Heat slopes for bleaching occurrence and positive severity did not both repeat in four coral samples in 2003-2019

We tested heat links to any bleaching and to bleaching percentage among positive Reef Check surveys. Four Atlantic and Pacific samples in 2003-2019 used 9,623 unique survey IDs. Models controlled country, year, month and depth. All four occurrence intervals included zero after correction for eight primary tests. Only early Atlantic positive severity had an interval above zero: 13.14 percentage points per four TSA degree-heating weeks, with interval 5.21 to 21.06. Its repeat-site interval included zero. A 1% report threshold changed which intervals excluded zero. These models did not establish clear links for both parts across samples. Selected visits, missing outcomes, outcome-based conditioning, site confounding and country dependence limit inference. The package gives aggregate results, code, source hashes, numerical checks and a figure. External record data remain outside the bundle.

coral bleaching · heat stress · conditional severity · occurrence · observation bias · regional replication