Prediction of potential fishing grounds for chub mackerel in the Northwest Pacific utilizing a combination of multivariable Gaussian mixture model and Bayesian approach

Abstract

Chub mackerel is an economically important pelagic fish species in the Northwest Pacific Ocean. Accurate forecasting of fishing grounds can greatly improve the fishery production capacity and support conservation-focused fisheries management. This study utilized China's fishery catch data for chub mackerel in the Northwest Pacific from 2014 to 2019, along with various marine environmental factors, including sea surface temperature (SST), chlorophyll concentration (CHL), eddy kinetic energy (EKE), sea surface height anomaly (SSHA), and the strength of the SST gradient (SSTG), to develop a predictive model for fishing grounds. Initially, two Gaussian mixture distributions were simulated to ascertain the distributions of environmental factors for "true" and "false" fishing grounds, respectively. Subsequently, a fishing ground prediction model was constructed using a Bayesian approach informed by the parameters of the Gaussian mixture models. Additionally, a novel training dataset division scheme was established based on cross-validation results to enhance the model's training and testing accuracy, as well as its generalization capability. The forecasting model was ultimately validated against actual fishery data collected from April to November 2020, revealing that the predicted fishing ground locations closely aligned with the actual observed locations for each month. The area under the curve (AUC) values ranged from 0.667 to 0.870, with an average AUC of 0.766 and a variance of 0.076, indicating a reasonable level of accuracy in the model's forecasts. The findings of this study contribute to a deeper understanding of the influence of oceanographic conditions on chub mackerel and offer a novel approach for predicting its fishing grounds.