Complex dynamics may limit prediction in marine fisheries |
| |
| Authors: | Sarah M Glaser Michael J Fogarty Hui Liu Irit Altman Chih‐Hao Hsieh Les Kaufman Alec D MacCall Andrew A Rosenberg Hao Ye George Sugihara |
| |
| Institution: | 1. Department of Fisheries, Virginia Institute of Marine Science, , Gloucester Point, VA, 23062‐1346 USA;2. Department of Biology, College of William & Mary, , Williamsburg, VA, 23187‐8795 USA;3. Northeast Fisheries Science Center, NOAA, , Woods Hole, MA, 02543‐1026 USA;4. Department of Marine Biology, Texas A&M University at Galveston, , Galveston, TX, 77553 USA;5. Department of Biology, University of Boston, , Boston, MA, 02215 USA;6. Institute of Oceanography and Institute of Ecology and Evolutionary Biology, National Taiwan University, , Taipei, 106 Taiwan;7. Southwest Fisheries Science Center, NOAA, , Santa Cruz, CA, 95060 USA;8. Union of Concerned Scientists, , Cambridge, MA, 02138‐3780 USA;9. Scripps Institution of Oceanography, University of California, San Diego, , La Jolla, CA, 92093‐0210 USA |
| |
| Abstract: | Complex nonlinear dynamics in marine fisheries create challenges for prediction and management, yet the extent to which they occur in fisheries is not well known. Using nonlinear forecasting models, we analysed over 200 time series of survey abundance and landings from two distinct ecosystems for patterns of dynamic complexity (dimensionality and nonlinear dynamics) and predictability. Differences in system dimensionality and nonlinear dynamics were associated with time series that reflected human intervention via fishing effort, implying the coupling between human and natural systems generated dynamics distinct from those detected in the natural resource subsystem alone. Estimated dimensionality was highest for landings and higher in abundance indices of unfished species than fished species. Fished species were more likely to display nonlinear dynamics than unfished species, and landings were significantly less predictable than abundance indices. Results were robust to variation in life history characteristics. Dynamics were predictable over a 1‐year time horizon in seventy percent of time series, but predictability declined exponentially over a 5‐year horizon. The ability to make predictions in fisheries systems is therefore extremely limited. To our knowledge, this is the first cross‐system comparative study, and the first at the scale of individual species, to analyse empirically the dynamic complexity observed in fisheries data and to quantify predictability broadly. We outline one application of short‐term forecasts to a precautionary approach to fisheries management that could improve how uncertainty and forecast error are incorporated into assessment through catch limit buffers. |
| |
| Keywords: | Complexity coupled human natural systems fisheries population dynamics forecasting models nonlinear dynamics prediction |
|
|
