Detailed mapping of coral bleaching events provides an opportunity to examine spatial patterns in bleaching over scales of 10s to 1,000s of km and the spatial correlation between sea surface temperature (SST) and bleaching. We present data for two large-scale (2,000km) bleaching events on the Great Barrier Reef (GBR): one from 1998 and another from 2002, both mapped by aerial survey methods. We examined a wide range of satellite-derived SST variables to determine which one best correlated with the observed bleaching patterns. We found that the maximum SST occurring over any 3-day period (max3d) during the bleaching season predicted bleaching better than anomaly-based SST variables and that short averaging periods (3–6days) predicted bleaching better than longer averaging periods. Short periods of high temperature are therefore highly stressful to corals and result in highly predictable bleaching patterns. Max3d SST predicted the presence/absence of bleaching with an accuracy of 73.2%. Large-scale (GBR-wide) spatial patterns of bleaching were similar between 1998 and 2002 with more inshore reefs bleached compared to offshore reefs. Spatial change in patterns of bleaching occurred at scales of ~10skm, indicating that reefs bleach (or not) in spatial clusters, possibly due to local weather patterns, oceanographic conditions, or both. Approximately 42% of reefs bleached to some extent in 1998 with ~18% strongly bleached, while in 2002, ~54% of reefs bleached to some extent with ~18% strongly bleached. These statistics and the fact that nearly twice as many offshore reefs bleached in 2002 compared to 1998 (41 vs. 21%, respectively) makes the 2002 event the worst bleaching event on record for the GBR. Modeling of the relationship between bleaching and max3d SST indicates that a 1C increase would increase the bleaching occurrence of reefs from 50% (approximate occurrence in 1998 and 2002) to 82%, while a 2C increase would increase the occurrence to 97% and a 3C increase to 100%. These results suggest that coral reefs are profoundly sensitive to even modest increases in temperature and, in the absence of acclimatization/adaptation, are likely to suffer large declines under mid-range International Panel for Climate Change predictions by 2050.