6 COMMERCIAL FISHERIES: DATA SOURCES AND MEASURES
A key stage in this study is establishing the benchmark information on zones 0-1 NM, 1-3 nm, 3-6 NM and 6-12 NM. This needs to done for each IFG plus Shetland. Ideally, estimates for catch, number of vessels by gear type, number of days by gear type would enable a comprehensive description of what is currently happening in these zones in each IFG. In addition estimates for the costs of gear conflict would be highly desirable.
The most important variable for the model is the value of catch for each zone in each IFG. A significant problem is that we do not have direct observations for catch by zone. There are however landings data and, even allowing for misreporting by fishermen and other errors, the belief within MS is that landings data are an accurate record of the fish coming ashore in each IFG area. Indeed, they can be checked against the records of registered buyers.
Although there are good landings data for each IFG, the landings data are not a good proxy for IFG catch. Even if it was a good proxy, there is still the problem of disaggregating the data to specific zones within each IFG (i.e. 0-1, 1-3 and 3-6 NM).
There are a number of reasons for the disparity between total landing in an IFG and catches in the IFG's water. For example, the landings in a "hypothetical home" IFG will include fish caught by:
- Home IFG's vessels fishing in neighbouring IFG waters.
- Home IFGs vessels fishing outside 6 miles.
- Neighbouring IFG vessels fishing outside 6 miles close to the home IFG's area.
- Neighbouring IFG vessels fishing neighbouring IFG's area but landing in the home IFG because of better port facilities, buyers, processing capacity.
- Nomadic vessels passing through the area.
Thus, the home IFG landing data could include a proportion of fish caught outside 6 miles or within 6 miles but in another IFG area. For the same reasons, a proportion of the recorded landings of other IFGs will include fish caught in the hypothetical home IFG. Thus IFG catch could be greater or less than recorded landings.
It has been necessary to use a range of data sources and procedure to compile the catch benchmarks. The method used depended on whether vessels were over 15 metres or 15 metres and under.
6.1 Fishing Activity and Catch by Vessels over 15 m
As described in Section 2, data from vessels over 15 m is collected by combining data from the Vessel Monitoring System ( VMS) positioning system and an electronic log produced by a specified Electronic Reporting System ( ERS).
VMS transmits a "ping" every two hours. This records the latitude and longitude of the vessel when the ping made. A key problem is identifying when a vessel is fishing as opposed to progressing to/from a fishing ground. By examining the distance covered between "pings" the VMS system provides an estimate of the speed of the vessel, and the speed of the vessel is largely determined by the activity it is undertaking. In port, the speed is zero and when travelling it is normally in excess of 5 knots. Thus any periods fishing can be identified by a vessel speed between 0 and 5 knots.
This approach is not perfect. A vessel proceeding carefully to and from a harbour may record a speed in that range and some trawls may be slightly faster than 5 knots. In coastal areas with strong tide flows, vessel steaming might occasionally be falsely identified as towing. Tests suggest that inaccuracies are limited, though the severity of the error will vary across geographical areas.
As described in the previous section, ERS provides details of the volume and value of catch landed. Vessel landings and VMS data can be linked through the vessel's identification details. In linking vessels' recorded landings to where it was caught, the approach taken by Marine Scotland is to allocate each vessel's total catch equally across every period's fishing. Thus if a vessel recorded 1500 kg over a week and there were 5 estimated "fishing" pings, each ping would be allocated a catch of 300 kg.
Marine Scotland allocates pings to ICES sub rectangles for further analysis at EU level. For the purposes of this study, it might have been possible to use this level of analysis. This would have required the matching of sub-rectangle to IFG's and also to zones (e.g. 0-1, 1-3 and 3-6 NM). This was rejected because the complex of sea lochs and zones within IFGs results in sub-rectangles overlapping land areas and IFGs, as well as zones within IFGs. Thus it was decided to allocate each ping to an IFG area using the ping's geographic co-ordinates and a GIS system.
6.1.1 VMS and IFG Zonal Catch Estimates for Vessels Over 15m
The GIS package utilised was Arc- GIS Desktop and the first task was to define the areas needed in the analysis. The Coastline was that defined by the Ordinance Survey ( OS) for the Strategi map series. The Buffer Function was then used to define polygons (shape files) 1 NM, 3 NM, 6 NM and 12 NM from the shore. Using the polygon boundaries and the IFG boundaries, new polygons were created defining zones within IFGs (e.g. the 0-3 NM zone in the Clyde and West IFG). Once the polygons were created any ping could be associated with a zone using the Spatial Join function.
The VMS data were supplied by Marine Scotland as six Access (mdb) data files for each of 2010, 2011 and 2012 and three combined files for vessels from the Rest of the UK for the same years. Some simple reformatting was required to produce dbf files that could be used in Arc- GIS.
After adding the file to the Arc database, the latitude and longitude data was used to locate each ping on the map using the Add XY function  . The figure below is an example of the visual output available by using the Spatial Join function to link fishing pings with IFG areas. It shows the output for the over 15 m demersal fleet.
Fig 6.1.1 IFGs and Recorded Fishing Locations for > 15 m Demersal Trawlers, 2012
Finally, data points outside the 12 NM limit were excluded and the resulting selected data file exported as a dbase file. The table below provides summary data on the Pings
|Number of VMS Pings||% of Pings|
|<1 NM||<3 NM||<12 NM||Total|
After allocation and selection the resulting single species data files were exported as a dbf file which, as a first step, can be analysed in Excel. However combined files are necessary because vessels identified as say nephrop trawlers can occasionally catch queen scallops on a trip. Excel does not handle large files particularly well and its statistical tools for producing say multi-layer cross tabs are limited. Consequently the files were read into SPSS, combined and then analysed. The tables produced in SPSS were then transferred to Excel for production of the final tables.
188.8.131.52 Adjustment for Port Entry
As discussed earlier, a key element in the analysis are the assumptions that a) vessels travelling between 0 and 5 km/h are "fishing" and that b) the catch is equal at all recorded locations. Visual inspection of the "fishing pings" shows considerable fishing effort along the course vessels normally take on approaches to ports. The figure below illustrates over 15 m demersal trawlers approaching Scalloway and Lerwick steaming at less than 5 km/h but almost certainly not fishing. These incorrect fishing pings are most probably recording vessels steaming slowly as they approach ports.
Fig 6.1.2 Pings from Demersal Vessels and Port Entry Routes
The result of misspecification of fishing pings is that the catch volume and value within the 1 mile zone will be over-estimated and outside the zone will be underestimated. To adjust for this problem the following procedure was adopted:Each major port was examined to see if fishing pings were coinciding with routes normally followed by vessels steaming slowly whilst approaching port. These approach routes were added to the map All "fishing pings" on or close to (within 200 m) of the route were identified The total catch value and volume for each species/gear type in each IFG of these selected pings was calculated and deducted from the estimated 1 NM catch values and volumes. Whilst all the catch value was removed from the 1 NM zone (the area most likely to have slow travel), only half the catch value and volume was deducted from the 3 NM zone and 75% from the 6 NM zone. In the absence of any other information, this was assumed to be reasonable. Finally the catch removed from the record had to be caught in some area. For this exercise it was assumed to have come from within the 12 NM/ IFG zone although it was recognised that in the case of deep sea ports like Peterhead or Fraserburgh this will be an overestimate.
In conclusion, we are confident that for the over 15 m vessels the estimates on the location of effort and catch are sufficiently accurate for the requirements of this study.
6.2 Fishing Activity and Catch for Vessels Under 15m
Whilst VMS data provides data for vessels over 15 m there was no systematic collection of data on the location of the fishing effort or catch of the under 15 m segment of the inshore fleet
Fortunately, Marine Scotland Science's recently completed ScotMap project had addressed this problem by undertaking a one-off survey of inshore operators. The project asked them to draw their fishing activity as a polygon on an electronic map and provide data on the % contribution of this fishing area to gross vessel earnings, information on those earnings (average annual gross over the years 2007-2011) This data was then used to calculate a monetary value for each polygon. For each polygon the primary, and where relevant the secondary target, species was identified and the value of landings from each polygon was estimated based on information provided by fishermen. Fishermen who move between different grounds might draw more than one polygon.
6.2.1 Estimating IFG Zonal Fishing Effort for Vessels Under 15 m
In order to obtain the co-operation of the fishermen, as part of the Scotmap data agreement brokered by MS, strict conditions were imposed on the release of data. This effectively prevented any access to the raw data, apart from a limited number of staff at Marine Scotland Science ( MSS) working on the Scotmap project. This meant that this study had no access.
In the absence of the raw data, the consultants provided MSS staff with a series of zonal boundaries for each IFG. For example zone 4_3 would be within 0-3 NM of the shore in IFG4. For each IFG area, MSS staff clipped the complete raw data to each distance zone (0-1 NM, 0-3 NM, and 3-6 NM) zone and provided a count of the number of vessels fishing within each zone.
Whilst the number of vessels is an indication of the location, intensity and type of effort in a distance zone, there are a number of problems. Individual operators might draw more than one polygon and report the catching of more than one species caught in each polygon. In addition, each polygon might also overlap the boundaries of this study's defined zones.
Thus a single vessel could appear in a number of zones. Unfortunately, without access to the raw data, we only know the number of vessels that appear in each zone. The number of days, or hours spent in each zone is unknown. Our only measure of effort is therefore the number of vessels counted in each zone, and this measure will double count some vessels. For example, the Scotmap survey had 1080 vessels responding. Summing the number of vessels identified in each of this study's zones, yields a total of 1461 vessels. This is an overestimation of 35.2% across the whole inshore area.
The overestimation will vary between IFG areas and gear types. The table below shows the implied overestimate associated with IFGs and gear types.
|N'rops||D'rsal||Scallops||N'rops||Other Shellfish||Scallops||Other Shellfish||Pelagic||Total|
|Total Vessel Count||198||35||28||286||862||50||5||201||1665|
|No of Scotmap Vessels||120||28||21||252||805||39||5||154||1080|
The table above for <15 m vessels is based on 1080 unique vessels in the Scotmap Survey. The vessel count of 1665 represents the number of vessels counted using each gear type in each IFG. The number of vessels counted across IFGs was 1424 (not reported in the table above). On this basis there would appear to be more vessel mobility between gear types than between IFGs.
For comparison, the table below gives a similar table based on VMS data for over 15 m. vessels.
|D'rsal||N'rops||Pelagic. Trawl||Other Trawl||Dredge||Pots||Total Vessel Count||Total VMS Vessels||% Over Sample|
|Total Vessel Count||388||364||57||54||108||31||1002||853||17.5%|
|Total VMS Vessels||141||170||28||33||52||16||440|
|% Over Sample||175%||114%||103%||63%||107%||93%||127%|
As can clearly be seen larger vessels are much more mobile with vessels, on average, covering 2.5 zones. Identifying the location of over 15 m vessels is not problematic because the VMS data is based on pings which occur every 2 hours and these provide a reliable locator of fishing effort.
The table below shows estimates of the number of vessels in the 1, 3, 6 and 12 NM zones by IFG Area in 2011. The zones are concentric i.e. the vessels in the 12 NM limit include those in the 0-6, 0-3 and 0-1 NM limits.
|South West||North West||Outer Hebrides||MF& NC||Orkney||East Coast||Total|
Whilst there are additional boats in the larger areas the numbers are small; most vessels are found at some stage to be fishing within the 1 NM limit. The relative importance of fishing within 0-1 is evident.
6.2.2 Estimating Zonal Catch Value for Vessels Under 15 m
The aim of Scotmap was to map fishing activity and value associated with fishing by under 15 m vessels. It produced Raster maps showing the distribution of vessel activity and fishing value by species in Scotland. The figure below shows typical output from the project which is found at http://www.scotland.gov.uk/Topics/marine/science/MSInteractive/Themes/ScotMap
Fig 6.2.1 Example of Raster Map produced by ScotMap Project
These maps were generated by
- Splitting the inshore area into very small rectangles.
- Taking each polygon in turn and allocating the vessel to any of the small rectangles contained within it
- Taking each polygon in turn and subdividing equally the value generated to each small rectangle found within the polygon.
- Aggregating the number of vessels and the value in each of the small rectangles.
- Producing a raster map by giving each value in the small polygon a colour e.g. 5 might be bright red, 3 a dark orange, 1 a light orange (see above Figure).
The process above means the Raster maps contain the aggregated information from the ScotMap project in a very dense format.
By splitting the raster into the IFG zonal polygons and summing the values in each of the small rectangles within the polygon it is possible to reverse the process and identify the value in each of this study's zones by species/gear-type.
Within Arc- GIS the "Zonal Statistics As Table" tool scans the raster values within any defined polygon and produces the sum of the raster values as one of the outputs. The table below shows the values generated by the procedure.
|1 mile||3 mile||6 mile||12 mile|
6.2.3 Coverage and Non-Response Adjustment for Vessels Under 15m
The ScotMap project was based on information volunteered by fishermen. There could therefore be an element of error in both defining the fishing area and in assessing the revenue generated, but this was thought to minimal. A far greater problem is non-response. The table below shows the estimated survey coverage of vessels and revenue based on port of registration.
|District Name||Vessel Coverage (i)||Average Landings 2010-11 Coverage (ii)|
|Aberdeen||63/81 (78%)||£ 4.02M /£ 4.39M (92%)|
|Anstruther||92/115 (80%)||£ 7.06M /£ 8.04M (88%)|
|Ayr||41/98 (42%)||£ 4.53M /£ 6.54M (69%)|
|Buckie||34/52 (65%)||£ 2M /£ 3.31M (60%)|
|Campbeltown||80/125 (64%)||£ 7.22M /£ 11.88M (61%)|
|Eyemouth||78/96 (81%)||£ 7.37M /£ 8.69M (85%)|
|Fraserburgh||77/96 (80%)||£ 2.33M /£ 4.35M (54%)|
|Kinlochbervie||11/20 (55%)||£ 0.56M /£ 1.05M (54%)|
|Lochinver||8/13 (62%)||£ 0.69M /£ 1.22M (57%)|
|Mallaig||19/39 (49%)||£ 1.11M /£ 2.38M (47%)|
|Oban||61/107 (57%)||£ 5.9M /£ 8.53M (69%)|
|Orkney||130/130 (100%)||£ 9.66M /£ 10.34M (93%)|
|Peterhead||41/45 (91%)||£ 1.42M /£ 1.51M (95%)|
|Portree||86/133 (65%)||£ 7.96M /£ 11.36M (70%)|
|Scrabster||50/75 (67%)||£ 2.53M /£ 3.28M (77%)|
|Stornoway||172/200 (86%)||£ 10.67M /£ 12.22M (87%)|
|Ullapool||47/85 (55%)||£ 3.67M /£ 5.48M (67%)|
|Total:||1090/1510 (72%)||£ 78.71M /£ 104.56M (75%)|
During their assessment of the coverage, MSS staff had identified the reported catch associated with the non-responding vessels from the landings data. MSS staff were able to provide the catch by species/gear type and by home port for all vessels not in the sample. The IFG figures for non-reporting vessels were then estimated by aggregating the data for the ports within the IFG (with the exception of Fraserburgh which was aggregated into the Moray IFG). The split into zones was based on the recorded information for similar IFG/Gear type combinations. This was then simply added into the data set. The finalised data was then sent for validation by the fishing community.
6.2.4 Final Validation of Results
All the re-estimates were sent to every Fishery Office for final checking. Two small anomalies were identified relating to pelagic catches in the Moray Firth and North Coast and in the East Coast IFG. These reasons for these related to recording procedures rather than computational outcomes. Appropriate adjustments were made.
6.2.5 The Special Case of Shetland under 15 m vessels
Inshore fishing in Shetland operates under the Shetland Regulating Order Regulated Fishery Order 2012. The Order is managed by the Shetland Shellfish Management Organisation ( SSMO) that regularly commissions studies of fish stocks and fishing activity around the islands.
One such study paralleled the ScotMap. Publications include Shape files identifying fishing grounds. An example is given in the figure below.
Although rather crude it was assumed that catch in each zone was proportionate to the area fished. Less crudely the total catch equates to the landings. Thus for each of the species types the percentage of the area in each zone was calculated and multiplied by total catch to give final estimates of the amount caught within 1, 3 and 6 NM of the shore.
6.3 Modelling Effort for Under 15 m Vessels
Measuring fishing effort for each zone within each IFG effort is a complex calculation involving vessel numbers, the size and power of those vessels and the length of time they spend in each zone. The only variable which measured effort for the under-15m vessels was the vessel counts. We endeavoured to develop an effort variable (equivalent vessel unit) which captured the amount of effort vessels devoted to each zone within each IFG. As the project progressed it became clear that apportioning the total equivalent vessel units between zones within IFGs was only possible by assuming a perfect correlation between revenue and effort. In effect the impact of, for example, a 10% change in effort would be a 10% change in the revenue. There is nothing to be gained by a "bottom up" modelling of effort if this requires the assumption of a perfect correlation.
The problem however is that for the under 15m fleet, the conversion of trawlers to creelers or the transfer of licences from trawling to creeling is based on the number of vessels involved. Rather than try to calculate revenue/vessel from our data, which would have required the calculation of a vessel equivalent, the revenue per vessel obtained from Seafish's Cost and Revenue Statistics was used. Estimates of the effort currently employed inshore are thus not used in the model.