The Coastal Reference Fleet 2007-2019

This report is a modified excerpt from the master thesis “A descriptive analysis of data from the Norwegian Coastal Reference Fleet and comparison with catch-per-unit-effort data from research surveys” by Hanne Hjelle Hatlebrekke, for the degree Master of Science in Fisheries Biology and Management at the University of Bergen in March 2021. The thesis was a collaboration between the University of Bergen and the Institute of Marine Research (IMR), research group Fisheries dynamics, and aimed to investigate the possibility of analysing data from the Norwegian Coastal Reference Fleet (CRF) using the statistical tool Pasgear II. In this report we present a general overview of the efforts and contributions from the CRF from the time period 2007-2019, and trends in CPUE for the species cod (Gadus morhua), haddock (Melanogrammus aeglefinus), and golden redfish (Sebastes norvegicus).

Fish is a renewable resource, providing long term food and income when managed sustainably. In Norway, this is mainly done by having quotas (Gullestad et al. 2014) along with species-specific minimum catch sizes and subsequent gear restrictions to prevent the catches of too-small individuals (Johnsen and Eliasen 2011; Gullestad et al. 2014; Gullestad et al. 2015). These quotas are set based on recommendations from the International Council for the Explorations of the Sea (ICES), where data from multiple data sources such as catch statistics and research surveys are analysed to predict future stock sizes (Nagelsen 2018). These surveys are time consuming and expensive to conduct, as well as only occurring at certain times of the year and often only for a few weeks. Although the methods used during the surveys are consistent in time and reproducible, the reduced temporal coverage may result in the data not being representative for all species due to seasonal variations in the distributions of a stock (Durant et al. 2008). In contrast to scientific surveys, the fisheries are active throughout the year. The Coastguard and the Directorate of Fisheries (DoF) collect biological and environmental samples from both landed catch and inspections at sea to use in assessments and research (Nedreaas et al. 2006; Ministry of Trade, Industry and Fisheries 2017). Similarly, IMRs  “catch sampling lottery” receive catch samples from the pelagic fishery (www.hi.no/fangstprover), and their rented catch sampling vessel as well some stationary personnel collect biological samples from port landings. Data on the fishing effort used to obtain the catches, however, are not reported, limiting the usability of the catch data as direct indices of fish abundance.

Catch per unit effort (CPUE) has been extensively used to describe population dynamics, distributions in space and time, and catch rates (Petitgas et al. 2003; Verdoit et al. 2003). It is important that a CPUE series has a good enough spatial coverage to ensure comparison of the same populations over time (Hillborn and Walters 1992). To interpret CPUE data, knowledge and understanding of the distribution of both the fish and the fishing effort is important (Paloheimo and Dickie 1964) as landings in an area depend on the local population dynamics and the available quotas and restrictions (Jakobsen and Lindkvist 2003). For example, as a fishery develops, effort might increase to best take advantage of the resources, and CPUE will also be high. After some time, CPUE will in theory start to decrease as the stock abundance decreases (Jennings et al. 2001), however this has been shown to not always be the case (Harley et al. 2001). Nevertheless, CPUE timeseries may provide more information about a fishery than landing statistics alone.

2.1 - Aims of the program

To improve estimates on discards (the part of the catch that is returned to the sea (Alverson et al. 1994)) and bycatch (the incidental catch of non-target species (Alverson et al. 1994)), to supplement the data collected by the government and research institutions, and to collect continuous and consistent samples, several programmes involving fishers self-sampling their catches are being developed world-wide (Nedreaas et al. 2006; Bell et al. 2017; Van Overzee et al. 2019). In 2005, a Coastal Reference Fleet (CRF) was established in Norway (Clegg and Williams 2020), consisting of vessels ≤15 m (Williams and Gundersen 2020), and generally operating within 12 nautical miles from the shore (Fangel et al. 2015). The CRF has since grown to include vessels using a variety of gears and targeting different species (Pennington and Helle 2011), often varying gear and fishing location based on their current target species and time of year (Berg 2019). In this report we present results for the years 2007-2019 since we consider the two first years being training and introduction years.

While the need for more biological data and information about the fish caught by Norwegian fishers was the initial intention of establishing a reference fleet, it has become increasingly important to explore the possibility of using fishery dependent data as indices of fish abundance for data poor populations both in Norwegian waters and in areas where survey data are not available for stock assessment worldwide. An external review of the IMR’ stock assessment processes recommended that analyses of commercial CPUE should be revisited to determine if useful timeseries can be generated, with appropriate adjustments for technology differences and research on the variations of catchability with time (IMR 2020).

The four main goals of the Norwegian Reference Fleet are to: (1) support fish stock assessments with biological data including length composition of catches (length and weight measurements for all species captured), age composition of catches (otoliths and scales collected), and quality control and facilitation of data for stock-assessment; (2) document the fishing effort and catch composition of total catches, including bycatch, discards and catches of non-commercial species, seabirds and sea mammals to provide data for the monitoring of biodiversity, fishing effort and catch per unit effort (CPUE) over time; (3) provide a platform for the collection of additional samples from fisheries, and (4) increase collaboration and strengthen dialogue between researchers and the fishing industry (Clegg and Williams 2020).

2.2 - Vessel selection

The CRF includes vessels using a variety of gears and targeting different species (Pennington and Helle 2011), often using different gears and fishing locations based on their current target species and time of year (Berg 2019). An overview of the fleet in 2016 can be found in Appendix 1. It provides biological and environmental data with a good statistical, spatial, and temporal coverage electronically, including species that might be hard to sample on scientific surveys such as deep-sea species, populations in shallower waters or closer to land, or generally rare species (Nedreaas et al. 2006). Furthermore, the data on discards reported by the CRF are the only consistent discards data in Norway (Berg 2019). As the vessels partaking in the CRF are also part of the entire Norwegian fishery, the data collected are assumed to be representable for the entire fleet and thus used as a proxy for the catch and effort in both research and for management purposes (ICES 2017).

By law, the bidding must be open for all vessels to apply, and the selection is based on the location of the home port (assuming this is where the majority of the fishing activity occurs) and the fishing pattern to ensure broad data coverage and representation (Fuglebakk et al. 2018; Clegg and Williams 2020; Williams and Gundersen 2020). The fishing pattern refers to how a fisher fishes with regards to gear, species, and effort, as well as the general composition of the CRF. The Norwegian coastline is divided into nine statistical areas (Figure 1) by the DoF to account for local differences in fish populations along the coast, and the IMR aims to have at least two vessels in the CRF from each area. The full names of these areas along with the shortened names used in this report are presented in Table 1. The contracts for the CRF last for four years (Clegg and Williams 2020; Williams and Gundersen 2020), however if the reported data is believed not to be representative of the catches, the contract may be terminated (Bjørge and Moan 2017). Thus, the data collected are assumed to be of good quality due to continuous training, with the scientists and the fishers having a trust-based partnership (Nedreaas et al. 2006).

The contracts for the CRF are generally made between the IMR and a specific vessel, however as quotas are per vessel, one boat may not have a large enough quota to operate throughout the year. Some skippers solve this by either having multiple vessels used in succession as the quotas are filled, or by cooperating with other fishers to catch the quotas for both vessels between them. Additionally, a fisher can borrow a vessel while they wait for their own to be repaired, for example. Provided that all vessels meet the same requirements as the original vessel in the contract, the skipper can register data from all boats used, although they cannot register data from multiple vessels fishing at the same time. In this study, the data are reported as being from the vessel holding the contract, as this most accurately represents the number of fishers participating in the CRF. This vessel is identified using an individual call sign, which is a permanent signature for a vessel.

Table 1. The Directorate of Fisheries’ statistical areas along the Norwegian coast.

2.3 - Data collection

Contracted vessels are equipped with the appropriate measuring instruments, and the crews receive training in how to sample the catches consistently, including taking individual length and weight measurements and collecting otoliths for age determination (Nedreaas et al. 2006; Clegg and Williams 2020; Williams and Gundersen 2020). Sampling methods are the same as those used on research surveys conducted by the IMR, however the protocol is slightly different regarding e.g. the number of samples and how often they are to be taken (Table 2) (Nedreaas et al. 2006; Clegg and Williams 2020; Williams and Gundersen 2020). Consistent methods allow the studying of long-term trends in commercially important stocks, as well as monitoring fleet behaviour and how technical advancements and regulations influence fishing effort and efficiency in real-time (Nedreaas et al. 2006; Pennington and Helle 2011). The fishers are encouraged to record everything they catch, including what is discarded at sea, by receiving payment for the number of taxa caught per fishing operation with the promise that these data will not be used as grounds for inspections (Clegg and Williams 2020). Each vessel is allocated a contact person from the IMR who visits at least once a year, and the IMR provides training and updates the equipment as necessary (Nedreaas et al. 2006; Clegg and Williams 2020; Williams and Gundersen 2020).

Table 2. Protocol for registration and sampling in the CRF (Clegg and Williams 2020).

2.4 - Pasgear II

Pasgear II is a customised data processing program for storing and analysing fisheries data (Kolding and Skålevik 2009). It has been used in numerous studies, e.g. Dincer and Bahar (2008) and Prchalova et al. (2013). Raw, organised data from the CRF database were directly imported to Pasgear II where they were further processed, and all figures (except Figure 1 and Appendices 1 and 5) were made using this software.

2.5 - Catch per unit effort (CPUE)

The basic assumption in fisheries theory is that catch (C) and stock abundance, or standing biomass ( ), are related by

where f is a measurement of the nominal fishing effort or intensity, and q is the catchability coefficient.

Rearranging to

where U is catch per unit effort (CPUE), which is the weight or number of a species caught per unit effort used to obtain the catch, and proportional to the biomass if q is constant  (Sparre and Venema 1998 ; Jennings et al. 2001).

In Pasgear II, CPUE is calculated as

where y is absolute effort, n is the number of samples, and W_i is the catch in weight or numbers.

The IMR frequently uses boat-days or gillnet-days as a measure of effort. The CRF registers data per fishing trip, which has a higher resolution than boat-days. In this report, CPUE is therefore defined as landed weight caught per fishing operation using gillnets, excluding gillnets targeting anglerfish (gear codes 4129 and 4149). A fishing operation is the deployment and subsequent hauling of a specific fishing gear, and the use of multiple mesh sizes equates to one fishing operation per mesh size. For CPUE to be calculated correctly, the effort term needs to include all efforts made to obtain a sample, not just the effort made when the species of interest occurred in the catch.

2.5.1 - Investigating CPUE from multiple geographical areas merged

The areas north of 62°N is a separate region of management for cod, haddock, and golden redfish, and will thus be a region of increased focus in this report. As the structure of the CRF is such that the vessels are signed on depending on which statistical area they are most active in, they are not evenly spread out along the Norwegian coastline. When looking at CPUE-values for multiple areas combined, the values were thus weighted based on the surface area of each statistical area. This was done by dividing each surface area with the largest surface area.

The cod stock is for management purposes considered divided at 67°N (ICES 2021). To compare the populations north and south of this latitude, the statistical areas East Finnmark (03), Troms (04), Vesterålen (05) and Vestfjorden (00) were combined to form the region north of 67°N, and Helgeland (06) and Møre-Trøndelag (07) were merged into one region between 62° and 67°N, also referred to as south of 67°N onwards. These regions were weighted separately, so the surface areas were divided by the largest surface area within each region.

2.6 - The Index of Relative Importance

Pasgear II includes a predefined analysis which calculates an Index of Relative Importance (IRI) (Kolding 1989) and provides an output with the total catch composition of the dataset provided raw data has been imported. This is a way of standardising the relative importance of a species in a catch based on the weight, the number of individuals, and the frequency of occurrence (the number of fishing operations where a taxon was registered) for each taxonomic group in the catches (Kolding 1989). The output can be given as a percentage of the total IRI, which is a measure of relative abundance of a taxon within the entire dataset and presents the most commonly observed species in order of importance, or as a rectangle (Kolding and Skålevik 2009).

In Pasgear II, %IRI is calculated as

where %W_i and %N_i is the percentage weight and number of each taxon of the total catch, %F_i is the percentage frequency of occurrence of each taxon in the total number of fishing operations, and S is the total number of taxa.

3.1 - Vessels

The number of vessels fluctuates from year to year due to e.g. project economy and vessels joining and leaving the program. The number of vessels has varied from 16 unique vessels in 2007 to 25 unique vessels in 2015, with an average of 21 vessels participating each year (Table 3).

Table 3. Number of unique vessels (N) contributing to the Norwegian Coastal Reference Fleet 2007-2019

Although each contract with a CRF vessel only lasts for four years, it is common for skippers to apply for a renewal of their contract and continue their participation as illustrated in Figure 2. This indicates a positive experience of the cooperation with the IMR. Note that a few missing months may only mean the vessel was not active during this period, and not that the contract ended or was terminated by either party. In total, 64 different vessels, usually with 1-3 people on board, have participated in the CRF between 2007-2019. A breakdown of vessel activity within each statistical area can be found in Appendix 2.

The number of vessels in the CRF fishing in each statistical area varied annually and between the areas (Figure 3). In general, there was an increase in the number of vessels in most areas until a peak in 2016, after which there was an overall decline (Figure 3A). Figure 3B shows the number of vessels with its homeport in each area, which is also the area where the majority of the fishing activity of a given vessel takes place; for example, there were no participating vessels with homeport in Vesterålen (05) in 2014 (Figure 3B), but several vessels with homeports elsewhere fished there that year (Figure 3A). Although the numbers in Figure 3B did not vary as much as in Figure 3A, the overall trends are similarly showing a general increase in the numbers of vessels with an additional rise in the movement between the areas towards the end of the study period.

Furthermore, Conversely, the number of unique vessels fishing in each statistical area and hence, spatial coverage of CRF self-sampling, also showed large seasonal variations, with a marked decrease in sample sizes for May, June, and July (Figure 4). There were additional differences in the activity of the CRF in each area with season, with highest effort in East Finnmark (03) and Vestfjorden (00) being the most fished-in areas during the first quarter of the year, and Skagerrak (09) attracting more vessels after July.

In general, the number of fishing operations per year in the CRF increased throughout the study period, peaking in 2016 (Figure 5). The largest increase was seen in Vestfjorden (00), where the seasonal skrei-fishery is important and thus attracting vessels from other areas. There was also an observed increase in Stad-Austevoll (28).

3.2 - Composition of gears

In every year, more than 60% of all fishing operations carried out by the CRF were using gillnets (Figure 6). From 2013 there was a proportional increase in the use of traps, and from 2014 a larger portion of fishing operations used longlines and shrimp trawls, while the relative use of Danish seines and fyke nets decreased. Interestingly, the type of fishing gear varied greatly both between and within each area (Figure 7), showing that the fishing pattern was not constant. In East Finnmark (03), gillnets were used almost exclusively in 2007 then not at all in 2019, having been replaced by Danish seines, longlines, and traps. Similar decreasing trends were observed in Helgeland (06), Stad-Austevoll (28), and Skagerrak (09), where gears such as traps, fyke nets, and shrimp trawls have become more common. In fact, years with abrupt changes in the types of gears used were observed in all statistical areas, apart from Møre-Trøndelag (07), where gillnets were used almost exclusively. Unspecified trawl was only used by one fisher three times in Skagerrak (09) in 2015.

3.2.1 - Gillnets

The proportional decrease in the use of gillnets seen in Figure 6 was also seen in the total number of gillnets used in each statistical area per year; Figure 8 shows a peak in numbers in 2010 and subsequent decrease towards 2015 followed by a somewhat higher and stable use. In some areas, such as Troms (04) and Møre-Trøndelag (07), the number was quite stable throughout the study period; in other areas such as Vesterålen (05) and south of Møre-Trøndelag, the numbers greatly varied.

The total soak time for gillnets did not show any clear trends over the study period (Figure 9); rather, it seems to have varied greatly within each area and between the years. In Vestfjorden (00) there has been a gradual increase in the accumulated hours, especially since 2014, whereas a decrease can be seen in Helgeland (06). Vesterålen (05) and Austevoll-Lindesnes (08) in particular had large variations in total soak time.

Gillnet as a category of gears is not a homogenous group, but one where the fishers use a variety of mesh sizes to regulate selectivity when fishing. In this study, mesh size groups (measured in bar length) were used in addition to unknown mesh size. These groups can be seen in Figure 10, which also shows that the mesh sizes used by the CRF varied both within and between the areas, as well as between the years. This Figure also illustrates the development in the use of gillnets in general in each area over time by looking at the total number of fishing operations in each year. Note that the y-axes are free to highlight individual trends within each area. In all areas, the relative use of each mesh size group varied throughout the study period; in East Finnmark (03), Austevoll-Lindesnes (28) and Skagerrak (09) mesh sizes have generally increased with time, whereas in Helgeland (06), Møre-Trøndelag (07), and Stad-Austevoll (28) mesh sizes have decreased. Additionally, the fishers have overall become better at recording the mesh sizes used instead of using “Unknown”.

In all areas, the number of fishing operations using gillnets showed seasonal variations in the local fisheries (Figure 11). There were relatively low levels of activity in all statistical areas during the summer months compared to the first as last quarters of the year, aside from in Vesterålen (05) and Møre-Trøndelag (07) where activities did not increase in the autumn, and in Skagerrak (09) where the peak in fishing operations occurred in June followed by lower numbers August and September. Furthermore, in Vestfjorden (00), Helgeland (06), and Møre-Trøndelag (07) the mesh sizes increased throughout the year; in East Finnmark (03) and Troms (04) mesh sizes decreased. Note that in Figure 11, the y-axes are free to highlight individual trends within each area.

3.3 - Catch composition

287 taxa were registered by the CRF between 2007-2019 from 29 137 individual fishing operations. A comprehensive list including the total numbers of individuals, the weight, and the % IRI can be found in Appendix 4. The most important species overall were cod, which occurred in 61% of the fishing operations over the course of the study period and accounted for 45% of the total weight and about 5% of the counted specimens, and saithe (Pollachius virens, L.), occurring in 53% of the operations and accounting for 19% of the total weight and 4% of the counts (Figure 12). Figure 12 further illustrates the taxa with the highest IRI, which are presented along with the proportions of numbers and weight registered. The size of the bar is proportional to the IRI for each species. Throughout the study period, haddock was found in 42% of the catches, however in much lower quantities than cod when looking at both numbers and weight. Golden redfish (marked with an asterisk in Figure 12) was caught in 15% of the catches, and barely accounted for 1% of the total weight and less than 0.5% of all counted individuals.

Figure 13 shows the % IRI for the 15 species with the highest value overall between 2007-2019. These species accounted for 98.3% of the total IRI, 90.7% of the total number of individuals, and 95.7% of the total registered weight. Cod became increasingly important between 2007-2012, followed by a slight decline. The % IRI for haddock was quite stable, with somewhat higher values in 2014 and 2015. Although it is hard to discern from the Figure, golden redfish had the highest values in 2009-2011. Overall, the species composition of the CRF varied between the years, with species such as the brown crab (Cancer pagurus, L.) being less important after 2009 and northern shrimp (Pandalus borealis, K.) having a higher % IRI from 2016 onwards.

3.4 - Catch per unit effort infor the CRF

To investigate the temporal development of catch per unit effort (CPUE) in the CRF, landed catches were weighed and the unit effort was set to per fishing operation using gillnets (excluding gear codes 4129 and 4149). There was no obvious temporal trend in CPUE of all species throughout the study period between the statistical areas (Figure 14). However, there was a steep decline in CPUE in East Finnmark (03) after 2008 with subsequent peaks in 2013 and 2014, and an increase to a peak in 2014 followed by a decrease in Troms (04). Vesterålen (05), Vestfjorden (00), and Stad-Austevoll (28) showed similar trends but with earlier peaks, whereas the remaining areas had quite stable CPUE-values throughout the study period.

CPUE of only landed cod showed similarities to that of all catches (Figure 15), especially in areas north of 62°N (Troms (04), Vesterålen (05), and Vestfjorden (00)) indicating that it is an important commercial species locally. At this scale, CPUE of cod was very low in the more southern areas of Stad-Austevoll (28), Austevoll-Lindesnes (08), and Skagerrak (09), and any trends are not discernible from the Figure. Landed haddock had high CPUE in Vestfjorden (00) and Møre-Trøndelag (07) at the start of the study period, and a peak in Troms (04) in 2016 (Figure 16). Overall, CPUE was quite variable in all areas. CPUE for golden redfish also showed variations, especially from 2007 to 2014, however a peak in Stad-Austevoll (28) in 2009 skews the y-axis making any trends hard to distinguish (Figure 17). This catch was from a vessel fishing both coastally and offshore and exceeding 20 m in length. This is larger than most coastal vessels, however it was included in the CRF to ensure coverage of this area and to reflect the diversity of the Norwegian fishing fleet. Also note the differences in the scaling of the y-axes in Figures 15, 16, and 17; as effort is the same for all species, it is evident that the catches of cod were much larger than those of haddock and golden redfish. After 2014, CPUE of golden redfish has been closer to zero.

3.4.1 - CPUE for cod, haddock, and golden redfish north of 62°N

The majority of the fishing on all three species occurs north of 62°N, so this region is managed separately from the areas further south. When comparing CPUE for cod, haddock, and golden redfish in this region, both cod and golden redfish showed increasing CPUE-values from the start of the study period until 2011, after which they both declined (Figure 18). Haddock had more variable values, but with an overall decline as well. Note that the species have separate y-axes.

ICES has recently recommended to manage the cod stock north of 62˚N as two management units, i.e., north and south of 67˚N (ICES 2021). These regions were therefore inspected separately in Figure 19. Note that there are two y-axes, highlighting the differences in the amount of cod caught in each region. The curves show a similar trend with an initial increase followed by decreasing CPUE-values, although the peak in the southern region came later than the northern peak.

3.4.2 - CPUE for cod, haddock, and golden redfish south of 62°N

As trends in CPUE for the statistical areas south of 62°N were not too obvious in Figures 14-17, these areas were inspected separately for each species. For cod, CPUE in Austevoll-Lindesnes (08) and Skagerrak (09) showed only slight variations with time, whereas there were three distinct peaks in CPUE in Stad-Austevoll (28) in 2009, 2013, and 2016 (Figure 20). For haddock, a general increase was observed in Stad-Austevoll (28) and Austevoll-Lindesnes (08), while CPUE in Skagerrak (09) decreased overall (Figure 21). The previously discussed peak in CPUE for golden redfish in 2009 is prominent in Figure 22 as well, however excluding Stad-Austevoll (28) only showed that the data for golden redfish was sparce before 2015, with almost no data after this (Figure 23).

4.1 - The development of the CRF

The CRF program has developed and changed over the course of the study period. The number of vessels participating has increased, and with it the amount of data collected each year. Additionally, the fleet has become more mobile; at the start of the program the IMR did not encourage the participation in seasonal fisheries as this would decrease the data collected in a vessel’s home area, however, the movement of vessels across statistical areas has become more frequent in the CRF over time. This has strengthened the data collected in areas such as Vestfjorden (00) and Vesterålen (05), where the skrei fishery takes place every spring and thus reflects the seasonality of the general fishery and the economically important seasonal fisheries in Norway  (Huse and Bakketeig 2018; Statistics Norway 2020) .

Compared to the entire Norwegian coastal fleet, the CRF is not very large. As the aim of having at least two vessels per statistical area is not being met every year, caution is advised when extrapolating the data on smaller scales. For example, when looking at the gear composition for the entire study period in Figure 6, the variations in gear types seem gradual; when focusing on each statistical area in Figure 7, however, some areas had sudden shifts and complete replacements of gear as seen in e.g. East Finnmark (03). With few vessels in each area, the signing of a vessel using different gears will affect the composition of data collected here considerably. This could be unintentional by the IMR in order to simply secure coverage of the area, or intentional like in Skagerrak (09) where a shrimp trawler was signed on to reflect this locally important fishery  (Ziegler et al. 2016) . Such sudden changes within an area have not been observed in the general fishery (Appendix 5), thus the changes in gear representation in the CRF does not always directly reflect or represent the overall fleet. Furthermore, in areas with seemingly little gear development over time, such as Møre-Trøndelag (07) where the vessels in the CRF mostly use gillnets, this does not match with the costal fleet overall as traps have become more common here over the course of the study period. These differences in the fishing pattern between the CRF and the entire coastal fishery will result in divergences in the overall catch compositions and the CPUEs, which could be further examined by comparing the catch composition of the CRF with sales notes from the coastal fisheries.

At the beginning of the study period, several catches were recorded using “Unknown” mesh size. This occurred when a fisher using different mesh sizes during the same fishing trip would record it as just one catch without including the mesh size. However, in most areas the frequency of these recordings has decreased over time. The detailed recording of each fishing operation, e.g. individual mesh sizes, number of nets used, and soak time, is a great strength of the CRF, as well as the amount and coverage of data in both space and time. Additionally, being the only reliable source of fishery dependent data on discards, the CRF data could increase the accuracy of assessments while being more cost-effective than collecting information using research vessels  (Pennington and Helle 2011) . Lastly, close cooperation between fishers and scientists appears to increase mutual understanding of the stocks, improving overall fisheries management  (Nedreaas et al. 2006) .

The species composition of the catches by the CRF showed variations across the study period. In all years, cod had the highest % IRI, which was expected as the cod fishery is the most valuable fishery in Norway (Statistics Norway 2020), however in 2007 saithe was almost equally as important. At the start of the study period, brown crab had quite high % IRI. This species is a common bycatch species in demersal gillnets and is thus often discarded at sea  (Savina et al. 2017) . The lower % IRI after 2009 does not necessarily mean lower catches of brown crab but that other species became more important in the catches. For example, the increased importance of northern shrimp from 2016 coincides with the larger proportion of fishing operations using shrimp trawls in Skagerrak (09) (Figures 6 and 7).

4.2 - Biological factors affecting catchability

CPUE is frequently used as an index of abundance, however it is commonly known that the assumptions this is based on may not always hold true (Beverton and Holt 1957; Harley et al. 2001). For CPUE to be proportional to the size of a population, catchability would have to be constant (Jennings et al. 2001). Unfortunately, catches are influenced by a multitude of factors (Crespin et al. 2008; Maunder et al. 2006), some of which can neither be measured nor controlled for. These factors can be grouped into biological and technological elements which interact with and influence each other.

Biological factors influencing catchability are difficult to control for, and are further affected by season, the age of the fish, the environment, and interspecies interactions  (Crespin et al. 2008) . For a fish to appear in a catch, it must be present and available on the fishing ground. For example, due to the low stock levels of golden redfish, the probability of catching this species will be lower than that of catching cod. Furthermore, the chances of catching cod in Vestfjorden (00) and Vesterålen (05) will be elevated in the spring due to the seasonal migration this species undergoes to spawn. This spawning time and location partially overlap with that for haddock, and large spawning aggregations will temporarily increase catchability (Bakketeig et al. 2017). In a fishery, the fishers will have experience and knowledge of the local conditions and seasonal abundances of the stocks, andstocks and will typically fish more where they anticipate higher catches of their target species.

The behaviour of a species, in both its environment and towards a given fishing gear, will directly affect its catchability  (Crespin et al. 2008) . Demersal species, such as cod and haddock, are often targeted using a bottom trawl  (Huse et al. 2000; Eide et al. 2003) , but might react to noise from the vessel towing the trawl and thus avoid it  (Huse et al. 2000) . This behaviour could change depending on the vessels as the noise levels of the engines vary. Additionally, although species dependant, fish that are similar are likely to cluster together. Consequently, fish caught in the same fishing operation are likely to be more similar to each other than to the remainder of the population, known as intra-cluster correlation (Cochran 1977). A day’s catch from a vessel in the CRF is therefore considered a cluster of fish, and not a random selection of individuals from the total catch, or a population of fish  (Nedreaas et al. 2006; Pennington and Helle 2011) . Thus, further v ariations in catches of e.g. golden redfish might be due to observed shoaling behaviour close to the seafloor during the day that disperses into the water column at night (Gauthier and Rose 2002). This change in activity and habitat preference would change the catchability of the species depending on the time of day. However, w hen calculating mean CPUE over an entire year, and looking at trends over long time scales, factors such as seasonality and individual fish behaviour will be evened out.

The size and morphology of a fish will affect their likelihood of getting caught depending on the specific gear used (Hamley 1975; Næsje et al. 2007). For example, the mesh size of a gillnet will determine if a fish will be able to swim right through, get caught, or make contact but not get entangled if the meshes are too small. However, the morphology of a species may increase the probability of entanglement (Hamley 1975; Næsje et al. 2007), e.g. the spikes on the dorsal fins of golden redfish might get tangled in the net even though the size of the individual is not targeted by the mesh sizes used. The catch of non-target species will decrease the available net area in which target species may get caught, reducing the efficiency of the net.

4.3 - Technological factors affecting catchability

Technological factors can largely be controlled for when planning a study or fishing operation, however they are also affected by seasonal and biological changes. Ideally, for abundance estimates, technological factors should be kept as standardised as possible to keep catchability constant, as each gear type has a different catchability coefficient. This is further influenced by the design, the size, the colour, and the material of each gear (Hamley 1975). For example, using a thinner twine in gillnets will increase the chance of entanglement if a fish comes in contract with the net, but if the twine is a colour which stands out from the background (such as red), the fish might see it and avoid the net altogether (Hamley 1975).

As fishery independent surveys data aim to provide a more accurate estimates of fish population characteristics and abundance indices to track stock size over time, the vessels, sampling design, and gears are mostly standardized consistent during scientific surveys. In the CRF, however, the composition of the fleet is constantly changing, and the vessels participating in 2019 were largely not the same ones as in 2007. This was also apparent when looking at the fishing patterns in the form of gear composition in each statistical area, which showed local changes with time, likely due to the exchange of vessels in the CRF. Thus, regarding standardised sampling for abundance, the variability of the CRF may not seem ideal; however, one of the aims of the CRF is to adequately represent the entire Norwegian coastal fishery. To fulfil this aim, a more varied fleet is desirable as the coastal fleet itself is very diverse, if the data collected from the CRF are to be representative of the whole coastal fishery along the Norwegian coast. Thus, several objectives must be balanced against each other when designing the composition of the CRF.

Focussing on gillnets, the total number of gillnets used showed some annual variations within each area (Figure 8), variations which were expected with the changes in the use of gillnets relative to other gear types (Figure 6). Although the number of gillnets deployed was recorded, the individual net length is not recorded by the CRF but is usually assumed to be 27.5 m. Gillnets that are longer or shorter than this will have different catchability coefficients due to the differences in surface area, however this would not be apparent from the data recorded by the CRF at present. Both the number of gillnets and the frequency of the different mesh size groups varied within each statistical area and between the areas over time. Gillnets are highly selective by e.g. adjusting the hanging ratio or mesh size to target different species and sizes (Hamley 1975; Næsje et al. 2007), however catch rates of any fish will depend on location, fishing depth, and soak time. In 2014, vessels in the costal fleet north of 62°N with a registered length of <11 metres were given an unlimited quota for cod by the DoF (DoF 2014). This could explain the increase in gillnets with mesh sizes of 80 to 110 mm, as these are often used to target cod (Berg 2019). There were also large seasonal differences in the mesh sizes used, for example the fishery targeting the angler takes place during the autumn, primarily using gillnets with mesh sizes >130 mm (Regulations on the practice of fishing in the sea 2005, § 34). These nets only need to be hauled every three days, as opposed to nets targeting cod and haddock, which have to be hauled every day according to regulations (Regulations on the practice of fishing in the sea 2005, § 28).

During a fishing operation, how the equipment is used further affects the probability of a certain catch. For passive gear, increased soak times will initially increase the number of specimens caught, however after some time the quality of the catch might decrease, and the remaining number of hooks on a longline or space in a gillnet will not be sufficient to secure a larger catch (Hamley 1975). Furthermore, the position of a gear could affect the size of the catch, by e.g. placing a gillnet perpendicular to the movement of a stock instead of along it or leaving it to soak overnight to catch nocturnally active species (Hamley 1975).

4.4 - Fishers’ effect on catchability

Some of the temporal changes in CPUE may be an effect of the changing vessel composition of the CRF. CPUE for cod in East Finnmark (03) showed a drop in 2015 (Figure 15), likely due to one vessel leaving the program and another one entering. These fishers might have different target species, fish during different parts of the year, or fish in locations that are far apart while still being within the same statistical area. Size of the boat and engine power will also influence CPUE (Bishop 2006). Additionally, a s fishing is an intentional activity based on years of individual experience, when and where it occurs is not random (Hamley 1975). Thus, catchability is rarely a constant, as each fisher may improve with time both from experience and by technological advancements. As previously noted, the marine environment is a dynamic habitat where certain conditions might result in aggregations of fish. With experience, fishers target these fishing grounds to take advantage of these naturally high densities (Harvey et al. 2001; Petitgas et al. 2003) and might obtain large catches even though the overall abundance of a species is low . Good communication between fishers and scientists and knowledge about the individual vessels in a fleet and how they operate is therefore essential when interpreting the CPUE data (Maunder and Punt 2004; Maunder et al. 2006).

The objectives will also affect the catches, as a fisher tries to maximise catches of their target species while minimising any bycatch as this will be a waste of time and resources. An example of this is the regulations around the golden redfish fishery. Since 2014, this fishery has been closed with a bycatch limit of maximum 10% (Appendix 6). A fisher’s catch composition is influenced by individual behaviour towards changes in regulations, e.g. aiming for the minimum or maximum allowable bycatch. Furthermore, fishers applying to partake in the CRF might be more interested in fisheries biology and conservation, and thus more likely to follow regulations to a stricter degree (Berg 2019). As seen in Figures 17, 22, and 23, golden redfish CPUE from the CRF has been low since this fishery closed in 2014, indicating that the fishers are trying to avoid catching it by using selectivity measures such as changing the fishing grounds and the mesh sizes of the gillnets. However, during surveys, the researchers are not trying to be selective, and comparisons with CPUE from survey data suggests that the stock levels of golden redfish were quite stable throughout the study perio d (Hatlebrekke 2021). For this species, survey data will therefore better reflect the status of the stock as it is assumed not as selective, while the fisheries behaviour suggest that they actively avoid catching golden redfish.

For the variety of reasons mentioned above, CPUE is not necessarily proportional to abundance. Oftentimes commercial CPUE remains high although abundance is declining  (Harley et al. 2001) , which is known as hyperstability (Hilborn and Walters 1992), and if not noticed thus may lead to biomass being overestimated while fishing mortality is underestimated (Crecco and Overholtz 1990). Within the CRF, different levels of sampling between the areas increases the variability and decreases the precision of the data, and as each vessel uses one primary gear or a small selection of gears, the data collected will be biased. However, when looking at comprehensive datasets over larger regions and longer time periods, this bias will be evened out and rather represent the diversity of the Norwegian coastal fisheries. To increase the resolution of the CRF data, the number of vessels would need to be increased as they are the primary sampling unit  (Pennington and Helle 2011) .

4.5 - Limitations

The CRF data base from 2007 to 2019 consisted of 295 372 individual catch records from 29 137 fishing operations. When working with such large datasets, there is an increased risk of human errors during recording, punching, and data handling. Similarly, misidentification of species or incorrect recording of data are likely to have occurred in the CRF. To prevent such mistakes, the IMR has organised workshops focussing on recognising and differentiating between e.g. different species of rays (Berg and Nedreaas 2021). These workshops might increase the taxonomic resolution of data collected afterwards. Moreover, previous studies have shown that data reporting from the vessels are not 100% consistent when compared with sales notes from landings. This could be due to not sampling all catches, or there being some technical error (Berg 2019). Additionally, as the data from the CRF are collected almost continuously, unexpected events such as illness or bad weather might prevent the hauling of gears. In addition to lowering the quality of the catch, the soak time recorded would be longer than usual. As the relationship between the IMR and the fishers is trust-based, these irregular data recordings are assumed to have been done with good reason. Furthermore, it is not possible to land all species at all locations along the coast. It is known that fishers might discard at sea what they know cannot be landed, instead of discarding it at port, as the fishers may end up having to pay to land unwanted species at the landing facilities. This could be prevented by increased local reception facilities, which would improve the data on species that are not commercially targeted (Berg and Nedreaas 2021).

Due to transitioning from the database Regfisk to Sea2Data at the IMR in 2014, some of the soak times for gillnets were incorrectly transferred. In Regfisk, a separate column was used when coding for durations longer than 99 hours, but unfortunately this column was lost in the conversion. Extensive effort has been made to rectify this in the CRF database, including during the present study, however it is likely that some errors remain undetected. Additionally, errors such as recording one hour as one minute have been detected and corrected by cross-referencing with individual length measurement forms by the IMR prior to this study.

The Norwegian CRF is a dynamic fleet, and over the course of the study period the fishing pattern including gear types and mesh sizes changed in the CRF. This was partly due to the increased participation in seasonal fisheries. The most common species caught by the CRF was cod, and the most common gear used was gillnets. Large variations in CPUE for cod, haddock, and golden redfish were observed, and trends in CPUE varied between the statistical areas. Throughout the study period, regulations for golden redfish have been variable and increasingly restricted, explaining both the dissimilarities in the trends and CPUEs close to 0 for the CRF after 2014. Biological and technological factors affecting the catchability in commercial fisheries are evened out by calculating annual mean CPUE, although caution is advised when using CPUE from fisheries dependent catches to infer abundance of a stock. The technical variations offer, however, insight into the catch composition and fishing pattern of the coastal fisheries, and further analyses of the data collected by the CRF could increase the understanding of the species and the coastal fleet as a whole.

5.1 - CRF into the future

We believe that involving fishermen in scientific data collection will give the end users of scientific advice, i.e., also the fishers, an ownership to the data and hence create more trust in making “the scientific map match the terrain”. We think that an ecosystem and a resource will benefit from co-management and -monitoring and joint data collection of common resources. In this way the CRF provides a cheap way to collect data with a great spatial and temporal coverage, as long as the skipper behaviour is understood, and technology creep is monitored.

The DoF proposes to expand the obligatory data collection from the whole fishing fleet. The purpose is to collect better position and activity data from the smallest fishing fleet. The DoF believes that the data from the fleet below 15 meters will provide increased knowledge of stocks along the coast about which there is currently little data and provide better management of these stocks. This is especially true for the stationary stocks. Among other things, the knowledge will enable the administration to detect danger signals at an earlier stage. The position and effort data will be of great value to the marine land management and in the industry's own work to meet the requirements of various certification schemes. The requirements for tracking and reporting will be introduced step by step. Today, the fishing fleet over 15 meters is required to track via VMS (Vessel Monitoring System) and do catch reporting and to report on where to land the fish and time of landing via ERS (Electronic Reporting System). The new requirements for position and activity data are proposed to be introduced step by step and from 1 July 2022, the requirements for tracking and reporting will apply to the fishing fleet between 11 and 15 meters. For the smaller vessels between 10 and 10.99 meters, the requirements will apply from 1 March 2023 and from and including 1 January 2024 for all vessels under 10 meters.

The CRF will be important and helpful during the implementation progress of this obligatory data collection from the whole Norwegian fishing fleet less than 15 meter. The CRF may e.g., be used as a test fleet for quality assurance and comparisons, and for app development for efficient data handling at sea.

An obligatory electronic reporting of catch and effort from all vessels less than 15 meters will provide a huge dataset of CPUEs, and the main purpose of a CRF may hence slightly change towards more boats with competence and focus on biological self-sampling, and eventually a sampling lottery among these.

Alverson, D.I., Freeberg, M.H., Pope, J.G. & Murawski, S.A. 1994, ‘A global assessment of fisheries bycatch and discards’, FAO Fisheries Technical Paper, vol. 339.

Bakketeig, I.E., Hauge, M. and Kvamme, C. (eds). 2017. ‘Havforskningsrapporten 2017’, Fisken og havet, særnr. 1–2017. Available from https://www.hi.no/filarkiv/2017/06/ havforskningsrapporten_2017.pdf.

Bell, R.J., Gervelis, B., Chamberlain, G., and Hoey, J. 2017. Discard Estimates from Self-Reported Catch Data: an Example from the U.S. Northeast Shelf. North American Journal of Fisheries Management 37: 1130–1144. doi:10.1080/02755947.2017.1350219.

Berg, H.S.F. 2019, June. Estimation of discard of cod (Gadus morhua) in Norwegian gillnet fisheries. University of Bergen.

Berg, H.S.F. and Nedreaas, K. 2021. Estimation of discards in Norwegian coastal gillnet fisheries – 2012-2018. Fisken og Havet 2021-1. ISSN:1894-5031. 95 pp. In Norwegian. Summary in English.

Beverton, R.J.H., and Holt, S.J. 1957. Ministry of Agriculture, Fisheries and Food: Fishery Investigations, Series II.

Bishop, J. 2006. Standardizing fishery-dependent catch and effort data in complex fisheries with technology change. Reviews in Fish Biology and Fisheries 16: 21-38. doi:10.1007/s11160-006-0004-9

Bjørge, A., and Moan, A. 2017. Revised estimates of harbour porpoise (Phocoena phocoena) bycatches in two Norwegian coastal gillnet fisheries. Biological Conservation, 161, 164-73.

Clegg, T., and Williams, T. 2020. Monitoring Bycatches in Norwegian Fisheries -Species registered by the Norwegian Reference Fleet 2015-2018. Rapport fra Havforskningen ISSN:1893-4536 2020-8.

Cochran, W.G. 1977, Sampling Technique, Third., John Wiley & Sons, Inc., New York City, New York.

Crecco, V., and Overholtz, W.J. 1990. Causes of density-dependent catchability for Georges Bank haddock Melanogrammus aeglefinus. Canadian Journal of Fisheries and Aquatic Science 47: 385–394.

Crespin, L., Choquet, R., Lima, M., Merritt, J., and Pradel, R. 2008. Is heterogeneity of catchability in capture-recapture studies a mere sampling artifact or a biologically relevant feature of the population? Population Ecology 50(3): 247–256. Springer. doi:10.1007/s10144-008-0090-8.

Dincer, A.C., and Bahar, M. 2008. Multifilament Gillnet Selectivity for the Red Mullet (Mullus barbatus) in the Eastern Black Sea Coast of Turkey, Trabzon. Turkish Journal of Fisheries and Aquatic Sciences 8(2): 355-359.

Directorate of Fisheries. 2014. J-2-2014: (Utgått) Forskrift om endring av forskrift om regulering av fisket etter torsk, hyse og sei nord for 62°N i 2014. Available from https://www.fiskeridir.no/Yrkesfiske/Regelverk-og-reguleringer/J-meldinger/Utgaatte-J-meldinger/J-2-2014.

Durant, J.M., Hjermann, D.Ø., Sabarros, P.S., and Stenseth, N.C. 2008. Northeast Arctic cod population persistence in the Lofoten–Barents Sea system under fishing. Ecological Applications 18(3): 662–669. John Wiley & Sons, Ltd. doi:10.1890/07-0960.1.

Eide, A., Skjold, F., Olsen, F., and Flaaten, O. 2003. Harvest Functions: The Norwegian Bottom Trawl Cod Fisheries. Marine Resource Economy 18: 81–93. doi:10.1086/mre.18.1.42629384.

Fangel, K., Aas, Ø., Vølstad, J.H., Bærum, K.M., Christensen-Dalsgaard, S., Nedreaas, K., Overvik, M., Wold, L.C., and Anker-Nilssen, T. 2015. Assessing incidental bycatch of seabirds in Norwegian coastal commercial fisheries: Empirical and methodological lessons. Global Ecology Conservation 4: 127–136. Elsevier. doi:10.1016/j.gecco.2015.06.001.

Fuglebakk, E., Otterå, H., Clegg, T., Dingsør, G. & Vølstad, J.H. 2018, 'Assessing representativeness and bias in vessel selection in commercian catch sampling programs conducted in collaboration with the industry.’ In Kennelly, S.J., & Borges, L. (Eds.) Proceedings of the 9th International Fisheries Observer and Monitoring Conference, Vigo, Spain, pp. 157–60.

Gauthier, S. and Rose, C.A. 2002. Acoustic observation of diel vertical migration and shoaling behavior in Atlantic redfishes, Journal of Fish Biology 61(5): 1135-1153. https://doi.org/10.1111/j.1095-8649.2002.tb02461.x.

Gullestad, P., Aglen, A., Bjordal, Å., Blom, G., Johansen, S., Krog, J., Misund, O.A., and Røttingen, I. 2014. Changing attitudes 1970–2012: evolution of the Norwegian management framework to prevent overfishing and to secure long-term sustainability. ICES J. Mar. Sci. 71(2): 173–182. Oxford Academic. doi:10.1093/icesjms/fst094.

Gullestad, P., Blom, G., Bakke, G. & Bogstad, B. 2015, ‘The “Discard Ban Package”: Experiences in efforts to improve the exploitation patterns in Norwegian fisheries’, Marine Policy, vol. 54, pp. 1–9.

Hamley, J. M. 1975. Review of gillnet selectivity. Journal of the Fisheries Board of Canada, 32(11), 1943-1969.

Harley, S.J., Myers, R.A., and Dunn, A. 2001. Is catch-per-unit-effort proportional to abundance? Canadian Journal of Fisheries and Aquatic Science 58(9): 1760–1772.

Hatlebrekke, H.H. 2021, March. A descriptive analysis of data from the Norwegian Coastal Reference Fleet and comparison with catch-per-unit-effort data from research surveys. University of Bergen.

Hilborn, R.. and Walters, C.J. 1992. Stock and recruitment. In Quantitative Fisheries Stock Assessment (pp. 241-296). Springer, Boston, MA.

Huse, G., and Bakketeig, I.E. 2018. Ressursoversikten 2018.

Huse, I., Løkkeborg, S., and Vold Soldal, A. 2000. Relative selectivity in trawl, longline and gillnet fisheries for cod and haddock. ICES Journal of Marine Science 57: 1271–1282. doi:10.1006/jmsc.2000.00813.

ICES. 2017. Anglerfish in ICES Subareas I and II. Available from https://www.ices.dk/sites/pub/Publication Reports/Expert Group Report/acom/2017/AFWG/13-AFWG Report 2017 Section 10_Anglerfish in ICES Subareas I and II.pdf [accessed 13 March 2021].

ICES. 2021. Report from benchmark workshop for Barents Sea and Faroese stocks (WKBarFar 2021). ICES Scientific Reports. 3(21). 205 pp. https://doi.org/10.17895/ices.pub.7920

IMR. 2020. External Review of the Norwegian Institute of Marine Research Stock Assessment Processes. Final Report from an expert group composed of W. Karp (chair), J. Ianelli, A. Rindorf, and G. Stefannson. www.hi.no. 29 pp.

Jakobsen, S.E. and Lindkvist, K.B. 2003. Offentlig politikk og regional næringsutvikling i fiskerinæringen, Samfunns- og Næringslivsforkning AS. Arbeidernotat.

Jennings, S., Kaiser, M.J. and Reynolds, J.D. 2001. Marine Fisheries Ecology, Blackwell Science Ltd, Berlin.

Johnsen, J.P., and Eliasen, S. 2011. Solving complex fisheries management problems: What the EU can learn from the Nordic experiences of reduction of discards. Marine Policy 35: 130–139.

Kolding, J. 1989. The fish resources of Lake Turkana and their environment - Thesis for the Cand. Scient degree in Fisheries Biology and Final Report of KEN 043 Trial Fishery 1986-1987. Dept. of Fisheries Biology, University of Bergen. 262 pp.

Kolding, J., and Skålevik, A. 2009. Introduction and manual to Pasgear 2, version 2.3. University of Bergen Department of Fisheries and Marine Biology High Technology Centre N-5020, Bergen, Norway.

Maunder, M.N., and Punt, A.E. 2004. Standardizing catch and effort data: a review of recent approaches. Fisheries Research 70: 141-159. doi:10.1016/j.fishres.2004.08.002

Maunder, M.N., Sibert, J.R., Fonteneau, A., Hampton, J., Kleiber, P., and Harley, S.J. 2006. Interpreting catch per unit effort data to assess the status of individual stocks and communities. ICES Journal of Marine Science 63: 1373-1385. doi:10.1016/j.icesjms.2006.05.008

Ministry of Trade, Industry and Fisheries. 2017. Generelt om ressurskontroll i fiskeriene. Available from https://www.regjeringen.no/no/tema/mat-fiske-og-landbruk/fiskeri-og-havbruk/ulovlig-fiske/generelt-om-ressurskontroll-i-fiskeriene/id622386/ [accessed 13 March 2021].

Nagelsen, V. 2018. Slik blir et kvoteråd til. Available from https://www.hi.no/hi/nyheter/2018/september/slik-blir-et-kvoterad-til [accessed 13 March 2021].

Nedreaas, K., Borge, A., Godøy, H., and Aanes, S. 2006. The Norwegian Reference fleet: co-operation between fishermen and scientists for multiple objectives. ICES: 1–12.

Næsje, T.F., Hay, C.J., Nickanor, N., Koekemoer, J., Strand, R. and Thorstad, E.B. 2007. Fish populations, gill net catches and gill net selectivity in the Lower Orange River, Namibia, from 1995 to 2001, Norwegian Institute for Nature Research, NINA Report 231, pp. 81. Available from https://munin.uit.no/bitstream/handle/10037/11395/article.pdf?sequence=3&isAllowed=y.

Van Overzee, H., Dammers, M., and Bleeker, K. 2019. Discard self-sampling of Dutch bottom-trawl fisheries in 2017-2018. (No. 19.024). Stichting Wageningen Research, Centre for Fisheries Research (CVO). doi:10.18174/510188.

Paloheimo, J.E., and Dickie, L.M. 1964. Abundance and fishing success. ICES Journal of Marine Science, 155: 152–163.

Pennington, M., and Helle, K. 2011. Evaluation of the design and efficiency of the Norwegian self-sampling purse-seine reference fleet. ICES Journal of Marine Science 68(8): 1764–1768. doi:10.1093/icesjms/fsr018.

Petitgas, P., Poulard, J.C., and Biseau, A. 2003. Comparing commercial and research survey catch per unit of effort: megrim in the Celtic Sea. ICES Journal of Marine Science, vol. 60, no. 1, pp. 66–76. doi: 10.1006/jmsc.2002.1321

Prchalova, M., Kubecka, J., Riha, M., Cech, M., Juza, T., Ketelaars, H.A.M., Kratochvil, M., Mrkvicka, T., Peterka, J., Vasek, M., and Wagenvoort, A.J. 2013. Eel attacks – A new tool for assessing European eel (Anguilla anguilla) abundance and distribution patterns with gillnet sampling. Limnologica 43(3): 194-202. doi: 10.1016/j.limno.2012.09.003.

Regulations on the practice of fishing in the sea / Forskrift om utøvelse av fisket i sjøen . 2005. Available from https://lovdata.no/dokument/SF/forskrift/2004-12-22-1878?q=Forskrift%20om%20ut%C3%B8velse%20av%20fisket [accessed 22 March 2021].

Savina, E., Krag, L.A., Frandsen, R.P., and Madsen, N. 2017. Effect of fisher’s soak tactic on catch pattern in the Danish gillnet plaice fishery. Fisheries Research 196: 56–65. Elsevier B.V. doi:10.1016/j.fishres.2017.08.009.

Sparre, P., and Venema, S.C. 1998. Introduction to tropical fish stock assessment. Part 1: Manual. FAO Fisheries Technical Paper 306(1): 1–423. Available from http://www.fao.org/3/W5449E/w5449e.pdf [accessed 29 January 2021].

Statistics Norway. 2020, Fiskeri (avslutta i Statistisk sentralbyrå). Available from https://www.ssb.no/fiskeri [accessed 22 March 2021].

Verdoit, M., Pelletier, D. and Bellail, R. 2003. Are commercial logbook and scientific CPUE data useful for characterizing the spatial and seasonal distribution of exploited populations? The case of the Celtic Sea whiting. Aquatic Living Resources, 16(6): 467-485. doi.org/10.1016/j.aquliv.2003.07.002

Williams, T., and Gundersen, S. 2020. Havforskningsinstituttets Referanseflåte - Årsrapport 2019/2020.

Ziegler, F., Hornborg, S., Valentinsson, D., Skontorp Hognes, E., Søvik, G., and Ritzau Eigaard, O. 2016. Same stock, different management: quantifying the sustainability of three shrimp fisheries in the Skagerrak from a product perspective. ICES Journal of Marine Science 73(7): 1806–1814. Oxford University Press. doi:10.1093/icesjms/fsw035.

7.1 - Appendix 1 – The Coastal Reference Fleet in 2016

7.2 - Appendix 2 – Contributing vessels per statistical area

7.3 - Appendix 3 – Vessel ID

Table i. Vessel ID and the corresponding unique call sign.

Tabell i. Vessel ID og korresponderende kallesignal.

7.4 - Appendix 4 – All taxa caught during the study period

Table ii. All taxa caught by the Norwegian Coastal Reference Fleet between 2007 and 2019 sorted by %IRI (Index of Relative Importance). % Number is primarily discarded specimens; % Weight (kg) is primarily landed catch.

Tabell ii. Alle arter og taksonomiske grupper som ble fanget av Kystreferanseflåten mellom 2007 og 2019 sortert etter %IRI (i ndeks for relativ viktighet). % Number er primært utkast; % Weight (kg) er primært beholdt fangst.

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

Table ii. Continued.

Tabell ii. Fortsettelse tabell ii

7.5 - Appendix 5 - Sales notes from the entire Norwegian coastal fleet

7.6 - Appendix 6 - Golden redfish regulations

Table iii. The regulations regarding the golden redfish (Sebastes norvegicus) coastal gillnet fishery in Norwegian waters north of 62°N (Regulations on the practice of fishing in the sea 2005, § 34).

Tabell iii. Reguleringer på kystfisket etter vanlig uer (Sebastes norvegicus) med garn i norske farvann nord for 62. breddegrad (Forskrift om utøvelse av fisket i sjøen 2005, § 34).