COD

Gadus morhua


Assessment report
Published by

Marine and Freshwater Research Institute, Iceland

Published

7 June 2024

General information

Cod Gadus morhua is widely dispersed in Icelandic waters, with higher abundance in north-western, northern, and north-eastern part of the shelf. Cod is considered demersal with moderately wide depth distribution which can vary from depths of few meters down to 600 m, occasionally even deeper. Adult cod has not much preference regarding bottom structure and can be found on various substrates; however, a large share of the cod juveniles prefer moderately sheltered, shallow kelp and seagrass environments. The ideal sea temperature for cod is around 4-7°C, nevertheless the temperature limits for this species are somewhat wider, and a significant proportion of the catch is taken where temperature is less than 2°C. Cod spawns all around Iceland by smaller regional spawning components, however the main spawning areas are situated in the south, southwest and west. Spawning starts early in the spring (March-April) on the main spawning grounds in the warmer waters in the south. In the past, spawning started later on in the colder waters in the north, but in recent years spawning time in the north has advanced significantly. North- and eastward pelagic egg- and larval drift mainly occurs clockwise to the nursery grounds situated in the north and north-eastern area. The adult stock takes feeding migrations to the deeper waters in the north-west and south-east, but part stays in the shallow domains to feed. Cod is the most important exploited groundfish species in Iceland.

Fishery

Due to wide spatial distribution of cod in Icelandic waters, the fishing grounds are scattered around the shelf and partially divided by gear type (Figure 5, Figure 6, Figure 7). Demersal trawl is the main fishing gear (Table 1, Figure 5). Main fishing grounds for demersal trawl are situated offshore in deeper relatively cold waters to the north-west, northeast, and east of the island. In recent years, the spatial distribution of demersal trawl fishery has been gradually contracting and aggregating at the previously mentioned trawl fishery hotspots (Figure 5). Longline accounts for the next largest portion of the catch in the cod fisheries and is widely distributed around the Icelandic shelf, with lowest reported catch in the south and southeast coast (Figure 6). The distribution pattern of the catches remains consistent between the years with occasional hotspots. Cod fisheries of the remaining fishing fleet, i.e. gillnets, demersal seine and jiggers, are widely distributed, but mainly take place in shallow waters (Figure 7).

Spatial distribution of the cod fishery has been relatively stable for the past years (Figure 8). Changes in depth and spatial distribution (Figure 3 and Figure 4) are partly caused by changes in gear composition (Figure 2). For cod, the average depth in bottom trawl is 230 m, longline 160 m, but 80 m for demersal seine and gillnets. Mixed fisheries considerations do also affect spatial distribution of the fisheries. For example, cod TAC (Total Allowable Catch) was 50-80% of the cod TAC from 2003-2008 leading to increased fisheries in areas where cod was abundant. For comparison, TAC for cod has been 15-20% of the cod TAC in recent years.

Cod catches increased from under 200 þúsund tonnum to 250 þúsund tonnes frin 1994 to 1999 (Figure 2). Over the next 10 years cod catches steadily declined and reach a low point in 2008, at about 150 þúsund tonnum. Catches increased again and and have fluctuated more recently within 200 - 250 thousand tonnes. Historically, gillnets and bottom trawl were the most important gear with bottom trawls seldom accounting for less than 40% of the catches. The proportion grew from 45% in 2010-2016 to roughly 55% more recently. Gill nets accounted for 20% of catches before 2001 but thereafter declined to around 7% more recently. Most of the bottom trawl catches were taken in the northwest, but gillnet catches were important in the south and west during spawning time. The share of gillnets has declined continuously in recent decades, while that of longlines has increased (Figure 2). Longline fisheries have the widest spatial distribution of the fleets targeting cod (Figure 6), although most of the catches come from the west and northwest. The proportion of catches caught by longlines before 2003 was close to 20%, which grew to 35% during the years 2005 - 2016. Introduction of large longliners with automatic baiting in recent decades has expanded the fishing area of longliners to deeper waters. Demersal seines have steadily accounted for roughly 5-7.5% of cod catches. In some areas, especially in the northwest and southeast, cod can be found in dense schools in certain hotspots, a fact exploited by captains when they want to catch large amount of cod in short time, e.g. just before landing. Condition and size of cod in different areas is also an issue regarding fishing areas, but all those factors are weighed against proximity to landing harbour.

In 2023, more than half of the cod catch was taken in bottom trawl (54%, around 26% on longlines, 6% by gillnets, 6% by jiggers, and 7% by demersal seine. The largest proportion of the catch in recent years was taken in the western and northwestern area, followed by the northeast and southwest areas. Cod was caught at similar depth as in previous years, but perhaps slightly more shallow (Figure 2).

The number of vessels accounting for 95% of the annual catch of cod in Icelandic waters reduced from around 1000 to 750 vessels in 1994-1999 (Figure 1). This reduction occurred despite annual catch increasing by almost 100 thousand tonnes. In 1999-2008, the number of vessels accounting for 95% of the cod catch reduced with reduced total catches to about 400 vessels. Since 2009 the number of vessels has remained relatively constant between 250 and 500, although the most recent years are marked by having the lowest numbers of vessels. This decline is noticeable in all the fleets (Table 1).

Figure 1: Cod in division 5a. Number of vessels (all gear types) accounting for 95% of the total catch annually since 1994. Left: Plotted against year. Right: Plotted against total catch. Data from the Directorate of Fisheries.

Figure 2: Cod in division 5a. Landings in tons and percent of total by gear and year

Table 1: Cod in division 5a. Number of Icelandic vessels landing cod, and all landed catch divided by gear type.
Year Nr. Demersal Seine Nr. Other Nr. Long Line Nr. Bottom Trawl Demersal Seine Other Long Line Bottom Trawl Total catch
2000 133 827 530 183 14930 17623 49946 103558 186057
2001 106 766 515 160 17015 17002 47172 99071 180260
2002 101 724 450 155 13584 19305 42405 87885 163179
2003 107 721 461 147 13375 16026 44654 88422 162477
2004 103 722 470 135 14228 14840 57397 95769 182234
2005 98 604 463 134 12770 8106 69444 84018 174338
2006 93 509 447 126 10358 5859 71037 82417 169671
2007 97 473 425 123 8711 4397 58943 71499 143550
2008 92 427 370 113 8441 4151 53843 58172 124607
2009 81 798 336 113 10370 8190 61005 79667 159232
2010 75 1008 286 111 8296 9372 57491 75609 150768
2011 65 1061 290 110 9106 12665 57711 73538 153020
2012 74 1099 305 118 9989 13417 67777 85265 176448
2013 71 1054 297 110 10092 15237 74835 101453 201617
2014 65 1012 292 110 10407 16355 77807 95830 200399
2015 67 943 266 103 11938 13957 79244 103530 208669
2016 60 956 246 99 15930 15299 84509 111016 226754
2017 67 832 221 94 15398 14945 75244 117891 223478
2018 63 813 201 83 15818 16221 78316 135030 245385
2019 44 794 190 81 14181 13592 78326 135661 241760
2020 43 839 158 81 16198 15884 68103 146788 246973
2021 53 841 145 87 17695 16014 69460 140773 243942
2022 60 850 120 83 16496 15172 62613 124950 219231
2023 61 867 99 81 15958 13286 54213 117211 200668

Figure 3: Cod in division 5a. Depth distribution of cod catches from bottom trawls, longlines, trawls and demersal seine from Icelandic logbooks

Figure 4: Cod in division 5a. Changes in spatial distribution of cod catches as recorded in Icelandic logbooks.

Figure 5: Cod in division 5a. Spatial distribution of catches by demersal trawls.

Figure 6: Cod in division 5a. Spatial distribution of catches by long lines.

Figure 7: Cod in division 5a. Spatial distribution of catches by gillnets, demersal seine, and jiggers.

Figure 8: Cod in division 5a. Spatial distribution of catches by all gears for selected years.

Data available

In general, sampling is considered good from commercial catches from the main gears (demersal seines, longlines, gillnets and trawls). The sampling does seem to cover the spatial and seasonal distribution of catches (see Figure 10 and Figure 11). In 2020, sampling effort was reduced substantially, on-board sampling in particular, due to the COVID-19 pandemic. Although this reduction in sampling continued through 2022, sampling operations have returned to normal and current samples are considered to be sufficiently representative of the fishing operations.

Figure 10: Cod in division 5a. Ratio of samples by month (bars) compared with proportion landings by month (solid black line) split by year and main gear types. Numbers of above the bars indicate number of samples by year, month and gear.

Figure 11: Cod in division 5a. Fishing grounds last year as reported in logbooks (contours) and positions of samples taken from landings (crosses) by main gear types.

Landings and discards

All landings in 5a before 1982 are derived from the STATLANT database, and also all foreign landings in 5a to 2005. The years between 1982 and 1993 landings by Icelandic vessels were collected by the Fisheries Association of Iceland (Fiskifélagið). Landings after 1994 by Icelandic vessels are given by the Icelandic Directorate of Fisheries. Landings of foreign vessels (mainly Norwegian and Faroese vessels) are given by the Icelandic Coast Guard prior to 2014 but after 2014 this are also recorded by the Directorate. Discarding is banned by law in the Icelandic demersal fishery. Measures in the management system such as converting quota share from one species to another are used by the fleet to a large extent and this is thought to discourage discarding in mixed fisheries. In addition to prevent high grading and quota mismatch the fisheries are allowed to land fish that will not be accounted for in the allotted quota, provided that the proceedings when the landed catch is sold will go to the Ministry funds. A more detailed description of the management system can be found on https://www.responsiblefisheries.is/seafood-industry/fisheries-management.

Figure 12: Cod in division 5a. Estimates of annual discards by gear (point estimates and 95% confidence interval). No estimates are available after 2017.

Length compositions

The length distribution of landed catch has shifted towards larger cod in the last ten years (Figure 13). The bulk of the length measurements is from the main fleet segments, i.e. trawls, longlines, gillnets and demersal seine (Table 2). The number of available length measurements by gear has fluctuated in recent years in relation to the changes in the fleet composition.

Figure 13: Cod in division 5a. Commercial length distributions by gear and year

Table 2: Cod in division 5a. Number of samples and length measurements from landed catch.
Year Bottom trawls Num. samples Bottom trawls Num. lengths Demersal seines Num. samples Demersal seines Num. lengths Gillnets Num. samples Gillnets Num. lengths Longlines Num. samples Longlines Num. lengths
2000 766 172132 23 3265 27 4517 124 29780
2001 1131 170398 79 13660 541 39836 281 39915
2002 1233 162365 328 19270 1032 46109 367 46589
2003 1131 114366 428 13648 1214 33432 444 67124
2004 1239 107977 433 18281 780 28705 746 86231
2005 1092 101166 844 18605 916 38072 873 108923
2006 859 79264 680 16333 976 34720 923 115548
2007 946 75259 1013 16850 833 23972 596 96760
2008 849 67630 654 14430 875 23796 562 105976
2009 884 76100 988 16078 763 24751 741 87231
2010 806 77979 757 11241 849 26467 986 81958
2011 596 64643 921 7443 652 29408 765 56099
2012 604 54037 748 8928 646 22778 1124 98415
2013 661 73855 694 2840 765 4272 630 83238
2014 531 46615 262 5340 453 27415 691 96774
2015 554 65641 1018 6858 767 6565 1037 84003
2016 493 57116 1031 7182 797 26568 1060 97164
2017 518 67512 1270 8287 311 7413 368 77691
2018 264 48111 1368 6545 1004 16636 395 74874
2019 451 81165 330 4970 43 5754 292 56710
2020 191 35494 581 3915 226 12606 84 13242
2021 325 53645 900 6468 11 1133 38 4333
2022 228 38180 510 5804 546 1893 52 11348
2023 150 23230 70 3882 162 9741 34 4187

Age compositions

Table 3 shows the number of otoliths samples and number of age readings divided by gear type and Figure Figure 11 shows the location of otolith sampling.

The age composition of the catch has shifted from younger to older fish in the last few decades (Figure 14), likely as a result of decreasing fishing pressure. The number of year classes contributing to the catches has increased in recent years as a result of the low fishing mortality in recent years (Figure 15).

Table 3: Cod in division 5a. Number of samples and otoliths collected from landed catch.
Year Bottom trawls Num. samples Bottom trawls Num. otoliths Demersal seines Num. samples Demersal seines Num. otoliths Gillnets Num. samples Gillnets Num. otoliths Longlines Num. samples Longlines Num. otoliths
2000 766 10034 23 885 27 1051 124 2223
2001 1131 9200 79 500 541 2241 281 2830
2002 1233 8619 328 905 1032 2644 367 2846
2003 1131 8146 428 814 1214 1906 444 3409
2004 1239 9053 433 736 780 1375 746 2483
2005 1092 6503 844 1343 916 1258 873 3874
2006 859 5720 680 1452 976 2107 923 3656
2007 946 6094 1013 1842 833 1745 596 3748
2008 849 5024 654 1181 875 1539 562 2883
2009 884 5418 988 1333 763 1720 741 1972
2010 806 5880 757 1121 849 1505 986 3516
2011 596 5403 921 1417 652 1197 765 2779
2012 604 5757 748 1334 646 1557 1124 3895
2013 661 6194 694 1041 765 1790 630 3302
2014 531 5104 262 747 453 1162 691 2096
2015 554 4937 1018 1686 767 1632 1037 2128
2016 493 5015 1031 2006 797 1674 1060 2183
2017 518 3818 1270 2189 311 908 368 1119
2018 264 2369 1368 2073 1004 1290 395 945
2019 451 2828 330 966 43 300 292 1237
2020 191 1847 581 1397 226 437 84 775
2021 325 2171 900 2304 11 200 38 750
2022 228 1264 510 788 546 972 52 362
2023 150 1511 70 588 162 200 34 440

Figure 14: Cod in division 5a. Catch at age from the commercial fishery in Iceland waters. Bar size is indicative of the catch in numbers and bars are colored by cohort.

Figure 15: Cod in division 5a. Catch at age from the commercial fishery in Iceland waters. Proportion biomass caught by year and age, bars are colored by cohort.

Weight at age in the catch

Mean weight at age in the catch (Figure 16) declined from 2001 to 2007, reaching then a historical low in many age groups. Mean weights at age in age groups 3 to 9have been increasing in recent years and are currently around the average weights observed over the period from 1985, while around 10% below average in older age groups.

Figure 16: Cod in division 5a. Mean weight at age in the catch from the commercial fishery in Icelandic waters. Bars are coloured by cohort.

Natural mortality

No information is available on natural mortality. For assessment and advisory purpose, the natural mortality is set to 0.2 for all age groups.

Catch, effort and research vessel data

Catch per unit of effort from commercial fisheries

Catch per unit of effort data (Figure 17) shows that for hauls where the catch is composed of more than 50 % cod the CPUE has been steadily increasing since 1990 for the main gear types. The CPUE from all catches from all gears is amongst the highest recorded. The catch per unit effort could not be estimated after 2021 as the effort data from several years were not available.

Figure 17: Cod in division 5a. Catch per unit of effort in the most important gear types. The dashed lines are based on locations where more than 50% of the catch is cod and solid lines on all records where cod is caught. A change occurred in the longline fleet starting September 1999. Earlier only vessels larger than 10 BRT were required to return logbooks but later all vessels were required to return logbooks. Data are excluded for years after 2021 all gears and 2022-2023 for gillnets and longlines due to effort data inconsistencies.

Icelandic survey data

The Icelandic spring groundfish survey (hereafter spring survey) has been conducted annually in March since 1985. In addition, the Icelandic autumn groundfish survey (hereafter autumn survey) was commenced in 1996. However, a full autumn survey was not conducted in 2011.

Figure Figure 18 shows both a recruitment index based on abundance of cod smaller than 55 cm, and trends in various biomass indices. Survey abundance by tow site (Figure 19) and changes in spatial distribution (Figure 20).

The total biomass index in the spring survey has been high but fluctuating and with a slight decline over the last decade according to the spring survey index. The total spring (SMB) and fall survey (SMH) measurements decreased significantly from the highest value observed in 2017 to the 2020 measurement, and have since increased again slightly (Figure 18). While the 2021 and 2022 spring survey measurement were on par with that observed in 2018 and 2019, the fall survey measurement in 2021 continued to decline, it being the lowest observed since 2004. The 2020 survey indices were substantially below expectations for size classes that constitute the bulk of the fishable biomass, a trend which continued in 2021 in autumn survey indices but not 2021 spring survey indices. In general, the two surveys have shown similar trends through time (Figure 18) but the contrast through the increase and decline since the late 2000s is greater in the fall survey. The discrepancy between the last two pairs of the spring (2021 and 2022) vs the fall biomass measurements (2020 and 2021) are the highest observed in the time series. A greater decline is therefore observed in the autumn survey biomass index (Figure 18), although this difference between the two survey measurements has declined in most recent years of the survey.

Cod in the spring survey in 2024 was caught all around Iceland, with catch hotspots in offshore waters in the north and southwest, and in shallow waters in the south (Figure 19). The catch on continental slope to the west was similar to the previous year. Spatial distribution of the total biomass index of the catch in the spring survey, shows that the NW and NE areas are dominating in all years (Figure 20). However, some temporal changes have been occurring in recent years with the catch in the NE area decreasing and increasing in the W and SE area. In 2024 there was increase in almost all areas except for the northern areas.

Spatial distribution of cod in autumn survey in 2024 was similar as in previous years (Figure 20). The majority of cod in the autumn survey has been caught on the traditional fishing grounds in the northwest and northeast (Figure 20).

Length distributions from both surveys illustrate quite clearly age groups division in the youngest age groups (Figure 21). Thereafter the division is not quite as clear, due to variability in individual growth and maturity, but some multimodal length distribution can be seen. The large year classes observable in the 2020 and 2021 length distributions are now beginning to enter the fishable stock.

Survey age-based indices of older fish are all relatively high in recent decade despite the pattern that several of the year classes showing high indices recently were showed low - moderate indices when younger (Figure 22). The 2020 spring survey anomaly are clearly apparent, e.g.for year-classes 2014 and 2015 that are around the long term average in 2019 (then ages 4 and 5) but roughly half of that in 2020 (then ages 5 and 6).

Figure 18: Cod in division 5a. Indices (total biomass, biomass > 55 cm, biomass > 80 cm and abundance ≤55 cm) in the Spring Survey (March) 1985 and onwards (line shaded area) and the autumn survey (point ranges).

Figure 19: Cod in division 5a. Location of cod in the 2024 March (SMB) and the 2023 Autumn (SMH) surveys, bubble sizes are relative to catch sizes, and crosses indicate stations where no cod was observed.

Figure 20: Cod in division 5a. Changes in geographical distribution of the survey biomass.

Figure 21: Cod in division 5a. Length disaggregated abundance indices from the March and Autumn surveys.

Figure 22: Cod in division 5a. Age disaggregated indices in the Spring Survey (left) and the autumn survey (rights). Note different scales on y-axes.

Stock weight at age

Mean weights in the spring survey for all ages of cod were below average during roughly 2000 – 2010. After this period, younger ages remained below average but older ages become mainly above average. The autumn survey shows a similar trend, but recent years have begun to show a switch toward lower weights at all ages (Figure 23).

Figure 23: Cod in division 5a. Stock weights from the March survey in Icelandic waters. Bars are coloured by cohort.

Stock maturity at age

Maturity-at-age data are shown in Figure 24. Those data are obtained from the groundfish survey in March. Maturity by age has been generally been below avarage at younger ages, indicating a shift toward older-maturiing fish in recent decades

Figure 24: Cod in division 5a. Maturity at age in the survey. Bars are coloured by cohort. The values are used to calculate the spawning stock.

Data analyses

Analytical assessment

A separable statistical catch at age model (sometimes referred to as MUPPET, described in Björnsson, Hjörleifsson, and Elvarsson (2019)) with four periods where the selection pattern is assumed to be constant. The last separable period is from 2007 to the present. The survey residuals are modeled as multivariate normal distribution to account for potential survey “year effects” - this being a feature in place since 2002. It is a statistical cohort model where fishing mortality can change gradually over time, constrained by a random walk. The same framework is used to carry the stock dynamics forward to evaluate reference points and HCR. This framework was benchmarked in 2021 through a harvest control rule evaluation (ICES (2021b)). The survey residuals are modelled as multivariate normal distribution to account for potential survey “year effects” - this being a feature in place since 2002.

Data used by the assessment

The assessment relies on four sources of data, that are described above. These are the two surveys, commercial samples and landings. The commercial data is used to compile catch at age data that enter the likelihood along with the survey at age from both surveys. Stock weights and catch weights at age are derived from the spring survey and catches respectively. The maturity data is similarly collected in the spring survey. Prior to 1985, when the spring survey started, stock weights and maturity at age were assumed constant at the 1985 values. A full description of the preparation of the data used for tuning and as input is given in the stock annex (see ICES (2021a)) and input data to the assessment are available on the MFRI website www.hafogvatn.is.

The input to the analytical age-based assessment are catch at age 1955–2023 (age 3 to 14) and ages 1 to 14 (from the 1985–2024 spring (often referred to as SMB in this report) and ages 3 to 13 from the 1996–2023 fall groundfish surveys (often referred to as SMH in this report). The method for deriving the catch at age is based on 20 metiers: two areas (north and south), two seasons (January–May and June–December) and five fleets (bottom trawl, longline, hooks (jiggers), gillnet and Danish seine).The reference biomass (4+) upon which the TAC in the fishing year is set is the sumproduct of the population numbers in the beginning of the assessment year and catch weights in that year. The catch weights are not known and hence need to be predicted from stock weights measured in the spring survey, based on a relationship between catch and stock weights observed in the previous year. The mean weight age in the catch (Figure 16) declined from 2001 to 2007, reaching then a historical low in many age groups. The weights at age have been increasing in recent years and are at or above the average in the most important age groups. The variation in the pattern of weight at age in the catches is in part a reflection of the variation in the weight in the stock as seen in the measurements from the surveys (Figure 23).

Diagnostics

The diagnostics (Figure 26) shows the large negative residuals in the spring survey 2020 for the most important age groups (ages 4 to 8), as well as smaller negative residuals for the surrounding years 2019 and 2021, especially for the fall survey. The spring survey residuals are rather high for age groups 10 years and older in recent years, and especially high for age 8 in 2023 and 9 in 2024. A summarised diagnostic of the observed vs predicted survey biomass (Figure 25) illustrates deviation between the model estimates and the point estimates. There are indications that interannual variability in survey measurements in both surveys has increased in recent years compared with that observed in the past.

Figure 25: Cod in division 5a. Aggregated model fit to the total biomass indices. Note that residual correlation is estimated (see text for further details).

Figure 26: Cod in division 5a. Residuals from the model fit to survey and catch data based on the both the surveys. Red circles indicate negative residuals (observed < modelled), while blue postive. Residuals are proportional to the area of the circles.

Model results

The results of this year’s assessment show that the spawning stock in this assessment year is estimated to be 0.377477 kt. Weight and maturity at age used in the calculation of SSB are presented in Table 5. The values estimated in recent years are higher than have been observed during the last five decades. The reference biomass is estimated to be 1075600 t in 2024 and the harvest rate 0.182 in 2023. Year classes since the mid-1980s are estimated to be relatively stable but with the mean around 35% lower than observed in the period 1955 to 1985.

Results of the assessment are provided in the stock summary in Table 5 and Figure 27, as well as on the MFRI website www.hafogvatn.is. The reference biomass has been steady in recent years. The first estimates of the 2021 and 2022 year classes indicate that they are somewhat low, but they will not begin to enter the reference biomass until 2025.

Estimated spawning stock biomass (SSB) has increased in recent years, although fluctuating, and it’s peak in 2017 was larger than in almost 60 years. Harvest rate has declined and is at its lowest value in the assessment period. Recruitment since 1988 has been substantially lower than the average recruitment in the period 1955–1985. The increase in SSB is therefore primarily the result of lower harvest rate. It is estimated that the current fishable biomass is composed of several poor years of recruitment (e.g., 2013 and 2016), but also several good recruitment years (e.g., 2015 and 2019), indicating that variability in biomass levels can be expected to continue.

Figure 27: Cod in division 5a. Summary from assessment. Dashed vertical line indicates the assessment year and yellow shaded region the uncertainty as estimated by the model.

The retrospective pattern of the assessment is shown in (Figure 28) along with the Mohn’s rho values. The default 5-year peels resulted in the following values: 0.033 for recruitment, 0.03 for harvest rate, and for spawning stock biomass. Leave-one-out analyses indicate that the assessment follows trends in the spring survey analysis the closest. Removal of the autumn survey series causes a slight increase in biomass estimates, removal of the spring survey series causes a more substantial decrease, as does following only the catch data (Figure 29).

Figure 28: Cod in division 5a. Analytical retrospective analysis of the assessment of cod with a 5 year peel.

Figure 29: Cod in division 5a. Comparison of assessment results where either the spring survey, the autumn survey, or both are omitted from the estimation.

Short term projections

Landings of Icelandic cod in 2023 are estimated to have been 217157 tonnes, the bulk taken by the Icelandic fleet. To perform short-term projections, estimates of catch for the current calendar year are needed. The projection this year assumes that the remainder of the TAC for the fishing year that ends on August 31. is caught along with the expected catch of the foreign fleet (3 kt). The projected stock status for the interim year and the output from the projection in Table 4.

Mean annual discard of cod over the period 2001-2012 was around 1% of landings in weight, but more recent data indicate that discarding may have increased slightly (MRI 2016). The method used for deriving these estimates assumes that discarding only occurs as high grading.

Following the management plan the advice for the coming fishing year (2024/2025) is based in the biomass of fish aged 4+ at the beginning of the calendar year (2024). Reference biomass is calculated using catch weights, so a deterministic projection of the growth of the catch in weight is needed. In recent years, the estimates of mean weights in the catch of age groups 3–9 in the assessment years have been based on a prediction from the spring survey weight measurements in that year using the slope and the intercept from a linear relationship between survey and catch weights in preceding year (for ages 10 and older the weights from the previous year are used).

Based on this, the mean weights at age in the catches in 2023 are predicted to be close to the long-term average for most age groups (Figure 16), even though the weights in the spring survey in those age groups are below or at the long term mean (Figure 23).

An alternative model based using all data from 1990 onwards to estimate and within each age group 3 to 9 (labelled ‘alt’) was explored in which the catch weight in the assessment year would be predicted using “each age” and the observed stock weights in the assessment year. This alternative model has in the past given much more plausible estimates of catch weights in last year’s assessment (2024), although this year, the reference biomass in the terminal year was very similar ((3.3407194 %). A retrospective analysis indicated that the overall predictive power of the alternative reference biomass was better (cv of 0.035 vs 0.050, bias -0.0020 vs -0.0049) using the alternative model (Figure 30, Figure 31). The alternative model was discussed within the NWWG 2022 that the alternative model could be an improvement over the current spaly weight prediction model. However, it was decided that before implementation, it would be beneficial for the method to be externally reviewed either as a working document appended to next year’s report, or through next benchmark, that for this stock will most likely occur in 2026 or 2027.

Figure 30: Cod in division 5a. Comparison of the short term prediction of catch weights from stock weights (blue), used in the calculation of reference biomass, using a regression across ages of catch weight on stock weight data observed in 2023 (red). Historical values and a regression based on all years of data represent the long-term average relationship in grey.

Figure 31: Cod in division 5a. Comparison of the short term prediction of reference biomass using the standard procedure (spaly) versus an alternate method (alt) to the realised value a year later (black line).

Table 4: Cod in 5.a. “Summary of the assessement and results of the short-term projections when the harvest control rule is applied. In 2024, the assessment includes estimates for SSB, B4+, and recruitment, whereas values for catch and harvest rate are based on intermediate year assumptions. Projections begin with 2025. Fishing year catches are only given for intermediate years and projections.”
Year Calendar Year Catch Fishing Year Catch SSB B4+ HR Recruit. (age 3)
1955 539486 726863 2091220 0.242 151032.0
1956 461780 588192 1818970 0.256 143774.0
1957 454375 575094 1640420 0.303 161483.0
1958 508429 690445 1650990 0.290 215103.0
1959 436801 639696 1580900 0.295 303945.0
1960 474805 583802 1657780 0.246 153744.0
1961 387897 465511 1430570 0.269 195934.0
1962 393690 505861 1464380 0.275 125273.0
1963 408335 460580 1298970 0.329 173269.0
1964 417289 420221 1211060 0.333 197608.0
1965 376136 323049 1053000 0.346 219590.0
1966 346153 295853 1063480 0.321 233032.0
1967 343542 280743 1139740 0.321 320099.0
1968 399163 248564 1242600 0.319 171262.0
1969 388876 354360 1335840 0.338 239766.0
1970 459027 354965 1332990 0.341 179674.0
1971 436786 253213 1083980 0.380 193088.0
1972 393200 225793 979053 0.386 142012.0
1973 361881 245417 831019 0.443 278016.0
1974 362136 188988 909809 0.402 187117.0
1975 361758 175181 891382 0.395 259291.0
1976 331284 146066 948599 0.361 369462.0
1977 337924 199274 1298580 0.257 144156.0
1978 329859 212640 1308980 0.271 223896.0
1979 357435 308024 1409790 0.291 237555.0
1980 423297 370325 1510730 0.301 141783.0
1981 451034 275957 1254480 0.326 145073.0
1982 386498 183093 987452 0.330 141066.0
1983 293409 144153 802754 0.358 227535.0
1984 282869 154353 913772 0.339 143502.0
1985 325493 169405 940339 0.373 140282.0
1986 366011 194963 866433 0.440 299764.0
1987 380240 145613 989352 0.386 250845.0
1988 372356 159668 978580 0.376 176924.0
1989 329697 161505 949862 0.363 97221.6
1990 320762 197665 816583 0.388 130695.0
1991 305691 156893 699315 0.399 113579.0
1992 267035 143236 566540 0.451 160730.0
1993 244266 114758 587416 0.344 129525.0
1994 171845 151164 568857 0.302 80894.8
1995 158645 172776 565297 0.313 142708.0
1996 178234 156771 679849 0.288 166140.0
1997 205340 191306 797531 0.289 92231.2
1998 250490 201671 739073 0.345 156564.0
1999 262623 178197 729258 0.334 76123.1
2000 230756 163610 590621 0.400 167692.0
2001 213547 159972 663695 0.329 155559.0
2002 195343 192648 713703 0.291 158211.0
2003 203923 189334 742223 0.298 180887.0
2004 232687 196477 811977 0.269 85207.6
2005 225649 226044 728900 0.278 154920.0
2006 199202 217507 691882 0.260 132604.0
2007 179458 202042 668181 0.233 95370.5
2008 151217 253816 672217 0.255 131337.0
2009 178911 234039 745824 0.234 117274.0
2010 168064 263775 794567 0.216 125821.0
2011 171098 322624 840655 0.224 165795.0
2012 195313 358291 961248 0.223 176293.0
2013 228333 380761 1084670 0.205 125603.0
2014 219466 348359 1091120 0.208 174820.0
2015 228543 455212 1173130 0.208 148771.0
2016 247259 397722 1227700 0.201 99953.9
2017 247293 525256 1155880 0.224 156956.0
2018 270171 517051 1203980 0.220 167527.0
2019 270377 463460 1152390 0.232 116943.0
2020 266545 415044 1045480 0.256 146136.0
2021 255143 404690 1102950 0.227 130425.0
2022 224560 412477 1090690 0.207 169403.0
2023 206944 392593 1136840 0.182 139550.0
2024 200600 211983 377477 1075600 0.194 112024.0
2025 212239 212939 404273 1069480 0.199 125549.0
2026 212737 212524 419259 1060550 0.201 142138.0
2027 212871 213754 428230 1074920 0.199 132529.0

Management

History

The Ministry of Industries and Innovation is responsible for management of the Icelandic fisheries and implementation of legislation. Cod was included in the ITQ system in 1984, but effort management was also implemented during the first years of the TAC system, partly to help those that thought they received an unfair share of the quota. This “additional effort” management system led to the catches exceeding TAC by 20-30% in the first years of the ITQ system.

In 1990 the law was changed, and effort management eliminated except for the smallest coastal fleet that was managed by fishing days. At the same time, many limitations of the quota transfer were released and the fishing year from 1 September to 31 August was introduced. These laws took effect on 1 September 1991. In the first years, advice by MRI (Marine Research Institute) was based on reducing F (Fishing mortality) by 40%. TAC exceeded advice during those years and catch exceeded TAC. The cod stock reduced rapidly in the early nineties due to low recruitment and high fishing mortality. The need for more strict control of fisheries was apparent and 2-3 years of work by a group of fisheries scientist lead to an adoption of HCR (Harvest Control Rule) for the fishing year 1995/96. The HCR led to significant reduction in fishing mortality.

Since the HCR was introduced, TAC has been set according to the HCR, but catch has exceeded TAC on the average, however by around 5% in recent years (Figure 32, Table 5). The main explanation for catch exceeding advice is that catch in the effort control system exceeded predictions, but the predicted catch is subtracted from the calculated TAC according to the HCR. The current effort control system for the small boats that started in 2009, includes TAC constraint so catches should not exceed TAC by large amount (1-2%). The TAC system does not include catches taken by Norway and the Faroe Islands by bilateral agreement. The level of those catches is known in advance but has until recently not been taken into consideration by the Ministry when allocating TAC to Icelandic vessels. There is no minimum landing size for cod in 5a, although revenues of landings below a minimum size value are heavily penalized to minimize profit. There are agreements between Iceland, Norway and the Faroe Islands relating to a fishery of vessels in restricted areas within the Icelandic EEZ. Faroese vessels are allowed to fish 5600 t of demersal fish species in Icelandic waters which includes maximum 1200 tonnes of cod and 40 t of Atlantic halibut.

A system of catch-quota balancing allowances are in place that allow for species quota transformations and transfers between years as illustrated in Figure 33. Quota transfers from other species to cod are not allowed, and net transfers from cod to other species have been relatively low in recent fishing years (Figure 33, upper). Net transfers of unused cod quota from one fishing year to the next have usually been in the range of 0-7%.

Harvest control rule

The primary essence of the rule is that the TAC for the next fishing year (starting 1. September in the assessment year and ending 31. August next year) is based on a multiplier on the reference biomass of four years and older in the assessment year (B4+). The rule has gone through some amendments and revisions over time. The last significant change occurred in 2007, when the harvest rate multiplier upon which the TAC for the next fishing season is based was changed from 0.25 to 0.20. The current rule has in addition a catch stabilizer. When the SSB in the assessment year is estimated to be above (220 kt) the decision rule is:

The TAC for the current fishing year (2022/2023) based on last year’s assessment was 211309.

Following the benchmark 2021 the reference biomass upon which the advice is based was approximately 20% lower in recent years than based on setting prior to the benchmark. This in part is reflected in somewhat higher recent harvest rate than intended although it is still within the range expected in the HCR simulation. During the benchmark, reference points and the definition of how harvest rate is presented were also updated.

Figure 32: Cod in division 5a. Comparison of the realised catches and the set TAC for the fishing operations in Icelandic waters.

Figure 33: Cod in division 5a. An overview of the net transfers of quota between years and species transformations in the fishery in 5a.

Table 5: Cod in division 5a. ICES advice and official landings. All weights are in tonnes.
Year ICES advice Catch corresponding to advice TAC ICES catch for the fishing year ICES catch for the calendar year
1988* National advice 300 000 350 000   377 554
1989* National advice 300 000 325 000   363 125
1990* National advice 250 000 300 000   335 316
1991* National advice 240 000 245 000   307 759
1991/1992 National advice 250 000 265 000 274 000 264 834
1992/1993 Reduce F by 40% 154 000 205 000 241 000 250 704
1993/1994 Reduce F by 40% 150 000 165 000 197 000 178 138
1994/1995 Reduce F by 50% 130 000 155 000 165 000 168 592
1995/1996 Apply catch rule 155 000 155 000 170 000 180 701
1996/1997 Apply catch rule 186 000 186 000 202 000 203 112
1997/1998 Apply catch rule 218 000 218 000 227 000 243 987
1998/1999 Apply catch rule 250 000 250 000 254 000 260 147
1999/2000 Apply catch rule 247 000 250 000 257 000 235 092
2000/2001 Apply catch rule 203 000 220 000** 221 000 236 702
2001/2002 Apply catch rule 164 000 190 000** 219 000 209 544
2002/2003 Apply catch rule 183 000 179 000** 202 000 207 246
2003/2004 Apply catch rule 210 000 209 000 227 000 228 342
2004/2005 Apply catch rule 205 000 205 000 217 000 213 867
2005/2006 Apply catch rule 198 000 198 000 207 000 197 202
2006/2007 Apply catch rule 187 000 193 000 191 000 171 646
2007/2008 Apply catch rule 152 000 130 000 143 000 147 676
2008/2009 Apply Fmax < 124 000 160 000^ 171 000 183 320
2009/2010 Apply Fmax < 135 000 150 000^^ 170 000 170 025
2010/2011 Apply catch rule 160 000 160 000 167 000 172 218
2011/2012 Apply catch rule 177 000 177 000 185 000 196 171
2012/2013 Apply catch rule 196 000 196 000 213 000 223 582
2013/2014 Apply catch rule 215 000 215 000 226 000 222 021
2014/2015 Apply catch rule 218 000 218 000 223 000 230 165
2015/2016 Apply catch rule 239 000 239 000 251 000 251 219
2016/2017 Management plan 244 000 244 000 237 644 243 945
2017/2018 Management plan 257 572 257 572 270 217 267 221
2018/2019 Management plan 264 437 264 437 265 385 263 025
2019/2020 Management plan ≤ 272 411 272 411 272 385 270 302
2020/2021 No advice requested*** - 256 593 272 137 265 740
2021/2022 Management plan ≤ 222 373 222 373  239 925 242 192
2022/2023 Management plan ≤ 208 846 208 846 219 803 218 181
2023/2024 Management plan ≤ 211 309 211 309
2024/2025 Management plan ≤ 213 214
  • Calendar year.
    ** Amended catch rule.
    *** Advice for 2020/2021 was issued by the Icelandic Marine and Freshwater Research Institute (MFRI; 256 593 tonnes).
    ^ Initial TAC set to 130 000 tonnes according to the catch rule, raised to 160 000 tonnes in January 2009.
    ^^ Set according to the catch rule.
    # The values refer to the catches in the second year within the specified fishing year.

On reference points

Prior to the 2021 benchmark the ICES reference points that matter for the advice (ICES (2021b)) were set the same as in the HCR. Other fishing pressure reference points were set based on the conventional F. In the 2021 benchmark there was a requirement that reference points should be set in accordance with the guidelines and that fishing pressure should be set in the same units as used in the HCR. Since this stock has been fished for quite a while at a rate that is close to that resulting in MSY, the \(B_{pa}\) reference point was based on the 5% percentile of SSB with the stabilizer in the HCR being ignored. The resulting value was 265 kt. This may not be the most optimum approach because the influence of incoming age 4 weigh quite high in the reference biomass, something that is actually ameliorated in the HCR that uses a buffer. If advice is based on no buffer it may be better to base the reference biomass not on catch weights but stock weights, because then the influence of age four would be reduced.

On the measure of fishing pressure

Given the push to define fishing pressure in the same units as used in the HCR one may need to consider how one should derive the harvest rate. For the Icelandic cod this is more cumbersome than normally because the advice is not for a calendar year but fishing year. It was decided to use the following metric in the summary (Table 5) as well as the table in the advice sheet:

\(HR_y = \frac{1/3*Catch_y + 2/3*Catch_{y+1}}{B_{4+, y}}\)

where the fractions represent the proportion of the catch of the fishing year taken in the different calendar year. This measure of fishing pressure is by no means the best one but reflects best the “intended” harvest rate as stipulated in the HCR.

Management considerations

All the signs from commercial catch data and surveys indicate that cod in 5.a is at present in a good state. This is confirmed in the assessment and the recent harvest control rule evaluation (ICES (2021b)). The stock is in a high state; however, highly variable recruitment patterns in the past decade indicate that the stock size is expected to fluctuate in the future. As the harvest control rule has a built-in cap, these fluctuations will be dampened in advice.

References

Björnsson, Höskuldur, Einar Hjörleifsson, and Bjarki ór Elvarsson. 2019. “Muppet: Program for Simulating Harvest Control Rules.” Reykjavik: Marine; Freshwater Researh Institute. http://www.github.com/hoski/Muppet-HCR.
ICES. 2021a. Stock Annex: Cod (Gadus morhua) in Division 5.a (Iceland grounds).” International Council for the Exploration of the Seas; ICES publishing. https://doi.org/https://doi.org/10.17895/ices.pub.18622199.
———. 2021b. “Workshop on the Re-Evaluation of Management Plan for the Icelandic Cod Stock (WKICECOD), ICES Scientific Reports. 3:30.” International Council for the Exploration of the Seas; ICES publishing. https://doi.org/https://doi.org/10.17895/ices.pub.7987.
MRI. 2016. Mælingar á brottkasti þorsks og ýsu (e. Measurments of discards of Cod and Haddock), 2014–2016, Reykjavik, Iceland.” Vol. 3. Marine; Freshwater Research Institute, Iceland; Marine Research Institute, Iceland. https://www.hafogvatn.is/static/research/files/fjolrit-183pdf.