Mismatch between VMS data temporal resolution and fishing activity time scales

Abstract

The use of Vessel Monitoring System (VMS) data to map fishing activity is challenged by the mismatch between the temporal resolution of position records (typically 2h) and the time scale of fishing activity in fisheries with short trips and short fishing operations such as purse seining for small pelagics. We analysed the first five years of VMS and logbook data for the Portuguese purse seine fleet, when 10min resolution VMS data were available, to evaluate bias and errors in fishing trips and fishing sets’ identification related to the mismatch. We adapted the standardised VMS analysis workflow for EU fleets to the characteristics of the fishery and developed a framework to quantify bias for different VMS-based products, by resampling the 10min VMS dataset at 20, 30, 60min and 2h intervals. For the Portuguese purse seine fishery, a 2h time interval resulted to 42% missed fishing trips compared to the 10min time interval data and a bias towards longer fishing trips. For trips that were correctly identified in the 2h resampled dataset, 7% of the fishing sets were missing and fishing locations were identified with an error of approximately 2.36km. The general spatial patterns of fishing operations – i.e. fishing grounds – were not significantly altered by the decrease of the data temporal resolution. Our framework is applicable to other fisheries and could become a useful tool for managers using VMS data.