PSPaul D. Spencer

Paul D. Spencer

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Paul D. Spencer has a total of 21 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Dynamic harvest control rules for fish stocks with time-varying productivity

    by Jeremy S Collie, Richard J Bell, Paul D Spencer, Rachel C Marshall, Cóilín Minto

    Biological reference points are benchmarks against which stock status can be assessed and are central to fisheries management advice. Calculation of non-proxy reference points depends on the demographic parameters of the stock, particularly stock-recruitment parameters. There is growing evidence of non-stationary, pre-recruit mortality parameters that would imply time-varying reference points. It may be possible, however, to develop robust (time-invariant) harvest control rules (HCRs) that can partially mitigate for dynamic mortality parameters via lagged biomass dependency. It is unclear under what conditions either approach would be preferable. To address this question, we use time-varying stock-recruitment relationships for three contrasting stocks and counterfactually manage plausible histories under: (1) constant proxy fishing mortality F target; (2) time-varying target F; (3) constant biomass-linked HCR; and (4) a time-varying biomass-linked HCR. In contrast to a previous study that found conservation gains with constant, status-quo management targets, we find that time-varying HCR policies were able to maintain higher levels of biomass during low-productivity periods and achieve higher yields during periods of high productivity. These benefits result from not fixing reference points to proxies that become outmoded when productivity varies. We find that a time-varying harvest policy may be advantageous when the productivity signal is strong and persistent over decadal time scales. The performance of a time-varying HCR will also be improved when there is a short time lag between assessing productivity and adjusting the fishing mortality rate. By contrast, a time-invariant HCR will be favored for stocks with weaker productivity signals relative to the short-term recruitment variability or longer assessment lags and possibly older recruitment ages. While some generalization is possible, we recommend that time-varying harvest policies be evaluated on a stock-by-stock basis using this approach.


  • Open Access

    Climate‐informed stock‐recruitment models identify potential environmental links for commercial fish stocks

    by Rachel C. Marshall, Jeremy S. Collie, Richard J. Bell, Paul D. Spencer, Cóilín Minto

    Several studies have demonstrated non‐stationarity in marine fish maximum per capita recruitment rate, which may be linked to regional environmental drivers. Exploring relationships between environmental drivers and time‐varying recruitment rate will provide insight into recruitment processes, and suggest how fisheries managers may adapt biological reference points and adjust harvest‐control rules in response to changes in the ecosystem state. To test for significant effects on the recruitment rate of 48 commercial United States fish stocks with nonstationary per‐capita recruitment, environmental indices from two Atlantic and three Pacific regions are incorporated into a state‐space dynamic Ricker stock‐recruitment model. We developed two model variants for incorporating climate into the state‐space model, one based on the climate variable itself and the other on its rate of change. Simulation tests indicated that if the true system has a climate driver, most of the time a climate‐enhanced model will detect it. However, the different model variants may have similar statistical support, such that model selection may be unable to distinguish the ecological hypotheses of the climate‐recruitment interaction. For 35 (73%) of the true stock time series, at least one environmental variable was identified as nominally significant for improving on the non‐climate dynamic model. The simulation results indicate that the identified environmental indices likely do impact recruitment of the given stocks, even if the form of that effect may not be clearly identified. Our results can now support more detailed stock‐specific analyses of fisheries dynamics, climate drivers, model specification and ultimately management action in the future.


  • Trait‐based climate vulnerability assessments in data‐rich systems: An application to eastern Bering Sea fish and invertebrate stocks

    by Paul D. Spencer, Anne B. Hollowed, Michael F. Sigler, Albert J. Hermann, Mark W. Nelson

    Trait‐based climate vulnerability assessments based on expert evaluation have emerged as a rapid tool to assess biological vulnerability when detailed correlative or mechanistic studies are not feasible. Trait‐based assessments typically view vulnerability as a combination of sensitivity and exposure to climate change. However, in some locations, a substantial amount of information may exist on system productivity and environmental conditions (both current and projected), with potential disparities in the information available for data‐rich and data‐poor stocks. Incorporating this level of detailed information poses challenges when conducting, and communicating uncertainty from, rapid vulnerability assessments. We applied a trait‐based vulnerability assessment to 36 fish and invertebrate stocks in the eastern Bering Sea (EBS), a data‐rich ecosystem. In recent years, the living marine resources of the EBS and Aleutian Islands have supported fisheries worth more than US $1 billion of annual ex‐vessel value. Our vulnerability assessment uses projections (to 2039) from three downscaled climate models, and graphically characterizes the variation in climate projections between climate models and between seasons. Bootstrapping was used to characterize uncertainty in specific biological traits and environmental variables, and in the scores for sensitivity, exposure, and vulnerability. The sensitivity of EBS stocks to climate change ranged from “low” to “high,” but vulnerability ranged between “low” and “moderate” due to limited exposure to climate change. Comparison with more detailed studies reveals that water temperature is an important variable for projecting climate impacts on stocks such as walleye pollock (Gadus chalcogrammus), and sensitivity analyses revealed that modifying the rule for determining vulnerability increased the vulnerability scores. This study demonstrates the importance of considering several uncertainties (e.g., climate projections, biological, and model structure) when conducting climate vulnerability assessments, and can be extended in future research to consider the vulnerability of user groups dependent on these stocks.


  • Open Access

    Modelling spatially dependent predation mortality of eastern Bering Sea walleye pollock, and its implications for stock dynamics under future climate scenarios

    by Paul D. Spencer, Kirstin K. Holsman, Stephani Zador, Nicholas A. Bond, Franz J. Mueter, Anne B. Hollowed, James N. Ianelli

    Arrowtooth flounder (Atheresthes stomias) are an important predator of juvenile walleye pollock (Gadus chalcogramus) in the eastern Bering Sea (EBS) shelf and have increased 3-fold in biomass from 1977 to 2014. Arrowtooth flounder avoid the summer “cold pool” (bottom water ≤2°C) and variability in cold pool size and location has affected their spatial overlap with juvenile walleye pollock. Developing a method to account for the relationship between climate change and pollock mortality can highlight ecosystem dynamics and contribute to better assessments for fisheries management. Consequently, spatially resolved predation mortality rates were estimated within an age-structured walleye pollock stock assessment population model (based on spatial information on diet and abundance from trawl surveys), along with the effect of sea surface temperature (SST) on pollock recruitment. Projections of SST and cold pool area to 2050 were obtained (or statistically downscaled) from nine global climate models and used within an age-structure population model to project pollock abundance given estimated relationships between environmental variables and predator and prey spatial distributions, pollock recruitment, and maximum rate of arrowtooth flounder consumption. The climate projections show a wide range of variability but an overall trend of increasing SST and decreasing cold pool area. Projected pollock biomass decreased largely due to the negative effect of increased SST on pollock recruitment. A sensitivity analysis indicated that the decline in projected pollock biomass would be exacerbated if arrowtooth flounder increased their relative distribution in the EBS northwest middle shelf (an area of relatively high density of juvenile pollock) in warm years.


  • Open Access

    Fisheries Science Vol. 76 No. 3

    by Alistair J. Hobday, Sung Lee, Kerim Aydin, Paul D. Spencer, Thomas K. Wilderbuer, Zhang Chang

  • Density‐independent and density‐dependent factors affecting temporal changes in spatial distributions of eastern Bering Sea flatfish

    by Paul D. Spencer

    The general warming of the eastern Bering Sea (EBS) and the wide range of abundance exhibited by several eastern Bering Sea flatfish motivated an examination of how density‐dependent and density‐independent factors may influence the spatial distributions of EBS flatfish. In this study, EBS trawl survey data from 1982 to 2006 were used to examine how temporal changes in the distributions of six flatfish species groups [yellowfin sole (Limanda aspera), rock sole (Lepidopsetta sp.), flathead sole (Hippoglossoides sp.), Alaska plaice (Pleuronectes quadrituberculatus), arrowtooth flounder (Atheresthes sp.), and Greenland turbot (Reinhardtius hippoglossoides)] are related to temporal changes in the location of the ‘cold pool’ (bottom water < 2°C), and how the area occupied by flatfish are related to the cold pool and population abundance. Rock sole and flathead sole distributions have generally moved northwest since 1982 and are significantly correlated with the movement of the cold pool, whereas arrowtooth flounder avoid the cold pool and their area occupied is inversely related to the size of the cold pool. The area occupied by arrowtooth flounder and rock sole are also significantly related to stock abundance. Multivariate statistical models indicate that the location of rock sole is more strongly related to stock abundance than to the cold pool, whereas the area occupied by arrowtooth flounder is more strongly related to the area of the cold pool rather than abundance. The temperatures occupied by several flatfish stocks indicate a substantial variability in suitable temperatures. These results suggest that a complex suite of density‐dependent and density‐independent factors may determine the response of EBS flatfish spatial distributions to increasing temperatures.


  • Application of a Kalman Filter to a Multispecies Stock Complex

    by P.D. Spencer, J.E. Ianelli

    Proceedings of 2003 symposium on managing newly developing and small-scale fisheries with limited data.