Science Advice for the Development of a Precautionary Framework for American Eel in Canadian Waters

Abstract

American Eel (Anguilla rostrata) forms a single panmictic population with a single spawning location over its continental range, extending from Greenland to northern South America. Within Canada, the American Eel is widely distributed over the six eastern provinces, and management is geographically fragmented. It has been assessed as Threatened by COSEWIC. To inform development of a precautionary approach framework for the management of anthropogenic mortality within localized regions, advice was requested on mortality reference points based on differences in life history traits within Canada, and a Canada-wide trend analysis. The modelling approach was developed and used to estimate a Canada-wide trend in American Eel abundance. Twelve standardized indices of abundance were combined into a single model using multiple transformation procedures and timeframes in the analysis. Twelve American Eel freshwater time series that have been previously identified as useful from four ecozones, defined based on differing life history characteristics, were included in the trend analysis up to 2018. Five time series representing yellow eel, from multiple geographic zones, had significant negative trends in catch through time. Each of these time series included data from before 1990. The remaining yellow eel time series, which all started after 1990, did not have significant trends in catch through time. One data set representing the elver stage had a positive but not significant trend and one data set representing silver eel had a negative but not significant trend. The Canada-wide trend produced negative trends from fits to data from the whole time series (1956-2018) and from 1980 to 2018. There was 100% likelihood of decline and 69.2 to 99.6% likelihood of a decline of at least 50% since 1980. When data were limited to 2000 to 2018, trends were less negative and not different from 0. In Canada, yellow eel abundances have been relatively stable over the last two decades, but statistically significant declines likely occurred previously in zones where data prior to 1990 are limited. Density-dependent matrix population models were used to identify mortality reference points (F) for American Eel sub-populations assuming that the sub-populations were stable and independent of each other. Prospective reference points were identified based on silver eel escapement (ESC) with the limit reference point and upper stock reference points set at 30% and 50% ESC, respectively. Mortality reference points (F30 and F50) were identified for elver and eel (yellow and silver eel) fisheries independently. Mortality reference point estimates were influenced by the density-dependence mechanisms included in the model. The simplest assumption, where all density-dependence acts in early life prior to the fishery activity, resulted in the most conservative reference points. Reference points for elver fisheries were consistent across zones with a fishing mortality of 1.2 and 0.67 representing the F30 and F50, respectively. Reference points for eel fisheries varied across zones: with a 350 mm minimum fishing size, F30 mortality ranged from 0.13-0.36 and F50 mortality ranged from 0.073-0.21. If density-dependence acts after the elver fishery activity, then the population may be more resilient to elver fishing mortality. However, mortality reference points (based on 30% and 50% ESC) will be close to excess mortality that would drive extinction, and the sub-population monitoring of elver abundance may not be able to detect the effects of elver fishing mortality until it is excessive. Population-specific mortality reference point estimates could be predicted based on silver eel escapement objectives from local estimates of turbine mortality, silver length, and minimum fishing size (if applicable). The presence of turbine mortality required fishery mortality reference points to be reduced to meet management objectives. When turbine mortality of silver eel was included in the models, under certain model assumptions, high cumulative turbine mortalities meant there was no scope for fishing mortality that would allow the target stock size to be achieved. Cumulative turbine mortalities represent the average effect of turbines across eel populations within a zone, including those that are unaffected by turbines and those that pass through multiple turbines. A meta-population model was used to investigate the effectiveness of managing local sub-populations in the context of the broader panmictic breeding population. The model explored alternative assumptions about leptocephali survival and dispersal, which had implications on predictions of how the population was structured among zones. Results indicated that setting fishery mortality reference points based on local sub-populations would be mostly consistent with achieving similar panmictic population objectives. However, uncertainty about leptocephali dispersal mechanisms suggests some sub-populations could be less likely to achieve population objectives. The sub-populations in zones at greater distances from the spawning grounds, were more likely to be impacted by excess mortality in other zones under some assumptions. This can also create conditions where the recovery of these zones is dependent on range-wide protections and recovery. This advice applies modelling approaches to develop reference points and a trend analysis for Canada. To create these models several assumptions were necessary due to unknowns in American Eel biology as well as data and environmental uncertainties. Efforts were made to scenario test many of the assumptions, but uncertainty remains. The models produced used existing data from literature sources and could not test all scenarios for the future (e.g., climate change), or all potential sources or consequences of mortality. Existing data sets for American Eel in Canada are unevenly dispersed both with respect to life cycle processes and geographic distribution. To better evaluate any future management practices eel monitoring programs targeting the main life cycle stages (i.e., elver, yellow, silver) throughout the species range will be required.