Book Review Climate Change Fisheries

North Sea fish stick to warming shallows rather than cooling off at depth

Rutterford, Louise A., et al. “Future fish distributions constrained by depth in warming seas.” Nature Climate Change (2015). DOI: 10.1038/nclimate2607

Sea surface temperature in the North Sea has risen four times faster than the global average over the past 30 years and is predicted to continue warming at a fast rate for the next 50 years (Fig. 1, right panel).  Due to the presence of highly valuable commercial fisheries in the North Sea, regional ecology has been monitored for some time now. Warming in the North Sea has been observed to change seasonality of fish spawning, plankton dynamics, and ecosystem interactions. Fish survey data indicates that many species are shifting towards higher latitudes and deeper waters to avoid warming temperatures. In addition, there have been substantial changes in the relative abundances of species.  Landings of fish species that prefer cooler temperatures have been cut in half, while warm-adapted species have increased by a factor of 2.5 since the 1980s.  These changes have made a marked impact on commercial fisheries for certain species.  Research on the impacts of warming have focused on historical survey data to describe past effects of warming. Process-based approaches have also been used to predict future abundances, but does not validate models based on historical data.  Process-based methods are limited because they depend on explicitly incorporating processes into the model (e.g., reproduction, metabolism, growth). The authors of this study used historical data to build a predictive model and determined the capability of the model to predict observed warming impacts.

They selected the 10 most abundant demersal (living on or near the bottom) fish species for the study.  Generalized additive models (GAMs) were generated using various sets of predictor variables (surface temperature, bottom temperature, level of fishing pressure) to predict distributions and abundances of fish earlier in the time series.  These were then adjusted to be able to predict known distributions and abundances later in the time series.  Finally, the models were used to predict changes over the next five decades. Winter and summer fish survey data was used.

Figure 1. Left panel: depth ranges and grid overlay of cells for which fish metrics and temperature was used.  Right panel: Sea surface temperature (SST) and near-bottom temperature (NBT) from summer and winter historical trends and forecasted 50 years from present day.
Figure 1. Left panel: depth ranges and grid overlay of cells for which fish metrics and temperature was used. Right panel: Sea surface temperature (SST) and near-bottom temperature (NBT) from summer and winter historical trends and forecasted 50 years from present day.
Figure 2.  Correlation between observed and predicted abundances using summer (left panel) and winter (right panel) survey data.  A value of 1 indicates perfect correlation.
Figure 2. Correlation between observed and predicted abundances using summer (left panel) and winter (right panel) survey data. A value of 1 indicates perfect correlation.

All models performed well when predicting observed fish abundances and distributions.  To determine the appropriate window of time for predictions, models were used to predict 10, 20 and 30 years into the future.  In both the winter (Fig. 2, left panel) and summer (Fig. 2, right panel), observed and modeled abundances were well correlated in 8 of the 10 species.  Models were then used to predict future distributions, abundance and temperature preference of those 8 species based on temperature forecasts.  Predictions indicated that species will tolerate warmer temperatures and maintain existing distributions – not shift distribution to escape warming temperatures.  This is in contrast to climate-envelope approaches which have suggested a poleward shift in distributions.  The authors suggest that previous methods do not account for importance of non-thermal habitats in fish distributions.  Notably, depth range did not vary between observed and predicted.  This suggests that distributions may be limited by depth-dependent habitat.  Species that were forced into deeper waters via surface warming were shown to decline in abundance.  Cold-water fish tended to exhibit greater depth range with warming temperature, but stopped significantly deepening after 1980.  This suggests that there is a limit to depth range and the authors do not expect ranges to deepen any further.

The results of the study emphasize the value of GAMs over process-driven approaches which depend on distinct mechanisms.  As data-driven models, GAMs encompass processes that are not always explicitly described in process-driven approaches.  The authors suggest that policy and management planning use data-driven models to make informed decisions.  With robust forecasts for ongoing ocean warming, it is necessary to accurately predict the effects on economically valuable North Sea fisheries.  It looks like we may see increased landings in warm-adapted fish species and declining landings of cold-water species.  This may lead to the expansion of new fisheries and the dissolution of others.

Leave a Reply

Your email address will not be published.