Rainbow Smelt Population Monitoring and Restoration on the Gulf of Maine Coast of Massachusetts
Bradford C. Chase, Matthew H. Ayer, and Scott P. Elzey
Rainbow smelt Osmerus mordax resources in Massachusetts have received little attention since the 1970s when sport fisheries attracted many anglers, resulting in large catches in numerous embayments. A recent survey of smelt spawning habitat found evidence that many smelt runs had declined since the 1980s (Chase 2006). Growing concern over the status of New England smelt resources prompted interest in assessing smelt populations and in developing restoration strategies. A fyke net survey of anadromous smelt on the Gulf of Maine coast of Massachusetts was initiated in 2004 to monitor annual spawning run demographics and to assist restoration experiments.
Fyke nets were set in six known smelt runs during the spring spawning seasons of 2005–2007. Four rivers were selected as annual population monitoring stations based on a range of expected run size and watershed characteristics. The Crane and North rivers, where recent restoration efforts improved habitat and documented a reestablished smelt run (Chase 2006), were selected to evaluate the methods and benefits of stocking smelt larvae. Nets were set in the intertidal zone below the downstream limit of smelt egg deposition. Three sets were made per week from March 7 to May 18 in each season. Captured fish were counted, measured, and released. Weekly samples of smelt were collected for aging and gametes in two rivers, Fore and Saugus, with the larger smelt runs. The collected smelt eggs were fertilized, incubated, marked with oxytetracycline (OTC), and stocked as yolk sac larvae. All smelt caught at the restoration sites will be inspected for marked otoliths. Analyses were conducted on size and age composition, and age comparisons were made to studies from the 1970s in the study area (Murawski and Cole 1978; Lawton et al. 1990).
The spawning run peak occurred between the last week of March and the first week of May, although smelt were captured during each week of sampling. The run onset was earlier and the duration longer than that in smelt runs at the northern end of their range (McKenzie 1964; Pettigrew et al. 2009). Each river displayed distinct characteristics of run peak and duration. The Fore River had a 10-week spawning season each year with a consistent peak during the second week of April. Contrary to expectations, the peak run in the Parker River (the most northern) occurred 3–4 weeks earlier than the peak in the Jones River (the most southern).
Fyke net catch rates varied widely among rivers, with the highest in the Fore River and the lowest in the two restoration rivers. Analyses of peak season catches (weeks 4–9) were made to assess the potential of using catch-per-unit-effort (CPUE) data as relative indices of abundance. The nominal CPUE data had high coefficients of variance, and transformed CPUE data also displayed high variance that could present difficulties for discerning annual trends in some rivers. Large year-to-year changes in CPUE were found with little evidence of annual synchrony among rivers. Stronger than average cohorts were evident in the Fore River for 2003 and in the Saugus River for 2005.