Warning This Web page has been archived on the Web.

Archived Content

Information identified as archived on the Web is for reference, research or recordkeeping purposes. It has not been altered or updated after the date of archiving. Web pages that are archived on the Web are not subject to the Government of Canada Web Standards, as per the Policy on Communications and Federal Identity.

Skip booklet index and go to page content

COSEWIC assessment and status report on the American Eel in Canada

Contribution of the St. Lawrence Eel Component – Landings Method

The second method is based on commercial landings.  The chain of assumptions is as follows.  Fisheries exploitation rates are assumed to be similar among regions, so landings are a linear function of standing stock biomass.  Production of silver eels, by weight, is assumed to be a linear function of standing stock biomass.  Since the relation between fecundity and body weight is roughly linear, egg production is taken to be a linear function of weight of silver eels produced.  Male silver eels are generally less numerous than females (Table 3), and they are much smaller (e.g. mean weight of silver males on Prince Edward Island is 11% of the mean weight of silver females; D.K. Cairns, DFO, unpubl.).  Hence male eels are generally a small fraction of total eel biomass.  For this reason total weight of silver eels produced is assumed to be the same as weight of female silver eels produced.  Since each parameter in the chain is assumed to be linearly related to the next, landings are assumed to be linearly related to egg production.

The period from 1970 to 1989, prior to the collapse in the Lake Ontario component, is used to calculate landings (Table 4).  The landings method indicates that FEA1 produces 26.4% of spawn output (Table 4). Based on relative size of landings, 5.9% of spawn output is from Ontario and 20.6% is from Quebec.

Table 4.  Relative contribution of American eel rearing regions to total egg production, based on the assumption that production of eggs is proportional to mean landings for 1970-1989.
RegionMean annual landings (t),
1970 - 1989
Percent of total
egg production
Portion ontarienne de l’AEED 11235,9 %
Portion québécoise de l’AEED 143120,6 %
Total de l’AEED 155426,4 %
AEED 2, AEED 3, AEED 4, AEED 546622,2 %
Canada1 01948,7 %
Littoral est des États-Unis1 07651,8 %
Total2 096100,0 %

Factors which reduce confidence in this analysis are listed below.

It is unlikely that exploitation rate by commercial eel fisheries was uniform across regions during the period covered (1970-1989).  Eel exploitation in North America is geographically heterogeneous.  FEA2 and FEA5, which have no or virtually no exploitation, are grouped with FEA3 and FEA4, where exploitation varies geographically.  The landings method assumes that overall exploiation rate of these areas combined is similar to local area exploitation rate.  No exploitation rates, even for local areas, are available for 1970-1989.  Exploitation rates have recently been estimated for Prince Edward Island, the Maryland portion of Chesapeake Bay, and the St. Lawrence estuary (ICES 2001, Caron et al. 2003), but data are insufficient to estimate exploitation rates over broad geographic areas.  It is plausible that overall exploitation rate in FEA1 is higher than the rate elsewhere, because all eels produced in FEA1 are subject to commercial fisheries during their exit to the spawning grounds, whereas other regions contain areas where no exploitation occurs.  If exploitation rate in the St. Lawrence is higher than elsewhere, the landings method will overestimate standing stock biomass relative to landings.  This could lead to an overestimate of the relative importance of St. Lawrence eels to the species' spawn output.

The landings method estimates that egg production from eels reared in the Mississippi is zero, because there are no landings there.  If the Mississippi produces eels which spawn in the Sargasso Sea, the landings method will overestimate percent contribution to total egg output from other areas, including FEA1.

The landings method treats all exploitation as equal.  Eels grow and are subject to natural mortality before becoming silver eels.  Growth and mortality schedules, which vary among regions, may impact the way exploitation affects silver eel production, leading to error in estimates of the relative importance of egg production.

The landings method treats Ontario and Quebec as discrete units.  However, a substantial portion of eels landed in Quebec completed their yellow (growth) phase in Ontario waters, and all or nearly all eels landed in Ontario spent their early years in Quebec waters.

In sum, both the discharge and the landings methods suggest that the St. Lawrence River basin has contributed a substantial portion of total spawn output by the American eel.  However, both methods are subject to large uncertainties.  The discharge method's assumption of a linear discharge-recruitment relation is particularly problematic.  Some sources of uncertainty suggest that it is more probable that the methods overestimate, rather than underestimate, the contribution of the St. Lawrence River basin to total spawn output.


Changes in Eel Abundance

COSEWIC uses change of abundance over three generations as a criterion for species classification. Mean generation time of female American eels reared in Canadian freshwaters is approximately 22 years. Mean generation time for female American eels reared in salt water is poorly known, but is probably roughly nine years, based on an estimate for Prince Edward Island marine habitats.  Using these mean generation times, and considering 2006 as "present," three generations ago encompasses the period 1940-1979. 

Table 5 compares means of American eel data series for the period of three generations ago (prior to 1980) to recent means.  Series include scientific indices and landings data.  Nine Canadian and one US data series are available for comparisons over three generations.  Three of the Canadian series are landings, three are from research electrofishing, and three are from non-electrofishing research surveys.  The single US series is landings.

Table 5.  Mean values of American eel data series.

Table 5.  Mean values of American eel data series.

Percent change between the period before 1980 and the 2000s ranged from -99.5 to +74.8 (Table 5). All of the four landings series showed negative change.  Five of the six survey indices were negative.  Comparisons from the 1980s to the 2000s, and from the 1990s to the 2000s, show mixed results.  Nine of 12 series showed negative change from the 1980s to the 2000s, and nine of 16 series showed negative change from the 1990s to the 2000s. 

Several factors limit the reliability of these series to indicate changes in American eel populations.  Fisheries landings are an indicator of minimum biomass, because landings can never exceed biomass.  Beyond that, landings are generally poor indicators of abundance, because they are subject to changes in fishing methods, regulations, and markets.  North American eel fisheries have been affected since 1970 by market factors, including competition in the European market from the advent of large-scale aquaculture production, and by tightened regulations.  Nevertheless, the value of eels in both Canada and the U.S. has increased 6-to-10-fold in inflation-corrected dollars (J.M. Casselman, Queen's University, unpubl.).  In the Atlantic coast of the Maritime Provinces, some eel fishing effort has been channelled into the elver fishery, which produces very small volumes of catch.  The only long-term series from the US is landings.  It is unclear what portion of the large decline (67.5%) in reported US landings between 1970-1979 and 2000-2003 stems from changes in fishing practices and intensity and what portion stems from changes in abundance.

There appears to be a relation between eel abundance and the North Atlantic Oscillation (Figure 9). Much of the data in the early period is from the 1970s. It is possible that abundance during that time was bolstered by favourable NAOI conditions.  If this is the case, then at least part of the decline between the early and the recent periods may be part of a natural cycle.

The most reliable long-term series are the fisheries-independent surveys in the St. Lawrence River system and the southern Gulf of St. Lawrence (Table 5).  Changes between early and recent periods for these series vary widely, from very steep declines (>90%) in Ontario, to smaller declines (-15.1%) in the St. Lawrence estuary, to high inter-river variability in the southern Gulf (from -87.9% to +74.8%).