Some observations on the impact of temperature and low flow on the onset of downstream movement of wild Atlantic salmon, Salmo salar L., smolts

A variety of proximate environmental triggers including water temperature and river flow have been correlated with the intensity and timing of salmon smolt migrations. In this study, the migratory timing and behaviour of wild salmon smolts leaving their natal stream were determined using a passive integrated transponder (PIT) antenna system at a study site on a tributary of the River Itchen, England. To examine the role of cumulative temperature, the onset of downstream movement of smolts was observed under natural control conditions (2000-2006). These control observations were then compared with a year when a managed severe low-flow event was imposed (2007). Under natural control conditions the onset of downstream movement was later when cumulative temperature was higher. The effect was stronger using the 30 days prior to 15 March (slope = 0.083, t(4) = 3.0, p = 0.04) than when using the cumulative temperature since 21 December (slope = 0.037, t(4) = 1.3, p = 0.26). Analysing the effects for each age group separately retained some evidence for a later migration for smolts aged 1+ (30 days prior to 15 March: slope = 0.069, t(4) = 1.5, p = 0.20; since 21 December: slope = 0.033, t(4) = 0.92, p = 0.41), while for 0+ smolts there was very little effect (30 days prior to 15 March: slope = 0.031, t(4) = 0.2, p = 0.88; since 21 December: slope = -0.010, t(4) = -0.5, p = 0.66). Taken as a whole, the results show no evidence of a negative relationship between cumulative temperature and the initiation of downstream movement of salmon smolts, and therefore appear contrary to previous studies. Migration of smolts under both control conditions and low flow was significantly correlated with sunset (p<0.01, r = 0.47, n = 170; p<0.01, r = 0.48, n = 36, respectively). The mean date of migration for 0+ smolts under low flow conditions was estimated to be 5.8 days earlier than in the control years (95% CI 0.8 to 10.9 days earlier). This difference was statistically significant (t(5) = -3.0, p = 0.031). By contrast, 1+ smolts did not migrate earlier under low flow (estimated 1.1 days earlier, t(5) = -0.27, p = 0.80, 95% CI 11.6 days earlier to 9.5 days later) and the migratory behaviour of both age groups appeared to be unaffected by the drought conditions. Crown Copyright (c) 2011 Published by Elsevier B.V. All rights reserved.