Conceived and designed the experiments: MW BG. Performed the experiments: MW. Analyzed the data: MW. Contributed reagents/materials/analysis tools: MW. Wrote the paper: MW BG. Other: Co-devised the analysis: BG.
The authors have declared that no competing interests exist.
Conservation of marine ecosystems will require a holistic understanding of fisheries with concurrent spatial patterns of biodiversity.
Using data from the UK Government Vessel Monitoring System (VMS) deployed on UK-registered large fishing vessels we investigate patterns of fisheries activity on annual and seasonal scales. Analysis of VMS data shows that regions of the UK European continental shelf (i.e. Western Channel and Celtic Sea, Northern North Sea and the Goban Spur) receive consistently greater fisheries pressure than the rest of the UK continental shelf fishing zone.
VMS provides a unique and independent method from which to derive patterns of spatially and temporally explicit fisheries activity. Such information may feed into ecosystem management plans seeking to achieve sustainable fisheries while minimising putative risk to non-target species (e.g. cetaceans, seabirds and elasmobranchs) and habitats of conservation concern. With multilateral collaboration VMS technologies may offer an important solution to quantifying and managing ecosystem disturbance, particularly on the high-seas.
For global commercial fisheries to maintain a sustainable future
Knowledge regarding the spatial ecology of non-target species of conservation concern (e.g. cetaceans, elasmobranchs, turtles and seabirds) is ever-growing from boat and aerial surveys
Creating a generalised, yet spatially and temporally explicit, understanding of fisheries effort with which to evaluate potential capture of target stocks and minimise putative risk to non-target species and habitats is far from trivial. Information on the at-sea distribution and behaviour of fishing vessels may be obtained from routine and opportunistic surveillance by enforcement agencies using boats and planes, but these approaches lack spatial and/or temporal coverage. Catch-book data can be used but are subject to potential biases in reporting
In the Europe Union, VMS operates on larger vessels of Member States fishing fleets (≥15 m overall length). Such vessels employ a range of fishing techniques to exploit demersal and pelagic fish species (e.g. dredging, beam trawling, pair-trawling, gill netting and longlining). These techniques have their respective degrees of selectivity for both their intended catch species but also non-target species and variable impacts on habitats. For example, small cetacean bycatch is commonly associated with bottom set gill-netting and pair trawling
Here we investigate the utility of data from the UK VMS to describe patterns of at-sea space use by large UK-registered fishing vessels. Such data may ultimately inform seascape scale conservation by feeding into marine spatial planning activities
Mapping of VMS data highlights considerable heterogeneity in space use (
The colour scale indicates the mean annual number of VMS derived data points within 9 km2 pixels, solid line circumscribes the UK declared fishing zone, broken line is 200 m depth contour. Regional labels: Western Channel (WA), Goban Spur (GS), Rockall (RK) and Northern North Sea (NI). b) Tonnes of fish (demersal and pelagic) landed by UK registered vessels from the shown ICES statistical reporting boxes. Total number of vessels registered at main UK fishing ports greater than 17 metres in overall length (filled circles). All vessels for Northern Ireland have been mapped to Belfast. c) Coefficient of variation of the mean annual distribution of fisheries activity, lighter colours indicate areas of greatest variability in space-use, darker areas indicate regions of consistent space-use on annual time-scales. d) Coefficient of variation of the mean monthly distribution of fisheries activity, lighter colours indicate areas of greatest variability in space-use, darker areas indicate regions of consistent space-use on monthly time-scales.
To validate the presented fishing patterns (
It is highly likely that VMS data plots fishing activity with a much greater degree of precision than inferences that could be made from catch-book data. Is this high resolution picture predictable across years and across seasons as would be needed for efficient design of spatially explicit management? When we spatially map coefficient of variation (CV) among years (
VMS was initially conceived to assist in the monitoring and control of fisheries activities and was legislated prior to changes in EU common fisheries policy
The importance of the identified centres of fisheries activity (i.e. Western Channel and Celtic Sea, Northern North Sea and the Goban Spur) can be explained from biological and physical oceanographic perspectives. These are regions where seafloor topography and currents set up physical features that act to support upwelling, enhanced mixing, input of nutrient rich waters, or aid the development and maintenance of frontal systems that aggregate biological matter
With the increased resolution of spatio-temporal patterns of fisheries a step improvement in knowledge of the spatial distribution of species and habitats of conservation concern is required. This requirement has been met, in part, by UK and EU funded research on small cetaceans
Although the VMS approach is a step forward in aiding the development of ecosystem management plans, there are a number of important caveats that must be considered in the interpretation of our findings, which suggest future directions for research. The fisheries activity maps are indicative of the spatial and temporal distribution of large UK-registered fishing vessels only. The patterns are therefore biased towards more offshore fishing activity and represent only a subset of the UK fleet. In addition, we only present data from the UK-registered fleet and not from other EU Member States operating in UK waters. The lack of these data does not detract from the utility of VMS data in providing a spatially and temporally explicit understanding of fisheries activity. Their absence does, however, highlight the need for integration with VMS data from other Member State vessels operating in UK domestic waters. A synoptic European view of fisheries activity will be essential for understanding the relationship between fisheries and migratory target and non-target species as they move seasonally between the waters of distant Member State.
The absence of metadata in the UK VMS on vessel gear type required us to use assumptions on movement speeds that most likely characterise fishing behaviour across several fishing methods employed by larger fishing vessels. In using a narrow range of speeds we believe we have been parsimonious in our estimation of when a vessel might be engaged in fishing. The common factor that a fishing vessel travels at slower speeds during fishing, gear deployment and retrieval, be it demersal or pelagic gear, provides a characteristic, albeit coarse, signal upon which to partition data. Expanding and contracting the width of the speed filter has the effect of widening or constricting the observed spatial patterns; what remain consistent are the identified centres of fisheries activity. Identification of these areas, their spatial range and their seasonality, provides important information for spatial management plans that could seek to manage fish stock extraction while mitigating risk to non-target species and habitats.
Not all fisheries techniques pose the same degree of risk to species and habitats of conservation concern, yet this lack of metadata does not prevent a coarse spatial interpretation of the putative risk posed to these groups as gear types, with their associated risks, are commonly deployed in known depths of water over particular habitat types. Moreover, non-target species adopt fairly predictable habitat utilisation patterns and physical habitats that represent areas of increased biodiversity can be mapped
Recent work to describe trawl intensity received by the seabed
Notwithstanding the caveats, the simple and coherent patterns of habitat occupation by fishing vessels presented here suggest that fishing activity could be managed on a more finely resolved spatial and temporal basis. Furthermore, with multilateral collaboration VMS technologies may offer an important solution to quantifying and managing ecosystem disturbance particularly on the high-seas, which has become evermore important as fisheries move into deeper
The Vessel Monitoring System (VMS) is an automated method of recording the location of fishing vessels at sea. The system consists of a tamper-proof installation onboard fishing vessels registered in the UK and was introduced under European Commission legislation (EC 686/97). Each unit consists of a global positioning satellite (GPS) receiver; a satellite transmitter and a power backup that will last approximately 72 hours
VMS data were obtained from the UK Sea Fisheries Inspectorate in 2005 (now the Marine and Fisheries Agency of the Department for Environment, Food and Rural Affairs). This dataset contained 5,788,188 records. Each record contained geographic coordinates in decimal degrees (World Geodetic System 1984 format) an accompanying time stamp in UTC and a vessel identification number. All received data were anonymous with respect to their vessel registration numbers, dimensions and administrative ports. The mean number of VMS records per year (see
Fishing trips were reconstructed as follows: a 5 km buffer zone was constructed around the coastline of Europe, this was used to determine when vessels were leaving or nearing ports. All records belonging to a vessel were assigned a logical flag (1 or 0) to indicate whether they were inside or outside this coastal buffer zone. The start and finish of a fishing trip was determined when a vessel moved out of and back into the zone with respect to time. Records occurring within the buffer zone were discarded. A speed filter was applied to remove improbable locations; this process removed locations necessitating travel speeds greater than 100 km hr−1 (∼55 knots) between time adjacent locations. The filter was triggered on 1,015 trips and removed 6,891 records.
Potential trips were discarded if they contained ≤3 VMS records, or were ≤6 hours in duration or had transmission breaks ≥5 days; removing 28,800; 12,121 and 168,549 records respectively (in total 3.6% of the original dataset). It is likely that these filters remove some legitimate fishing trips of short duration and may underestimate near-shore fishing effort. However, they were required to minimising the degree of visual supervision needed to manage this large dataset while maximising retention of VMS data. Post filtering the dataset contained 56,434 fishing trips (see
The modal frequency of record transmission was 2 hours (see
A speed rule was used to distinguish fishing from steaming or near-stationery movement. It was necessary to construct derived speeds for all VMS records as prior to 1-1-2006 transmission of speed and heading was not mandatory
The upper and lower speed thresholds for determining fisheries activity were influenced by the frequency distribution of vessel speeds (see
Fisheries activity was gridded at a spatial resolution of 9 km2 (3 km by 3 km pixel) by summing the number of VMS derived data points coincident to each pixel over monthly and annual scales.
a) Number of VMS records (x104) per year, b) number of vessel identification numbers active each year (filled bars) and cumulative increase in vessel identification numbers appearing each year in the VMS dataset (empty bars), c) frequency histogram of time elapsed (hours) between transmission of time adjacent records for all vessels in the 5 year VMS dataset, d) frequency histogram of transmitted and derived speeds (filled and empty bars respectively) for 3,126,042 VMS records, and e) frequency histogram of transmitted and derived headings (filled and empty bars respectively) for 3,126,042 VMS derived data points.
(0.03 MB DOC)
Data handling/filtering process applied to the VMS dataset.
(0.05 MB DOC)
Mean annual maps of fishing activity (vessels moving ≥3 and ≤10 km h 1) for the period 2000–2004. Maps show the mean number of data points at each pixel, where darker colour indicates greater number of visits by vessels travelling at speeds most likely to indicate fisheries activity.
(4.12 MB EPS)
Mean monthly maps of fishing activity (vessels moving ≥3 and ≤10 km h 1) for the period 2000–2004. Maps show the mean number of data points at each pixel, where darker colour indicates greater number of visits by vessels travelling at speeds most likely to indicate fisheries activity.
(9.28 MB EPS)
We thank K. Porter (Department for Environment, Food and Rural Affairs) for assistance in obtaining VMS data. We thank S. Northridge (Sea Mammal Research Unit, UK) and J. Reid (Joint Nature Conservation Council, UK) for constructive criticism during the analytical phase of this work. We thank L. Hawkes for comments on drafts of the manuscript.