General information
Starry ray is by far the most abundant elasmobranch species in Icelandic waters. It has a widespread distribution over the Icelandic shelf and upper slope at depths from 20-1000 m but is most common at 30-200 m. In Icelandic surveys, starry ray is rarely caught larger than 70 cm but most commonly at 30-50 cm. Reproduction is believed to occur to some extent throughout the year, however mainly during summer.
The fishery and landings
Starry ray is abundant in Icelandic waters and is a common bycatch in variety of fishing gears. Catches of starry ray are taken all around Iceland but mostly within Faxaflói in the southwest (Figure 1). The increased landings since the 1990s are partly related to an increased retention, compensating for a lower abundance of the D. batis complex. However, fishing regulations are likely responsible for the high proportion of landings from Danish seine in the nineties. Since 2007, landings are mainly reported from the longline fishery (Figure 2). Reported landings increased from 500 tonnes in 2007 to more than 1700 tonnes in 2012. Thereafter, landings have shown a steady decline and in 2022 they didn´t exceed 300 tonnes. Last year there are drastic decline in landings from the longline fishery (Figure 2). A large proportion of the landings is for local consumption linked to the yule season. This is reflected in the strong seasonality in landings; most landings are reported from September-November each year (Figure 3).
Survey data
Distribution and biomass indices
Starry ray is a frequent catch in MFRI spring (IS-SMB) and autumn surveys (IS-SMH). Seasonal differences in distributional patterns have been noted, with starry ray much less abundant on the shelf in IS-SMH than in IS-SMB. In IS-SMB, starry ray is found at 86% of all stations, but at about 50% of stations in the IS-SMH (Figure 4).
In MFRI groundfish surveys, starry ray is most abundant in the N and the NW (Figures Figure 5 and Figure 6). In IS-SMB there is a high abundance on the shelf off N-Iceland and in near-shore areas in the south and southeast (Figure 5 a,c,e). In IS-SMH, the main distribution is on the shelf break and starry ray is almost absent from the southern area (Figure 5 b,d,f). Seasonal migration could to some extend explain these seasonal differences in distributional patterns. However, the large seasonal difference in occurrence and catches, especially in the smallest length groups (>30 cm, Figure 5 c,d) could also be partly explained by differences in survey gear (size and weight). Starry ray is a frequent bycatch in several other MFRI surveys.The coastal shrimp survey occurs at various time periods in fjords and near coastal areas and starry ray is widely distributed within the survey areas (Figure 6 a). Similarly, starry ray is a frequent bycatch in the gillnet survey occurring early April each year (Figure 6 b).
In general, estimates of total biomass of starry ray in IS-SMB show a declining trend over the survey period 1985-2024 with few exceptions such as the estimate from 2023 which is the highest since 2004 (Figure 7).The biomass index in IS-SMB has decreased from 19 000 (average 1985-2000) to 14 000 (average 2001-2024). Decreasing trend is in particularly notable for large fish (≥50 cm) in years 1993-2008. Since 2010 the index for large fish has remained relative stable. Estimated biomass of juveniles (≤20 cm) in IS-SMB showed large variations in years 2003-2013 but appears to be stable with increasing trend in last decade. In IS-SMH, total biomass are overall lower than in IS-SMB and in particular, small individuals (≤20 cm) are hardly caught.
In IS-SMB the highest proportion of catch is taken in areas off NW-, NE- and SE- Iceland and the reduction in biomass is most prominent in these areas. In IS-SMH, the highest proportion of catch is taken in areas off NW- and NE-Iceland; the areas where a reduction in abundance has taken place (Figure 8).
Life history information
Length distributions from surveys indicate that most specimens are <60 cm . Mean size varies from 35-49 cm depending on surveys (Figure 9). The length distribution is negatively skewed as the proportion of large fish decreases quite abruptly (Figure 9) which is likely due to morphological attributes of the species.
Mean length in the spring survey is the lowest in all six surveys and considerably lower than mean length in IS-SMH (overall mean 35 and 40 cm, respectively). The proportion of larger fish decreases quite abruptly after reaching 50 cm (Figure 10 and Figure 11). In IS-SMB, the mean length has decreased from 38 cm (average 1996-1998) to 35.8 cm (average 2019-2023) (Figure 10). On the other hand, in IS-SMH the mean length has varied (from 38 cm to 43 cm) over the period without any specific direction (Figure 11).
The sex ratio is 1:1 in the spring survey, but in the autumn survey the ratio is skewed towards females (male:female ratio 1:1.57). Males are on average larger than females (40.5 cm and 38.8 cm, respectively). Data on maturity is sampled in the autumn survey allowing for calculations of maturity ogives. Length-at-50%-maturity (L50) is 43.3 cm and 41.9 cm for males and females, respectively (Figure 12). Anecdotal information suggests that starry ray undertakes seasonal migrations related to egg-laying activity. Recently, both surveys have started to sample data on egg case distribution, but trawl survey data may provide useful information on catches of viable skate egg cases and/or nursery grounds.
Stock assessment
Last year was the first attempt made to perform an analytical assessment of starry ray in Icelandic waters, based recommendations from ICES (2022). The starry ray is considered a data limited stock and follows the ICES framework for such (category 3.1, ICES 2021). A stochastic surplus production model in continuous time (SPiCT; Pedersen and Berg, 2017) is one of the official assessment methods for stocks in this category. The model quantifies observation and process errors and estimates stock status and reference levels with associated confidence intervals. SPiCT estimates MSY based reference levels, which can be used to calculate quantities relevant for fisheries management and ICES recommends using the 35th percentile for all quantities (Mildenberger et al., 2021)
Input data
The model synthesizes information from input priors, landings series from 1991-2022 and survey indices from the Icelandic Groundfish survey (IS-SMB) from 1985-2024). Priors used for the model were the intrinsic growth rate, r ̅ , and the medium initial biomass depletion, P ̅ and the n is fixed at 2 to resemble the Schaefer production curve (ICES 2021).
Table.1 Starry ray. Priors in model
| Priors | Value | Standard deviation |
|---|---|---|
| r | log(0.4) | 0.04 |
| p | log(0.5) | 0.25 |
Results
The output from the model is shown below in table 2 and 3. Model results are shown in Figure 13, the model diagnostics in Figure 14 and the analytical retrospective analysis in Figure 15. Following the checklist for the acceptance of SPiCT model (Mildenberger et al., 2021), one minor issue was found i.e. the Shapiro test indicate some non-normality in the residuals. However, slight violations of these assumptions do not necessarily invalidate model results. Apart from that issue, there are no violations of model assumptions based on one-step ahead residuals, the production curve is realistic (B/K = 0.5) (Figure 13) and the patterns in the retrospective analysis are consistent (Figure 15) . BMSY is estimated at 12.3 kt (Table 3). Annual estimates of B/BMSY and F/FMSY is shown in table 4.
Table 2. Starry ray. Summary of model results
| Estimate | 95% upper CI | 95% lower CI | |
|---|---|---|---|
| alpha | 72.620 | 46924.76 | 0.112 |
| beta | 0.058 | 102.94 | 0.000 |
| r | 0.148 | 0.253 | 0.086 |
| rc | 0.148 | 0.253 | 0.086 |
| rold | 0.148 | 0.253 | 0.086 |
| m | 909.25 | 1043.88 | 791.98 |
| K | 24649.96 | 37540.31 | 16185.82 |
| q | 0.000 | 0.000 | 0.000 |
| sdb | 0.002 | 1.354 | 0.000 |
| sdf | 0.481 | 0.630 | 0.367 |
| sdi | 0.152 | 0.193 | 0.120 |
| sdc | 0.028 | 44.011 | 0.000 |
Table 3. Starry ray. Summary of model results. Estimates of reference points
| Reference points | Estimate | 95% upper CI | 95%lowerCI |
|---|---|---|---|
| Bmsy | 12325.0 | 18770.2 | 8092.9 |
| Fmsy | 0.074 | 0.127 | 0.043 |
| MSY | 909.2 | 1043.9 | 792.0 |
Table 5. Starry ray. Estimates of B/BMSY and F/FMSY with 95% confidence intervals from the SPiCT model.
| Year | 95% lower CI | B/Bmsy | 95%upper CI | 95% lower CI | F/Fmsy | 95%upper CI |
|---|---|---|---|---|---|---|
| 1985 | 0.9877 | 1.2513 | 1.5851 | 0.0238 | 0.2722 | 3.1109 |
| 1986 | 1.0368 | 1.2942 | 1.6156 | 0.0283 | 0.2687 | 2.5503 |
| 1987 | 1.0758 | 1.3350 | 1.6567 | 0.0343 | 0.2604 | 1.9778 |
| 1988 | 1.1097 | 1.3742 | 1.7016 | 0.0421 | 0.2483 | 1.4654 |
| 1989 | 1.1426 | 1.4118 | 1.7443 | 0.0526 | 0.2342 | 1.0426 |
| 1990 | 1.1776 | 1.4481 | 1.7807 | 0.0696 | 0.2178 | 0.6814 |
| 1991 | 1.2164 | 1.4832 | 1.8086 | 0.1118 | 0.2033 | 0.3697 |
| 1992 | 1.2574 | 1.5156 | 1.8270 | 0.1435 | 0.2276 | 0.3611 |
| 1993 | 1.2943 | 1.5432 | 1.8401 | 0.1222 | 0.1971 | 0.3178 |
| 1994 | 1.3285 | 1.5681 | 1.8508 | 0.2489 | 0.3972 | 0.6340 |
| 1995 | 1.3032 | 1.5192 | 1.7711 | 0.8307 | 1.2925 | 2.0110 |
| 1996 | 1.2436 | 1.4364 | 1.6590 | 0.7885 | 1.2215 | 1.8922 |
| 1997 | 1.1981 | 1.3765 | 1.5815 | 0.7469 | 1.1530 | 1.7802 |
| 1998 | 1.1578 | 1.3264 | 1.5195 | 0.7340 | 1.1269 | 1.7301 |
| 1999 | 1.1274 | 1.2896 | 1.4751 | 0.6523 | 1.0007 | 1.5353 |
| 2000 | 1.1098 | 1.2679 | 1.4486 | 0.5992 | 0.9210 | 1.4155 |
| 2001 | 1.0958 | 1.2509 | 1.4279 | 0.6068 | 0.9376 | 1.4488 |
| 2002 | 1.0710 | 1.2229 | 1.3962 | 0.8943 | 1.3737 | 2.1100 |
| 2003 | 1.0006 | 1.1491 | 1.3197 | 1.1077 | 1.6826 | 2.5560 |
| 2004 | 0.9521 | 1.1009 | 1.2728 | 0.8024 | 1.2223 | 1.8620 |
| 2005 | 0.9423 | 1.0917 | 1.2649 | 0.5148 | 0.7882 | 1.2069 |
| 2006 | 0.9603 | 1.1113 | 1.2860 | 0.3648 | 0.5630 | 0.8689 |
| 2007 | 0.9866 | 1.1407 | 1.3190 | 0.3059 | 0.4754 | 0.7388 |
| 2008 | 1.0131 | 1.1721 | 1.3560 | 0.3265 | 0.5112 | 0.8002 |
| 2009 | 1.0285 | 1.1917 | 1.3806 | 0.4491 | 0.7018 | 1.0967 |
| 2010 | 1.0275 | 1.1922 | 1.3833 | 0.5637 | 0.8822 | 1.3808 |
| 2011 | 1.0154 | 1.1798 | 1.3708 | 0.6816 | 1.0729 | 1.6888 |
| 2012 | 0.9811 | 1.1418 | 1.3287 | 1.0279 | 1.6082 | 2.5162 |
| 2013 | 0.9099 | 1.0639 | 1.2439 | 1.2209 | 1.9011 | 2.9603 |
| 2014 | 0.8454 | 0.9981 | 1.1784 | 1.1846 | 1.8528 | 2.8979 |
| 2015 | 0.7851 | 0.9400 | 1.1256 | 1.0967 | 1.7207 | 2.6997 |
| 2016 | 0.7414 | 0.9006 | 1.0940 | 1.1003 | 1.7310 | 2.7233 |
| 2017 | 0.7069 | 0.8720 | 1.0756 | 0.6827 | 1.0790 | 1.7053 |
| 2018 | 0.7209 | 0.8925 | 1.1051 | 0.4057 | 0.6615 | 1.0785 |
| 2019 | 0.7371 | 0.9165 | 1.1396 | 0.5968 | 0.9675 | 1.5683 |
| 2020 | 0.7261 | 0.9120 | 1.1457 | 0.6563 | 1.0649 | 1.7279 |
| 2021 | 0.7263 | 0.9198 | 1.1649 | 0.6542 | 1.0714 | 1.7548 |
| 2022 | 0.7296 | 0.9320 | 1.1905 | 0.2712 | 0.4445 | 0.7286 |
| 2023 | 0.7691 | 0.9853 | 1.2624 | 0.1443 | 0.2469 | 0.4226 |
| 2024 | 0.8074 | 1.0385 | 1.3358 | 0.1536 | 0.2937 | 0.5617 |
| 2025 | 0.8431 | 1.0891 | 1.4068 | 0.0936 | 0.2937 | 0.9219 |
Quality of the assessmment
More tests, particularly the choice of priori distributions, may be considered. Retrospective pattern is often related to uncertainty about the shape n parameter and fixing it or constraining it using prior, often reduce the retrospective patterns. Shorter landing series was used (1991-2023) because the artefact of initial states is possible if catch series starts earlier than index (Maguire and Berg, 2020). Also, reporting on less valued/no valued species such as starry ray was inaccurate in the hayday of landing reports. There is a high seasonality in landings data that doesn’t reflect biological attributes of the species but rather increased demand of the fish in the last quarter of the year (MFRI technical report 2024). This could be explored for estimation of discard rate but discard rates are not known for starry ray in Icelandic waters. Studies elsewhere suggest resilience of this species to discard and thus relatively low discard mortality (Ellis et al. 2017, Knotek et al. 2019). Thus, discard mortality or survival of starry ray in Icelandic waters should be estimated.
Refernces
Ellis, J. R.,McCully Phillips, S. R. and Poisson, F. 2017. A review of capture and post-release mortality of elasmobranchs. Journal of Fish biology. 90(3): 653-722.
ICES (2021). Benchmark Workshop on the development of MSY advice for category 3 stocks using Surplus Production Model in Continuous Time; SPiCT (WKMSYSYSPICT). ICES Scientific Reports. Report. https://doi.org/10.17895/ices.pub.7919
ICES. 2022. ICES technical guidance for harvest control rules and stock assessments for stocks in categories 2 and 3. In Report of ICES Advisory Committee, 2022. ICES Advice 2022, Section 16.4.11. https://doi.org/10.17895/ices.advice.19801564
Knotek, R., Kneebone, J., Sulikowski, J., Curtis, T., Jurek, J., and Mandelman, J. 2019. Utilization of pop-up satellite archival transmitting tags to evaluate thorny skate (Amblyraja radiata) discard mortality in the Gulf of Maine groundfish bottom trawl fishery. ICES Journal of Marine Science. 77(1). 256-266.
Maguire, JJ and Berg CW. 2020. A SPiCT ASSESSMENTS OF THE NORTH ATLANTIC SHORTFIN MAKO SHARK. ICCAT. Collect. Vol. Sci. Pap. ICCAT, 76(10): 156-163.
Mildenberger, T.K., Kokkalis, A., Berg, C.W. 2022. Guidelines for the stochastic production model in continuous time (SPiCT). https://raw.githubusercontent.com/DTUAqua/spict/master/spict/inst/doc/spict_guidelines.pdf
Pedersen, M.W., Berg, C.W., 2017. A stochastic surplus production model in continuous time. Fish and Fisheries, 18: 226-243.