Forecasts of Climate and Fishing Impacts on the Pelagic Ecosystem: A Comparison of Two Ecosystem Models

A recently completed study comparing two ecosystem modeling approaches found broad agreement between models on the impacts of fishing and climate change in the central North Pacific (CNP). The work was a collaboration between PIFSC scientists Phoebe Woodworth-Jefcoats, Jeff Polovina, and Evan Howell, and their colleague Julia Blanchard from the University of Sheffield.

The study compared a species-based model and a size-based model. Both models found that while total fish catch was projected to increase with increasing fishing mortality, the catch of large fish (> 15 kg) was greatest at moderate levels of fishing mortality (F = 0.4). Additionally, both models projected a decline in catch at all sizes and at all levels of fishing mortality in response to climate change. However, the magnitude of this decline varied between models, with the species-based model projecting declines of 13 — 21% and the size-based model projecting declines of 46 — 57%. When climate change and fishing mortality perturbations were examined in combination, changes in total fish catch and catch of small fish (< 15 kg) largely followed changes in fishing mortality. The relative impact of climate change vs. fishing mortality on catch of large fish varied by model.

Both ecosystem models used the same input climate scenario, represented through densities of small and large phytoplankton. These densities were projected by the NOAA Geophysical Fluid Dynamics Laboratory's prototype Earth System Model 2.1 (ESM2.1), which is a coupled climate and biogeochemical model. Over the 21st century, ESM2.1 projects reduced nutrient input to surface waters in the North Pacific, and as a result densities of both small and large phytoplankton are projected to decline.

The two ecosystem models used in the study represent two fundamentally different approaches to ecosystem modeling. The species-based model relies on detailed diet and trophic interactions, while the size-based model is built upon the premise that predation in the pelagic environment is largely determined by body size. That two such different models project similar responses to climate change lends confidence to our ability to model such impacts at a broad scale.

Disparities in the magnitude of catch decline in response to climate change raise interesting questions regarding which aspects of climate change will have the greatest impact on the CNP. When the declines in catch are examined in conjunction with the phytoplankton densities used to represent the impacts of climate change, the species-based model suggests that the CNP food web is more sensitive to changes in total phytoplankton density. Conversely, the size-based model suggests that changes in the size structure of CNP phytoplankton community will have a greater impact on catch. The two ecosystem models also suggest differing levels of ecosystem sensitivity to fishing and climate change when both factors are combined. The size-based model projects the catch of large fish to be most sensitive to climate impacts while the species-based model projects climate impacts to dominate at lower F values (0.4 and 0.6) and fishing impacts to dominate at higher F values (0.8 and 1.0). Differences such as these present an ideal test for an integrated size- and species-based ecosystem model.

Using mean fish catch for the period 2081—2100 as a measure, the graphs show percent change in the mean catch at each level 
        of fishing mortality relative to (a) F = 0.2, without and with climate change (effect of increasing fishing mortality), (b) 
        no climate change (effect of climate change at each F value, and (c) F = 0.2 and no climate change (effect of both 
        increasing fishing mortality and climate change).  From left to right, bars in each cluster indicate increasing fishing mortality: 
        F = 0.2, 0.4, 0.6, 0.8, 1.0.
Using mean fish catch for the period 2081—2100 as a measure, the graphs show percent change in the mean catch at each level of fishing mortality relative to (a) F = 0.2, without and with climate change (effect of increasing fishing mortality), (b) no climate change (effect of climate change at each F value, and (c) F = 0.2 and no climate change (effect of both increasing fishing mortality and climate change). From left to right, bars in each cluster indicate increasing fishing mortality: F = 0.2, 0.4, 0.6, 0.8, 1.0.