Annual heat exposure keeps a positive 1982-2025 trend under 0.10 C SST changes in a one-degree OISST sample with two seasonal baselines

We tested annual positive SST excess above seasonal 90th percentile thresholds in 29,120 fixed OISST v2.1 pixels at one-degree spacing, 60 S to 60 N. We used all days or only five-day heatwave events. Bounds include every input change of at most 0.10 C and the rebuilt baseline thresholds. For 1982-2025, both dose measures retain positive area-weighted OLS slopes with both 1983-2012 and 1991-2020 baselines. The smallest lower slope is +0.727 C day per decade; the source slopes range from +6.271 to +10.448. At 0.14 C, only event dose with the earlier baseline keeps a positive bound. Through 2021, the shared guarantee falls to 0.05 C. Some tropical bounds are weaker. A feasible 0.5 C step reverses both dose slopes with both baselines after thresholds are rebuilt. We also prove that joining qualifying runs across gaps of at most two days cannot change positive threshold-excess dose. Exhaustive event tests, rebuilt-threshold tests and aggregate audits pass. These are conditional sensitivity results, not confidence intervals or evidence that actual errors obey the tested limits. One product, a fixed point sample and no physical attribution limit inference. This dose differs from standard cumulative intensity above the seasonal mean. Code, proof, source hashes, aggregate evidence and reproduction steps are included.