Masting is where individual plants within a species have highly variable, but synchronised, flowering within their population, a strategy which leads to increased fitness. Two important factors controlling the size of each year’s flower crop in masting plants are temperature and resources. Under climate change, temperature cues plants respond to will be altered and may result in plants flowering regularly, which would remove the benefits of masting (e.g. predator satiation). Likewise, climate change can alter the resources available to plants for growth and flowering, potentially increasing their vulnerability to biotic and abiotic factors and also changing their flowering regime.

In this study, I set out to test the response of masting plants to temperature and resource cues to compare two previously hypothesised models for predicting the size and timing of masting events. 1) The temperature one year prior to flowering (T1) + resources (Monks et al., 2016). 2) The temperature difference (DT) between one year prior and two years prior to flowering (T1 – T2 = DT)(Kelly et al., 2013). I used two alpine snow tussocks (Chionochloa pallens and C. macra) and a mountain daisy (Celmisia lyallii) at Mt Hutt in Canterbury. Temperature cues were manipulated by transplanting plants to different altitudes on Mt Hutt and near sea level to University of Canterbury (UC). These manipulations allowed measurement of flowering responses to cooler and warmer temperatures, including extreme warming at UC. Resource cue experiments involved adding NPK fertiliser to naturally occurring tussocks at different altitudes on Mt Hutt and to potted transplants at UC. Plant responses (flowering and vegetative growth) were compared to control (unfertilised) plants at each site.

I constructed temperature series for Mt Hutt. The best estimates for gaps in on-site records were usually from NIWA VCSN data, though Darfield and Christchurch temperatures were also good. Temperature cues significantly explained the flowering response of plants in long observational datasets, with DT cues better than T1 (absolute) temperatures, supporting the DT model in these three species. In transplants moved to different temperatures, DT temperatures again gave significant predictors for Chionochloa but T1 better predicted for Celmisia, although the Celmisia dataset was small as none flowered at Mt Hutt. Fertiliser application produced no increase in flowering after one year for study species, but there was an increase in vegetative growth (new tillers or rosettes), suggesting that temperature cues are more important than resource cues for flowering in these species. Flowering did increase in the second year for fertiliser C. pallens plants which had been at UC for four years and fertiliser C. lyallii plants which had been at UC for 3 years, showing that resources do have an important, albeit secondary, role. With the temperature difference explaining flowering better in these species it implies that if global temperatures rise the variability of mast events should be maintained, and this should be true for other masting species where the DT model explains observed flowering. However, in the more extreme temperatures at UC some plants displayed abnormal flowering and growth responses suggesting climate change may have some quite negative consequences for masting species.