Since the late 19th century, global average temperatures have risen significantly, impacting various biological processes, particularly plant phenology. Changes in flowering time can have far-reaching ecological consequences, such as disruptions in pollination due to asynchrony between flowering plants and pollinators, and impacts on trophic levels due to differential responses of prey and predators to warming temperatures. The study of these impacts has involved experimental warming of plots, field studies across temperature gradients, and analysis of herbarium collections. Herbarium specimens, though subject to collection biases, offer a valuable long-term perspective on phenological changes, provided large sample sizes are used and potential biases are addressed. Prior studies have shown varied responses in flowering time to temperature increases, ranging from 2-6 days earlier per 1°C increase, with some species even blooming later. This variation may be attributed to differences in species' responses to seasonal temperatures (winter or spring) and plant functional traits. This study aimed to use herbarium specimens to examine long-term changes in flowering time in relation to historical temperature data in eastern Pennsylvania, analyzing the effects of annual, winter, and spring temperatures and the influence of plant functional traits on the response to warming.

Previous research on climate warming's effect on flowering phenology has employed various methods, including experimental warming, studies across temperature gradients (elevation and latitude), and the analysis of herbarium specimens. While experimental warming and field studies offer controlled environments or natural gradients, they are limited in their ability to provide long-term historical perspectives. Herbarium specimens offer this long-term perspective, but researchers must acknowledge and mitigate biases stemming from non-random collection practices and potential inconsistencies in phenophase assessment. Existing studies using herbarium data have reported shifts in flowering time ranging from 2-6 days earlier per 1°C increase in temperature, although significant interspecific variation exists, with some species showing no or even opposite responses. This variation could be due to differing responses to seasonal temperatures (winter or spring) and to plant functional traits.

This study utilized herbarium specimens from four counties in eastern Pennsylvania, spanning from 1884 to 2015. Historical climate data were obtained from the National Oceanic and Atmospheric Administration (NOAA) using data from multiple weather stations within the region, averaging daily temperatures to calculate annual (YearTAVG), winter (WinterTAVG), and spring onset (SpringTAVG) average temperatures. Herbarium specimens from Muhlenberg College, The Academy of Natural Sciences, and The Morris Arboretum were used, with digital specimens sourced from the Mid-Atlantic Megalopolis Project. The researchers selected insect-pollinated species with at least 30 specimens, removing duplicates with identical dates and locations, and classifying species based on native status, growth form (woody vs. herbaceous), fruit type (dry vs. fleshy), and blooming season (spring vs. summer). Collection dates were converted to day of the year. Linear mixed-effects models were used to assess the relationship between temperature (YearTAVG, WinterTAVG, SpringTAVG) and flowering date, with species treated as a random effect. Additional models included binary variables (native status, growth type, fruit type, seasonality) and their interactions with temperature to analyze differences in phenological responses among categories. Non-parametric bootstrapping was used when necessary to handle non-normal residuals or heteroscedasticity. Species with flowering periods exceeding five months were excluded in some analyses to focus on the majority with well-defined shorter blooming periods.

Across all species, plants flowered 2.26 days earlier per 1°C increase in annual average temperatures and 2.93 days earlier per 1°C increase in spring onset average temperatures. Winter temperatures did not significantly influence flowering phenology. The relationship between temperature and flowering phenology was not significantly different between native and non-native species. Woody species flowered earlier than herbaceous species only in response to increasing annual temperatures. There was no difference in the phenological response between species with dry fruits and those with fleshy fruits. Spring-blooming species exhibited a significantly greater phenological response to warming yearly average temperatures than summer-blooming species (2.27 days earlier per 1°C increase). Excluding species with extended flowering periods did not substantially alter these results.

The findings demonstrate that climate change is significantly advancing flowering phenology in eastern Pennsylvania. The observed average advance of 2.26 days earlier per 1°C increase in yearly average temperature aligns with previous studies, although it's on the lower end of the reported range. This could be due to the relatively undeveloped nature of the study area prior to the mid-1990s and the absence of specimens from urban or industrial sites, potentially minimizing heat island effects. The stronger response to spring temperatures compared to annual temperatures underscores the importance of seasonal temperature variations. The lack of a significant response to winter temperatures suggests the possibility of counteracting effects of winter and spring warming, with warmer winters potentially delaying flowering in some species. The lack of difference in phenological response between native and non-native species contradicts expectations based on the potential for greater phenotypic plasticity in non-native species. This may be due to the underrepresentation of invasive non-native species in the study or the influence of adaptive selection. The unexpected stronger response in woody plants compared to herbaceous plants might be linked to the complex interplay of temperature and photoperiod cues for flowering in woody species. The greater advance in spring-blooming species compared to summer-blooming species aligns with previous research and can be explained by the earlier influence of warming on spring phenological processes and a reduced risk of late-winter damage.

This study provides evidence of climate change's significant impact on flowering phenology, highlighting the importance of considering both annual and seasonal temperatures and species-specific functional traits. Future research should investigate the underlying mechanisms driving the variation in responses among species, exploring the interplay of seasonal temperatures, other climate factors (precipitation, extreme events), and urban heat island effects. Phylogenetic analyses and consideration of species-specific flowering cues are also essential for a comprehensive understanding of climate change's impact on plant phenology.

Potential limitations of this study include potential underestimation of the phenological response in recent years due to declining herbarium collection rates. The relatively limited sample sizes for some categories could also affect the statistical power of the analyses. The study primarily focused on insect-pollinated species, excluding other pollination types, which could limit the generalizability of findings.