Development of a prey-predator species distribution model for a large piscivorous fish: A case study for Japanese Spanish mackerel Scomberomorus niphonius and Japanese anchovy Engraulis japonicus

One of the major challenges of species distribution models (SDMs) is incorporating the influence of prey on the predator. An ensemble modeling approach, the prey-predator species distributabion model (PPSDM), uses environmental variables (sea surface temperature, bathymetry, sea surface salinity, and sea surface current) and a biotic variable (predicted prey distribution: Japanese anchovy) to predict a piscivorous fish (Japanese Spanish mackerel) habitat in the China Seas. PPSDM comprised selected preferable single algorithms from nine algo-rithms, including tree-based models, machine learning models, regression models, an envelope-style method, and a classification model. Compared with that obtained when using habitat without the Japanese anchovy distri-bution, the regression models, machine learning models and a classification model have significant improve-ments. However, tree-based models have only slightly improved. The PPSDM results well reflected the ontogenetic migration of Japanese Spanish mackerel. We found that the wintering ground was distributed in regions with temperatures between 11 and 17 degrees C and salinities between 33.6 and 34.1. The bottom topography was also closely correlated with Japanese Spanish mackerel, and the suitable habitat was at 50-100 m isobaths as topography determined the position of isotherms (11, 17 degrees C), salinity isolines (33.6, 34.1), and fronts (tem-perature, salinity) in winter. The relationships among different climatic indices, temperature, and Japanese Spanish mackerel were analyzed. We found that the East Asian monsoon significantly influenced the Japanese Spanish mackerel habitat by affecting temperature variation. In the future, the suitable habitat of Japanese Spanish mackerel is projected to decrease by 21%, 28%, 36%, and 42% under RCP2.6, RCP4.5, RCP6.0, and RCP8.5, respectively. The Japanese Spanish mackerel in the northern Yellow Sea is projected to experience the 'cul de sac' effect, which would drive this species toward local extinction. The Japanese Spanish mackerel in the China Seas will probably shift northeastward through the Tsushima Strait, and the Sea of Japan may form the main habitat at the end of the century.

Automated summary

One major challenge in species distribution models (SDMs) is incorporating the influence of prey on the predator. This study utilizes a prey-predator species distribution model (PPSDM) to predict the habitat of Japanese Spanish mackerel in the China Seas. The PPSDM considers environmental variables such as sea surface temperature, bathymetry, and sea surface current, as well as a biotic variable, the predicted prey distribution (Japanese anchovy). The results show the importance of wintering ground temperature, salinity, and bottom topography in determining the habitat of Japanese Spanish mackerel.