Proposals for optimizing sea lamprey passage through a vertical‐slot fishway

Abstract

Movement of sea lamprey through a vertical-slot fishway (built at the Coimbra dam, in central Portugal) was continuously monitored between 2013 and 2016 by a video recording system. Visual count data were used to quantify the overall successful movements and identify seasonal and circadian patterns of fishway use. Explicative models (Boosted Regression Trees) were used to study the relationship between frequency of successful movements during the species migration peak and environmental variables. The aim of the study was to identify predictors that may be related with successful sea lamprey upstream migration through the fishway. Collected information was used to support further management recommendations for optimizing the fishway performance. During the 4-year study, near 50,000 lampreys successfully negotiated the Coimbra dam fishway reaching the upper stretch of river Mondego.Migratory peak occurred between April and May with an increase in passages during the night period (between 11 p.m. and 6 a.m.). It was observed an increase in use frequency that was related with the combined effect of flow and temperature. Higher passages were achieved when Coimbra dam discharge flows are lower than 100 m 3 /s and temperature between 15 and 19°C. Flow discharges higher than 150 m 3 /s seemed to inhibit successful sea lamprey passage. In dry years, when flows are almost constantly lower than 50 m 3 /s, temperature was the most important factor influencing fishway use.

Fish pass use by shads (Alosa alosa L. and Alosa fallax [Lacépède, 1803]): Implications for monitoring and management

“…The monitoring approach employed in this study revealed crucial information regarding shad upstream migration, and this was accomplished without manipulating individuals and with no associated mortality or impact in behaviour, proving itself adequate to fulfil the goals established. It also confirmed that relatively large numbers of shad use the vertical slot fish pass at Coimbra dam, proving this structure built in 2011 can also be considered a successful habitat rehabilitation measure for shads as well as for sea lamprey (Pereira et al, 2017, 2019). During the study period (2013–2017), over 26,000 Alosa sp.…”

“…(e.g., Acolás et al, 2004, 2006) and ecological modelling of spatial distribution (e.g., Nachón et al, 2016; Taverny et al, 2012; Trancart et al, 2014). BRTs have proven to be useful and to produce better results than other modelling techniques in the understanding and prediction of spatial distribution of fish (e.g., Bangley et al, 2020; Ferreira et al, 2013; Segurado et al, 2015); however, their application in the study of temporal distribution is less frequent (Escalle et al, 2016; França & Cabral, 2015; Pereira et al, 2019; Pittman & Brown, 2011). Furthermore, to our knowledge, the use of BRTs to model fish pass use and the environmental factors influencing its successful negotiation is a recent approach to these problems and may bring new insights on the matter (Pereira et al, 2019), especially due to its capacity to deal with missing data, outliers, different types of variables and non‐linear relationship between response and explanatory variables, which render the use of other statistical tools inadequate (Elith et al, 2008; Escalle et al, 2016).…”

“…These characteristics were particularly designed to ensure high passage effectiveness for Alosa sp. and sea lamprey (Pereira et al, 2019).…”

“…The respective peak of migration was identified, and of the total number of migrating fish observed in the window, the first and last 10% of all migrating fish were discarded to increase the resolution power during the peak of the spawning migration (i.e., 80% of total Alosa sp. movements), hence focusing the analysis on the variables influencing fish pass use for the majority of shad (Pereira et al, 2019).…”

“…BRTs combine the regression trees' capacity to accommodate missing data, outliers and different types of predictor variables, with the improved model accuracy obtained by using a boosting algorithm (Elith et al, 2008). This flexible tool chooses relevant variables, produces accurate fit functions and identifies and models interactions, making them highly useful to analyse environmental data (Elith et al, 2008; Pereira et al, 2019).…”