Multispecies survey design for assessing reef-fish stocks, spatially explicit management performance, and ecosystem condition

Principles of statistical sampling design were used to guide refinement of a 30-year multispecies fishery-independent diver visual survey of population abundance and size structure of more than 250 exploited and non-target fishes in the Florida coral reef ecosystem. Reef habitat features and no-take marine reserves (NTMRs) were used to partition the 885 km2 sampling domain into sub-areas (or strata) to control the variation of fish density. For the period 1999–2008, survey precision of population density and abundance (CV, coefficient of variation, ratio of standard error to mean) ranged from 7% to 20% for the majority of 13 primary exploited species in the Florida Keys and Dry Tortugas regions. Population sustainability metrics like species average length in the exploited life stage were comparable between our fishery-independent survey and fishery-dependent catch-sampling. The survey design also performed well for non-target fishes, yielding CVs between 6% and 15% for population density for the majority of 36 species. Sampling efficiency was improved over time via an iterative learning process by which past survey data was used to refine the stratification and allocation schemes of future surveys. We show how survey data are used to support multispecies stock assessments, evaluate the effectiveness of NTMRs, and assess ecosystem condition for the reef fish community.