Challenges for Diadromous Fishes in a Dynamic Global Environment

Climate Change May Have Affected Growth and Life History in Atlantic Salmon Juveniles over the Past 30 Years

Etienne Rivot, Charle Perrier, Céline Jouanin, Jean-Marc Roussel, Pascal Riera, Jean-Pierre Porcher, and Jean-Luc Baglinière


Atlantic salmon stocks are declining in European and North American rivers (ICES 2006). French rivers are located at the southern edge of the species distribution in Europe, providing a good opportunity to study life history variants in a changing environment (local and global) where unfavorable conditions may occur rapidly. In these rivers, almost all juveniles smoltify at age 1 (Smolt1) or age 2 (Smolt2), and adults return to rivers to spawn after 1 or 2 sea winters (1SW or 2SW). In the present work, we focus on the age at smoltification as an integrated parameter of changes occurring during the juvenile period in rivers. More precisely, variations in the mean age at smoltification (MAS) were studied in 23 rivers of the Armoricain Massif (western France) belonging to three geographical regions: northern Brittany, southern Brittany, and lower Normandy.

Archived scales of 24,500 adult salmon caught by anglers from 1972 to 2005 were read to estimate smolt age and sea age of each individual (Baglinière 1985). Logistic regressions indicated that MAS decreased significantly over the study period whatever the geographical region and sea age. Moreover, the analysis of autocorrelation in the time series of residuals highlighted the existence of medium-term fluctuations that were synchronous across the rivers, suggesting a response to a global environmental forcing. The proportion of Smolt1 and average monthly air temperature showed significant increasing time trends over the period (Figure 1). Linear correlation between series of residuals (proportion of Smolt1 and air temperature after removing the linear trend) was positive and highly significant (R2 = 0.55, P = 0.006) despite reduction of the degrees of freedom due to high autocorrelation in the series.