Standardising long-term eel (Anguilla anguilla) fyke net survey data reveals covariate effects and improves estimates of declining relative abundance

Eel recruitment and stock have been in decline since at least the mid 1980s. So far, much research has focussed on conservation objectives and estimating silver eel production in response to the EU Regulation (EC 1100/2007) and most surveys have been on waters subject to exploitation and/or stocking. The aim of this study is to examine a 35 -year yellow eel survey dataset to determine relative abundance trends while accounting for survey design, and important physical and operational covariates. Chains of ten, or sometimes five, fyke nets were fished at fixed sites in four lakes on a catchment in the west of Ireland not impacted on by exploitation or stocking. Captured eels (10,474) were counted for each trap, and eel weight was recorded for each chain of nets (5515 net nights fished). Data were analysed using a generalised additive model (GAM) that included smoothed trends over time and covariates day-of-year, depth, gradient, site and net chain. Yearly trends and between-chain variability were significant for all locations with trap depth and gradient being important for some lakes. Standardised trends by number and mass declined markedly in both tidal areas - 84.7% and - 89.5% by count and - 93.5% and - 89.5% by weight; significant declines (- 39.2% by count and - 54.1% by weight) were found in one of the freshwater lakes with the other, the one farthest upstream, having had changes over time but no significant difference between the start and end. This study provides a framework for analysing long -term fyke net catch data for eel, which may be useful for the international analysis of eel survey data.