Direct estimation of natural mortality rates for littoral marine fishes using populational data from a marine reserve

 Visual censuses conducted in a marine reserve (Medas Islands) were used to estimate the natural mortality rates (M) for five common fish species in the NW Mediterranean Sea (Coris julis, Diplodus annularis, D. sargus, Serranus cabrilla and Symphodus roissali). Visual censuses of these same five species were also performed at three sites in unprotected areas of the coast where both commercial and sport fishing activity was normal. Censuses were conducted over a 3 year period. Estimates of M in the 3 years displayed scant seasonal or interannual variation, which may mean that the populations were in equilibrium during that period. The results of this study showed that the relationships between M and the growth parameters and maximum life span were unclear, and considerable caution is therefore recommended when using indirect methods of estimating M based on those parameters. For certain species the values of M were equal to or greater than the estimated total mortality in the exploited areas. The virtual absence of piscivorous predators in the unprotected area as a consequence of the high level of fishing in that area contrasts with the high abundance of such predators in the marine reserve. Since predation is the main contributor to M, estimated mortality in the unprotected areas is attributable nearly entirely to fishing. It is suggested that M may vary according to alterations taking place in conditions in the ecosystem inhabited by a species and thus that use of a value of M for a pristine population cannot be extrapolated to exploited areas. Received: 18 January 2000 / Accepted: 14 July 2000