612 



Fishery Bulletin 97(3), 1999 



fishermen may feel greater equity with fewer restric- 

 tions on the number of participants or their catches. 

 Moreover, reserves have the potential to reduce vari- 

 ability in catches from year to year and to enhance 

 conservation of species and ecosystems. Fishing is 

 not the only threat to marine ecosystems, though, 

 and fisheries regulations are not sufficient to pro- 

 tect these systems (Allison et al., 1998). 



To our knowledge, no field study has yet examined 

 the effects of population growth potential or fishing 

 mortality on reserve benefits. In part, such studies 

 are made difficult by the uncontrolled nature in 

 which reserves are established. 



Relation of fishery benefits to reserve size 



Real-world fisheries span a range from lightly fished 

 to heavily overfished, and the optimal reserve size 

 will depend on the fishing mortality as well as the 

 population growth potential of the target species. 

 Because many fisheries involve multiple species with 

 widely divergent population growth potentials, choos- 

 ing a single best reserve size may be difficult. More- 

 over, key aspects of the life history of marine fish, 

 the larval phase in particular, remain a mystery. Be- 

 cause of these gaps in knowledge, it would be difficult 

 to make an accurate prediction of the optimal reserve 

 size even in a well-studied single-species fishery. 



Although our research sheds doubt on the use of a 

 universal reserve proportion, it does lend support for 

 the use of large reserve systems under certain cir- 

 cumstances. In the two real-world cases where the 

 necessary information existed, our models predicted 

 that reserves should encompass 75-809^ of the man- 

 agement area. These proportions are enormous and 

 may be unrealistic for several reasons. First, the 

 short-term economic losses from closing 80'7c of a 

 management area would be large, although our mod- 

 els predict that the recovery time for such heavily 

 overfished fisheries would be rapid (Sladek Nowlis 

 and Roberts, 1997). Second, the political challenges 

 of establishing such large reserves would be a formi- 

 dable barrier Finally, we do not stand firmly behind 

 these predictions because their accuracy is depen- 

 dent on parameter values that are poorly understood. 

 Nevertheless, consistent results across several spe- 

 cies suggest that reserves encompassing 40'7( or more 

 of a heavily fished management area could produce 

 substantial fisheries benefits. 



Though rare, at least one large reserve system does 

 exist. The Mombasa Marine National Park closed 

 over 60'7c of local fishing grounds (McClanahan and 

 Kaunda-Arara, 1996). This example fits nicely with 

 our model's assumptions because levels of fishing 

 effort remained similar in the fishing grounds be- 



fore and after the closure. After two years, total yields 

 had not surpassed those prior to reserve establish- 

 ment (McClanahan and Kaunda-Arara, 1996). How- 

 ever, catch per unit of effort had increased dramati- 

 cally and total yields showed potential for future in- 

 creases. In this case and others involving extensive 

 use of marine reserves, our research encourages an 

 adaptive approach that reflects the lack of knowl- 

 edge about fish life histories and the high degree of 

 uncertainty in these complex biological systems. 



Relation of reserve size to catch variability 



Our model supported Bohnsack's (1996) hypothesis 

 that catch variability will decrease with increasing 

 reserve size. Our models predicted decreases in catch 

 variability across a variety of levels of environmen- 

 tal variability and fishing mortalities. Our results 

 also complement other studies that showed that re- 

 serves could reduce catch variability,'^ decrease the 

 likelihood of bad years (Lauck et al., 1998), and in- 

 crease the persistence of fisheries vulnerable to over- 

 fishing (senior author's unpubl. data). 



To our knowledge, no field study has yet examined 

 the effects of reserves on catch variability. Although 

 they may be confounded by variability in fishing ef- 

 fort, the necessary data should be practical to collect 

 before and after reserve creation. 



Assumptions revisited 



As with all models, one must be careful in interpret- 

 ing the results of this one. It is based on parameter 

 values that in some incidences — larval survivorship 

 in particular — are poorly understood. However, the 

 model's predictions are qualitatively robust to param- 

 eter errors, meaning that its general predictions hold 

 true across a wide range of values and a wide vari- 

 ety of species. Our assumptions regarding the move- 

 ment of adults and larvae were far more critical in 

 influencing the conclusions we have drawn here. 



Our assumptions regarding adult movement have 

 wide applicability. Many fisheries target sessile or- 

 ganisms such as harvested kelp (Bustamente and 

 Castilla, 1990), slow-moving organisms including 

 many invertebrates (Davis and Dodrill, 1980; Davis 

 and Dodrill, 1989), and organisms with high site- 

 specificity such as many reef fish (Polunin and Rob- 

 erts, 1996). All of these systems are likely to approxi- 

 mate our assumptions of no adult movement. This 

 model is not universally applicable, as highly mobile 

 and migratory species, including many pelagic fisher- 



^ Mangel, M 1998. Environmental Studies Board, University 

 of California, Santa Cruz, CA 95064. Unpubl. data. 



