Frugal conservation: What does it take to detect changes in fish populations?

Many conservation projects have to develop practical and feasible ways to detect changes in populations. We compare the ecological information obtained using intensive monitoring of coral-reef fish populations with that which would be procured from less exhaustive sampling. At each of four sites in the Philippines, we surveyed all non-cryptic fish species in eight transects every month for up to 3 years. We first tested for changes across years and seasons in fish density, size and species richness. We then compared these results obtained by using all data to those obtained in simulated protocols that restricted effort in ways that were relevant in practical marine resource management. We demonstrate the potential for reducing time in the water (number of surveys or number of transects) or in the number of species surveyed (proportions of those that were easily identified or locally fished) with little loss of information. Bimonthly surveys detected most of the overall but few of the seasonal trends that were identified with monthly sampling; far fewer trends were detected with less frequent surveying. Similarly, most trends were detected when at least four transects were carried out monthly. Most trends were also detected when all possible easily-identified species or at least 75% of fished species were considered. Indeed, ability to detect overall trends remained high even when only fished species were considered in four transects monthly. We conclude that some selective sampling protocols can be sufficiently sensitive to detect important trends, and that the choice of protocol will depend on the objective of the research or management.