基于船舶轨迹的海陵岛渔船航次时空特征分析OA

The Automatic Identification System(AIS)provides extensive vessel-tracking data for monitoring illegal fishing,forecasting central fishing grounds,quantifying fishing efforts,and protecting the marine environment.Existing research on AIS data has mainly focused on vessel movement patterns and spatiotemporal activity intensity;however,vessel type classification and activity analysis have rarely been associated with voyages.Focusing on the importance of voyages in vessel behavior analysis,this study proposed a voyage-based ship classification method wherein a multidimensional feature set was constructed to integrate the overall shape of the voyage,motion characteristics,and geographic background information,enabling a more accurate and interpretable classification and facilitating the pattern discovery of fishing vessel behaviors.Using soft voting based on probability averaging to integrate the XGBoost and LightGBM models,a better balance between accuracy and minority class recognition capability was achieved.Based on the distinction between fishing and non-fishing vessel voyages,the spatial and temporal characteristics of fishing vessel voyages on Hailing Island in Guangdong Province were systematically analyzed.The following results were obtained:1)the accuracy of the voyage-based ship classification model reached 0.930.The model's decisions mainly rely on the ship size,sailing speed,trajectory shape,water depth,distance from the shore,and time preference.Among these,the contribution of ship size and speed to make the decision exceeded 63％,rendering them the most important features;meanwhile,spatial environment,trajectory patterns,and time-related behaviors played a supplementary role.2)The number of fishing vessels and voyages showed evident monthly variations:the lowest numbers were observed during the South China Sea fishing moratorium,while the primary and secondary peaks occurred in the month before the moratorium began and in the first month after it ended.3)The departure times of fishing vessels were relatively concentrated between T 03:00-05:59 and T 16:00-16:59,while return times were concentrated between T 13:00-15:59 and T 06:00-06:59;consequently,two short-duration operation patterns were established:early morning departure with same-day return in the afternoon,and evening departure with an early morning return the next day.4)Voyages mainly involved nearshore fishing,and those lasting 6-12 hours and involving sailing distances of within 100 km accounted for more than 46.82％and 72.20％of the total number of voyages,respectively.5)More than 75％of such voyages were local trips between Hailing Island and other parts of Yangjiang City,with only a small number travelling to neighboring cities;this indicated that the fishery network was highly localized.6)The fishing operation areas were concentrated in the nearshore region south of Hailing Island,with water depths up to 40 m and a maximum operation density of 12.065 h/km2.From the core area outward,the operation density decreased rapidly with increasing distance.This study reveals that fishing vessels on Hailing Island are highly dependent on nearshore resources and significantly affected by fishing moratorium policies,suggesting that current fishery management policies have begun to show positive effects on resource protection;however,there exists a need to further promote structural adjustments in the fisheries industry and encourage the development of distant-water fishing and marine aquaculture to alleviate the pressure on nearshore resources.