| Automatic Identification System is a digital information service system based on satellite positioning,computer network,data communication and information processing to realize vessels' identification,monitoring and communication.It can obtain the location of the fishing vessel in near real time,and send dynamic information such as latitude and longitude,course,speed,time and static information such as ship name,power and tonnage,etc.In recent years,the development of AIS satellite technology has expanded its application in the analysis of offshore fishing grounds.The ocean-going fishing vessels send data quickly and accurately to the ground station via satellite which can be used to detect the fishing activities,analyze spatio-temporal changes in fisheries and monitoring of illegal fishing operation.This study used satellite AIS data from 7-12,2017,fishing vessel attribute data,fishery production data from 9-11,2017,and sea surface temperature.Firstly,statistics on fishing vessels in the Central and Western Pacific were made based on fishing type and regional distribution.Secondly,combining with the log data and fishing vessel trajectory,the AIS data of the tuna longline fishing vessels is labeled into four states: hooking,hauling,drifting and streaming,in which the course,speed and trajectory characteristics were analyzed.Thirdly,the identification and accuracy of the fishing activities were verified by the threshold and support vector machine model.Then according to the speed threshold,the points of the whole study area were collected.The method of combining the global Moran index and hotspot analysis was used to identify the location of the central fishery,which also was verified with the fishery production data.Finally,the fishery changes from 7-12,2017 were discussed and their relationship with SST were analyzed.The results could be summarized as follows:1)According to the statistics of the vessels in the Western and Central Pacific Fisheries Commission(WCPFC),there are a total of 3,223 fishing vessels in this area.The types of operations mainly include longline,purse seine,squid fishing and trolling.The number of tuna longline fishing vessel is 2,119.The length of fishing vessel is mainly distributed in the range of 15~30m and 45~55m.2)The tuna longline fishing vessels generally start at around 5 am;the speed of the net is significantly negatively correlated with the catch,and the correlation coefficient is-0.427;the speed of the fishing vessel is 6.5~8.5 knots when the hook is released which is similar to the streaming,and the speed is 3~6 knots when hauling.The course when it is fishing showed obvious reverse characteristics.It occupies a certain proportion in each range when drifting.When vessel is streaming,the course shows a bimodal distribution and the difference of peaks is 180°,which shows the round-trip characteristics.The forward and backward speed difference and course difference are distributed near 0 in each condition,indicating the consistency of the speed and course.3)Compared with the threshold method which accuracy is 67.1%,the SVM model based on speed,course,forward and backward speed difference,course difference and time can effectively identify the fishing status of each point.The SVM identification accuracy is 80.09%,with the mean square error is 0.16,and its result is closer to the fishing intensity map generated by the raw maker data.4)Fish resources exhibited spatial correlation and showed significant concentration distribution pattern.The tuna resource hotspot areas gathered a high proportion of fishing effort,forming the tuna longline fishing ground in the western and Central Pacific.There was more than one central fishing ground each month.The grounds in July were roughly same as that in August,except that some had a trend to expand northward or outward.In September the largest fishing ground moved eastward to the area near 180?,and other grounds had different degrees of diffusion or shrink than before;from October to December the central fishing grounds were intensively distributed in the Hawaii islands near 170?E-180?E and 155?W,and the average of fishing effort was more than that between July and September.5)Related to different tuna species,the locations and SST of central fishing grounds were in three sections: cold tongue area near Hawaii islands where SST was 25.5-27.5?,mainly for Thunnus obesus and Thunnus albacares;warm pool area near 10?S-10?N where SST was 28-30.3?,mainly for Thunnus obesus and Katsuwonus pelamis;the areas in the south of equator that cold front and warm front intersected where SST is 25-29?,mainly for Thunnus alalunga.The innovation of this research is mainly reflected as followed:1)Using AIS information to analyze the operation characteristics of tuna longline fishing vessels and the correlation between speed and fishery production data,which provides reference for studying the behavior of fishing vessels.2)Based on the speed,course,forward and backward speed and course difference and time,the SVM model is constructed to complete the identification of the fishing states,which fills the gap in the current domestic research.3)Proposing a method based on AIS data to find the central fishing grounds by the Global Moran's Index and Hot Spot Analysis,which provides a new idea for studying fishing ground under the condition of insufficient catch data,and supplies the identification method of the central fishing grounds. |
PDF Full Text Request