Improving estimates of population status and trend with superensemble models. Anderson, S. C., Cooper, A. B., Jensen, O. P., Minto, C., Thorson, J. T., Walsh, J. C., Afflerbach, J., Dickey-Collas, M., Kleisner, K. M., Longo, C., Osio, G. C., Ovando, D., Mosqueira, I., Rosenberg, A. A., & Selig, E. R. FISH AND FISHERIES, 18(4):732-741, WILEY, 111 RIVER ST, HOBOKEN 07030-5774, NJ USA, JUL, 2017.
doi  abstract   bibtex   
Fishery managers must often reconcile conflicting estimates of population status and trend. Superensemble models, commonly used in climate and weather forecasting, may provide an effective solution. This approach uses predictions from multiple models as covariates in an additional ``superensemble'' model fitted to known data. We evaluated the potential for ensemble averages and superensemble models (ensemble methods) to improve estimates of population status and trend for fisheries. We fit four widely applicable data-limited models that estimate stock biomass relative to equilibrium biomass at maximum sustainable yield (B/BMSY). We combined these estimates of recent fishery status and trends in B/BMSY with four ensemble methods: an ensemble average and three superensembles (a linear model, a random forest and a boosted regression tree). We trained our superensembles on 5,760 simulated stocks and tested them with cross-validation and against a global database of 249 stock assessments. Ensemble methods substantially improved estimates of population status and trend. Random forest and boosted regression trees performed the best at estimating population status: inaccuracy (median absolute proportional error) decreased from 0.42 -0.56 to 0.32 -0.33, rank-order correlation between predicted and true status improved from 0.02 - 0.32 to 0.44 - 0.48 and bias (median proportional error) declined from - 0.22 - 0.31 to - 0.12 - 0.03. We found similar improvements when predicting trend and when applying the simulation-trained superensembles to catch data for global fish stocks. Superensembles can optimally leverage multiple model predictions; however, they must be tested, formed from a diverse set of accurate models and built on a data set representative of the populations to which they are applied.
@article{ ISI:000404555100008,
Author = {Anderson, Sean C. and Cooper, Andrew B. and Jensen, Olaf P. and Minto,
   Coilin and Thorson, James T. and Walsh, Jessica C. and Afflerbach, Jamie
   and Dickey-Collas, Mark and Kleisner, Kristin M. and Longo, Catherine
   and Osio, Giacomo Chato and Ovando, Daniel and Mosqueira, Iago and
   Rosenberg, Andrew A. and Selig, Elizabeth R.},
Title = {{Improving estimates of population status and trend with superensemble
   models}},
Journal = {{FISH AND FISHERIES}},
Year = {{2017}},
Volume = {{18}},
Number = {{4}},
Pages = {{732-741}},
Month = {{JUL}},
Abstract = {{Fishery managers must often reconcile conflicting estimates of
   population status and trend. Superensemble models, commonly used in
   climate and weather forecasting, may provide an effective solution. This
   approach uses predictions from multiple models as covariates in an
   additional ``superensemble{''} model fitted to known data. We evaluated
   the potential for ensemble averages and superensemble models (ensemble
   methods) to improve estimates of population status and trend for
   fisheries. We fit four widely applicable data-limited models that
   estimate stock biomass relative to equilibrium biomass at maximum
   sustainable yield (B/BMSY). We combined these estimates of recent
   fishery status and trends in B/BMSY with four ensemble methods: an
   ensemble average and three superensembles (a linear model, a random
   forest and a boosted regression tree). We trained our superensembles on
   5,760 simulated stocks and tested them with cross-validation and against
   a global database of 249 stock assessments. Ensemble methods
   substantially improved estimates of population status and trend. Random
   forest and boosted regression trees performed the best at estimating
   population status: inaccuracy (median absolute proportional error)
   decreased from 0.42 -0.56 to 0.32 -0.33, rank-order correlation between
   predicted and true status improved from 0.02 - 0.32 to 0.44 - 0.48 and
   bias (median proportional error) declined from - 0.22 - 0.31 to - 0.12 -
   0.03. We found similar improvements when predicting trend and when
   applying the simulation-trained superensembles to catch data for global
   fish stocks. Superensembles can optimally leverage multiple model
   predictions; however, they must be tested, formed from a diverse set of
   accurate models and built on a data set representative of the
   populations to which they are applied.}},
Publisher = {{WILEY}},
Address = {{111 RIVER ST, HOBOKEN 07030-5774, NJ USA}},
Type = {{Article}},
Language = {{English}},
Affiliation = {{Anderson, SC (Reprint Author), Univ Washington, Sch Aquat \& Fishery Sci, Seattle, WA USA.
   Anderson, Sean C.; Cooper, Andrew B.; Walsh, Jessica C., Simon Fraser Univ, Sch Resource \& Environm Management, Burnaby, BC, Canada.
   Jensen, Olaf P., Rutgers State Univ, Inst Marine Coastal Sci, New Brunswick, NJ USA.
   Minto, Coilin, Galway Mayo Inst Technol, Marine \& Freshwater Res Ctr, Galway, Ireland.
   Thorson, James T., NOAA, Natl Marine Fisheries Serv, Northwest Fisheries Sci Ctr, Fisheries Resource Anal \& Monitoring Div, Seattle, WA 98112 USA.
   Afflerbach, Jamie, Univ Calif Santa Barbara, Natl Ctr Ecol Anal \& Synth, Santa Barbara, CA 93106 USA.
   Dickey-Collas, Mark, Int Council Explorat Sea, Copenhagen, Denmark.
   Dickey-Collas, Mark, Tech Univ Denmark DTU, DTU Aqua Natl Inst Aquat Resources, Charlottenlund, Denmark.
   Kleisner, Kristin M., NOAA, Natl Marine Fisheries Serv, Ecosyst Assessment Program, Northeast Fisheries Sci Ctr, Woods Hole, MA 02543 USA.
   Longo, Catherine, Marine Stewarship Council, London, England.
   Osio, Giacomo Chato; Mosqueira, Iago, European Commiss, DG Joint Res Ctr, Directorate Sustainable Resources D, Unit Water \& Marine Resources D 02, Ispra, Italy.
   Ovando, Daniel, Univ Calif Santa Barbara, Bren Sch Environm Sci \& Management, Santa Barbara, CA 93106 USA.
   Rosenberg, Andrew A., Union Concerned Scientists, Cambridge, MA USA.
   Selig, Elizabeth R., Conservat Int, Arlington, VA USA.}},
DOI = {{10.1111/faf.12200}},
ISSN = {{1467-2960}},
EISSN = {{1467-2979}},
Keywords = {{CMSY; data-limited fisheries; ensemble methods; multimodel averaging;
   population dynamics; sustainable resource management}},
Keywords-Plus = {{SEASONAL CLIMATE FORECASTS; MULTIMODEL SUPERENSEMBLE; EXTINCTION RISK;
   FISHERIES; SELECTION; ENSEMBLE; OCEANS}},
Research-Areas = {{Fisheries}},
Web-of-Science-Categories  = {{Fisheries}},
Author-Email = {{sean.anderson@dal.ca}},
ResearcherID-Numbers = {{Dickey-Collas, Mark/A-8036-2008}},
Funding-Acknowledgement = {{Gordon and Betty Moore Foundation}},
Funding-Text = {{Gordon and Betty Moore Foundation}},
Number-of-Cited-References = {{62}},
Times-Cited = {{0}},
Usage-Count-Last-180-days = {{6}},
Usage-Count-Since-2013 = {{6}},
Journal-ISO = {{Fish. Fish.}},
Doc-Delivery-Number = {{EZ2QR}},
Unique-ID = {{ISI:000404555100008}},
OA = {{No}},
DA = {{2017-08-17}},
}

Downloads: 0