One of the most robust metrics for assessing the effectiveness of protected areas is the temporal trend in the abundance of the species they are designed to protect. We surveyed coral-reef fish and living hard coral in and adjacent to a sanctuary zone (SZ: where all forms of fishing are prohibited) in the World Heritage-listed Ningaloo Marine Park during a 10-year period. There were generally more individuals and greater biomass of many fish taxa (especially emperors and parrotfish) in the SZ than the adjacent recreation zone (RZ: where recreational fishing is allowed) - so log response ratios of abundance were usually positive in each year. However, despite this, there was an overall decrease in both SZ and RZ in absolute abundance of some taxa by up to 22% per year, including taxa that are explicitly targeted (emperors) by fishers and taxa that are neither targeted nor frequently captured (most wrasses and butterflyfish). A concomitant decline in the abundance (measured as percentage cover) of living hard coral of 1-7% per year is a plausible explanation for the declining abundance of butterflyfish, but declines in emperors might be more plausibly due to fishing. Our study highlights that information on temporal trends in absolute abundance is needed to assess whether the goals of protected areas are being met: in our study, patterns in absolute abundance across ten years of surveys revealed trends that simple ratios of abundance did not.