Mean fish differences fell at six Moorea sites between 2006–2007 and 2016–2018, but one shore omission reversed the change

We tested whether recorded fish communities became more alike at six fixed Moorea fore-reef sites. We used two fixed survey bands per site, the same divers, and open MCR LTER data. The primary comparison had 29,689 fish in 2006–2007 and 2016–2018. Mean Bray–Curtis distance between site count shares fell from 0.5331 to 0.4969, a change of −0.0363 or −6.8%. Eleven of 15 site pairs had a decrease. Each diver alone and equal-count samples kept the mean decrease. A multinomial finite-count correction retained 97.7% of the squared-distance change. However, leaving out the southwest shore gave an increase of +0.0451. Family grouping also gave an increase. Leaving out the most common taxon left almost no change. Mean nearby coral cover had returned, but recovery differed among sites. The result supports a modest decrease in the full fixed panel. It does not establish a common change across Moorea or a causal effect of coral loss. Three shores and two early years limit wider inference. The package has unchanged source objects, provenance, all count cells and pair results, figures, code, and offline repeat instructions. An independent calculation and a clean replay passed.

reef fish · spatial diversity · Moorea · coral recovery · count shares · survey counts · reproducibility