General enquiries on this form should be made to:randd.defra.gov.uk/...Document=11219_MF1217_Evid4_F… · Web viewGeneral Enquiries on the form should be made to: Defra, Procurements

General Enquiries on the form should be made to:Defra, Procurements and Commercial Function (Evidence Procurement Team)E-mail: [email protected]

Evidence Project Final Report

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The Evidence Project Final Report is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra websiteAn Evidence Project Final Report must be completed for all projects.

This form is in Word format and the boxes may be expanded, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code MF1217

2. Project title

Low-cost VMS data analysis: Assessment and applications

3. Contractororganisation(s)

CefasPakefield RoadLowestoftSuffolkNR33 0HT

54. Total Defra project costs £ 47,609(agreed fixed price)

5. Project: start date................. 01/07/2010

EVID4 Evidence Project Final Report (Rev. 06/11) Page 1 of 30

mailto:[email protected]

end date.................. 31/08/2012


6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing Evidence Project Final Reports contractors should bear in mind that Defra intends that

they be made public. They should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the Evidence Project Final Report can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the intelligent

non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

Vessel Monitoring Systems (VMS) have been effectively used onboard vessels over 15 metres in length for fisheries management. Increasing numbers of offshore developments present challenges for the sustainable management of the marine environment. Development of an ecologically coherent network of marine protected areas puts further pressure on existing marine industries. The inshore fishing fleet in England faces particular challenges and access to fishing grounds is essential to the families and communities it supports. Spatial management of inshore fishing activities using VMS technology is seen as a solution to management of fishing activities near, or within, marine protected areas.

Rolling out VMS to the inshore fleet has specific challenges; including fitting the equipment on the vessels, as well as requiring improved monitoring ability. Defra funded project MF1214 set out to develop VMS technology that meets the requirements of the inshore fishing fleet. This project complements this work by assessing the solution provided from an end-user perspective.

Overall the results proved promising. VMS technology is now available that allows frequent monitoring of fishing vessels at low cost. The challenges using mobile phone networks for data communication were overcome to provide continuous reporting of vessel position. The quality of the data reported was found to be good, although minor issues remain outstanding with the latest VMS unit provided by Succorfish.

Low-cost communication allows frequent reporting of vessel position, essential for inshore fisheries management, where activities take place over small areas and areas requiring spatial management can be as small as several hundred square metres. The optimal ping interval varies depending on the application, but needs of most applications would be met with a ping interval between 1 and 5 minutes. Larger ping intervals would reduce confidence in reconstructed tracks and could allow vessels to deviate into closed areas

Traditional VMS technology does not differentiate fishing from non-fishing events. MF1214 set


out to develop such a system, but delays during development meant only a single vessel was equipped and subject to evaluation. Results highlighted a number of issues and limitations of the adopted approach, whereby sensors were attached to the gear itself. Further work testing an alternative approach, whereby sensors are attached to the net drum or pot hauler, is recommended as it could offer significant benefits.

Analysis of vessel speed during fishing activity confirmed the assumptions made as part of traditional VMS processing. Vessel speed can be used as a proxy for fishing activity. Most trawling activity was observed between 1 and 5 knots. Most vessels using static gear operated at speeds less than 7 knots during fishing activities. Whereas speed as a proxy is acceptable for providing guidance on where activities take place. For enforcement purposes requiring absolute evidence, speed based rules may not be sufficient.

Low-cost communication could increase data volumes over 120 times compared to traditional VMS systems. Dealing with these large volumes of data has resource implications for managers. Online portals are suitable for real-time observation purposes, but in-depth analysis of the data would benefit from using desktop mapping software with bespoke software tools to reduce the burden on the managers.

The opportunities presented by fishing vessels collecting environmental data for the scientific community and skippers were demonstrated. The volume of data that would become available could provide benefits to ocean modelling and skipper’s ability to target species.

The low-cost VMS could bring significant benefits for the management of fishing activities near, or within, marine protected areas. The work has demonstrated that solutions are now available to address the specific needs and challenges facing the inshore fleet. The analyses undertaken provide guidance to managers in deciding optimal settings to meet the management requirements.

Project Report to Defra


8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Exchange).

1. Introduction

1.1 Overview

In recent years, the use of Vessel Monitoring System (VMS) data has found increasing application in fisheries management and research. Using the Global Positioning System (GPS) network of satellites, VMS systems are able to track fishing vessels with high levels of accuracy. Within Europe, the European Commission introduced legislation in 1997 for the monitoring of European fishing vessels for enforcement purposes using VMS (EC, 1997). Since 1 January 2000, all fishing vessels over 24 metres overall length are required to carry VMS technology and report their position at least every 2 hours. Over time, the requirement extended to smaller vessels and since 2005 all vessels over 15 metres overall length were required to carry an approved VMS system. In 2009 the European Commission published further regulations requiring all vessels over 12 metres length to carry a VMS device from 1 st January 2012. The actual implementation in the UK was delayed whilst type approval was underway.

The European Commission introduced VMS technology primarily for control and enforcement purposes, but the data have found increasing uptake elsewhere. The increasing need for spatially explicit data on where human activities take place in the marine environment has led to VMS data becoming increasingly used in a wide range of applications, such as stakeholder mapping, situating of marine protected areas (MPA), marine planning or understanding fisher's behaviour (Marchal et al., 2007; Witt & Godley, 2007; Stelzenmuller et al., 2008; Lee et al., 2011).

Compared to other methods available to map fishing activity, VMS is often considered to provide high resolution outputs. Whereas this might be true when assessing fishing effort in a national or regional context, the temporal resolution of the data is insufficient to meet the requirements of fisheries enforcement within sensitive marine protected areas (Lambert et al., 2012). One of the obstacles to improving the temporal resolution of the VMS data is the cost of transmitting position information via satellite communication.

The inability to accurately monitor where fishing activities occur requires buffer zones around sensitive features to deal with this uncertainty. The ability to prove fishing activity can occur close to sensitive features, without impacting upon them, may help fishermen to gain access to parts of closed areas that do not contain vulnerable habitats. This has proven an attractive proposition to the fishing industry, whilst at the same time, marine managers and statutory nature conservation bodies (SNCB) are willing to consider this option.

In June 2012 there were 5,684 registered and licensed fishing vessels in the United Kingdom, of which 2,900 were registered in England. A breakdown by vessel length is provided in Table 1 and illustrates the large proportion of vessels under 15m currently not subject to VMS monitoring. These smaller vessels often work closer inshore and alternative communication systems are available as identified in Caslake (2009). These alternative communication systems also allow for increased temporal resolution of the VMS data without significantly increasing the communication costs.

Table 1: Registered and licensed fishing vessels in the United Kingdom and England (MMO, 2012).

Vessel Length United Kingdom England VMS


<12m 4,754 (83.6%) 2,558 (88.2%) No VMS12-15m 277 (4.9%) 172 (5.9%) VMS required

since 01/01/2012 (delayed)

>15m 653 (11.5%) 170 (5.9%) VMS required since 01/01/2005

1.2 Policy background

The UK Government recognises the importance of a prosperous and efficient fishing industry, whilst at the same time providing protection to the marine environment (Defra, 2010). Fisheries 2027 details the current fisheries policy and identifies a sustainable fisheries sector as essential to delivering the Government's vision of clean, healthy, safe, productive and biologically diverse oceans and seas. As part of this, fish stocks will be managed within biological safe limits, and the importance of smaller fishing vessels is recognised for their social, cultural and economic contribution to coastal communities.

Under agreements such as the OSPAR Convention and the Convention of Biological Diversity, the UK has committed to establishing an ecologically coherent network of marine protected area (Natural England, 2012). The network will be made up of Special Areas of Conservation (SACs) and Special Protection Areas (SPAs) which are designated under the European Habitats and Birds Directive. In addition, the network consists of Sites of Special Scientific interest (SSSI) and RAMSAR sites. Finally, Marine Conservation Zones will be introduced under the Marine and Coastal Access Act to protect nationally important habitats (Figure 1).

Figure 1: MPA components in English waters.

In 2012, over 24% of English inshore waters were under European site protection (Defra, 2012). In addition, 127 MCZs have been recommended by the four MCZ Regional Projects for English waters (Figure2). The management of marine protected areas may require restrictions on the type of activities that can take place within the area. While fishing activities may continue in some area, in other areas the conservation objectives may require some, or all, fishing activities to be excluded. In particular bottom trawling activities are considered to have the potential to impact vulnerable seabed habitats and species. Scallop dredging is believed to have a more marked effect than beam trawling, which in turn has a greater effect than otter trawling (Kaiser et al., 2003). Within the fishing industry, there is considerable concern that the introduction of an MPA network will considerably reduce the fishing opportunities.

Stakeholders have welcomed the initiative to assess the role of VMS in managing fishing activities within MPAs.


Figure 2: Location of recommended Marine Conservation Zones.

1.3 Aims and objectives

This project complemented another Defra funded project: MF1214 - Further development and piloting of low cost Vessel Monitoring Technology on English fishing vessels. Whereas MF1214 was primarily concerned with developing a specification for a vessel monitoring system to meet the requirements of the inshore fishing fleet, this project set out to assess whether such system can meet end-user requirements.

Key requirements for a vessel monitoring system include reliability, repeatability, security and accuracy. To that extent, this project set out to review whether the data produced are sufficiently reliable to allow the activities to be monitored adequately. Similarly, the repeatability and accuracy of the data would be assessed to provide end user confidence in the system.

It is widely recognised that, for inshore applications such as access within MPAs, a higher temporal resolution of VMS data is required. While the highest possible logging frequency, providing the most accurate representation of a vessel's movements, may be desirable, this has to be balanced against the volume of data collected and requiring analysis. This project would therefore review appropriate logging frequency ranges for the different end-users.

Vessel monitoring systems are able to provide continuous reporting of a vessel's location. However the main interest in monitoring a vessel's activity is when it is engaged in fishing activities. Current VMS devices fitted onboard vessels over 15 metres overall length do not have any sensors to report whether a vessel is fishing or not. Speed based rules are therefore used as a proxy to indicate whether a vessel is fishing or not (Lee et al., 2010; Gerritsen & Lordan, 2011; Skaar et al., 2011). MF1214 therefore looked at robust options to determine whether a vessel is fishing or not. A solution was proposed by the hardware developer and this project would assess the performance of such system against traditional speed based rules and other validation approaches.

The benefits of the increased temporal resolution of the low-cost VMS system would be demonstrated by comparing results from the VMS used on vessels over 15 metres. A low-cost VMS device was fitted to two vessels also equipped with a traditional VMS unit. The aim of the comparison would be to assess the performance of the low cost VMS unit, but also to provide insights in the validity of the assumptions made when using VMS data from the vessels over 15 metres overall length to develop fishing effort data layers.

A final aim of the project was to use the environmental data collected during the low-cost VMS trials and assess their value to the marine research community.


2. Materials and methods

2.1 Study area

MF1214 chose the southwest of England as the study area for this work. The majority of vessels had their home port along the southcoast of Cornwall, Devon and Dorset. This allowed the project to trial the newly developed technology in a real-life management scenario, such as the Lyme Bay and Torbay candidate Special Area of Conservation (cSAC). Following an invitation to fishing vessels to participate in this trial, 31 vessels were selected. The home port of the vessels and the number of vessels participating in the trial per port is illustrated in Figure 3. The 31 vessels covered a range of vessels lengths, from less than 6 to over 15 metres, and covering different static and mobile gear types.

Figure 3: Location of the home port and number of participating vessels in the trial per port.

2.2 Low-cost VMS technology and data

This project made use of low-cost VMS technology developed under project MF1214. A summary is provided here for reference. Full details on the development process and technical specification can be found in the MF1214 Final Report (Rossiter & Caslake, 2012).

Following a tender process, a joint bid by Selex and Succorfish was awarded the contract to supply 31 low-cost VMS units. The SC1 unit from Succorfish was a tried and tested unit which had been used in many terrestrial and defence tracking applications. Unlike traditional fishing vessel monitoring systems sending positional information using satellite communication means, the SC1 unit sends its data over the terrestrial based mobile phone network (Figure 4). The benefit of this technology is that communication costs are considerably lower and therefore data can be sent more frequently, improving the temporal resolution of the VMS data. However, unlike satellite communication which is available worldwide, the SC1 unit could only report the position of this fishing vessel if it was located within range of the mobile phone communication network.

From the start, Succorfish recognised this shortcoming and as part of their tender included a commitment to develop an enhanced unit, SC2, which would overcome the issues identified as part of the initial trial, as well as ensuring continuous data reporting.


Figure 4: (Left) Traditional VMS system using satellite communication; (Right) VMS using mobile phone network for data communication.

The first SC1 devices were fitted in June 2010, with installation on all 31 vessels completed in August 2010 (Figure 5). Although it was originally anticipated that the enhanced SC2 unit would be deployed from January 2011, technical and logistical delays meant that the first SC2 unit was only installed by April 2012. In addition, issues were experienced with the initial data from the new SC2 units. As a result, time available to analyse data from SC2 units was limited and this report therefore focuses on the outcomes of the SC1 data analysis.

The logging frequency was initially set at 10 minutes but was reduced to 1 minute to provide a high resolution dataset for reference purposes. Within the system specification, there was a tolerance for the VMS pings to be sent up to 10 seconds after the 1 minute interval.

Figure 5: Example of SC1 installation onboard one of the vessels participating in the trial. Vessel: Nil Desperandum - DH390.

2.3 Fishing activity sensors

A fishing vessel which is steaming has no impact on sensitive seabed habitats, as opposed to a vessel towing fishing gear over the seabed within a sensitive area. Technical solutions to distinguish fishing from non-fishing activities were therefore investigated under project MF1214. The development of the enhanced SC2 unit included the ambition to integrate a sensor which would provide this detail. In the spirit of low-cost technology, a trial was undertaken of a system using radio frequency identification (RFID) tags and sensors. The system is based on a proximity sensor and a tag. The principle of the technology is illustrated in Figure 6. When attached to fishing gear, each time the gear goes overboard or comes back onboard, the tag would pass near a sensor, sending a signal to the VMS unit reporting the gear has passed near the sensor. Alternatively, the tags could be attached to net drums or pot haulers. The circular movement of the winch would send a sequence of signals to the VMS unit from which the activity of the vessel can be determined. Delivery of the SC2 VMS units was delayed significantly and issues were experienced when the system was first delivered. As a result, only data from a single vessel, for a very short period of time, were available for analysis within this project.


SC1

Figure 6: Schematic illustrating the principles of the RFID solution.

Recognising the need to understand when vessels were fishing and not fishing, and while awaiting the delivery of the RFID system, vessels were provided data storage tags (DST). Cefas Technology Limited G5 DSTs provided pressure and temperature information. DSTs were attached to fishing gear in a protective housing (Figure 7). Data were downloaded manually by Seafish staff and provided to Cefas for analysis. When fishing gear was deployed, the pressure data showed an increase, allowing the differentiation of fishing and non-fishing activity. Overall, the G5 tags performed well, although some units failed due to the conditions to which they were exposed during fishing operations.

Figure 7: (Left) Cefas Technology Limited G5 data storage tags; (Right) DST in protective housing attached to fishing gear.

2.4 Data analysis

All data analysis was undertaken by experienced Cefas staff. Cefas has, for many years, analysed VMS data from vessels over 15 metres overall length and this project built on the experience and tools developed during that time. A data import tool was modified to deal with the SC1 data and visualise the data in a geographic information system (GIS). The data could then be filtered for duplicates, trip identifiers added and tracks generated, using the previously developed tools. The tool also allowed for fishing effort layers to be derived using basic speed rules.

Data analysis was undertaken using ArcGIS 9.3.1 mapping software and the statistical computing package R. Analysis for logging frequency and track offsets made use of the VMSTools package in R (Hintzen et al., 2012). The tool was used to compare the offset between the real track (based on 1 minute VMS polling) and tracks generated using straight line interpolation at different VMS ping intervals.

Data for the period September 2010 - May 2011 were primarily used for the analysis of the SC1. Data from May 2012 was used for the analysis of the SC2 unit.


Figure 8: Customised tools for import and initial data analysis of the SC1 VMS data.

3. Results


3.1 VMS pinging frequency

Reduced cost and improved data resolution were identified as key requirements for an inshore VMS system. Although the Succorfish SC1 unit is able to report vessel location, speed and heading at intervals of just a few seconds, the quantity of data produced at such high frequency results in a large and complicated data analysis which may not be necessary to meet the requirements of the end user. Therefore, although frequent VMS pings may be desirable, a balance has to be found between volume of data and end-user requirements. The assessment of the optimal ping ranges for different end-users is explored through a number of approaches.

3.1.1 Deviation from real vessel track

Figure 1 illustrates the effect of increasing ping intervals on the potential of a vessel to deviate between two VMS pings. At one minute intervals (red dots) the vessel reports its position frequently and has very little scope to move far from the actual track drawn between the VMS pings. As the ping interval increases, the uncertainty in the actual track followed by the vessel increases, and the path followed by the vessel can increasingly be further away from the straight line interpolation between the two reported positions. The optimal VMS ping frequency should therefore be determined by the spatial accuracy requirements of the reconstructed track. Based on the example below, if it is not acceptable for a vessel to be a distance X15 away from the actual track, then a ping frequency has to be chosen of less than 15 minutes.

Figure 9: Schematic illustrating influence of increasing ping interval on vessel tracks.

We calculated the maximum distance a vessel can deviate from a straight line interpolation between two VMS pings by increasing its speed to inform the assessment of optimal ping frequency. Figure 10 presents the results from the analysis and demonstrates that, at low ping intervals, the maximum distance a vessel can deviate by increasing its speed between two VMS pings is in the order of tens of metres. These graphs present theoretical results, where a vessel could increase its speed from 1 knot to 6 knots and make abrupt turns. In reality, when a vessel is engaged in fishing activities, its ability to make such sudden changes to its speed and course are limited. The outcomes from this assessment should therefore be treated as a worst case scenario and in real-life situations the ability to travel a distance away between two VMS pings is likely to be less. Assuming that the real-world situation is half of the theoretical model, Table 2 presents the distances a vessel can remove itself from the straight line interpolation between two VMS pings in a given time interval. In real world situations we assume it is less likely that a vessel will increase its speed by 3 or 5 knots. Where the vessel starting speed is low e.g. 1 knot, if the vessel were to increase speed by 0.5 knots between pings, the vessel would only be able to deviate 9 metres from the straight line interpolation. Doubling the ping interval will double this distance. At a speed increase of 1 knot and a ping interval of 1 minute, the distance away from the straight line interpolation can be 13 metres. If the initial speed of the fishing vessel is 3 knots, the 0.5 knot increase means that the vessel can be up to 14 metres away from the straight line interpolation. At 120 minute intervals, the minimum ping interval as required by EU legislation, it can be seen that the vessel can be one, or several, kilometres away from the straight line interpolation track.


Figure 10: Modelling of vessel distance from interpolated track between 2 VMS pings at different reporting intervals. Compared against field observations from SC1 data (black line).

Using the approach presented by Lambert et al. (2012), we considered the 1 minute interval data to be the true positional track of the fishing vessel. Using the functionality of the VMSTools package we extracted the average offset from the real track to the interpolated straight line track. The result is included as the black line in Figure 10 and results are also included in Table 2. The field based results show that they are generally lower than the theoretical distance during that time period. The results are also in line with the findings from Skaar et al. (2011) who calculated average deviations of 500 metres at 1 hour ping intervals and 1,200 to 1,700 metres deviation at 2 hours ping interval. The 50% reduction on the theoretical distance is confirmed as a realistic assumption in this instance. Based on the tolerance specified by end-user requirements, this allows optimal VMS ping ranges to be set.

Table 2: Maximum distance away (in metres) from straight line interpolation between two VMS pings for different vessel speed increases relative to original speed of 1 or 3 knots. SC1 column shows actual distance based on field observations.

1 knot 3 knots SC10.5 kt increase 1 kt increase 0.5 kt increase 1 kt increase data

1 min 9 13 14 20 n/a2 min 17 27 28 41 105 min 43 67 70 102 21

10 min 86 134 139 204 7015 min 129 200 209 306 14030 min 259 401 417 612 39660 min 518 802 835 1,225 758

120 min 1,035 1,604 1,669 2,450 914

3.1.2 Vessel track reconstruction

Another approach to assess appropriate VMS ping intervals was based on the ability to reconstruct the original vessel track. The analysis was undertaken separately for static and mobile gear vessels, as they manoeuvre differently during fishing operations. A dataset for one month was used, consisting of 22,289 VMS pings from a mobile gear vessel and 38,445 VMS pings from a static gear vessel. This was further reduced to exclude data whilst in port or covering extensive steaming distances. This dataset was then sub-sampled to 1, 2, 5, 10, 15, 30, 60 and 120 minutes. Separate track lines were generated for each fishing trip based on a straight line interpolation principle. Interpolated track lines could be considered in combination with larger ping intervals (Hintzen et al., 2010; Russo et al., 2011), but such approach was not tested as part of the current assessment. The total line length for each track was calculated for the different VMS ping intervals.

Figure 11 presents the results for a static gear vessel operating off the south Devon coast and the results are summarised in Table 3. The graph shows a rapid decline in the ability to reconstruct the vessel's track accurately as the VMS ping interval increases. At a ping interval of 30 minutes, the reconstructed vessel track based on straight line interpolation is less than 50% the length of the vessel track reconstructed from the 1 minute VMS ping interval. Similarly, Figure 12 presents the results for the fishing vessel using mobile gears. The results are very similar, although there is more variation between the different vessel tracks


generated. The summary results presented in Table 4 are very similar to the summary results for the static gear vessel. For applications where it is important to be able to reconstruct the vessel's track, e.g. enforcement purposes or assessments of seabed area impacted by fishing activities, the higher VMS ping rate is therefore essential.

Figure 11: Track length at different ping intervals relative to full density track length for a static gear vessel. Vertical red lines indicate selected ping intervals.

Figure 12: Track length at different ping intervals relative to full density track length for a mobile gear vessel. Vertical red lines indicate selected ping intervals.

Table 3: Summary of track reconstruction ability at different ping intervals for a static gear vessel.

Ping interval

Percentage of true track length

1 100%2 96.2%5 88.2%10 77.6%15 68.9%30 45.1%60 26.5%

Table 4: Summary of track reconstruction ability at different ping intervals for a mobile gear vessel.

Ping interval

Percentage of true track length

1 100%2 97.1%5 86.7%10 66.7%15 53.5%30 26.8%60 21.9%


120 18.8% 120 16.4%3.1.3 Relation to management features

Whereas the previous sections dealt with the more practical differences of the different logging frequencies, another approach to recommend optimal ping frequencies was explored in relation to management features. The use of high density VMS data is envisaged as part of a solution for managing fishing activities near, or within, marine protected areas. From an enforcement perspective, the VMS ping frequency therefore needs to be sufficient to provide evidence that the vessel did, or did not, enter a restricted area. Evidence that a vessel entered a restricted area would require at least one VMS ping within the area, and therefore the VMS ping frequency would need to be proportional to the size of the management features. Figure 13 illustrates the distance travelled by a fishing vessel at a speed of 6 knots for different time intervals. Six knots is considered the fastest speed at which a mobile gear vessel would operate (Lee et al., 2010). The potential distance travelled is compared against a typical marine conservation zone reference area, the average recommended marine conservation zone diameter, and a one square kilometre box for reference purposes.

Figure 13 shows that only at one- or two- minute intervals can at least one ping be guaranteed within the smallest of the management features considered in this example: the MCZ reference area of 500 metres diameter. For the larger, average sized, MCZ, it can be seen that a ping frequency of up to one hour would be able to guarantee at least one ping within the area should a fishing vessel travel across the area. Figure13 allow managers to assess optimal ping frequency based on the size of the management features of interest.

Figure 13: Distance travelled by a vessel at speed of 6 knots in relation to different MPA sizes.

3.1.1 Fishing effort mapping

A final assessment of the impact of different VMS ping regimes was made in the context of fishing effort mapping. In this case the positional accuracy of the track is less relevant as the importance lies in accurately representing the spatial intensity of the activity.

Using a subset of data from vessels operating south of Lizard Point, Cornwall, the entire VMS dataset was summarised at a resolution of 0.01 degrees (approximately 700 by 1,300 metres) and 0.05 degrees (approximately 3.5 by 5.5 kilometres) (Figure 14 and Figure 15 respectively). Both datasets show that, as the time interval increases, the correct identification of the area where fishing vessels operate is reduced. For the smaller grid size the reduction is significant, although this can partly be explained by the grid resolution being unsuitable for data collected at such a large ping intervals. Another observation that can be made from these comparisons is the relative increase in intensity of areas where activity is highest. As the ping interval increases, the effort from areas with low activity levels is redistributed to those areas where most of the activity takes place. As such the increased ping interval reduces the extent of the area affected. For impact studies the first trawl is often considered the most damaging (Hiddink et al., 2006) and the


optimal ping interval and grid size have to be considered carefully.

Finally, Figure 16 demonstrates the impact of the different cell sizes themselves on the resulting fishing effort maps. Firstly, the larger grid cell size portrays the area affected as being significantly larger than does the smaller cell size. Secondly, the larger cell size also accumulates an increased amount of activity within it. These two facts combined give the perception of intensive activity over large areas and the end-users should be aware of this perception when managing activities using VMS data.

Figure 14: VMS effort summary maps using cell resolution of 0.01 degrees at different ping intervals.


Figure 15: VMS effort summary maps using cell resolution of 0.05 degrees at different ping intervals.


Figure 16: Comparison between different grid cell sizes and apparent intensity of fishing activities.

3.2 SC1 data assessment3.2.1 Reliability

The technical specification provided to the technical delivery partner, in this case the partnership between Selex and Succorfish, required the system to send position reports at regular intervals. When the vessel was set to ping at 60 seconds interval, the system was required to submit a report at least every 70 seconds. Figure 17 presents a ping interval histogram for a single SC1 unit for a vessel which mainly operates within inshore waters. When displaying the graph using a normal linear scale (Figure 17 inset) it can be seen that the vast majority of the VMS pings were recorded within 60 to 70 seconds of each other. To visualise the remainder of the data, Figure 17 presents the data on a logarithmic scale and demonstrates that there is a decrease in number of VMS pings with increasing ping interval. For this particular vessel, 93.2% of VMS pings had a ping interval between 60 and 70 seconds. 5.9% had a ping interval of 70 to 80 seconds, together accounting for over 99% of all VMS pings.

Figure 18 presents similar results for the entire SC1 fleet between August 2010 and May 2011. Over 85% of the VMS records were found to have a ping interval of less than the required 70 seconds. There was a decrease in the number of VMS pings as the time interval increased.


Figure 17: Actual ping interval histogram derived from a single SC1 unit. Inset shows same graph without logarithmic scale on the frequency axis.

Figure 18: Percentage of VMS pings within defined ping intervals for the entire SC1 fleet.

The results for the entire SC1 fleet had a lower proportion of pings within the required 70 second ping interval. This difference was explored through detailed investigations of the VMS data. One source of this difference was found to be as a result of regular skipping of a VMS ping. Although not observed for all SC1 units, this artefact was observed in 5 of the SC1 units for an unknown reason. Figure 19 demonstrates how 10 VMS pings are received at the required 60 second interval, but the 11th VMS ping is only received after more than 120 seconds. This artefact was observed regardless of distance from shore and in different geographical locations. The issue was observed in only a few units throughout the trial and is therefore believed to be hardware or firmware related.


Figure 19: Skipping of regular 1 minute VMS pings.

3.2.2 Spatial coverage

A second issue affecting the reliability of the ping interval is that of the communication network. The SC1 relies solely on mobile phone networks for the communication of its data. The system makes use of a roaming SIM card, using any of the available networks, rather than being limited to a single provider. For terrestrial applications this solution has proven very reliable, but for marine applications this leads to some limitations. Figure 20 demonstrates that closest to shore the proportion of VMS pings exceeding the 70 second interval is relatively low, often less than 1%. However, as the distance from shore increases, the proportion of VMS pings with an interval over 70 seconds in each grid cell increases. Beyond 12 nautical miles from shore the mobile phone network coverage seems to reduce rapidly and increases the number of grid cells with a large proportion of ping intervals exceeding 70 seconds. Figure 21 provides a more detailed assessment of the SC1 reliability relative to the distance from shore. Few VMS pings are received beyond 16 nautical miles, while between 10 and 16 nautical miles off shore the proportion of VMS pings with intervals exceeding 70 seconds increases. Ninety percent of the grid cells with a low proportion of VMS pings intervals over 70 seconds (less than 5%) can be found within 12 nautical miles from shore. Ninety-five percent of grid cells with a low proportion of VMS pings intervals over 70 seconds (less than 5%) can be found within 14 nautical miles from shore, whereas the 99% range is found at 23.0 nautical miles offshore.


Figure 20: VMS regular ping reliability illustrated as the percentage of VMS pings over the required 70 second period.

Figure 21: (Left) Proportion of VMS pings exceeding 70 seconds in relation to distance from coast.; (right) Cumulative frequency distribution of the distance to coast for grid cells where proportion of pings > 70 seconds is <0.05 (i.e. grid cells where 95% of the pings are ‘good’).

Although the mobile phone network coverage mainly causes data loss further offshore, it should be noted that is some near-shore locations poor network coverage also resulted in data loss (Figure 22). The complex topographic nature of the South Devon coastline is one such area where the mobile phone signal does not reach certain areas close to shore. This is confirmed by the predicted coverage maps provided by the mobile phone providers in the UK. As an example, the Vodafone predicted coverage is presented in Figure 23 and confirms "Limited coverage" in the area where data loss had occurred. The network coverage was also found to vary from day to day depending on weather and atmospheric conditions.

Reliance on an incomplete and unpredictable communication network is a major limitation for the effective monitoring of vessel movements using these methods. The SC1 unit is only able to send data in real-time, therefore when the network signal is poor, or absent, no data are sent. A recommendation was therefore made for the SC2 unit to store position data in its internal memory and release the data when the vessel returns within mobile phone coverage.


Figure 22: Poor network coverage causing data loss in nearshore areas.

Figure 23: Predicted Vodafone coverage for the area presented in Figure 22. (Source: http://www.vodafone.co.uk/vodafone-uk/personal-coverage/index.htm - 11th August 2012).

3.2.3 Accuracy assessment

A final part of the reliability assessment was to review the accuracy of the SC1 data. No field tests were undertaken comparing the SC1 with a known high accuracy GPS receiver. Instead, basic comparisons were undertaken to provide an indication of the accuracy of the SC1 data rather investigate the absolute accuracy.

Firstly, accuracy of the SC1 unit included a review of the stability and accuracy of the GPS signal. The data for vessels whilst in port was reviewed. Where it was expected that the vessel was moored alongside a pontoon or quay, the data were investigated further. Figure 24 presents the results for a single vessel moored overnight. The vessel was moored alongside a pontoon and the reported GPS positions were all within 5 metres of the central position. The accuracy of the GPS network is specified at less than 7.8m at 95% confidence. The data prove the SC1 system operates within these accuracy limits, although occasional outliers were observed in the dataset.

Figure 24: Accuracy assessment of the GPS signal received by the SC1 unit.

A second assessment included comparison of reported speed and heading values against calculated speed and heading. The calculated speed was obtained by measuring the distance between data points and using the time difference between them. Heading was derived between successive VMS pings. Where a gyrocompass provides heading information, vessel heading is not necessarily the same as the direction of


travel of the vessel. However, the heading information from the SC1 data is derived from GPS signal and should therefore be similar to the direction of travel of the vessel, or to the calculated direction between two data points.

A comparison of the heading data for a single vessel during March 2011 is presented in Figure 25 and shows a good correlation between the reported and calculated heading values. A similar correlation between speed reported by the SC1 unit and the calculated speed is presented in Figure 26. Again, the correlation between both datasets was good, confirming the accuracy of the speed values from the SC1 unit.

Figure 25: Correlation between reported and calculated heading from SC1 unit in March 2011 (r 2 = 0.502).

Figure 26: Correlation between reported and calculated speed (in knots) from SC1 unit in February 2011 (r2 = 0.838).

3.3 SC2 data assessment


The SC2 assessment was undertaken on a very small sample of data and the results in this section should therefore be treated with care. Delays in delivery of the SC2 unit meant that data delivery was delayed by more than one year. Once installed on a few vessels, teething problems caused further difficulties in obtaining a suitable dataset for this assessment. The data used in this section was primarily derived from a single vessel participating in the MF1214 trial. In addition, Succorfish made available data from SC2 units onboard other vessels not part of the MF1214 trial for assessment purposes. These latter data were incorporated to undertake certain aspects of the assessment, but for confidentiality reasons cannot be presented in their original form in this report.

3.3.1 Reliability

A first assessment was based on the SC2 unit installed on a fishing vessel operating from the East Devon coast. The SC2 unit was designed to overcome the loss of data when the vessel moves outside mobile phone network coverage, by storing the data in its internal memory and releasing it when mobile phone coverage returns. As a result, large ping intervals should no longer affect the data resulting from SC2 units. The SC1 system reported on average over 85% of the VMS pings within the required 70 seconds. However, data from the SC2 unit suggested only 66% of VMS pings were received within 70 seconds (Figure 27). A further 32% of pings were within 70 to 130 seconds. This reduction is not believed to be the result of poor mobile phone network coverage, as very few pings had a very long ping interval, often associated with vessels moving outside mobile phone network coverage.

Further investigation of the data reveals that the majority of the VMS pings exceeding 70 seconds where within 70 to 80 seconds from the previous report (Figure 28). Some SC1 systems showed every tenth ping having a larger ping interval. This was similar to issues observed in few SC1 units (Figure 19), however for the SC2 system tested, it seemed to be the case for every third ping (Figure 29). This suggests the source is more likely to be the SC2 hardware or firmware that needs further refinement, rather than any issues related to loss of data as a result of mobile phone coverage.

Figure 27: Percentage of SC2 pings within ping interval ranges.

Figure 28: Detail of the number of SC2 pings at 10 second ping interval ranges.


Figure 29: Skipping of regular 1 minute VMS pings for SC2 unit.

When investigating the SC2 data in more detail for all units, a lot of records with duplicate timestamps were found to present in the dataset. The histogram for one of the units not part of the MF1214 trial showed a significant proportion of VMS pings with a time difference of less than 1 second (Figure 30). While it might be expected that duplicates in time would also be duplicates in space, the data from the SC2 unit showed that in many cases VMS pings released within the same second were hundreds of metres apart (Figure 31). At present it is not known why the SC2 data record these duplicates. For enforcement purposes confidence is needed in the time and location details provided by the system. Manufacturers have been made aware of this issue and are seeking to resolve this as part of further improvements to the system.

Figure 30: Full time difference histogram showing large proportion of duplicate records (time difference = 0).


Figure 31: Mismatch in position between duplicate data points in time.

3.3.2 Spatial coverage

The SC2 unit was designed with the aim of resolving one of the major issues with the SC1 unit, the dependence on continuous mobile phone coverage for data reporting. The SC2 unit was therefore designed to store reports in its internal memory during periods where no connection could be made to the mobile phone network. When the SC2 unit would connect again to the mobile phone network, the data stored in its memory would be transferred to the land-based server. Only once the data transfer was verified would the data be removed from the internal memory.

In absence of an extensive SC2 dataset, no coverage map as presented in Figure 20 for SC1 data could be developed from SC2 records. The data for the single MF1214 trial vessel was plotted on top of the SC1 coverage map to see whether any data were collected in areas where data had previously been lost (Figure32). Unfortunately, the SC2 data were collected within 12 nautical miles from shore, where previously few issues were experienced with data loss. Based on the sample dataset available it could therefore not be assessed whether the unit was indeed correctly storing and releasing the data held in its internal memory.

Figure 32: Assessment of spatial coverage of SC2 data in relation to SC1 coverage (Background data as presented in Figure 20).


Succorfish made data available from a vessel equipped with an SC2 unit, operating elsewhere within UK waters. The vessel was 260 miles from its port and at least 80 miles from land. Mobile phone operators coverage charts show poor coverage in the nearby areas and therefore coverage would not extend out to the location where the vessel operated. The vessel plot demonstrates that no data were lost and that the system performed as expected. The initial SC2 results therefore seem very encouraging in terms of achieving the objective of storing VMS data whilst out of range and releasing on return.

Figure 33: Continuous SC2 data reporting whilst vessel operating offshore, outside mobile phone coverage. Location and coastline not shown due to data confidentiality.

3.3.3 Accuracy assessment

A final part of the reliability assessment was to review the accuracy of the SC2 data. No field tests were undertaken comparing the SC2 with a known high accuracy GPS receiver. Instead, basic comparisons were undertaken to provide an indication of the accuracy of the SC2 data rather than investigate the absolute accuracy. Figure 34 presents the data whilst the vessel equipped with a SC2 unit was in port. The accuracy of the GPS network is specified at less than 7.8m at 95% confidence. The data show that the majority of VMS pings are within 7.8 metres of the average position. A few data points were observed outside the 7.8m buffer, which may be attributed to the degraded GPS signal due to tall buildings along the quayside. It was also noted that the data provided was reduced to 1 metre resolution in latitudinal direction. The data suggest that overall the SC2 system provides accurate GPS positions, although occasional outliers were observed and can be expected in the dataset.

Figure 34: Assessment of the accuracy of the GPS signal received by the SC2 unit.

A second assessment included comparison of reported speed and heading values against calculated speed EVID4 Evidence Project Final Report (Rev. 06/11) Page 27 of 30


References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.Caslake, R. (2009) SR617 Seafish VMS project report. SeaFish Final Report, 18p.

Defra (2010) Food 2030. Defra, London, 81p.

Eastwood, P. D., Mills, C. M., Aldridge, J. N., Houghton, C. A., and Rogers, S. I. (2007) Human activities in UK offshore waters: an assessment of direct, physical pressure on the seabed. ICES Journal of Marine Science, 64: 453–463.

EC (1997) Commission Regulation (EC) No. 1489/97 of 29 July 1997 laying down detailed rules for the application of Council Regulation (EEC) No. 2847/93 as regards satellite-based vessel monitoring systems. Official Journal of the European Union, L202: 18-13.

Gerritsen H. and Lordan C. (2011) Integrating vessel monitoring systems (VMS) data with daily catch data from logbooks to explore the spatial distribution of catch and effort at high resolution. ICES Journal of Marine Science, 68 (1), 245-252.

Hiddink, J.G., Jennings, S., Kaiser, M.J., Quelros, A.M., Duplisea, D.E. and Piet, G.J. (2006) Cumulative impacts of seabed trawl disturbance on benthic biomass, production, and species richness in different habitats. Canadian Journal of Fisheries and Aquatic Sciences, 63: 721-736.

Hintzen, N.T., Piet, G.J. and Brunel, T. (2010) Improved estimation of trawling tracks using cubic Hermite spline interpolation of position registration data. Fisheries Research, 101: 108–115

Hintzen, N.T., Bastardie, F., Beare, D., Piet, G., Ulrich, C., Deporte, N., Egekvist, J. and Degel, H. (2012) VMSTools: open-source software for the processing, analysis and visualization of fisheries logbook and VMS data. Fisheries Research, 115-116: 31-43.

JNCC (2008) UK guidance on defining boundaries for marine SACs for Annex I Habitat sites fully detached from the coast. Online: http://jncc.defra.gov.uk/pdf/SACHabBoundaryGuidance_2008Update.pdf [Last visited: 31st July 2012]

Kaiser, M.J., Collie, J.S., Hall, S.J., Jennings, S. and Poiner, I.R. (2003) Impacts of fishing gear on marine benthic habitats. In: Responsible Fisheries in the Marine Ecosystem. CABI Publishing, Wallingford, pp. 197-217.

Lambert, G. I., Hiddink, J. G., Hintzen, N. T., Hinz, H., Kaiser, M. J., Murray, L. G., and Jennings, S. (2012) Implications of using alternative methods of vessel monitoring system (VMS) data analysis to describe fishing activities and impacts. ICES Journal of Marine Science, 69: 682–693.

Leblond, E., Berthou, P., Laurans, M., Woerther, P. and Quemener, L. (2008) The Recopesca project: a new example of participative approach to collect in-situ environmental and fisheries data using fisheries vessels of opportunity. Presentation at ICES Annual Science Conference 2008, 22-26 September, Halifax (Canada). ICES CM 2008/R:16.

Lee, J., South, A. B. and Jennings, S. 2010. Developing reliable, repeatable, and accessible methods to provide high-resolution estimates of fishing-effort distributions from vessel monitoring system (VMS) data. - ICES Journal of Marine Science, 67: 1260-1271.

Marchal, P., Poos, J-J. and Quirijns, F. (2007) Linkage between fishers' foraging, market and fish stock density: examples from some North Sea fisheries. Fisheries Research, 83: 33-43.

Mills, C. M., Townsend, S. E., Jennings, S., Eastwood, P. D., and Houghton, C. A. (2007) Estimating high resolution trawl fishing effort from satellite-based vessel monitoring system data. – ICES Journal of Marine Science, 64: 248–255.


MMO (2012) UK fishing vessel lists - 1st June 2012. Online: http://www.marinemanagement.org.uk/fisheries/statistics/vessel.htm [Last visited: 20th June 2012]

Natural England (2012) Marine Protected Areas. Online: http://www.naturalengland.org.uk/ourwork/marine/mpa/default.aspx [Last visited: 31st July 2012]

Piet, G. J., Quirijns, F. J., Robinson, L., and Greenstreet, S. P. R. (2007) Potential pressure indicators for fishing, and their data requirements. ICES Journal of Marine Science, 64: 110–121.

Reese, D. C., O'Malley, R. T., Brodeur, R. D., and Churnside, J. H. (2011) Epipelagic fish distributions in relation to thermal fronts in a coastal upwelling system using high-resolution remote-sensing techniques. – ICES Journal of Marine Science, 68: 1865–1874.

Rossiter, T. and Caslake, G. (2012) Further development and piloting of low cost Vessel Monitoring Technology on English fishing vessels. Defra Project MF1214, Final report: 32p.

Russo, T., Parisi, A. and Cataudella, S. (2011) New insights in interpolating fishing tracks from VMS data for different métiers. Fisheries Research, 108 (1): 184-194.

Skaar, K. L., Jørgensen, T., Ulvestad, B. K. H. and Engås, A. (2011) Accuracy of VMS data from Norwegian demersal stern trawlers for estimating trawled areas in the Barents Sea. ICES Journal of Marine Science, 68: 1615–1620.

Stelzenmüller, V., Rogers, S. I. and Mills, C. M. 2008. Spatio-temporal patterns of fishing pressure on UK marine landscapes, and their implications for spatial planning and management. – ICES Journal of Marine Science, 65: 1081–1091.

Vanstaen, K. and Silva, T. (2010) Developing a National Inshore Fisheries Data Layer from Sea Fisheries Committee and Marine Management Organisation Data. Final Report to Defra, MB0106. Cefas, Lowestoft, 115p.

Witt, M.J. and Godley B.J. (2007) A step towards seascape scale conservation: using vessel monitoring systems (VMS) to map fishing activity. PLoS One 2: 10. 10


General enquiries on this form should be made to:randd.defra.gov.uk/...Document=11219_MF1217_Evid4_F… · Web viewGeneral Enquiries on the form should be made to: Defra, Procurements

Documents