be refined to improve understanding of regional biodiversity.

This project examines one such case and discovers multiple hotspots (and cold spots) of biodiversity throughout the region.

  • Andrew B. Cooper, Simon Fraser University,
  • Andrew Solow, Woods Hole Oceanographic Institute Marine Policy Center,
  • Additional contributions from:
  • Andy Rosenberg, University of New Hampshire
  • Ernst Linder, University of New Hampshire
  • Lisa Barry, graduate student, University of New Hampshire
  • Funded by: Gulf of Maine Area Program
Project Detail:

The goal of this project was to explore the Northeast Fisheries Science Center (NEFSC) trawl survey data using statistical methods to create an estimate of the spatial distribution of species richness across the Gulf of Maine. Our project had two phases.

Phase 1: Developing a biodiversity index for the Gulf of Maine. In our pilot study we applied two statistical approaches to the NEFSC data. In the first, we assumed that the trawl survey adequately sampled the biodiversity of the gulf. We calculated a biodiversity index for each trawl, then smoothed between the trawls to generate an estimated biodiversity index for the whole Gulf of Maine.

In the second statistical approach, we assumed that the trawl survey adequately sampled the individual species in the gulf. We then calculated the spatial distribution of the species across the entire Gulf of Maine, and from the spatial distribution of species we generated a biodiversity index.

Our most profound result was that these two methods produced very different biodiversity indices for the Gulf of Maine. Because the causes of these divergent results were unclear, we elected to expand our research to include multiple years and to adjust the details of our analysis and see if we could reach any satisfying conclusions.

Phase 2: Quantitative approaches to assessing the spatial distribution of species richness in the Gulf of Maine. Our goal in this second phase was to develop statistical methods to specifically estimate the spatial distribution of species richness in the Gulf of Maine from NEFSC trawl survey data. Our particular challenges included behaviors of fish that result in their patchy distribution through the gulf, and the habitat heterogeneity of the Gulf of Maine itself. Our method is based on simple local estimation combined with local environmental estimation to address habitat heterogeneity. As we progressed through our research, we also developed a means by which to estimate the goodness of our estimation.

Preliminary results indicate that species richness is clustered into hotspots and can vary by more than a factor of two throughout the Gulf of Maine. The coefficient of variation about the estimates ranges from 5% to almost 20%.