Effects of phenological shifts across temperatures
The effect of parasitoids on average host survival depended on the relative emergence time of parasitoids, and its interaction with warming (Fig. 2A, & Fig. S4B; Table S4; note that warming alone did not alter average host survival). Probability of host survival was at or near zero when host and parasitoids emerged simultaneously and increased significantly through the first four days of delay, followed by a plateau between four and six days (P > 0.05) of parasitoid delay (Fig. 2A). Increased development rates in elevated temperatures magnified differences in host survival across phenological delays for the first four days of phenological delays: the largest increase in host survival occurred between 2- and 4-days (OR = 22.36, P < 0.001) in ambient temperatures, compared to 0- and 2-days (OR = 31.11, P < 0.001) in warming (Fig. 2A). Interestingly, even with late phenological delays, parasitoids continued to inflict mortality on hosts, as rates of host survival rarely exceeded 80% of what was attained in trials without parasitoids.
Like host survival, parasitism rates differed significantly across phenological delays, with rates of parasitism declining as phenological delay increased (P < 0.05 for all phenological delays). Warming significantly reduced parasitism rates, but the decline in parasitism was not coupled with increased host survival. When emergence phenologies matched, host survival was at or near 0% in both temperature treatments, but rates of successful parasitism ultimately exceeded 50% in ambient temperature, compared to a maximum of 10% under warming. Consequently, even though hosts were developing faster in elevated temperatures, differences in parasitism rates across phenological delays were diminished (Fig. 2B & Fig. S4A; Table S5), since the chance of parasitism was already less than 5% with just a two-day delay. Together, these results suggest that warming reduced the effects of phenological shifts on parasitism rates due to significant declines in parasitoid performance and, to a lesser extent, because of increased development rates in hosts.