Four Atlantic and Pacific coral samples gave uncertain slope differences between baseline survey dates in 1998-2019

We tested whether the date of a baseline survey changed the link between coral cover and later cover change. HeatCRD version 1 gave 357 annual three-survey sequences at 114 sites. Four Atlantic and Pacific period samples used the same records for two models. One model used the survey that starts the change interval. The other used the survey one year earlier. Models controlled year, month, source and survey-date heat. Shared-survey minus earlier-survey slopes were -5.91, -1.01, -3.05 and -0.32 cover change points per ten baseline cover points. All four corrected intervals included zero. Equal-sequence weights and Reef Check records kept the contrast signs. Only ten sequences had endpoint heat of at least four degree-heating weeks. Few sites, selected repeat records, common methods and limited heat data restrict inference. The contrast cannot isolate survey error from real ecological change. We supply code, selected source fields, aggregate results, source hashes and numerical checks.

coral cover · observation bias · baseline survey · HeatCRD · marine heat · regional comparison · null result