COSEWIC assessment and status report on the Atlantic wolffish Anarhichas lupus in Canada
Population Size and Trends
For evaluation by COSEWIC, far and away the most important data for marine fishes are those which document numerical decline. There is a wealth of such data. They come from regular scientific surveys conducted by government agencies and expressly designed to monitor changes in abundance of demersal fishes. The same data can also be used to to generate secondary but also useful information on possible changes in the number of fish per trawl, the mean size of the fish and the range and habitat.
Data from random-stratified scientific survey trawls off eastern Newfoundland (Atkinson, 1994) were provided by Canada’s Department of Fisheries and Oceans and summarized by Villagarcía (1995; see also Haedrich and Barnes, 1997). The surveys are intended mainly to assess the size of commercial fish stocks, but they also catch most species in the demersal fish community (Brown et al., 1996). The number of individual trawls made (the number of stations) in any one year can be well over a thousand. Subsequently, in July to September 2000, DFO made large amounts of trawl survey data and analyses available from all Canadian areas specifically for this wolffish assessment.
The number/tow (which fishery biologists refer to as the “catch per unit effort” or CPUE) from scientific surveys is used as an index of population size. Over the period from 1978 to 1993, this index was calculated as the total number of Atlantic wolffish caught off Newfoundland in a year divided by the total number of stations sampled at appropriate depth and temperature ranges for the species in that year. The appropriate depth and temperature ranges for the Atlantic wolffish are determined using the niche axis approach developed by Fischer and Haedrich (2000) and represent the ranges of those two environmental parameters within which the wolffish is most likely to be encountered, i.e. 100 to 400 m for depth and greater than or equal to -0.5°C for temperature.
The scientific survey data from Newfoundland indicate a sharp decline in the Atlantic wolffish population. In 1978, the number/tow averaged 10.5 individuals caught in each tow. The following year it dropped more than 25% to 7.1 individuals per tow. It
Figure 4. Top - Bars: CPUE (Σno./Σstn in each year) for Atlantic wolffish, Anarhichas lupus, caught in tows within appropriate depth and temperature ranges, 1978-93; Line, X’s - same, all stations (data provided by DFO in September 2000). Bottom - mean CPUE ± std error for all positive tows. Eastern Newfoundland, 1978-1993, DFO Fall Survey Data.
then continued to decline steadily, and by 1993 had fallen to just 0.96 individuals per tow. Over the full period from 1978 through 1993, 16 years or about two wolffish generations, the number/tow declined by 91% (Figure 4). Kulka and DeBlois (1997) note the same general decline for all Newfoundland statistical areas in the period from 1987 to 1995. Since 1995, scientific survey data that include information on Atlantic wolffish have been to be gathered by DFO. For the Scotian Shelf and Gulf of St. Lawrence, recent catch rates are comparable to ones observed earlier (Figure 5) and are low. In the northern Gulf, Atlantic wolffish are also not very abundant (DFO 2000, in Litt.); the average catch rate for 1990-99 is 0.42 fish per tow. As the graphs clearly show, the abundance of Atlantic wolffish in these more western regions has been low throughout the whole of the past 20 years, and has not shown the decline apparent off Newfoundland where the wolffish was once much more abundant than it is today (Figure 5) and where, as the ECNASAP map (Figure 2) shows and Mahon et al. (1998) indicate, this species assumes its greatest importance in the Northwest Atlantic.
Figure 5. Number of Atlantic wolffish, Anarhichas lupus, caught per tow for all tows in the Gulf of Maine (USA) Scotian Shelf and southern Gulf of St. Lawrence, 1970-1999. Circles: ECNASAP data; X’s, heavy line: data provided by DFO in July 2000. Note that Y-axis scales differ, and are much smaller than comparable survey catch rates off eastern Newfoundland(Figure 2).
DFO’s main population assessment tool is the STRAP computer program. This analysis takes catches from at least two trawls within defined strata, scales them according to the total area of the stratum (within which the species is assumed to be uniformly abundant), and calculates an estimated number of fish presumed present. To get a total, those numbers are summed across all strata where the fish was encountered. The size of a single stratum can range from 30 to 2,817 square nautical miles (average = 697) and, because a single survey trawl sample covers about 0.009 square nautical miles and there are on average 2 trawls per stratum, the scale-up is prodigious (Schneider et al., 1999).
The summed estimates from STRAP analysis results for Atlantic wolffish in all Canadian waters across the period 1978 to 1999 are shown in Figure 6 (see also Table 1). Because the sampling protocol changed in 1995, values from Newfoundland areas after 1994 are divided by a correction factor for comparability. For demersal species like wolffish, that factor ranges in general from 3.1 for adults to 10.7 for juveniles (Bundy et al., 2000). Calculated for the Atlantic wolffish alone using Bundy et al.’s (2000) formula, the factor is 4.85. The STRAP results also indicate a large decline in the wolffish population, down 87% over the 16 years from 1978 to 1994 (about two wolffish generations, and with no change in sampling proocol). From 1978 to 1999 (a little less than 3 wolffish generations, 1995-99 Newfoundland catches adjusted by 4.85), the decline indicated is 83%. Despite the debatable value of STRAP’s absolute numbers, the annual estimates turn out to be well-correlated (r = 0.93) with the simple metric we prefer, the number/tow.
Figure 6. Twenty years of STRAP analyses for Atlantic wolffish, Canadian waters. Solid line, filled circles: Engels trawl estimates. From 1978 to 1994 (16 years, i.e. about two wolffish generations) there was no change in sampling method, and the decline is 87%. Dotted line, small dots: corrected estimates from Campelen trawls using the factor 4.85 (see p. 5), 1995-1999. Based on information provided by DFO in September 2000. Straight line shows the COSEWIC endangered criterion.
From 1978 to 1993, the relative frequency of high survey catch rates off Newfoundland declined, and conversely low catch rates were on the rise (Figure 7). In 1978, catches of five or less wolffish in a tow constituted less than 40% of all catches; by 1984 the frequency of low catches (£5) had risen to more than 70%, and in 1993 it was almost 90%. The increased frequency of low catch rates is another indication of declining population density.
The trend to a declining mean size of fish, considered together with the declining numbers, is yet another indication that wolffish populations are in trouble. These trends vary according to area, but all are down (Figure 8). The average size of Atlantic wolffish from the Scotian Shelf and Southern Gulf of St. Lawrence, while showing inter-annual variation, declined overall by 50% or more from the mid-1980s to the present; the trend lines are essentially the same. Off Newfoundland, mean wolffish size declined from around a kilo in 1978 to near 700 gm in 1993 (Figure 8). Fish of that size are probably not yet mature. In the Northern Gulf, the average size in the period 1990-99 is less than 500 gm.
Figure 7. Percent frequency of large and small catches of Atlantic wolffish, Anarhichas lupus, by catch rate classes of 5, off eastern Newfoundland, 1978-1993. DFO Nfld Fall Survey Data.
Figure 8. Mean body size ( Σwt/Σno. in each year) of Atlantic wolffish, Anarhichas lupus, from Canadian waters (Circles: Newfoundland; X’s: Scotian Shelf; Squares: Southern Gulf of St. Lawrence), 1978-2000. Lines show the trends in each dataset; the Scotian Shelf and Southern Gulf lines are almost identical. Information provided by DFO in August and September 2000.
To the north in Greenland waters, scientific trawl surveys (Rätz, 1997; Möller and Rätz, 1999) have been done differently than in the US and Canada, so CPUEs are not directly comparable. Biomass and abundance of Atlantic wolffish off East Greenland increased somewhat from 1982 to 1998, but off West Greenland (the region nearest to Canada), biomass has declined by about an order of magnitude and numbers are down by about half. Mean age in the whole Greenland population in 1998 (= 4 yr) is half of what it was in 1982 (Möller and Rätz, 1999), so the picture appears to be quite similar to that seen off Newfoundland. To the south, Atlantic wolffish have never been very abundant in US waters. There, wolffish did experience a significant decline in the mid-70s (Figure 5), several years before one of much greater magnitude in Atlantic Canada.
- Date Modified: