Storm timing showed a surface temperature contrast but no clear bleaching contrast in pooled reef records from 2003 to 2019

We linked Reef Check surveys to dated tropical storm tracks to test whether prior storms had lower bleaching associations than later storms. A model with site, ocean-year and ocean-month effects used 6,448 survey days at 1,452 sites in three oceans from 2003 to 2019. The prior-minus-later bleaching contrast was -0.92 percentage points, with a 95% interval of -3.25 to +1.42. All five estimable ocean-period bleaching intervals included zero after correction for six tests. The later Indian Ocean contrast could not be estimated. A secondary model of weekly surface temperature anomalies gave a pooled contrast of -0.397 C, with an interval of -0.557 to -0.236 C. This contrast remained negative in pooled sensitivity checks. The only clear adjusted regional surface result included zero after removal of windows near unknown storm intensity. The surface result fits a cooling explanation. It does not establish a causal cooling effect or coral protection. Selected visits, missing intensity, few regional clusters, coarse surface temperature and limited high-heat samples restrict inference. The package gives the dated linkage dataset, missing-value evidence, raw source hashes, all estimates, code, figures and numerical checks. It does not measure pollution or species loss.

coral bleaching · tropical cyclone · storm timing · sea surface temperature · observation bias · negative control · replication · Reef Check · IBTrACS