The initiation of plant flowering and insect activity are often triggered by environmental cues, and unusual weather conditions could disrupt plant-pollinator interactions by reducing temporal overlap of plant and pollinator phenologies. If environmental conditions affect phenological modification and consequently plant reproductive fitness, plant populations would be expected to show ecotypic adaptation to their respective environments and thus vary in their risk of phenological mismatch. In this study, we examined the interaction between the self-incompatible spring ephemeral Corydalis ambigua and its main pollinators — overwintered bumblebee queens — at five sites with sequential snowmelt timings around Hokkaido, Japan.
C. ambigua flowers soon after snowmelt, and has to receive pollination service by bumblebee queens within their short flowering duration (4–8 days) to produce viable seed. During years with early snowmelt and warm spring temperatures, phenological mismatch between C. ambigua and bumblebee queens occurred, and seed set was decreased. We compared flowering progress and bumblebee visitation between C. ambigua in early- and late-snowmelt sites to identify ecotypic variation and determine whether the risk of phenological mismatch varies between sites.
In a common garden experiment, transplanted C. ambigua showed sequential flowering that mirrored the snowmelt sequence of their site of origins, suggesting that ecotypic adaptation of flowering traits to local snowmelt conditions may have occurred to ensure phenological matching with bumblebee emergence. Bumblebee visitation varied greatly between years and sites, with low bumblebee abundance in 2022 compared to 2023, and higher bumblebee abundance at late-snowmelt sites than at early-snowmelt sites. Interactions between site and year effects significantly influenced C. ambigua seed set, indicating that C. ambigua populations experience site-specific pollen limitation that varies annually.
Early-snowmelt sites have larger annual variation in snowmelt timing and generally lower bumblebee abundance than late-snowmelt sites, thus risk of phenological mismatch may be higher in early-snowmelt sites compared to late-snowmelt sites. This study suggests that the risk of phenological mismatch may differ between plant populations due to ecotypic variation and site-specific factors, which could lead to site variation in population dynamics under climate change. Further studies would be needed to determine whether C. ambigua populations can adapt quickly to track changes in climate.