Accurate prediction of the central fishing grounds of chub mackerel is substantial for assessing and managing marine fishery resources. Based on the high-seas chub mackerel fishery statistics and multi-factor ocean remote-sensing environmental data in the Northwest Pacific Ocean from 2014 to 2021, this article applied the gravity center of the fishing grounds, 2DCNN, and 3DCNN models to analyze the spatial and temporal variability of the chub mackerel catches and fishing grounds. Results:1) the primary fishing season of chub mackerel fishery was April-November which catches were mainly concentrated in 39°∼43°N, 149°∼154°E. 2) Since 2019, the annual gravity center of the fishing grounds has continued to move northeastward; the monthly gravity center has prominent seasonal migratory characteristics. 3) 3DCNN model was better than the 2DCNN model. 4) For 3DCNN, the model prioritized learning information on the most easily distinguishable ocean remote-sensing environmental variables in different classifications.

Pacific sardine (Sardinops sagax) is a commercially important species and supports important fisheries in the Northwest Pacific Ocean (NPO). Understanding the habitat distribution patterns of Pacific sardine is of great significance for fishing ground prediction and stock management. In this study, both single-algorithm and ensemble distribution models were established through the Biomod2 package for Pacific sardine by combining the species occurrence data, sea surface temperature (SST), sea surface height (SSH), sea surface salinity (SSS) and chlorophyll-a concentration (Chla) in the NPO during the main fishing season (June–November) from 2015 to 2020. The results indicated that the key environmental variables affecting the habitat distribution of Pacific sardine were the SSH and SST. The suitable habitat area for Pacific sardine showed significant monthly changes: the suitable habitat range in June was larger than that in July and August, while the suitable habitat range gradually increased from September to November. Furthermore, the monthly geometric centers of habitat suitability index (HSI) for Pacific sardine presented a counterclockwise pattern, gradually moving to the northeast from June, and then turning back to the southwest from August. Compared with single-algorithm models, the ensemble model had higher evaluation metric values and better spatial correspondence between habitat prediction and occurrence records data, which indicated that the ensemble model can provide more accurate prediction and is a promising tool for potential habitat forecasting and resource management.

The chub mackerel (Scomber japonicus) is one of the most influential small pelagic fish in the Northwest Pacific Ocean, and accurate modeling approaches and model selection are critical points in predicting the Scomber japonicus fishing grounds. This study investigated the changes in catches and fishing days on no moonlight and bright moonlight days (2014-2022) and compared the differences in predictive performance between the LightGBM and RF models on three datasets under the two modeling approaches [those based on the light fishing vessels operational characteristics (Approach one) and those not (Approach Two)]. The results were as follows: 1) Stronger moonlight intensity (e.g., full moon) can limit the fishing efficiency of light fishing vessels, with most years showing a trend of a higher percentage of fishing days on bright moonlight days than catches percentage, i.e., no moonlight days resulted in higher catches with lower fishing days; 2) Compared to Modeling Approach Two, under Modeling Approach one, RF model achieved better predictive performance on dataset B, while the LightGBM model achieved better predictive performance on both datasets A and B; 3) Overall, the Approach One achieved more satisfactory prediction performance, with the optimal prediction performance on the complete dataset C improved from 65.02% (F1-score of the RF model, Approach Two) to 66.52% (F1-score of the LightGBM model, Approach Two); 4) Under the optimal modeling approach (Approach One) and the optimal model (LightGBM model), the differences in the importance of the variables on dataset A (no moonlight days) and dataset B (bright moonlight days) were mainly centered on the environmental variables, with CV, SLA, and SSS being the most important in dataset A, and CV, DO, and SLA being the most important in dataset B. This study provides a more scientific and reasonable modeling undertaking for the research of light purse seine fishing vessels, which is conducive to guiding fishermen to select the operating area and operating time of the Scomber japonicus fishery more accurately and comprehensively and realizing the balanced development of fisheries in terms of ecology and economy.