Heat links to bleaching reports at multiples of five did not repeat in four Atlantic and Pacific samples in 2003-2019

We tested whether positive coral bleaching reports at multiples of five had a common link to heat stress. We used 3,362 Reef Check surveys with percentages above zero and below 100 in two oceans and two periods in 2003-2019. Such values formed 7-20 percent of reports. Separate country, year, month, depth and severity-band models gave heat slopes of +9.26, -7.94, -1.64 and -1.98 percentage points per four TSA degree heating weeks. Only early Pacific had a four-test country-grouped interval below zero. Early Atlantic had a positive site-effect estimate. Thus the tested associations did not repeat across all four samples. Values at multiples of five can result from correct measurements, averaging or reporting conventions. These data do not identify rounding error or a causal heat effect on reporting. Selected positive reports, missing surveys, within-band severity and country dependence limit inference. The package gives code, input hashes, aggregate results, numerical checks and a figure. External record data remain outside the publication.

coral bleaching · heat stress · observation bias · numeric form · regional replication · measurement audit