This study was conducted in a subalpine meadow at 3,000 m a.s.l., approximately 4 km north of the Rocky Mountain Biological Laboratory (RMBL) in Gothic, Colorado. Phenological events were tracked in three focal species: Delphinium nuttallianum (Ranunculaceae), Valeriana edulis (Caprifoliaceae), and Potentilla pulcherrima (Rosaceae).
This study implemented a factorial design of warming and early snowmelt treatments to disentangle the potentially independent effects that snowmelt timing and temperature have on flowering and fruiting phenology. We used four climate treatments: control, early snowmelt, warming, and combined warming x early snowmelt. There were fourteen 7 x 7 m plots in total, and treatments were arranged in a split plot design. Seven plots received an early snowmelt manipulation, and seven were snowmelt-control plots. Warming and control sub plots were placed within each of the 14 plots (3 of each per plot). Ambient air temperatures were increased via passive warming open-top-chambers (OTC), made out of plexiglass (sensu Marion et al. 1997). The chambers were hexagonal in shape and ca. 1 m2 and 0.5 m tall. Each control sub plot was also ca. 1 m2. Controls for Valeriana were not in sub plots and instead were scattered individuals within the plot (these plants were already tagged as part of a pre-existing demographic study). Warming chambers were deployed after the snow had melted in all plots to avoid affecting the date of snowmelt. The OTCs were set-up on 17 – 19 June 2019 and remained in place through 22 August 2019. Early snowmelt treatments used a 9 x 9 m black shade cloth to advance spring snowmelt timing within the plot, plus a 1 m buffer on all sides to reduce edge effects. Snowmelt timing was manipulated by placing black shade cloths directly on the snow surface, thereby decreasing albedo and increasing heat absorption. Snowmelt treatments were initiated in the early spring (27 April), and were removed over the course of one week as the snow melted in each plot (4 June – 10 June). Mounted time lapse cameras recorded snowmelt timing for four quadrants within each plot. Snowmelt timing was defined as at least 50% visible bare ground within the quadrant.
A data logger was mounted in the center of each plot from 3 July–22 August, recording the air temperature every 15 min (HOBO data loggers, Onset Computer Corporation, USA). Additional data loggers rotated between Valeriana individuals for one week at a time throughout the season. We used an interpolation model with early season temperature data from a weather station ca. 4 km away (billy barr, unpublished data) to fill in temperatures between snowmelt timing (early to mid-June, depending on the plot) and the date on which we began collecting temperature data (3 July), as well as temperatures for weeks when there was no data logger stationed at an individual Valeriana.
Before reproductive structures had emerged, five individuals from Delphinium and five from Potentilla were tagged in each sub plot. Valeriana was part of a larger demographic study, thus numbers of Valeriana individuals per plot varied. Individual plants within each plot were followed throughout the growing season to determine the timing and duration of flowering and fruiting. Flowering was recorded when flowers were fully open, and the reproductive structures were mature. The number of open flowers were counted on each individual three days per week. For Valeriana, the number of flowering stalks was counted because Valeriana produces many small flowers within each inflorescence. Fruiting was recorded when fruits became fully ripe, indicated by a change in fruit color or in fruit drying. Flowering duration and fruiting duration were defined as the length of time (in days) between the first and last flower or fruit, respectively. Peak flowering and peak fruiting were defined as the day on which 50% of flowers or fruits were counted on an individual (following Høye et al., 2007; Iler et al., 2013).
There are well-established effects of plant size on reproductive output, and size can also affect phenology (Forrest, 2014; Ollerton and Lack, 1998). We therefore accounted for plant size when considering how climate change affects plant phenology and reproductive output. Size was measured after flowering started, under the assumption that plants were done growing for the season once they started allocating resources to reproduction. Mature fruits for Delphinium and Potentilla were collected, and seeds per individual were counted in the lab. At the end of the season, some individual Potentilla plants still had immature fruit, so we excluded fruit duration for Potentilla from our analysis because the end fruiting date was unknown. For Valeriana, seed counts were estimated in the field following Petry et al. (2016): seeds were counted in a subset of the inflorescence, then the number of these subsets needed to fill the entire inflorescence was estimated visually.
