Longspine thornyhead (Sebastolobus altivelis) COSEWIC assessment and status report: chapter 7

Population sizes and trends

Information sources

The trawl observer program archives the earliest BC records on the longspine thornyhead fishery. Trawl catch information is stored in DFO’s PacHarvTrawl database. Prior to 1996, fishermen did not target this deep-water species, and dockside agencies recorded landed specimens as “thornyheads”. The directed fishery for longspine thornyheads started off the WCVI in a region dubbed “Beginner’s Ledge”, so-called due to its ease of access by all existing trawlers. Market demand from Japan drove prices higher and management actions spurred development of the fishery northward (Figure 7). Trawl catch history by area is detailed in Table 3; the annual catch (trawl + longline) and quota history of this species is detailed in Table 4.

Figure 7.  Spatio-temporal evolution of the longspine thornyhead fishery from its initial start on Beginner’s Ledge (horizontal band at UTM North 5400-5450) off WCVI in 1996. Total catch and mean CPUE are indicated for each 4-month interval. Source: Schnute et al. (2004).

Figure 7.  Spatio-temporal evolution of the longspine thornyhead fishery from its initial start on Beginner’s Ledge (horizontal band at UTM North 5400-5450) off WCVI in 1996. Total catch and mean CPUE are indicated for each 4-month interval. Source: Schnute et al. (2004).
Table 3.  Annual (fishing year) total catch (kept + discarded) by the trawl fishery of longspine thornyhead (tonnes) in PMFC areas along the BC coast (3CD ≈ west coast of Vancouver Island, 4B ≈ Strait of Georgia, 5AB ≈ Queen Charlotte Sound, 5CD ≈ Hecate Strait, 5E ≈ west coast of the Queen Charlotte Islands, UNK =Unknown, CST = coastwide). Catches are rounded to the nearest tonne; entries marked ‘--’ indicate no recorded catch. Data reside in PacHarvTrawl. Fishing years run from April to March, unless otherwise noted.
Year 3C 3D 4B 5A 5B 5C 5D 5E UNK CST
UNK
---
---
---
---
---
---
---
---
28
28
19961
466
396
---
0
2
0
0
1
---
867
19972
185
107
---
0
0
---
---
1
---
293
1997
361
203
---
7
2
0
0
1
---
575
1998
431
392
---
6
1
0
1
8
---
839
1999
141
751
---
0
1
---
0
19
---
912
2000
163
513
---
54
31
0
0
144
---
905
2001
185
271
---
28
22
0
0
144
---
650
2002
219
249
---
48
27
1
---
116
---
660
2003
132
165
---
53
22
0
2
73
---
448
2004
137
98
---
6
6
0
1
55
---
304
2005
38
31
---
0
1
0
---
13
---
83
Total
2,458
3,176
0
203
115
2
5
576
28
6,564

1 Feb-Dec;
2 Jan-Mar

Table 4.  Annual (fishing year) total catch (kept + discarded) by the trawl and hook and line (HL) fisheries of longspine thornyhead (tonnes) along the BC coast. Historical quotas are reported from various management plans. Values are rounded to the nearest tonne; entries marked ‘--’ indicate no recorded catch or quota. Data reside in the PacHarvTrawl and PacHarvHL databases. Fishing years run from April to March, unless otherwise noted.
Fishing
Year
Trawl Catch
HL
(t)
Halibut
Total Trawl Quota
HL
(t)
Halibut
Total
19961
867
---
0
867
---
654 C
---
654
19972
293
---
---
293
225
---
---
225
19973
575
---
---
575
860
900
---
1,760
1998
839
---
---
839
861
39
---
900
1999
912
---
---
912
855
45
---
900
2000
905
0
---
905
404 V + 425 L
29 C
30 C
889
2001
650
2
---
650
405 V + 425 L
27 C
28 C
885
2002
660
0
0
660
405 V + 230 X
27 C
28 C
690
2003
448
---
---
448
405 V + 230 X
26 C
28 C
690
2004
304
---
---
304
405 V + 230 X
26 C
28 C
690
2005
83
NA
NA
83
405 V + 230 X
27 C
28 C
690
Total
6,536
2
0
6,536
7,000
1,800
170
8,973

1 Feb-Dec;
2 Jan-Mar for Trawl;
3 Jan 97 – Mar 98 for HL
V west coast Vancouver Island
L exploratory quota for fishing north of line 230° True from Lookout Island
X experimental quota for fishing north of 50°30'N
C quota for longspine and shortspine thornyheads combined

Although this directed fishery was technically new in 1996, the standards of DFO’s New Emerging Fisheries Policy[1] were not well-defined at that time. One of the guiding principles of the NEFP states: “Information on the abundance, distribution, and productivity of the target species (must be) identified as the key scientific requirement for development of precautionary management strategies.” Distribution of longspine thornyhead is fairly well-known given the nature of the observer program (100% coverage). Additionally, DFO management has demanded the collection of synoptic biological samples, including otoliths for ageing. While the latter are of limited use due to the lack of a feasible ageing protocol, they represent a tremendous information base once this limitation is overcome. All biological information is stored in DFO’s GFBio database. At present this fishery is not profitable due to factors listed below. Consequently, pressure on the population has lessened.

Abundance

Absolute abundance remains unknown. Relative abundance estimates appear in the following sections and in Schnute et al. (2004). From 1996 to 2005, the commercial fleet removed approximately 6,564 t of longspine thornyhead biomass from BC coastal waters. This equals 57.6 million fish, assuming a conversion rate of 0.114 kg/fish.

Fluctuations and trends

Observed commercial trawl

Haigh et al. (2005) analyze longspine thornyhead catch/effort data from the DFO PacHarvTrawl database using a general linear regression model (GLM) assuming a log-normal distribution. The analysis uses data from April 1, 1996, coinciding with 100% independent observer coverage on all major trawl operators. The analysis excludes tows from depths less than 500 m due to the potential of species misidentification with shortspine thornyheads Sebastolobus alascanus. Three discrete fisheries for longspine thornyheads exist – the west coast of Vancouver Island called “WCVI” (PMFC regions 3C and 3D), the outer bank of Queen Charlotte Sound called “Tidemarks” (PMFC regions 5A and 5B), and the west coast of the Queen Charlotte Islands called “Rennell” (PMFC region 5E).

A GLM analysis for each fishery produced declining trends of relative CPUE (catch per unit effort) beginning from the first year of each fishery (Figure 8). Linear regression through the index points (Table 5) transformed by natural logs yields annual trend as r = eb - 1, where b = the fitted slope (Schnute et al. 2004). The accumulated relative change over N observations is  RN = eb(N-1) - 1. A comparison of the trends shows little difference between WCVI (r = 0.0761, R9 = -0.469, 1996-2004) and Tidemarks (r = -0.0900, R5 = -0.314, 2000-2004), in spite of the paucity of data for Tidemarks. The estimated CPUE decline in Rennell (r = -0.201, R5= -0.593, 2000-2004) greatly exceeds the declines in the other two areas.  A coastwide index comprising the regional indices weighted by bottom area between the 500 and 1,600 m isobaths (Haigh et al. 2005) yields an annual decline r= -0.0825 and an accumulated decline R9 = -0.498.

Figure 8. Comparison of CPUE indices for the three areas analyzed.  Each series has been standardized relative to the geometric mean over the period 2000/01 to 2004/05.  The error bars show 95% confidence limits. Source: Haigh et al. (2005).

Figure 8.  Comparison of CPUE indices for the three areas analyzed.  Each series has been standardized relative to the geometric mean over the period 2000/01 to 2004/05.  The error bars show 95% confidence limits. Source: Haigh et al. (2005).

WCVI longspine thornyhead survey

DFO in collaboration with the Canadian Groundfish Research and Conservation Society (CGRCS) conducted a WCVI survey targetting longspine thornyheads annually from 2001 to 2003 (Starr et al. 2002, Starr et al. 2004, Krishka et al. 2005). Survey depths ranged from 500 to 1,600 m, while the areal extent ranged from 48°05'N to 50°30'N. Schnute et al. (2004) assessed this survey and compared it with commercially available indices. Their findings indicate that during the period of the survey (2001-2003), the abundance index exhibited no significant trend.

Table 5.  Annual index values from GLM analyses on commercial CPUE data from three fishing areas (Haigh et al. 2005), standardized to the mean of the 2000-2004 values. The coastwide index comprises the regional indices weighted by bottom area between the 500 and 1,600 m isobaths (WCVI = 8,506 km², Tidemarks = 2,908 km², Rennell = 3,162 km²; Haigh et al. 2005). Calculated parameters summarize index trends: b = slope of linear regression line describing ln / = a + by , where I = index value, y = year; r = annual relative growth rate = e b - 1; R = accumulated relative change over N-1 periods = e b(N-1) - 1, where N = number of indices I.
Year WCVI Tidemarks Rennell Coast
1996
1.73103
---
---
1.73103
1997
1.42119
---
---
1.42119
1998
1.31118
---
---
1.31118
1999
1.19443
---
---
1.19443
2000
1.18994
1.22008
1.35363
1.23146
2001
0.99237
1.04646
1.29210
1.06818
2002
1.01706
0.94322
1.00389
0.99947
2003
0.95420
1.01830
0.78692
0.93070
2004
0.84643
0.77194
0.56347
0.77019
Trend Parameters
b
-0.07917
-0.09428
-0.22488
-0.08607
r
-0.07612
-0.08997
-0.20138
-0.08247
R
-0.46920
-0.31417
-0.59323
-0.49768

Trend summary

The relative survey indices from 2001-2003 appear consistent with the relative CPUE indices computed from commercial catch for this assessment and with an analysis of the CPUE information for the west coast of Vancouver Island in the most recent PSARC (Pacific Science Advisory Review Committee) assessment of longspine thornyhead (Schnute et al. 2004) (Figure 9).  The latter analysis is based on the same data but on different analytical methods than the analysis shown in Figure 8.

Haigh et al (2005) showed no overall trend in proportion of commercial sets with zero catches in the three time series listed above (west coast of Vancouver Island, 1996-2004; Tidemarks 2000-2004; Rennell 2000-2004).

Figure 9.  Comparison of WCVI CPUE indices from the most recent PSARC assessment of longspine thornyhead (Schnute et al. 2004), longspine survey indices (as in Fig. 8), and new indices calculated in 2005 for PMFC areas 3C+3D (as in Fig. 8).  All indices have been set so that the geometric mean from 2001/02 to 2003/04 equals 1. Source: Haigh et al. (2005).

Figure 9.  Comparison of WCVI CPUE indices from the most recent PSARC assessment of longspine thornyhead (Schnute et al. 2004), longspine survey indices (as in Fig. 8), and new indices calculated in 2005 for PMFC areas 3C+3D (as in Fig. 8).  All indices have been set so that the geometric mean from 2001/02 to 2003/04 equals 1. Source: Haigh et al. (2005).

The three series of relative CPUE indices derived from commercial catch and effort data (Figure 8) show rates of decline of 8% per year for WCVI, 9% per year for Tidemarks, and 20% per year for Rennell. These rates expressed as total declines yield 47% over 8 years (1996-2004) for WCVI, 31% over 4 years (2000-2004) for Tidemarks, 59% over 4 years (2000-2004) for Rennell. Few other species experience directed fishing effort below 800 m, hence the CPUE index is considered suitable for tracking abundance. However, CPUE may be influenced by factors related to the fishery in addition to abundance changes. Some of these factors cited by Schnute et al. (2004) are:

  • Fishermen have experienced a recent increase in sablefish bycatch when fishing for longspines, especially in the north. Without adequate sablefish quota, skippers must seek out fishing opportunities where tows are less productive for sablefish and, consequently, longspines (abundance of the two species is positively correlated). This behaviour depresses the index.
  • In the early years of the fishery, observers did not always sample to determine the species split between shortspine and longspine thornyheads, relying instead on information from the factory. More recent samples attempt to identify the complete species composition of each tow. This change in behaviour has possibly introduced a bias across years. The improvement in taxonomic ability would depress the index if observers in earlier years tallied shortspines as longspines.
  • Fuel costs have increased substantially. The fishery on longspines ranks high in fuel consumption among all the groundfish fisheries, with tow durations in the range 4-12 h. Higher fuel costs and lower profit margins tend to discourage directed or exploratory fishing on the resource. This behaviour increases the incidence of tows where longspines are not the dominant organism, and consequently depresses the index.
  • The price of thornyheads has declined substantially in the last few years, partly due to an increase in the Canadian dollar relative to the US and Japanese currencies. A reduced profit margin tends to discourage directed fishing. Again, fewer tows where longspine thornyhead is predominant depresses the index.

The three survey abundance indices obtained from 2001 to 2003 for the WCVI population of longspine thornyheads show no trend over the three years of the survey (slope=+0.1%). The trend from the WCVI survey is a more reliable abundance indicator over these three years than CPUE trends from fishery-dependent data, although it covers a much shorter time period.

The prediction of a “three generational” change for this population based on results from short time series remains highly uncertain.  However, the substantial decline in survey indices over a short period, combined with the life history characteristics of this species, suggest that abundance has undergone a substantial reduction which may be difficult to reverse or manage.

Rescue effect

Bordering populations in Washington and Alaska could act as population sources given the planktonic nature of the larvae, although there is no direct evidence to support this scenario. Along the continental USA, the 2005 estimates of spawning stock and total biomass were 75,049 t and 162,642 t, respectively (Fay 2006). The ratio [spawning stock biomass]:[pristine (unfished equilibrium) biomass] was estimated at 0.71, suggesting a lightly exploited population.  A fishery management regime including TACs (total allowable catches) adjusted to prevent overfishing is in place in the Washington-California area.

[1] New Emerging Fisheries Policy.

Page details

Date modified: