Modelling past migrations to determine efficient management rules favouring silver eel escapement from a large regulated Floodplain Lake

As human activities caused a dramatic decline of European eel population since the 1970 s, the European Union has set targets to ensure a 40% escapement to the sea of the silver eel biomass by considerably reducing anthropogenic impact. Thus, human obstacles to fish migration like dams and hydropower plants should enable efficient management measures ensuring safe passage for eels during the migration. In order to provide a quick and efficient assessment of eel conservation measure applied to the sluice gates of a large floodplain lake, we implemented a novel evaluation method using predictions on past migration seasons when no management measure were applied. For this purpose, we collected acoustic telemetry monitoring data over three migration seasons and fitted a predictive model based on Boosted Regression Trees (BRTs) to describe the influence of environmental parameters on migration. The water level difference over two days proved to be decisive, along with early migration occasions in the season, as an increase of at least 10 cm water level was associated with a sharp increase of migration probability. We then used the BRT model to predict migration occasions at the dam over 8 past seasons and forecast the impact on escapement of management measures if they had been applied. Thanks to this original prospective assessment of dam management measures, we identified an enhanced opening decision rule, capable of increasing the eel escapement for each year of the study. The management measure is particularly efficient during the years with poor hydrological conditions (i.e. droughts), that is the periods with the lowest initial escapement rates. Finally, efficient management measures to increase silver eel escapement were based on increasing the number of gate opening days by only a few days per year (c.a. 15%). However, the management measure only focused on silver eel migration, so that further investigations should consider the impacts of other threats occurring during the life cycle. This study also provides a highly operational approach for fast evaluation of conservation measures, avoiding lengthy and expensive monitoring campaigns of classical ex-post assessments.