BO: development of spatial data analysis for (pulse) fisheries data

Samenvatting

Due to a decrease in the area available for fishing, competition between fishing vessels for space and resources is expected to increase. These changes are driven mainly by other human activities that take up space, such as the construction of new offshore wind parks and the designation of marine protected areas. Research on pulse fishing has led to a unique situation where detailed spatial fisheries data was collected, allowing for the study of the behaviour of an innovative fishing method with low ecological impact, especially in a setting in which space is becoming increasingly limited. This project aims to further the development of statistical tools available to analyse spatial fisheries data, using Vessel Monitoring System (VMS) data of the Dutch tickler chain (TBB) and pulse (PUL) beam trawl fleets in the period 2017-2021, when the innovative PUL gear was used by part of the Dutch beam trawler fleet and was being studied. First, weekly fishing grounds for each vessel were identified using spatial GAMs. For these GAMs, different threshold levels of fishing activity were tested to see which level gave the most accurate representation of a fishing patch. The GAMs were then used to study fishers’ behaviour, such as the number of patches exploited during a trip, the size of exploitation patches, the length of stay in a patch, and the aggregation of vessels, and to test whether the pulse fleet and traditional fleet’s behaviour differ significantly. The effects of this behaviour on the efficiency of the fishery were studied, specifically on the catch per unit effort (CPUE) of the beam trawler fleet’s main target species, plaice and sole. We found that in all years TBB distribution is much more diffuse and spread out over the northern part of the North Sea, whereas PUL distribution is mostly restricted to the southern North Sea. Through the years, the PUL fishing activity dwindles. Simultaneously, the TBB activity shifts southward, towards the area where PUL fishing was previously concentrated. For TBB, we observe a strong decrease in plaice CPUE as length of stay increases, as well as when the number of vessels increases. For PUL, plaice CPUE seems less affected by length of stay and number of vessels. Sole CPUE seems in general less affected by the time spent in a patch and the number of vessels in a patch. For both TBB and PUL, CPUE decreases as the length of stay increases (though not as strongly as for plaice and TBB). For TBB, sole CPUE actually increases as the number of vessels in a patch increases. For PUL, the CPUE only slightly decreases as the number of vessels increases. An aspect that was not considered in this study, is the difference in CPUE inside and outside of a patch. We compared the length of stay in a fishing patch for different peel levels. With higher peel levels, and thus higher threshold values, the mean length of stay, maximum length of stay, mean surface of fishing ground and mean number of cores all decrease. With the lowest thresholds, there is so much overlap between weeks that vessels are quickly identified as being in the same area (on average for 14.7 consecutive weeks). Using the knowledge gained in this study, we recommend the use of a relatively high threshold to identify fishing grounds, where the surface area is most in line with other studies of resource patches. It would also be advisable to use a set threshold level of fishing effort, rather than a percentage of the max effort, to allow for easier comparison between vessels and weeks. Although the method is very suitable for studying spatial distributions on relatively small datasets (i.e. one vessel in one week per model), the method does not work well when applied to larger datasets (whole fleets and whole years), as information in lost when scaling up. Because of the large number of sequential calculations necessary for these analyses, this type of work is very suited for parallel computing. For the future application of this method, more detailed comparisons to other spatial fisheries analysis methods, such as the one used by Rijnsdorp et al. (2022), are needed. The research on pulse fishing has led to a unique situation in data collection, allowing us to gain knowledge on an innovative fishing method with low ecological impact, especially in a setting in which space is becoming increasingly limited. The method developed here provides a sound and relatively easy to use analysis of fishing grounds. Such methods are increasingly important for a number of reasons: Spatial conflict in the marine sphere is increasing due to (new) human activities at sea, the composition of the (Dutch) fleets is changing rapidly, and the distribution of (commercially relevant) fish species is likely to shift due to climate change. Gaining a detailed understanding of spatial fisheries patterns is crucial for effective management of marine resources and a sustainable fishing industry.