Figure 7. Mean projected changes in the priming period (Dp, day), measured as the number of days between the first HS of a year and the start of the continuous MHW, by the end of 21st century (2071-2100) relative to the historical period (1985-2014). The mean states from the first to last columns are simulations by GFDL-ESM4, CESM2-WACCM and MRI-ESM2 and those from the first to last rows refer to the simulations under SSP1-2.6, SSP2-4.5, SSP3-7.0 scenarios, respectively.
Unlike the duration, peak intensity and accumulated heat stress, a larger fraction of the ocean is projected to experience decreases in the heating rate (HRc) and priming period (Dp), mostly in the tropics and subtropics (Figure 3-7). While most of the decreases in HRc are small (< 0.04 °C·day-1), the decreases in the subtropical Pacific in GFDL-ESM4 under SSP3-7.0 are up to 0.20 °C·day-1, which is roughly twice the mean heating rate in the historical period (Figure 2a, 6c). The declines of HRc correspond with the greater rate of increase in Dc relative to that in HSpeak. In all three models, the duration of priming period is projected to decrease in the western tropical Pacific, in conjunction with the increases in Dc or the duration of continuous heat stress (Figure 3, 7). In GFDL-ESM4, the largest decreases are as high as 150 day and approximately 60 day larger than with the other two models. In contrast, there are also increases in Dp with most below 30 day, although several large increases in Dp occur in the equatorial Indian and Atlantic Ocean in each of the three models (e.g., the increase >150 day under SSP3-7.0). These cases involve a long Dp followed by a short and mild MHW event, which does not fit the definition of priming in that a real priming period needs to occur prior to severe MHW event (Figure 7).
4.3 Projected changes in warm-season MHW properties over global areas of coral reef and kelp systems, and role of model biases