4. Another approach would be to look at the possibility of rotating harvest regimes such as have been described for territorial or sedentary resources (see Caddy and Seijo, 1998). Here, a constant proportion of all productive unit areas are closed in a given year. Simply for illustration, Table 6 shows the results of several simulations where spatially 'patchy' non-migratory resources are subject to rotations of different duration. The optimum rotation cycle (years) is based on a number of factors, notably the maximum yield, for each of six combinations of the von Bertalanffy K and natural mortality rate M that might apply to a population, and the conclusion reached is that rotation periods longer than six years may be suboptimal in situations where recruit distribution is very patchy:
Table 6. Some close to optimal (integer year) rotation periods (years) for non-migratory resources (from Caddy and Seijo, 1998)
|
M |
K |
||
|
0.1 |
0.3 |
0.6 |
|
|
0.1 |
6 |
6 |
6 |
|
0.3 |
6 |
6 |
6 |
|
0.6 |
4 |
2 |
2 |
Naturally, dispersion will reduce the effectiveness of protection offered by rotating closures, and the simulations done here would have to be repeated with feasible levels of interchange of individuals.
Although apparently less 'water tight' as a mechanism than for non-migratory species, some areas of seasonally high aggregation may be sufficiently protected by a rotating/closure mechanism to ensure spawning occurs undisturbed. Evidently the importance of dispersal between unit areas will depend on the size of these areas in relation to the mean annual dispersal distance of fish, a factor that could be analysed by simulation, in combination with tagging experiments to test various migratory hypotheses.
More relevant perhaps for migratory species is the idea of using the 'gauntlet model' of sequential risk at spatially-differentiated fishing sites to determine acceptable levels of survival to maturity. An index such as that shown in Figure 13, modified from the SPARCLE model presented by Kleiber (1996), could provide some indication of how mortality is fluctuating through time, even if an accurate growth curve may not exist for the species.
Figure 13 modified from the above author suggests that, if a gauntlet model applies and has been confirmed by tagging, and there are fishing sites 1,2,3...n distributed along a known migration route, log ratios of catch rates for a cohort t could be calculated for each fishing season whose values might be adjusted by effort or catch control at each location so as to remain above a certain limit, such that:
[ln(CPUE)t+DT,n/ln(CPUE)t,1] > = Rcrit
A general comment applies to all of the above hypothetical management mechanisms based on spatial criteria: they may become more widely used as telemetry systems for locating vessels make oceanic MCS procedures more feasible.
