Warming spring temperatures and decreased winter precipitation have advanced the date of snowmelt over much of the western US, so increased frost damage may seem paradoxical as an effect of climate change, but at RMBL a trend for earlier snowmelt (initiation of growing season) has not been accompanied by a change in date of last spring frost, resulting in increased exposure of plants. My study of wildflower phenology, initiated in 1973, and of demography (since 1998), shows the influence of climate change, regional and global, on population biology of aspen sunflower. Although many studies now document recent changes in phenology, this study may be among the first showing a subsequent effect on demography. I collect data on the phenology and abundance of flowering by counting all flowers in 30 2x2m plots every other day for the growing season. These data are then correlated with environmental data (snowmelt dates, temperature), which explain much of the variation in flower phenology and abundance.
Results/Conclusions

I have observed frost damage to leaves, flowers, or fruits of 15 species. The number of Helianthella capitulae flowering in plots has ranged from 1 to 4,448 since 1975. Since 1992, the proportion of flower buds killed by frost has ranged from 0 – 100%. Over the past eight years bud mortality has been zero in one year; in the other seven it has ranged from 65 – 100%. The probability and degree of frost damage are correlated with the date of snowmelt (r² = 0.547, p = 0.001); late snowmelt delays the growing season so that there are no sensitive buds at the time of last hard frost. In years with early snowmelt there is no significant correlation between the number of unfrosted heads and snowmelt date, but for years with later snowmelt there is a significant correlation between these variables (r² = 0.363, p = 0.008), indicating a threshold effect for snowpack. Low flower production is related to low recruitment; the two highest recruitment years followed years with high flower production. Demographic data collected in an annual census of individually tagged plants in permanent plots show that populations are declining. The influence of regional climate change from the phase change of the North Pacific Oscillation is suggested by the fact that the means for % inflorescences killed by frost are 36.1% for the seven years prior to the 1998 phase change and 76.6% for the nine years since then.