Discovering Early Birds in Trawlers through VMS Data Analysis

Abstract

Vessel Monitoring Systems (VMS) records trajectory data of fishing vessels, which provides a new tool for fishery research. Previous research distinguished the activities for fishing vessels and analyzed the fishing effort distribution. This paper proposes a nontrivial question that whether there are some early birds in trawlers for fishing. Early birds refer to trawlers which not only arrive in the fishing area earlier than other trawlers but also are followed by a large portion of trawlers. We present a system to discover the early birds through VMS data. After defining following relation quantitatively on VMS dataset, we present the system to discover the early birds of trawlers. The system contains two steps. The first is to construct following relation through VMS dataset. The second is to discover the early birds which are followed by a large portion of trawlers from the following relation. We exploit the VMS dataset of 1750 otto trawlers from September, 2015 to May, 2017 to verify our approach. Thirty early birds are discovered from the first twelve months’ VMS data, which are followed by 77% trawlers in later nine month. Moreover, the fishing region of these 30 early birds in one week for the last 9 month are always containing more than 80% of the fishing density of all the trawlers in the corresponding next week, which confirms the early birds as the fishing region indicators.