A one-month tolerance recovered 38-50 percent of lost next-year coral matches in four Atlantic and Pacific samples, 2003-2018

We audited calendar-month matching for next-year coral records in the Atlantic and Pacific in 2003-2018. We built 9,162 month and depth cells from 10,423 Reef Check hard-coral rows in a public archive. Of 8,744 current cells, 3,751 have a next-year record in the same depth band. Exact-month matching retains 36-39% under equal returning site-year weights in four ocean-period samples. Adjacent calendar months recover 38-50% of the lost strict matches. Country bootstrap intervals describe uncertainty in these rates. Removing month wraparound and changing depth bands gives similar gains. Strict-match loss after high bleaching differs clearly from lower bleaching in two samples, but this contrast does not repeat across all four. It is unadjusted and can reflect site composition. Wider windows permit seasonal differences and do not guarantee unbiased ecological estimates. Missing archive records, selected return sites, broad depth bands, and unknown visit logs limit use. Code, source hashes, cells, match labels, all estimates, and independent checks are included.

coral bleaching · observation bias · survey timing · record linkage · calendar-month matching · regional comparison