Technical Comments

Response to Comment on “Impacts of historical warming on marine fisheries production”

See allHide authors and affiliations

Science  16 Aug 2019:
Vol. 365, Issue 6454, eaax7170
DOI: 10.1126/science.aax7170

Abstract

Szuwalski argues that varying age structure can affect surplus production and that recruitment is a better metric of productivity. We explain how our null model controlled for age structure and other processes as explanations for the temperature-production relationship. Surplus production includes growth, recruitment, and other processes and provides a more complete description of food production impacts than does recruitment alone.

In our recent paper (1), we used temperature-linked surplus production models to measure the influence of ocean warming on 235 marine fish and invertebrate populations and to explore the drivers and consequences of the resulting mixture of positive, negative, and neutral influences. In response to this analysis, Szuwalski (2) makes the following three arguments: (i) Changes in age structure resulting from fishing can also affect the productivity of exploited populations; (ii) the null model that we used was inadequate for simulating the case of no correlation between temperature and surplus production; and (iii) recruitment is a better metric for examining environmentally driven (i.e., time-varying or nonstationary) productivity because it integrates the effects of time-varying age structure. We show below why these concerns either are misdirected or do not undermine the conclusions of our paper.

First, we agree that changing age structure is an additional process affecting fisheries productivity. Szuwalski’s carefully designed examples help to illustrate this point and show how autocorrelated residuals are generated by ignoring the impacts of age structure. We also note that age structure is only one of many such processes that potentially have an impact on fisheries productivity, and a narrow focus on age structure risks overlooking the potentially equally important impacts of primary productivity, oxygen concentration, pH, predator-prey interactions, and other processes. It is for this exact reason that we used a null model to test whether our observations could be expected in the absence of a relationship between sea surface temperature (SST) and surplus production. Szuwalski appears to have misunderstood our null model approach, which already accomplished the analysis that he called for—namely, to “preserve the endogenous pattern in production and randomize the exogenous driver of SST.” The purpose and specification of our null model are also consistent with classical definitions of this approach in ecology. For example, Gotelli and Graves (3) provide the following definition in their Null Models in Ecology textbook:A null model is a pattern‐generating model that is based on randomization of ecological data or random sampling from a known or specified distribution. The null model is designed with respect to some ecological or evolutionary process of interest. Certain elements of the data are held constant, and others are allowed to vary stochastically to create new assemblage patterns. The randomization is designed to produce a pattern that would be expected in the absence of a particular ecological mechanism. (p. 3)To build our null model, we randomized the temperature time series to disassociate the observed alignment of temperature and productivity. We went further than most null models and also preserved the trend and autoregressive properties of the original temperature time series. In addition, we maintained the surplus production time series in its original order, thereby preserving any endogenous patterns in production. Our null model therefore simulated the temperature-productivity patterns that would be expected if there were no relationship between temperature and productivity. When the model was refit using the randomized (null) temperature time series, the influence of warming was significantly dampened relative to the model fit to the true temperature time series. This test successfully confirmed that the observed temperature-production relationship was highly unlikely in the absence of a true correlation between temperature and production. It also confirmed that any potential endogenous patterns in surplus production (such as those caused by changes in age structure) were unlikely to explain our results.

In addition, we agree that recruitment is a useful metric for investigating time-varying productivity, but note that it is not a replacement for surplus production in analyses. In particular, surplus production is important because it (i) captures the cumulative effects of the environment on recruitment, somatic growth, and natural mortality, and (ii) allows us to calculate the impact of environmental change on maximum sustainable yield. Maximum sustainable yield is more easily translated into impacts on human welfare than is recruitment because the former is directly related to food production, whereas the latter is only indirectly linked. Ocean warming has reduced and will continue to reduce fish growth rates and body sizes (4, 5), resulting in decreased yield per recruit (6). Because recruitment-based assessments of the impact of climate change on productivity fail to account for these large negative impacts on growth rates, we expect recruitment-based assessments of climate impacts to be positively biased. These positive biases may explain the divergent conclusions between our surplus production analyses and the conclusions of the recruitment analysis by Szuwalski (7). Analyses of surplus production offer a powerful alternative because they integrate the cumulative impacts of climate change on recruitment, growth rates, and mortality.

In summary, we argue that (i) changing age structure is one of many unobserved processes affecting fisheries productivity, and we accounted for such influences appropriately in our analysis; (ii) our null model is already consistent with both Szuwalski’s request and classical definitions in ecology, and the null model confirmed that the observed relationship between temperature and production was highly unlikely in the absence of a true relationship between temperature and production; and (iii) analyses of environmentally driven surplus production are advantageous in comparison to analyses of recruitment because surplus production integrates cumulative impacts on all life history processes (including recruitment) and documents these impacts in terms of maximum sustainable yield. Thus, the conclusions of our paper remain unchanged. Ocean warming has driven declines in marine fisheries productivity and the potential for sustainable fisheries catches in many places, with other regions experiencing increases. These impacts are likely to become more negative in the future. Preventing overfishing and developing management strategies that are robust to temperature-driven changes in productivity are essential if society is to maintain and rebuild the capacity for global wild-capture fisheries to supply food and support livelihoods in a warming ocean.

References

View Abstract

Subjects

Navigate This Article