Mark-recapture models for determination of mortality, migration and growth in Pomatomus saltatrix (Teleostei)

Abstract

This study primarily attempts to develop models to estimate population dynamic parameters from mark-recapture data. Model implementation is illustrated using data collected from the South African Pomatomus saltatrix fishery . The models developed allow for the estimation of mortality, survival and migration rates in exploited fish stocks. A growth model is also developed which simultaneously estimates growth parameters as well as validates the hard structure banding using age-length and markrecapture data. There are number of advantages to these models . Given appropriate datasets the mark-recapture models developed in this study can be applied to others species of interest. The models can be modified easily e.g. the growth model can incorporate growth functions other than the von BertalanfIy model. The models can be programmed into a spreadsheet which facilitates the estimation of parameter variances using likelihood profile or bootstrapping methods and allows the testing of model assumptions based on simulations. A general mortality model is developed and is illustrated with mark-recapture data from the P. saltatrix fishery. The model provides an estimate of the average fishing mortality for the Cape and KwaZulu-Natal and is then extended to include movement between the Cape and KwaZulu-Natal. It utilises mark-recapture data from the Sedgwick's-ORI Tagging Programme as well as effort and catch data from the National Marine Linefish System (NMLS). Estimates of annual fishing mortality rates in KwaZulu-Natal are derived from the model which takes into account immigration of P. saltatrix into KwaZulu-Natal from the Cape as well as emigration from KwaZulu-Natal to other areas including the Cape. The average fishing mortality rate was estimated to be 0.27 year" between 1984 and 1993 in the Cape and KwaZulu-Natal combined. This is likely to be underestimated because of non-reporting of tags, shedding of tags and tag-induced mortality. The model is shown to be robust for estimating the average fishing mortality rate and exploitation rate only when annual variability in fishing mortality is small during the study period. The second model to quantify migration into and out of KwaZulu-Natal waters suggested that the whole adult Cape stock migrates into KwaZulu-Natal during winter. Further, this whole stock is available to fishing in KwaZulu-Natal although there is probably large exchange between inshore and offshore areas and, in the latter zone, P. saltatrix is inaccessible to shore-based fishing. Large fishing mortality rates for the years 1987 to 1993 were estimated in KwaZulu-Natal. These large fishing mortality rates may prevent the return migration of P. saltatrix to the Cape and the model predicts that possibly less than 4% actually return to the Cape. An age and growth study based on otolith readings was also undertaken. Validation of the growth banding as annual was confirmed by developing a model that estimated growth parameters using age-length data and simultaneously estimating times-at liberty of tagged individuals based on arbitrarily chosen band deposition periodicities. It is shown that the assumption of annual banding led to the best prediction of periods of liberty of tagged individuals with small coefficients of variations in the parameter estimates. However, since only a few tagged animals were used in the analysis more research is needed to verify the robustness of this technique for use on other fish. The growth of P. saltatrix in the present study was found to be faster than that of a previous study in South Africa. A modified delay-difference model was developed to estimate relative biomass and relative catch based on observed mean body weights and effort indices. For the period 1956 to 1972 the model predicts that there was a decline in P. saltatrix abundance with corresponding declines in mean weight of the catch. Although during this period there was a general decline in fishing mortality, the fishing mortality was sufficiently high for growth overfishing to occur. It was found that during the 17 year period there was a 44% reduction in biomass which is similar to an estimate in another study. Catch during the period was annually variable but generally declined with time especially in the later years. The decline in average weight harvested and the variable but lower catches during this period are consistent with observations by fishers . An evaluation of the present closed season for P. saltatrix in conserving egg production was performed. It showed that better conservation of egg production is possible by shifting the present closed season (September to November) to extend over the October to December season but this may adversely affect the tourism industry in KwaZulu-Natal. Shortening the present closed season by one month (September) does not affect egg production but increases present yield levels. This study suggests that the closed season may not be useful in terms of reducing the fishing mortality rate on P. saltatrix as fishers may be encouraged to fish harder in the open season to make up for the lost yields of the closed season. Moreover, lengthy closed seasons may also increase fishing mortality because fishers tend to fish harder in the months open to fishing. Assuming no large annual recruitment variations the P. saltatrix stock is presently optimally exploited as current fishing mortality rates are just below the MSY or optimum yield levels.