By Fiona Gell and Callum Roberts

Some scientists point out, rightly, that most studies of marine reserves employ designs which cannot unequivocally deliver a verdict on whether they work. Many studies compare a single reserve with one or more control sites. Since in some cases (but certainly not all), reserves were chosen because they have good quality habitats, this leaves open the possibility that differences detected are habitat rather than protection effects. Similarly, changes over time in measures of reserve performance may be due to habitat or background environmental changes.

The strongest study design for reserves research is considered to be before-after-control-impact-pairs analysis (BACIP). Here several reserves (three or more) are paired with several control locations, and data collected at intervals before (ideally three or more times) and after protection. In this way, the effects of protection can be separated from those of habitat. Sites adjacent to reserves may receive spillover and will not be adequate controls. So to settle questions of spillover we need several sets of reserve-adjacent area-control site triplets. Our difficulties do not end there. Reserves can potentially export larvae tens of kilometers away, so sites within that supply envelope may also be affected by the reserve and will not represent true controls. Conditions and habitats in control and reserve sites must be matched closely, but as distance between them increases, conditions may diverge. Good controls are very difficult to find.

There are also human problems. Few funding organizations will support collection of several years of pre-protection data. Scientists also find it hard to maintain control over the design of reserve experiments. Management plans are often modified, reserve boundaries changed, and protection poorly implemented. It is hardly surprising then that almost no studies have achieved this level of design sophistication. Furthermore, almost none collect data on fishing effort, which is essential to interpreting findings. Without such data it is impossible to know whether absence of an effect is because reserves don’t work or is just due to lack of protection.

Some people suggest that fishers’ resistance to reserves will diminish or disappear when scientists produce better quality evidence, but we doubt this. Fishers are most often convinced of the usefulness of reserves through the experience of other fishers. This makes an all-round picture of how reserves have affected fishing, the wider community and the ecosystems, of more relevance than statistical tests. However, skeptical fishery managers and decision makers may be won over by stronger science. That said, we find it paradoxical that many managers place more faith in management tools whose performance has not been subject to the level of critical scrutiny they demand of reserves. This is not to say that seeking such a high standard of proof is not necessary for reserves. The next generation of studies must strive for it. But we should also demand the same evidence of efficacy for other fishery management tools. The poor state of the world’s fisheries suggests these tools are not performing as intended.

(This piece was excerpted, with permission from Elsevier, from: Gell F.R. and Roberts C.M. 2003. Benefits beyond boundaries: the fishery effects of marine reserves. Trends in Ecology and Evolution 18:448-455. Copyright 2003 Elsevier)

For more information:

Fiona Gell, Port Erin Marine Lab, University of Liverpool, Port Erin, Isle of Man, UK. Email:

Callum Roberts, Environment Department, University of York, Heslington, York, YO10 5DD, United Kingdom. Tel: +44 1904 434066; E-mail: