ArticlePDF Available

Herbarium specimens can reveal impacts of climate change on plant phenology; a review of methods and applications



Studies in plant phenology have provided some of the best evidence for large-scale responses to recent climate change. Over the last decade, more than thirty studies have used herbarium specimens to analyze changes in flowering phenology over time, although studies from tropical environments are thus far generally lacking. In this review, we summarize the approaches and applications used to date. Reproductive plant phenology has primarily been analyzed using two summary statistics, the mean flowering day of year and first-flowering day of year, but mean flowering day has proven to be a more robust statistic. Two types of regression models have been applied to test for associations between flowering, temperature and time: flowering day regressed on year and flowering day regressed on temperature. Most studies analyzed the effect of temperature by averaging temperatures from three months prior to the date of flowering. On average, published studies have used 55 herbarium specimens per species to characterize changes in phenology over time, but in many cases fewer specimens were used. Geospatial grid data are increasingly being used for determining average temperatures at herbarium specimen collection locations, allowing testing for finer scale correspondence between phenology and climate. Multiple studies have shown that inferences from herbarium specimen data are comparable to findings from systematically collected field observations. Understanding phenological responses to climate change is a crucial step towards recognizing implications for higher trophic levels and large-scale ecosystem processes. As herbaria are increasingly being digitized worldwide, more data are becoming available for future studies. As temperatures continue to rise globally, herbarium specimens are expected to become an increasingly important resource for analyzing plant responses to climate change.
Submitted 22 November 2017
Accepted 16 March 2018
Published 3 April 2018
Corresponding author
Casey A. Jones,
Academic editor
Frank Berninger
Additional Information and
Declarations can be found on
page 11
DOI 10.7717/peerj.4576
2018 Jones and Daehler
Distributed under
Creative Commons CC-BY 4.0
Herbarium specimens can reveal impacts
of climate change on plant phenology; a
review of methods and applications
Casey A. Jones and Curtis C. Daehler
Department of Botany, University of Hawaii at Manoa, Honolulu, HI, United States of America
Studies in plant phenology have provided some of the best evidence for large-scale
responses to recent climate change. Over the last decade, more than thirty studies
have used herbarium specimens to analyze changes in flowering phenology over time,
although studies from tropical environments are thus far generally lacking. In this
review, we summarize the approaches and applications used to date. Reproductive
plant phenology has primarily been analyzed using two summary statistics, the mean
flowering day of year and first-flowering day of year, but mean flowering day has
proven to be a more robust statistic. Two types of regression models have been
applied to test for associations between flowering, temperature and time: flowering day
regressed on year and flowering day regressed on temperature. Most studies analyzed
the effect of temperature by averaging temperatures from three months prior to the
date of flowering. On average, published studies have used 55 herbarium specimens
per species to characterize changes in phenology over time, but in many cases fewer
specimens were used. Geospatial grid data are increasingly being used for determining
average temperatures at herbarium specimen collection locations, allowing testing
for finer scale correspondence between phenology and climate. Multiple studies have
shown that inferences from herbarium specimen data are comparable to findings from
systematically collected field observations. Understanding phenological responses to
climate change is a crucial step towards recognizing implications for higher trophic
levels and large-scale ecosystem processes. As herbaria are increasingly being digitized
worldwide, more data are becoming available for future studies. As temperatures
continue to rise globally, herbarium specimens are expected to become an increasingly
important resource for analyzing plant responses to climate change.
Subjects Plant Science, Climate Change Biology
Keywords Climate change, Phenology, Herbarium specimens, Methods
Carl Linnaeus pioneered the study of phenology when he outlined methods for investigating
associations between flowering and climate in the 1700s (Linnaeus, 1751;Von Linné, 2003;
Puppi, 2007). Around 1850, Charles Morren introduced the term ‘‘phenology’’ to describe
his observational studies of yearly flowering (Morren, 1853;Demarée & Rutishauser, 2009).
Early field studies of plant phenology have been thoroughly reviewed by Van Schaik,
Terborgh & Wright (1993),Fenner (1998) and Forrest & Miller-rushing (2010). Long-term
observations in field studies have provided a valuable resource for analyzing phenological
How to cite this article Jones and Daehler (2018), Herbarium specimens can reveal impacts of climate change on plant phenology; a re-
view of methods and applications. PeerJ 6:e4576; DOI 10.7717/peerj.4576
responses to recent climate change (Walther et al., 2002;Parmesan & Yohe, 2003). A
growing need for historical data that allows for the exploration of ecological implications
of climate change prompted researchers to look to herbarium specimens. A few phenology
studies such as Borchert (1996) and Rivera & Borchert (2001) used herbarium specimens
to study flowering periodicity, but not in the context of climate change. The first study
to use herbarium specimens to understand phenological responses to climate change was
published in 2004 by Primack et al. (2004).Primack et al. (2004) analyzed 372 specimen
records (1885–2002) and found flowering had advanced approximately eight days over the
last century. Between 2004 and 2017, more than 30 studies were published using herbarium
specimens to examine changes in phenology in response to climate change.
The most common approach found in studies using herbarium specimens follows the
procedure set by Primack et al. (2004). This can be summarized as collecting Julian dates
from herbarium specimens, collecting long-term temperature data from an independent
source, and then using regression analyses to analyze correlations between Julian dates,
temperatures and time (Primack et al., 2004;Miller-Rushing et al., 2006;Gallagher, Hughes
& Leishman, 2009;Robbirt et al., 2011;Gaira, Dhar & Belwal, 2011;Molnár et al., 2012;
Panchen et al., 2012;Park, 2012;Primack & Miller-rushing, 2012;Li et al., 2013;Calinger,
Queenborough & Curtis, 2013;Hart et al., 2014;Rawal et al., 2015;Park & Schwartz, 2015).
Primack et al. (2004) recorded the date of collection from each herbarium specimen and
then extracted Julian dates from the collection dates. A Julian date is a value between 1 and
365 corresponding to the day of year when the specimen was collected. Linear regression
models are also the most widely used statistical models in field studies investigating
flowering phenology (Zhao et al., 2013).
An early criticism of using herbarium specimens was that plant parts preserved as
herbarium specimens might not have been collected during their peak flowering season,
potentially biasing interpretations (Lamoureux, 1973). Daru, Van der Bank & Davies (2017)
also found spatial, temporal, trait, phylogenetic, and collector biases among herbarium
specimen samples. Daru, Van der Bank & Davies (2017) concluded that while some of
these biases can be accounted for using statistical approaches, future herbarium collections
should focus on filling large gaps in the data. Other studies have found that large sample
sizes afforded by herbarium specimens, and the use of mean flowering times (mean of
Julian dates), could yield valid inferences, even if specimens were not collected at the time
of peak flowering (Primack et al., 2004;Bertin, 2015). Collector bias and plant size choice
have also been overcome by statistical analyses when mean flowering times were used as
the variable of interest, rather than the date of first-flowering (Robbirt et al., 2011;Davis et
al., 2015).
Most of the studies we reviewed used two types of linear regression models to show
evidence of associations between phenology and climate change (Table 1). These studies
regressed flowering day on temperature (82%) and flowering day on year (64%) (Table 1).
These studies have primarily been conducted with specimens from herbaria in temperate
latitudes such as the Eastern Himalayas (Gaira, Dhar & Belwal, 2011;Li et al., 2013;Gaira
et al., 2014;Hart et al., 2014), Southern Australia (Gallagher, Hughes & Leishman, 2009;
Rawal et al., 2015), Northern Europe (Robbirt et al., 2011;Diskin et al., 2012;Molnár et al.,
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 2/16
Table 1 Methods of studies. The column ‘‘Flw Day Temp’’ represents studies that conducted a type of regression analysis with flowering day
(Flw Day) as the dependent variable and temperature average (temp) or year as the independent variable. The ‘‘1¯x’’ symbol represents studies that
analyzed a difference in the mean flowering day between historic and current time period groups rather than using a type of regression analysis.
Species Specimens Specimen
per species
Authors Year Geographic region (flw temp) (flw year)
1 117 117 Gaira et al. 2011 Eastern Himalayas x
1 N/A N/A Gaira et al. 2014 Eastern Himalayas x x
1 192 192 Robbirt et al. 2011 Northern Europe x
5 158 32 Rawal et al. 2015 Southern Australia x x
5 540 108 Diskin et al. 2012 Northern Europe x x
20 371 19 Gallagher et al. 2009 Southern Australia x x
20 1,108 55 Davis et al. 2015 North America x x
28 1,587 57 Panchen et al. 2012 North America x x
36 460 13 Hart et al. 2014 Eastern Himalayas x
>37 372 10 Primack et al. 2004 North America x x
39 216 6 Lavoie & Lachange 2006 North America x
39 5,424 139 Molnár et al. 2012 Northern Europe x
41 909 22 Li et al. 2013 Eastern Himalayas x x
42 142 3 Miller-Rushing et al. 2006 North America x x
43 N/A N/A Primack & Miller-Rushing 2012 North America x
87 N/A N/A Neil et al. 2010 North America x
141 5,053 36 Calinger et al. 2013 North America x
186 30,000 161 Bertin 2015 North America 1¯x
370 1,125 3 Searcy 2012 North America 1¯x
1,185 5,949 5 Park 2012 North America x
>1,700 19,328 11 Park 2014 North America x
24,105 823,033 34 Park & Schwartz 2015 North America x x
2012), and North America (Primack et al., 2004;Lavoie & Lachance, 2006;Miller-Rushing et
al., 2006;Primack & Miller-Rushing, 2009;Neil, Landrum & Wu, 2010;Panchen et al., 2012;
Park, 2012;Primack & Miller-rushing, 2012;Searcy, 2012;Calinger, Queenborough & Curtis,
2013;Park, 2014;Park & Schwartz, 2015;Bertin, 2015;Davis et al., 2015). Although studies
by Borchert (1996) and Zalamea et al. (2016) analyzed flowering periodicity in tropical
plants using herbarium specimens, we found no study to date that has used herbarium
specimens to analyze effects of recent climate change in a tropical region. In this review,
we examined how studies chose sample sizes, flowering specimens, temperature averages
and geographical scale in their analyses. We also examined how these studies validated the
use of herbarium specimens and we provide suggestions for methods to be used in future
Survey methodology
Between 2015 and 2017, we compiled and reviewed studies that used herbarium specimens
to assess climate change and flowering phenology. We searched Web of Science (1900—
present), JSTOR (1665—present) and Google Scholar for studies containing the terms
herbarium, specimen, phenology, and climate change. The methods of each study
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 3/16
were reviewed for; sample size, determining flowering status of specimens, approach
to determining temperatures, geographic variation, and any validations of the use of
herbarium specimens (e.g., comparisons to field observations). Studies and methods were
then categorized and a synthesis of each category is discussed; sample sizes and regression
methods were also summarized (Table 1).
Specimen sample sizes
Sample size, or the number of specimens used per species, varied across studies (Table 1).
The minimum number of specimens used per species was occasionally as low as two or three
records (Searcy, 2012). Some studies using herbarium data have set a minimum number of
herbarium specimens per species or a minimum time range for collections in order to more
accurately estimate phenologies and change over time. Calinger, Queenborough & Curtis
(2013) and Gallagher, Hughes & Leishman (2009) set a minimum of 10 specimens in order
to meet statistical assumptions of different models. Molnár et al. (2012) eliminated a species
from analyses because collections only yielded dates across an eight year time span. Park
& Schwartz (2015) eliminated species with records that spanned less than three years. Neil,
Landrum & Wu (2010) organized species into functional groups (spring ephemerals, spring
shrubs, fall ephemerals, winter-spring ephemerals, and winter-spring shrubs) in order to
overcome the problem of low sample sizes for each species but found that responses of
individual species varied greatly within functional groups.
Several studies found sample size had a greater influence on first-flowering estimates
than on mean flowering estimates. Miller-Rushing & Primack (2008) used field data and
found that small sample sizes led to biased estimations of first-flowering dates, but mean
flowering day was not biased by sample sizes. Moussus, Julliard & Jiguet (2010) investigated
sample sizes by simulating 10 known phenological estimators, such as mean flowering day
and first-flowering date. After comparing known phenological shifts from simulated sample
data with shift estimations from models using the same data, Moussus, Julliard & Jiguet
(2010) concluded that first-flowering dates were inaccurate because they showed much
greater differences in comparisons than mean flowering day. Low sample sizes prompted
Bertin (2015) to provide a detailed analysis of how sample size affected mean, median,
range, early flowering and late flowering summary statistics. In random simulations
comparing sample sizes, mean flowering day values deviated less than five days for species
with as few as four samples (Bertin, 2015). Bertin (2015) concluded that the mean was
a more robust measure of phenology than other estimators of early flowering. Bertin
(2015) also showed that by increasing the sample size to 20, mean flowering times deviated
only one to two days. A recent study by Pearse et al. (2017) used a Weibull distribution
to estimate the start of the process of flowering rather than using only first-flowering
observations. Pearse et al. (2017) showed that by controlling for differences in sampling,
first-flowering, peak-flowering (median) and cessation of flowering show similar changes
over time in response to climate change. The model used by Pearse et al. (2017) was also
shown to be consistent with changes in mean-flowering from a separate sample using an
early time period.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 4/16
Larger sample sizes may be required if phenology varies across a species’ geographic
range. In order to analyze species distributions using herbarium specimens, Van Proosdij
et al. (2016) found that the minimum number of herbarium specimens sampled should be
between 14 and 25 depending on the geographical range of the species. The Van Proosdij et
al. (2016) study used simulated species to assess the minimum herbarium samples required
for acceptable model performance in both virtual and real study areas. Some species with
narrow geographical ranges could be modeled with as few as 14 herbarium records while
wide ranging species could be satisfactorily modeled with a minimum of 25 records (Van
Proosdij et al., 2016). Based on these studies, we recommend caution when interpreting
results from samples sizes with fewer than 30 records (Miller-Rushing & Primack, 2008;
Moussus, Julliard & Jiguet, 2010;Bertin, 2015). The average sample size across studies in
this review was about 55 records per species (Table 1). We also recommend using the mean
flowering day of year rather than averages of first flowering dates (Calinger, Queenborough
& Curtis, 2013;Gallagher, Hughes & Leishman, 2009;Pearse et al., 2017).
Determining flowering status of specimens
Some studies have simply recorded the presence or absence of flowers from herbarium
specimens as an indicator of flowering, but other studies have used more detailed criteria
to assess flowering status on specimens. Haggerty, Hove & Mazer (2012) provided a primer
to assist researchers with collecting data from herbarium specimens. Haggerty, Hove &
Mazer (2012) suggested researchers assign a phenophase for each specimen, such as pre-
flowering, first-flowering or peak flowering. Haggerty, Hove & Mazer (2012) also noted
that researchers must assume the stem on the herbarium sheet represents the flowering
phenophase for the entire plant. Past studies, such as Diskin et al. (2012), have used a
scoring system from 1 to 5 to categorize phenophase stages raging from ‘‘no flowers’’
to ‘‘end of fruiting’’ on each specimen. Diskin et al. (2012) categorized flowering as 50%
of buds open on the specimen. Calinger, Queenborough & Curtis (2013) also categorized
flowering as 50% of flower buds in anthesis to ensure that the samples were in peak
flowering. For a species with an inflorescence, Davis et al. (2015) only counted specimens
as flowering if greater than 75% of flowers were open. Standardization of phenological
terms remains a core challenge of mining phenological data (Willis et al., 2017). Initiatives
such as the Plant Phenology Ontology (PPO) working group are currently structuring
phenological terms for more uniform application across studies (Willis et al., 2017).
Studies in temperate regions have used varying methods to determine flowering status
for species with long flowering durations. For example, Molnár et al. (2012) and Bertin
(2015) excluded species that flowered outside of the peak flowering season of the region,
defined as the period from late-spring to early-summer. Molnár et al. (2012) removed one
species because its peak flowering date was in September and focused on 40 other taxa that
had flowering peaks from in spring and early-summer. The excluded species was a strong
outlier and it was suggested that autumn climate events may affect species differently than
spring climate events (Molnár et al., 2012). Park (2012) also removed outlier records when
flowering records fell outside the peak regional flowering season. Flowering records before
Julian day 45 and after Julian day 310 were removed from analyses to reduce biases caused
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 5/16
by winter flowering species. Additionally, Park (2012) removed records that were 150 days
after the median flowering date for each species to reduce errors caused by any second
flowerings that can happen in autumn months. Several other studies removed taxa with
long flowering durations to reduce variance among species. Bertin (2015) excluded native
weedy species with flowering durations from spring to fall. Gallagher, Hughes & Leishman
(2009) only used species with a flowering duration of less than three months. Panchen et
al. (2012) chose to use only species with clear beginning and ending points to investigate
long and short flowering duration. Panchen et al. (2012) found that plants with shorter
flowering durations required smaller sample sizes to produce significant results when
regressing flowering day on year.
Other studies such as Calinger, Queenborough & Curtis (2013) and Lavoie & Lachance
(2006) disregarded the effect of flowering duration and noted the results of Primack
et al. (2004), which reported no bias associated with long or short flowering durations
when mean estimations are analyzed. Plants in tropical regions often have long flowering
durations (Van Schaik, Terborgh & Wright, 1993;Fenner, 1998), but as long as flowering
is not continuous throughout the year, methods applied to temperate regions should
also yield valuable insight into effects of climate change on phenology in the tropics.
While studies using herbarium specimens to analyze long-term changes have been limited
to temperate regions, future studies could use circular statistics to analyze long-term
phenological changes in tropical regions (Fisher, 1993;Morellato, Alberti & Hudson, 2010).
Circular statistics have been used to analyze flowering phenology in several tropical field
studies, but these studies lacked long-term climate change analyses (Novotny & Basset,
1998;Morellato et al., 2000;Cruz, Mello & Van Sluys, 2006;Rogerio & Araujo, 2010;Tesfaye
et al., 2011;Nadia, Morellato & Machado, 2012;Nazareno & Dos Reis, 2012;Staggemeier,
Diniz-Filho & Morellato, 2010;Carvalho & Sartori, 2015;Kebede & Isotalo, 2016).
Averaging temperatures
The foundational study by Primack et al. (2004) examined temperature averages from three
calendar months prior to the specimen flowering date, with the assumption that flowering
date is a function of temperatures experienced in past months. Field investigations such
as Fitter et al. (1995) have shown temperature averages from different sets of months
preceding flowering affected flowering phenology in different ways. More recently, Calinger,
Queenborough & Curtis (2013) chose to regress the month of flowering with temperature
averages from each of the eleven months prior to flowering. They found that temperature
averages from three months prior to the date of flowering showed the strongest correlations
with flowering (Calinger, Queenborough & Curtis, 2013). Robbirt et al. (2011) investigated
three sets of temperature averages over three month intervals and also found that three
months prior to flowering had the most predictive power. Similarly, Rawal et al. (2015)
regressed flowering on temperature averages for each species from 1, 3, 6, 9, and 12 months
prior to flowering, because responses can vary by species. Rawal et al. (2015) also found
that mean temperatures three months prior had the greatest influence on flowering time
for all species.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 6/16
Other studies have used average temperatures from spring months because spring
temperatures generally have the most predictive power for flowering date (Miller-
Rushing & Primack, 2008;Primack & Miller-Rushing, 2009;Robbirt et al., 2011;Calinger,
Queenborough & Curtis, 2013;Park, 2014;Park & Schwartz, 2015). Bertin (2015) found an
interesting trend that supported the effect of spring temperatures: the earlier a species’ mean
flowering time occurred in the spring, the more the species’ mean dates had shifted toward
an earlier day of year over time. Robbirt et al. (2011) also found the highest correlations of
flowering day with spring temperature averages across March, April and May. Calinger,
Queenborough & Curtis (2013) found significant changes in flowering in response to average
spring temperatures (February–May) but not in response to summer temperatures (June–
September). Gaira, Dhar & Belwal (2011) found the highest correlations between flowering
and temperatures in earlier months from December–February in a Himalayan perennial.
As an alternative to using mean monthly temperatures, Diskin et al. (2012) investigated
the averages of temperature anomalies, or deviations from the overall long-term mean,
for 2, 3, and 6 month periods from January to June and found averages from six months
prior to flowering had the strongest correlations. Park (2014) used temperature averages
across three month periods from early spring to late summer and found a similar trend.
Temperature averages were organized into early, mid, and late seasonal classes within
the months of February–October. Park (2014) found warming temperatures had affected
species in the early spring class more than other classes. Park & Schwartz (2015) also used
early, mid and late seasonal classes for spring and summer and found that mid-season
phenology events should be modeled differently than early or late season events. Hart et al.
(2014) used annual temperatures and temperatures from each season (spring, summer, fall,
and winter) and found significant correlations for annual and fall temperature averages,
but with opposite effects. Hart et al. (2014) discussed that warmer fall temperatures may
delay the chilling requirement for Rhododendron species, resulting in a delay in flowering
while warmer annual temperatures will lead to advances in flowering overall. Other studies
found annual temperature means were as useful as spring temperatures. Davis et al. (2015)
found similar results between spring and annual temperature averages and used annual
averages in analyses. Gallagher, Hughes & Leishman (2009) also used annual temperature
means for analyses and explained that seasonal means were correlated with annual means.
We recommend investigating the effect of temperature by analyzing averages from
multiple sets of months prior to flowering for each species rather than using only
one fixed spring interval or only annual temperatures (Diskin et al., 2012;Robbirt et al.,
2011;Calinger, Queenborough & Curtis, 2013). Caution should be taken when analyzing
temperature averages from the same months prior to flowering for all species when
flowering month varies by species. For example, when analyzing the effect of temperature
averages from three months prior for all species, Calinger, Queenborough & Curtis (2013)
found that for many species, flowering was correlated with temperatures three months
earlier, yet for species with an earlier mean flowering day in April, January temperatures
(three months prior to flowering) did not predict flowering date; instead, temperature
averages from the months of February, March and April were better predictors for those
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 7/16
Geographic variation
Among species that have broad geographic ranges, differences in climate in different parts
of the species’ range can complicate attempts to correlate a species’ flowering day with
temperature. Several methods have been used to account for climate variability across
a species’ range. An early study by Lavoie & Lachance (2006) investigated the effects of
climate variation on the phenology of Coltsfoot (Tussilago farfara L.) across a range of
about 10,000 km2in Quebec, Canada. Temperature data from 88 meteorological stations
were averaged together across this range. To account for early snow cover melt in the
southern part of this range, flowering dates from individuals in southern locations were
normalized with individuals in northern locations by subtracting extra periods of snow
cover from individuals in the north. The adjusted dates indicated flowering occurred 33
days earlier over the last century while original (unadjusted) dates indicated flowering
occurred 19 days earlier over the last century.
While the study by Lavoie & Lachance (2006) adjusted actual dates for analyses, more
recent studies mostly account for climate variation using georeferenced climate data at
various scales. Calinger, Queenborough & Curtis (2013) accounted for climate variation
across Ohio by using temperature averages from ten US Climate Divisions across the
state, each about 8,000 km2. A total of 344 Climatic Divisions were established across the
contiguous United States in 1895 in order to monitor climate records more accurately.
These divisions have now accumulated about 100 years of climate records (Guttman &
Quayle, 1995). A later study by Park (2014) used average temperatures across the US county
where each specimen was collected.
Other studies accounted for climate variation across longitude, latitude, or elevation.
Robbirt et al. (2011) analyzed the geographical effect of longitude and found that flowering
occurred 4.86 days earlier per degree of longitude in a westward direction across the
southern coastal counties of England. A later study by Bertin (2015) used Hopkins’
bioclimatic law to normalize dates on specimens. Hopkins (1918) generally stated that for
every increase in a degree of latitude, or increase of 121.92 m elevation, the life history events
of plants and animals were delayed by four days. Bertin (2015) found consistencies with
Hopkins’ bioclimatic law using latitude and elevation and chose to normalize flowering
dates by adding expected phenological deviations from both latitude and elevation. Gaira,
Dhar & Belwal (2011) also analyzed climate variation using elevation when temperature
data were not available, assuming a 6.5 C change in temperature per 1,000 m change in
elevation in the Himalayan region.
Other studies used temperature averages across large regions. Li et al. (2013) used
temperature data that was averaged from 36 meteorological stations across the
Tibet Autonomous Region. Molnár et al. (2012) used temperature averages from 10
meteorological stations across Hungary and stated that the data were statistically
indistinguishable across stations (93,030 km2). Park & Schwartz (2015) averaged
temperatures from 13 stations across South Carolina, USA (82, 931 km2). A later
study by Robbirt et al. (2014) used temperature averages from an area between Bristol,
Preston, and London, across the United Kingdom (17,000 km2). Robbirt et al. (2014)
used geographical divisions called Watsonian vice-counties specifically delineated for the
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 8/16
purposes of collecting scientific data, much like the US Climate Divisions. Robbirt et al.
(2014) found temperature averages were sufficient because climate variation across the
Watsonian vice-counties used in their study did not significantly differ.
In order to more accurately estimate temperature averages across a region, recent studies
have used Geographical Information Systems (GIS) to project finer-scale climate layers
across a region and extract temperature data from specific Global Positioning System (GPS)
points. Gallagher, Hughes & Leishman (2009) referenced GPS locations for each specimen
and extracted the temperature averages at specimen GPS points from a gridded map of
temperature averages across Australia (5 km2resolution). Hereford, Schmitt & Ackerly
(2017) also extracted climate data from 176 collection locations in order to analyze species
distributions and phenology. Rawal et al. (2015) used the nearest data point from gridded
climate averages across Victoria, Australia. Edwards & Still (2008) analyzed the climate
envelopes of grasses by assigning GPS points to herbarium specimen locations in order
to extract temperature averages from gridded climate maps (250 m2resolution). Kosanic
et al. (2018) manually geo-referenced locations using herbarium specimen localities and
provided a methodology for assigning GPS coordinates when analyzing species distributions
and phenology. Standardizing methods for geo-referencing localities of herbarium records
without GPS coordinates could allow for more specimen data and larger sample sizes.
Bloom, Flower & DeChaine (2018) developed a comprehensive protocol for standardizing
spatial accuracy of geo-referenced specimen localities for species distributions.
Future studies of phenology could benefit from such geo-referencing methods because
several phenology studies only included data from specimens with GPS coordinates. Studies
using GPS data are able to account for climate variation with higher resolution, although
accuracy still depends on the underlying empirical data and modeling approach used to
generate GIS climate layers.
We recommend using the most spatially precise temperature data available, such as
climate divisions (Calinger, Queenborough & Curtis, 2013;Robbirt et al., 2014) rather than
state or region averages (Li et al., 2013;Park & Schwartz, 2015). Using GPS specimen
data to identify local climate conditions from GIS climate layers (Gallagher, Hughes &
Leishman, 2009;Edwards & Still, 2008) is also now generally more precise and convenient
in comparison to making generic and coarse-scale corrections for latitude, longitude or
elevation (Gaira, Dhar & Belwal, 2011;Robbirt et al., 2011;Bertin, 2015). If temperature
averages from larger areas are used, we recommend testing for climate variability across
smaller divisions before using averages across the larger area (Lavoie & Lachance, 2006;
Molnár et al., 2012;Robbirt et al., 2014).
Validation: herbarium specimens versus field observations
Field data are often combined with herbarium specimen data in analyses, allowing for
comparison and sometimes allowing for validation of conclusions based on herbarium
data (Primack et al., 2004;Miller-Rushing et al., 2006;Bertin, 2015). Primack et al. (2004)
used herbarium specimens for historic data and field observations for current data and
combined the two in analyses. Studies by Miller-Rushing et al. (2006) and Bertin (2015)
also compared herbarium specimen data with field observations. Miller-Rushing et al.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 9/16
(2006) found that phenology inferences from herbarium specimens alone differed from
the combined data by only about one day.
An early study by Borchert (1996) found that herbarium specimen data produced slightly
longer flowering durations than field data, but noted that durations were mostly similar
overall. Borchert (1996) and Rivera & Borchert (2001) found phenology data from field sites
largely overlapped that of herbarium specimens with only slight differences. The negligible
differences between herbarium specimen data and field data in these studies helped
justify the use of herbarium specimen data to analyze phenology in more recent studies.
Nevertheless, several more recent studies specifically compared phenology estimates from
field data to those made from herbarium specimens.
Bolmgren & Lonnberg (2005) compared herbarium specimen data directly to field data
and found the two data sets were overall highly correlated with only minor differences.
For example, herbarium specimens showed a slightly earlier mean flowering for spring-
flowering plants than field data, but the difference was not significant (Bolmgren &
Lonnberg, 2005). Later studies by Robbirt et al. (2011) and Davis et al. (2015) also primarily
focused on testing the validity of using herbarium specimen data. Robbirt et al. (2011) used
a principal axis regression analysis to compare herbarium derived peak-flowering dates
with field derived peak-flowering dates and found a high degree of correlation. Robbirt
et al. (2011) discussed how the high degree of correlation between herbarium and field
data also supports the notion that geographically different records will not significantly
alter the robustness of either data set. A study by Davis et al. (2015) used a paired t-test to
compare mean first-flowering day between herbarium specimens and field data and found
no statistical difference. Davis et al. (2015) concluded that both specimen and field data
could be combined and used as a whole.
In order to increase sample sizes, Molnár et al. (2012) added about 2,000 field
observations to about 5,000 herbarium records, resulting in 70% herbarium records
for the study. Similarly, Panchen et al. (2012) added about 2,000 field records to about
1,500 herbarium records, for a total of 43% herbarium records for the study. Searcy (2012)
combined herbarium specimen and field data and then split the combined data into
two time periods (1863–1935 and 1994–2008). Herbarium specimen data may provide
some advantages over field data. Bolmgren & Lonnberg (2005) and Primack et al. (2004)
noted that using herbarium specimens conserves time and resources, especially when
species are located in difficult to access geographical areas such as mountain peaks or
islands. Herbarium specimens are also collected over a greater period of time from a
larger geographical area while field data are often from specific localities over a shorter
time period (Primack et al., 2004;Bolmgren & Lonnberg, 2005;Bertin, 2015;Davis et al.,
2015). Herbarium specimens also provide long-term records that are widely accessible for
multiple studies. Despite criticisms, herbarium specimen data have been shown to produce
similar enough results to field data that herbarium specimen data are now widely accepted
in phenological studies.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 10/16
The use of herbarium specimens for the investigation of flowering phenology has grown
considerably during the past decade. As efforts to produce digital copies of specimens and
label information have amassed large datasets, new approaches for analyzing responses to
climate change are rapidly becoming available. Although small sample sizes have often been
used in early studies of phenology, various factors, such as wide geographic range, may
require larger sample sizes. Based on recent validations, estimations of mean-flowering
should be used rather than first-flowering because estimates of first-flowering are more
sensitive to sampling. Statistically modeling the start of the flowering process appears
to be another promising approach to investigating how climate change has affected the
beginning of a flowering cycle (Pearse et al., 2017). The use of GPS data appears to be the
way forward for the advancement of methods in the study of phenology. GPS point data
allow for correspondence with higher resolution temperature data in climatically diverse
geographical regions. Studies using herbarium specimen data will continue to help us
understand the impact of recent climate change on plant reproductive phenology. Other
aspects of plant phenology that can be analyzed using herbarium specimens, such as fruit
ripening and spring leaf emergence, have important implications for higher trophic levels,
which may include rare animals dependent on plant resources (Everill et al., 2014;Zohner
& Renner, 2014;Mendoza, Peres & Morellato, 2016). Studies using herbarium specimens
have become an asset for long-term climate change vulnerability assessment. These
studies have begun to analyze the effects of climate change on community composition
(Miller-Rushing & Primack, 2008;Park, 2014), species distribution (Hereford, Schmitt &
Ackerly, 2017;Kosanic et al., 2018), coevolved plant pollinator relationships (Molnár et al.,
2012;Robbirt et al., 2014), functional groups (Miller-Rushing & Primack, 2008;Panchen et
al., 2012;Calinger, Queenborough & Curtis, 2013;Munson & Long, 2017), and phylogenetic
relationships (Bolmgren & Lonnberg, 2005;Molnár et al., 2012;Primack & Miller-rushing,
2012). Future studies investigating phylogenetic signals and plasticity are needed to further
improve our understanding of adaptation and resilience to climate change. As temperatures
continue to rise globally, herbarium specimens will continue to be crucial resources for
analyzing phenological responses to climate change.
The authors received no funding for this work.
Competing Interests
Curtis C. Daehler is an Academic Editor for PeerJ.
Author Contributions
Casey A. Jones conceived and designed the experiments, performed the experiments,
analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or
tables, authored or reviewed drafts of the paper, approved the final draft.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 11/16
Curtis C. Daehler prepared figures and/or tables, authored or reviewed drafts of the
paper, approved the final draft.
Data Availability
The following information was supplied regarding data availability:
The research in this article did not generate any data or code (literature review).
Bertin RI. 2015. Climate change and flowering phenology in Worcester County, Mas-
sachusetts. International Journal of Plant Sciences 176:107–119 DOI 10.1086/679619.
Bloom TDS, Flower A, DeChaine EG. 2018. Why georeferencing matters: introducing a
practical protocol to prepare species occurrence records for spatial analysis. Ecology
and Evolution 8:765–777 DOI 10.1002/ece3.3516.
Bolmgren K, Lonnberg K. 2005. Herbarium data reveal an association between
fleshy fruit type and earlier flowering time. International Journal of Plant Sciences
166:663–670 DOI 10.1086/430097.
Borchert R. 1996. Phenology and flowering periodicity of neotropical dry forest
species: evidence from herbarium collections. Journal of Tropical Ecology 12:65–80
DOI 10.1017/S0266467400009317.
Calinger KM, Queenborough S, Curtis PS. 2013. Herbarium specimens reveal the
footprint of climate change on flowering trends across North-Central North
America. Ecology Letters 16:1037–1044 DOI 10.1111/ele.12135.
Carvalho S, Sartori ALB. 2015. Reproductive phenology and seed dispersal syndromes of
woody species in the Brazilian Chaco. Journal of Vegetation Science 26:302–311.
Cruz D, Mello MAR, Van Sluys M. 2006. Phenology and floral visitors of two sympatric
Heliconia species in the Brazilian Atlantic forest. Flora 201:519–527.
Daru BH, Van der Bank M, Davies TJ. 2017. Unravelling the evolutionary ori-
gins of biogeographic assemblages. Diversity and Distributions 66:1014–1016
DOI 10.1111/ddi.12679.
Davis CC, Willis CG, Connolly B, Kelly C, Ellison AM. 2015. Herbarium records
are reliable sources of phenological change driven by climate and provide novel
insights into species’ phenological cueing mechanisms. American Journal of Botany
102:1599–1609 DOI 10.3732/ajb.1500237.
Demarée GR, Rutishauser T. 2009. Origins of the word phenology [Abstract 291]. Eos
Transactions American Geophysical Union 90.
Diskin E, Proctor H, Jebb M, Sparks T, Donnelly A. 2012. The phenology of Rubus
fruticosus in Ireland: herbarium specimens provide evidence for the response of
phenophases to temperature, with implications for climate warming. International
Journal of Biometeorology 56:1103–1111 DOI 10.1007/s00484-012-0524-z.
Edwards EJ, Still CJ. 2008. Climate, phylogeny and the ecological distribution of C4
grasses. Ecology Letters 11:266–276 DOI 10.1111/j.1461-0248.2007.01144.x.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 12/16
Everill PH, Primack RB, Ellwood ER, Melaas EK. 2014. Determining past leaf-out times
of New England’s deciduous forests from herbarium specimens. American Journal of
Botany 101:1293–1300 DOI 10.3732/ajb.1400045.
Fenner M. 1998. The phenology of growth and reproduction in plants. Perspectives in
Plant Ecology, Evolution and Systematics 1:78–91 DOI 10.1078/1433-8319-00053.
Fisher NI. 1993. Statistical analysis of circular data. Cambridge: Cambridge University
Fitter AH, Fitter RSR, Harris ITB, Williamson MH. 1995. Relationships between
first-flowering date and temperature in the flora of a locality in central England.
Functional Ecology 9:55–60 DOI 10.2307/2390090.
Forrest J, Miller-rushing AJ. 2010. Toward a synthetic understanding of the role of
phenology in ecology and evolution. Philosophical Transactions of the Royal Society
B: Biological Sciences 365:3101–3112 DOI 10.1098/rstb.2010.0145.
Gaira KS, Dhar U, Belwal OK. 2011. Potential of herbarium records to sequence
phenological pattern: a case study of Aconitum heterophyllum in the Himalaya.
Biodiversity and Conservation 20:2201–2210 DOI 10.1007/s10531-011-0082-4.
Gaira KS, Rawal RS, Rawat B, Bhatt D. 2014. Impact of climate change on the flow-
ering of Rhododendron arboreum in central Himalaya, India. Current Science
Gallagher RV, Hughes L, Leishman MR. 2009. Phenological trends among Australian
alpine species: using herbarium records to identify climate-change indicators.
Australian Journal of Botany 57:1–9 DOI 10.1071/BT08051.
Guttman NB, Quayle RG. 1995. A historical perspective of US. Climate divisions.
Bulletin of the American Meteorological Society 77:293–303.
Haggerty BP, Hove AA, Mazer SJ. 2012. Primer on herbarium-based phenological
research. Guide for college and public audiences for understanding the use of
preserved plants in climate change research. California Phenology Project.
Hart R, Salick J, Ranjitkar S, Xu J. 2014. Herbarium specimens show contrast-
ing phenological responses to Himalayan climate. Proceedings of the Na-
tional Academy of Sciences of the United States of America 111:10615–10619
DOI 10.1073/pnas.1403376111.
Hereford J, Schmitt J, Ackerly DD. 2017. The seasonal climate niche predicts phenology
and distribution of an ephemeral annual plant, Mollugo verticillata.Journal of Ecology
105:1323–1334 DOI 10.1111/1365-2745.12739.
Hopkins AD. 1918. Periodical events and natural law as guides to agricultural research
and practice US Department Of Agriculture Weather Bureau. Monthly Weather
Review 643:5–42.
Kebede M, Isotalo J. 2016. Flowering and fruiting phenology and floral visitation of four
native tree species in the remnant moist Afromontane forest of Wondo Genet, South
Central Ethiopia. Tropical Ecology 57:299–311.
Kosanic A, Anderson K, Harrison S, Turkington T, Bennie J. 2018. Changes
in the geographical distribution of plant species and climatic variables on
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 13/16
the West Cornwall peninsula (South West UK). PLOS ONE 13:e0191021
DOI 10.1371/journal.pone.0191021.
Lamoureux CH. 1973. Phenology and reproduction of Hawaiian plants, a preliminary
report. Island ecosystems. US International Biological Program, Technical report
number 24.
Lavoie C, Lachance D. 2006. A new herbarium-based method for reconstructing the phe-
nology of plant species across large areas. American Journal of Botany 93:512–516
DOI 10.3732/ajb.93.4.512.
Li Z, Wu N, Gao X, Wu Y, Oli KP. 2013. Species-level phenological responses to ‘‘global
warming’’ as evidenced by herbarium collections in the Tibetan Autonomous
Region. Biodiversity and Conservation 22:141–152 DOI 10.1007/s10531-012-0408-x.
Linnaeus C. 1751. Philosophia botanica. Stockholm: Apud Godofr. Kiesewetter.
Mendoza I, Peres CA, Morellato LPC. 2016. Continental-scale patterns and climatic
drivers of fruiting phenology: a quantitative neotropical review. Global and Planetary
Change 148:227–241 DOI 10.1016/j.gloplacha.2016.12.001.
Miller-Rushing AJ, Primack RB. 2008. Global warming and flowering times in Thoreau’s
Concord: a community perspective. Ecology 89:332–341 DOI 10.1890/07-0068.1.
Miller-Rushing AJ, Primack RB, Primack D, Mukunda S. 2006. Photographs and
herbarium specimens as tools to document phenological changes in response to
global warming. American Journal of Botany 93:1667–1674
DOI 10.3732/ajb.93.11.1667.
Molnár VA, Tökölyi J, Végvári Z, Sramkó G, Sulyok J, Barta Z. 2012. Pollina-
tion mode predicts phenological response to climate change in terrestrial or-
chids: a case study from Central Europe. Journal of Ecology 100:1141–1152
DOI 10.1111/j.1365-2745.2012.02003.x.
Morellato LPC, Alberti LF, Hudson IL. 2010. Applications of circular statistics in plant
phenology; a case studies approach. In: Hudson IL, Keatley MR, eds. Phenological
research; Methods for environmental and climate change analysis. Dordrecht: Springer,
Morellato LPC, Talora DC, Takahasi A, Bencke CC, Eliane C, Zipparro VB. 2000.
Phenology of Atlantic rain forest trees: a comparative study. Biotropica 32:811–823.
Morren C. 1853. Souvenirs phénologiques de l’hiver 1852–1853. Bull de L’Académie
Royale Des. Sciences, des Lettres et des Beaux-Artes de belgique 20:160–186.
Moussus J, Julliard R, Jiguet F. 2010. Featuring 10 phenological estimators using
simulated data. Methods in Ecology and Evolution 1:140–150
DOI 10.1111/j.2041-210X.2010.00020.x.
Munson SM, Long AL. 2017. Climate drives shifts in grass reproductive phenology across
the western USA. New Phytologist 213:1945–1955 DOI 10.1111/nph.14327.
Nadia TL, Morellato LPC, Machado IC. 2012. Reproductive phenology of a Northeast
Brazilian mangrove community: environmental and biotic constraints. Flora
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 14/16
Nazareno AG, Dos Reis MS. 2012. Linking phenology to mating system: exploring the
reproductive biology of the threatened palm species Butia eriospatha.Journal of
Heredity 103:842–852.
Neil KL, Landrum L, Wu J. 2010. Effects of urbanization on flowering phenology in the
metropolitan phoenix region of USA: findings from herbarium records. Journal of
Arid Environments 74:440–444 DOI 10.1016/j.jaridenv.2009.10.010.
Novotny V, Basset Y. 1998. Seasonality of sap-sucking insects (Auchenorrhyncha,
Hemiptera) feeding on Ficus (Moraceae) in a lowland rain forest in New Guinea.
Oecologia 115:514–522.
Panchen ZA, Primack RB, Aniśko T, Lyons RE. 2012. Herbarium specimens, pho-
tographs, and field observations show Philadelphia area plants are responding to
climate change. American Journal of Botany 99:751–756 DOI 10.3732/ajb.1100198.
Park IW. 2012. Digital herbarium archives as a spatially extensive, taxonomically
discriminate phenological record; a comparison to MODIS satellite imagery. Inter-
national Journal of Biometeorology 56:1179–1182 DOI 10.1007/s00484-012-0521-2.
Park IW. 2014. Impacts of differing community composition on flowering phenology
throughout warm temperate, cool temperate and xeric environments. Global Ecology
and Biogeography 23:789–801 DOI 10.1111/geb.12163.
Park IW, Schwartz MD. 2015. Long-term herbarium records reveal temperature-
dependent changes in flowering phenology in the Southeastern USA. International
Journal of Biometeorology 59:347–355.
Parmesan C, Yohe G. 2003. A globally coherent fingerprint of climate change impacts
across natural systems. Nature 421:37–42 DOI 10.1038/nature01286.
Pearse WD, Davis CC, Inouye DW, Primack RB, Davies TJ. 2017. A statistical estimator
for determining the limits of contemporary and historic phenology. Nature Ecology
& Evolution 1:1876–1882 DOI 10.1038/s41559-017-0350-0.
Primack D, Imbres C, Primack RB, Miller-Rushing AJ, Del Tredici P. 2004. Herbarium
specimens demonstrate earlier flowering times in response to warming in Boston.
American Journal of Botany 91:1260–1264 DOI 10.3732/ajb.91.8.1260.
Primack RB, Miller-Rushing AJ. 2009. The role of botanical gardens in climate change
research. New Phytologist 182:303–313 DOI 10.1111/j.1469-8137.2009.02800.x.
Primack RB, Miller-rushing AJ. 2012. Uncovering, collecting, and analyzing records
to investigate the ecological impacts of climate change: a template from Thoreau’s
Concord. BioScience 62:170–181 DOI 10.1525/bio.2012.62.2.10.
Puppi G. 2007. Origin and development of phenology as a science. Italian Journal of
Agrometeorology 12:24–29.
Rawal DS, Kasel S, Keatley MR, Nitschke CR. 2015. Herbarium records identify
sensitivity of flowering phenology of eucalypts to climate: implications for species
response to climate change. Austral Ecology 40:117–125 DOI 10.1111/aec.12183.
Rivera G, Borchert R. 2001. Induction of flowering in tropical trees by a 30-min reduc-
tion in photoperiod: evidence from field observations and herbarium specimens.
Tree Physiology 21:201–212 DOI 10.1093/treephys/21.4.201.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 15/16
Robbirt KM, Davy AJ, Hutchings MJ, Roberts DL. 2011. Validation of biological
collections as a source of phenological data for use in climate change studies: a
case study with the orchid Ophrys sphegodes. Journal of Ecology 99:235–241
DOI 10.1111/j.1365-2745.2010.01727.x.
Robbirt KM, Roberts DL, Hutchings MJ, Davy AJ, Gardens RB, Tw S. 2014. Report
potential disruption of pollination in a sexually deceptive orchid by climatic change.
Current Biology 24:2845–2849 DOI 10.1016/j.cub.2014.10.033.
Rogerio FR, Araujo AC. 2010. Flowering phenology and pollination of ornithophilous
species in two habitats of Serra da Bodoquena, Mato Grosso do Sul, Brazil. Annals of
the Brazilian Academy of Sciences 82:843–855.
Searcy KB. 2012. Changes in the flora of the Mount Holyoke Range, Hampshire
Co., Massachusetts over the past 150 years (1860–2010). Rhodora 114:113–132
DOI 10.3119/11-06.
Staggemeier VG, Diniz-Filho JAF, Morellato LPC. 2010. The shared influence of
phylogeny and ecology on the reproductive patterns of Myrteae (Myrtaceae). Journal
of Ecology 98:1409–1421.
Tesfaye G, Teketay D, Fetene M, Beck E. 2011. Phenology of seven indigenous tree
species in a dry Afromontane forest, Southern Ethiopia. Tropical Ecology 52:229–241.
Van Proosdij ASJ, Sosef MSM, Wieringa JJ, Raes N. 2016. Minimum required number
of specimen records to develop accurate species distribution models. Ecography
39:542–552 DOI 10.1111/ecog.01509.
Van Schaik CP, Terborgh JW, Wright SJ. 1993. The phenology of tropical forests—
adaptive significance and consequences for primary consumers. Annual Review of
Ecology and Systematics 24:353–377 DOI 10.1146/
Von Linné C. 2003. Linnäus’ philosophica botanica. New York: Oxford University Press
(Translation by S Freer).
Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee TJC, Fromentin
JM. 2002. Ecological responses to recent climate change. Nature 416:389–395
DOI 10.1038/416389a.
Willis CG, Ellwood ER, Primack RB, Davis CC, Pearson KD, Gallinato AS, Yost JM,
Nelson G, Mazer SJ, Rossington NL, Sparks TH, Soltis PS. 2017. Old plants, new
tricks: phenological research using herbarium specimens. Trends in Ecology &
Evolution 32:531–546 DOI 10.1016/j.tree.2017.03.015.
Zalamea AP, Munoz F, Stevenson PR, Paine CET, Sarmiento C, Sabatier D, Heuret
P. 2016. Continental-scale patterns of Cecropia reproductive phenology: evidence
from herbarium specimens. Proceedings of the Royal Society B: Biological Sciences
278:2437–2445 DOI 10.1098/rspb.2010.2259 .
Zhao MF, Peng CH, Xiang WH, Deng XW, Tian DL, Zhou XL, Yu GR. 2013. Plant phe-
nological modeling and its application in global climate change research: overview
and future challenges. Environmental Reviews 21:1–14 DOI 10.1139/er-2012-0036.
Zohner CM, Renner SS. 2014. Common garden comparison of the leaf-out phenology
of woody species from different native climates, combined with herbarium records,
forecasts long-term change. Ecology Letters 17:1016–1025 DOI 10.1111/ele.12308.
Jones and Daehler (2018), PeerJ, DOI 10.7717/peerj.4576 16/16
... To test whether climate change is associated with changes to the proportion of CL flowers, we ran two models: (1) a temporal model testing whether the proportion of CL flowers has changed across the past century, and (2) a climate model testing whether the proportion of CL flowers is associated with precipitation and temperature. For studies that leverage long-term data to ascertain the effect of climate change on phenology via correlative approaches, it is standard practice to analyze the effect of time separately from the effect of climate, given autocorrelation between year and climate in the era of anthropogenic climate change (Jones and Daehler, 2018). We ran each of these models as a generalized linear model (GLM) with the proportion of CL flowers as the response variable and a quasibinomial family to accommodate a proportion as the response. ...
... We also find that V. sororia plants have advanced the timing of both CL and CH flowers across this time period (Figure 5A, B). To (Visser and Both, 2005;Byers, 2017;Jones and Daehler, 2018). Our study builds upon these well-documented changes to when reproduction occurs (Visser and Both, 2005;Byers, 2017;Jones and Daehler, 2018) by demonstrating that climate change can also potentially affect how reproduction occurs. ...
... To (Visser and Both, 2005;Byers, 2017;Jones and Daehler, 2018). Our study builds upon these well-documented changes to when reproduction occurs (Visser and Both, 2005;Byers, 2017;Jones and Daehler, 2018) by demonstrating that climate change can also potentially affect how reproduction occurs. Prior studies have suggested that species with mixed mating systems may be sensitive to anthropogenic global change (e.g., Jones et al., 2013) which, if true, is of particular consequence given that mixed mating occurs in approximately 42% of floral species worldwide (Goodwillie et al., 2005). ...
Full-text available
Premise: Although the balance between cross‐and self‐fertilization is driven by the environment, no long‐term study has documented whether anthropogenic climate change is affecting reproductive strategy allocation in species with mixed mating systems. Here, we test whether the common blue violet (Viola sororia; Violaceae) has altered relative allocation to the production of potentially outcrossing flowers as the climate has changed throughout the 20th century. Methods: Using herbarium records spanning from 1875 to 2015 from the central United States, we quantified production of obligately selfing cleistogamous (CL) flowers and potentially outcrossing chasmogamous (CH) flowers by V. sororia, coupled these records with historic temperature and precipitation data, and tested whether changes to the proportion of CL flowers correlate with temporal climate trends. Results: We find that V. sororia progressively produced lower proportions of CL flowers across the past century and in environments with lower mean annual temperature and higher total annual precipitation. We also find that both CL and CH flower phenology has advanced across this time period. Conclusions: Our results suggest that V. sororia has responded to lower temperatures and greater water availability by shifting reproductive strategy allocation away from selfing and toward potential outcrossing. This provides the first long‐term study of how climate change may affect relative allocation to potential outcrossing in species with mixed mating systems. By revealing that CL flowering is associated with low water availability and high temperature, our results suggest the production of obligately selfing flowers is favored in water limited environments.
... The digitization process itself may introduce further error or biases for phenology or other data derived from digitized herbarium specimens (Groom et al. 2019). Due to these limitations, supplementing phenology data derived from herbarium specimens with other types of phenology data is proving to be potentially valuable approach (Davis et al. 2015;Spellman and Mulder 2016;Jones and Daehler 2018). ...
... Specimens with only the flowering phenophase were retained and used in our analyses due to the sparsity of other phenological stages represented in the herbarium material. As almost all flowering specimens had most of the flowers open, we assumed flowering phenophase to mainly represent peak flowering (Jones and Daehler 2018). A recent study has shown that scoring flowering time as two or more flowering phases does not provide much additional insight in phenology studies using herbarium specimens (Ellwood et al. 2019). ...
... Average monthly temperatures for March, April, and May of the year of flowering were used in the analysis. Combining multiple months into a single climatic variable (i.e., average spring temperature) has previously been found to produce significant relationships between phenological events and climate and is thus commonly used based on the assumption that flowering DOY is affected by temperature in the months preceding flowering (Sparks et al., 2000;Menzel 2003;Primack et al. 2004;Jones and Daehler 2018). Using average temperatures also give results that are comparable to more complex degree day models (Balser, 2016) but has the added advantage of being more easily understood, which is an advantage in presenting the results to the public. ...
Full-text available
Phenology has emerged as a key metric to measure how species respond to changes in climate. Innovative means have been developed to extend the temporal and spatial range of phenological data by obtaining data from herbarium specimens, citizen science programs, and biodiversity data repositories. These different data types have seldom been compared for their effectiveness in detecting environmental impacts on phenology. To address this, we compare three separate phenology datasets from Denmark: (i) herbarium specimen data spanning 145 years, (ii) data collected from a citizen science phenology program over a single year observing first flowering, and (iii) data derived from incidental biodiversity observations in iNaturalist over a single year. Each dataset includes flowering day of year observed for three common spring-flowering plant species: Allium ursinum (ramsons), Aesculus hippocastanum (horse chestnut), and Sambucus nigra (black elderberry). The incidental iNaturalist dataset provided the most extensive geographic coverage across Denmark and the largest sample size and recorded peak flowering in a way comparable to herbarium specimens. The directed citizen science dataset recorded much earlier flowering dates because the program objective was to report the first flowering, and so was less compared to the other two datasets. Herbarium data demonstrated the strongest effect of spring temperature on flowering in Denmark, possibly because it was the only dataset measuring temporal variation in phenology, while the other datasets measured spatial variation. Herbarium data predicted the mean flowering day of year recorded in our iNaturalist dataset for all three species. Combining herbarium data with iNaturalist data provides an even more effective method for detecting climatic effects on phenology. Phenology observations from directed and incidental citizen science initiatives will increase in value for climate change research in the coming years with the addition of data capturing the inter-annual variation in phenology.
... With many herbaria dating back to some 200 yr, and hundreds of millions of specimens worldwide, herbaria are a tremendous treasure for studying phenology changes both long term and large scale. Previous studies have indeed found strong patterns of long-term phenology changes in herbarium data (Primack et al., 2004;Miller-Rushing et al., 2006;Davis et al., 2015;Willis et al., 2017;Lang et al., 2019;Park et al., 2019;reviewed by Jones & Daehler, 2018), and they have also demonstrated that phenology trends estimated from herbarium data are comparable to those from field observations (Davis et al., 2015;Jones & Daehler, 2018;Miller et al., 2021). However, almost all previous studies have been carried out in the USA, and there has been little work so far on herbaria and plant phenology in Europe (but please refer to Robbirt et al., 2011;Diskin et al., 2012;Molnar et al., 2012). ...
Today plants often flower earlier due to climate warming. Herbarium specimens are excellent witnesses of such long‐term changes. However, the magnitude of phenological shifts may vary geographically, and the data are often clustered. Therefore, large‐scale analyses of herbarium data are prone to pseudoreplication and geographical biases. We studied over 6000 herbarium specimens of 20 spring‐flowering forest understory herbs from Europe to understand how their phenology had changed during the last century. We estimated phenology trends with or without taking spatial autocorrelation into account. On average plants now flowered over 6 d earlier than at the beginning of the last century. These changes were strongly associated with warmer spring temperatures. Flowering time advanced 3.6 d per 1°C warming. Spatial modelling showed that, in some parts of Europe, plants flowered earlier or later than expected. Without accounting for this, the estimates of phenological shifts were biased and model fits were poor. Our study indicates that forest wildflowers in Europe strongly advanced their phenology in response to climate change. However, these phenological shifts differ geographically. This shows that it is crucial to combine the analysis of herbarium data with spatial modelling when testing for long‐term phenology trends across large spatial scales.
... Researchers recently started to use the vast collections of plant specimens available from the world's herbaria for phenological analyses (Willis et al., 2017;Pearson et al., 2020). Pressed, preserved plant specimens collected over centuries and around the globe can serve as snapshots of plant phenology contributing to a wealth of research relating to the timing of phenological events (Jones and Daehler, 2018). Although most herbarium specimens were not collected to detect phenological stages, phenological data from specimens have proven reliable and irreplaceable for understanding plant phenology (Davis et al., 2015). ...
Full-text available
Climate change represents one of the most critical threats to biodiversity with far-reaching consequences for species interactions, the functioning of ecosystems, or the assembly of biotic communities. Plant phenology research has gained increasing attention as the timing of periodic events in plants is strongly affected by seasonal and interannual climate variation. Recent technological development allowed us to gather invaluable data at a variety of spatial and ecological scales. The feasibility of phenological monitoring today and in the future depends heavily on developing tools capable of efficiently analyzing these enormous amounts of data. Deep Neural Networks learn representations from data with impressive accuracy and lead to significant breakthroughs in, e.g., image processing. This article is the first systematic literature review aiming to thoroughly analyze all primary studies on deep learning approaches in plant phenology research. In a multi-stage process, we selected 24 peer-reviewed studies published in the last five years (2016–2021). After carefully analyzing these studies, we describe the applied methods categorized according to the studied phenological stages, vegetation type, spatial scale, data acquisition- and deep learning methods. Furthermore, we identify and discuss research trends and highlight promising future directions. We present a systematic overview of previously applied methods on different tasks that can guide this emerging complex research field.
... Despite the scarce metadata, historical specimens are of great importance in studies of different biological disciplines and processes including taxonomy (e.g., typification; Wagenitz, 2001;Vorontsova & Knapp, 2010;Borges & al., 2018), biogeography and climate change biology (Case & al., 2007;González-Orozco & al., 2016), biodiversity (Nelson & This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 2018), phenology (Jones & Daehler, 2018;Yost & al., 2019;Pearson & al., 2020), nature conservation (Nualart & al., 2017), invasiveness of species (Crawford & Hoagland, 2009), population genomics (e.g., loss of genetic diversity; Cozzolino & al., 2007), as well as historical sciences, e.g., in studies about botanical collecting practices, biographies of certain botanists and the practice of specimen exchange and acquisition in the past (Groom & al., 2014;Ghorbani & al., 2018). Especially the steady advances in high-throughput sequencing and ancient-DNA extraction protocols increase the applicability of historical specimens in DNA-based studies (Bieker & Martin, 2018). ...
Three generations of the Wendland family, Johann Christoph Wendland (1755–1828), Heinrich Ludolph Wendland (1792–1869) and Hermann Wendland (1825–1903), were outstanding court gardeners at the Royal Gardens of Herrenhausen (Hanover, Germany). The “Wendlands” not only enriched the botanical diversity of the gardens through plants obtained in exchange, purchased or collected during their own journeys, but they also achieved a good scientific reputation by publications on diverse botanical themes including the description of new taxa. Moreover, there exists a herbarium of approximately 16,600 plant specimens related to these gardens, the Wendlands and their publications. Numerous papers dealing with the Wendlands and the Royal Gardens of Herrenhausen have been published. The Herrenhausen Herbarium, however, also known as the Wendland Herbarium, which was donated to the University of Göttingen in 1969, was so far neither digitized nor revised. Here, we describe the history and the current digitization of this herbarium, which includes specimens collected between 1780 and 1857. It consists of three major parts: the Herrenhausen Herbarium arranged according to the Linnaean System (13,035 specimens), the palm collection (1069 specimens) and smaller collections (specimens collected by Hermann Wendland in Central America, collections of Carl Hoffmann from Costa Rica and a part of Jakob Friedrich Ehrhart's herbarium, who was a student of Linnaeus; approximately 2500 specimens). The systematic digitization of this historical herbarium as a whole revealed a seemingly unconventional arrangement of the specimens, which we explain here for the correct interpretation of collection data. Furthermore, and despite the meagre information on the herbarium labels, we were able to identify 260 specimens from a trip of Heinrich Ludolph Wendland to Switzerland in 1820. By comparing the specimen labels with entries in his diary and travel report, we were able to retrace detailed information on localities and dates, providing historical biodiversity information. Also, the historical identification of these specimens was revised. A list of collectors represented in the Herrenhausen Herbarium is provided, which includes famous names such as Linnaeus and his students Afzelius, Bergius, Ehrhart, Schreber, and Thunberg.
... A synthesis of methods and applications of these has been published recently. Jones and Daehler (2018) highlight the increasing importance of herbaria following their digitization. Spellman and Mulder (2016) used phenological observations collected in a participatory science program to evaluate phenological models developed on the basis of herbarium samples. ...
... To avoid the potential confounding effect of supplemental watering of specimens on Museum grounds, we used phenological observations only from the unmanaged King Canyon site. Although herbarium specimens are increasingly used in phenological research (Willis et al. 2017;Jones and Daehler 2018;Pearson et al. 2021), these observations are only associated with a single date, and are highly left censored (meaning that it is only known that flowering occurred some time before the observation date), ultimately containing much less information than direct and regular observations of phenological transitions, such as the weekly observations made by Desert Museum staff. Because of this issue, in combination with a low number of flowering herbarium specimens from the Desert Museum grounds, we limit our phenological observations to the direct and regular observations of live, naturally occurring plants by Desert Museum staff. ...
Climate warming changes the phenology of many species. When interacting organisms respond differently, climate change may disrupt their interactions and affect the stability of ecosystems. Here, we used global biodiversity facility occurrence records to examine phenology trends in plants and their associated insect pollinators in Germany since the 1980s. We found strong phenological advances in plants but differences in the extent of shifts among pollinator groups. The temporal trends in plant and insect phenologies were generally associated with interannual temperature variation and thus probably driven by climate change. When examining the synchrony of species-level plant–pollinator interactions, their temporal trends differed among pollinator groups. Overall, plant–pollinator interactions become more synchronized, mainly because the phenology of plants, which historically lagged behind that of the pollinators, responded more strongly to climate change. However, if the observed trends continue, many interactions may become more asynchronous again in the future. Our study suggests that climate change affects the phenologies of both plants and insects and that it also influences the synchrony of plant–pollinator interactions.
Climate change is widely known to affect plant phenology, but little is known about how these impacts manifest in the widespread sagebrush ecosystem of the Western US which supports a number of wildlife species of concern. Shifts in plant phenology can trigger consequences for the plants themselves as well as the communities of consumers that depend upon them. We assembled historical observations of first flowering dates for 51 species collected in the 1970 and 80s in a montane sagebrush community in the Greater Yellowstone Ecosystem and compared these to contemporary phenological observations targeting the same species and locations (2016‐2019). We also assembled regional climate data (average spring temperature, day of spring snowmelt, and growing degree days) and tested the relationship between first flowering time and these variables for each species. We observed the largest change in phenology in early spring flowers, which as a group bloomed on average 17 days earlier, and as much as 36 days earlier, in the contemporary data set. Mid‐summer flowers bloomed on average 10 days earlier, nonnative species 15 days earlier, and berry‐producing shrubs 5 days earlier, while late summer flowering plants did not shift. The greatest correlates of early spring and mid‐summer flowering were average spring temperature and day of snowmelt, which was 21 days earlier, on average, in 2016‐2019 relative to the 1973‐1978. The shifts in flowering phenology that we observed could indicate developing asynchronies or novel synchronies of these plant resources and wildlife species of conservation concern including: Greater sage‐grouse, whose nesting success is tied to availability of spring forbs; grizzly bears ‐‐ which rely heavily on berries for their fall diet; and pollinators. This underscores the importance of maintaining a diverse portfolio of native plants in terms of species composition, genetics, phenological responsiveness to climatic cues, and ecological importance to key wildlife and pollinator species. Redundancy within ecological niches may also be important considering that species roles in the community may shift as climate change affects them differently. These considerations are particularly relevant to restoration and habitat‐enhancement projects in sagebrush communities across western North America.
Herbarium specimens provide a critical source of phenological data that can be used to identify the direct and indirect drivers of variation in flowering date within and among species. Specimen-based phenological research in California has been accelerated by digitization efforts such as the California Phenology Network, which has scored and archived the phenological status of over 1.4 million specimens to date. Using this new data source in the Consortium of California Herbaria's CCH2 data portal, we obtained data from 993 specimens of the iconic California Poppy, Eschscholzia californica Cham., along with climate data representing all collection sites. Our goal was to determine how long-term and interannual climate variation affect flowering dates, and whether the magnitude of phenological sensitivity to climate varies across the species' range. We found that specimens collected from chronically warm or dry sites flowered relatively early, and flowering date was more sensitive to long-term mean temperature than to long-term mean precipitation. Independent of these effects of long-term conditions, flowering date in E. californica was sensitive to interannual variation in seasonal precipitation, but the direction of this effect depended on the season in which the precipitation occurred. Specimens sampled from sites experiencing warmer-than-average springs in the year of collection flowered 2.73.3 days earlier for every 1C increase in spring temperature relative to long-term mean spring temperature. The magnitude of these effects, however, varied across the range of E. californica, with greater sensitivity to temperature in relatively cooler regions and no discernible sensitivity in relatively warm regions. Consistently, California Poppies exhibited significant phenological advancement over the last 120 years, but this advancement was restricted to the cooler portions of its range. Our results provide one of the first accounts of intraspecific variation in both phenological sensitivity to climate and the magnitude of phenological shifts over time, and demonstrate that, for a single species, location- or population-specific estimates of phenological sensitivity or of temporal trends in phenology might not accurately predict phenological responses to climate change in other locations throughout its range. In this study, we highlight the utility and promise of herbarium specimens for addressing novel questions about the phenological responses of plants to climate trends.
Full-text available
Recent climate change has had a major impact on biodiversity and has altered the geographical distribution of vascular plant species. This trend is visible globally; however, more local and regional scale research is needed to improve understanding of the patterns of change and to develop appropriate conservation strategies that can minimise cultural, health, and economic losses at finer scales. Here we describe a method to manually geo-reference botanical records from a historical herbarium to track changes in the geographical distributions of plant species in West Cornwall (South West England) using both historical (pre-1900) and contemporary (post-1900) distribution records. We also assess the use of Ellenberg and climate indicator values as markers of responses to climate and environmental change. Using these techniques we detect a loss in 19 plant species, with 6 species losing more than 50% of their previous range. Statistical analysis showed that Ellenberg (light, moisture, nitrogen) and climate indicator values (mean January temperature, mean July temperature and mean precipitation) could be used as environmental change indicators. Significantly higher percentages of area lost were detected in species with lower January temperatures, July temperatures, light, and nitrogen values, as well as higher annual precipitation and moisture values. This study highlights the importance of historical records in examining the changes in plant species’ geographical distributions. We present a method for manual geo-referencing of such records, and demonstrate how using Ellenberg and climate indicator values as environmental and climate change indicators can contribute towards directing appropriate conservation strategies.
Full-text available
Species Distribution Models (SDMs) are widely used to understand environmental controls on species’ ranges and to forecast species range shifts in response to climatic changes. The quality of input data is crucial determinant of the model's accuracy. While museum records can be useful sources of presence data for many species, they do not always include accurate geographic coordinates. Therefore, actual locations must be verified through the process of georeferencing. We present a practical, standardized manual georeferencing method (the Spatial Analysis Georeferencing Accuracy (SAGA) protocol) to classify the spatial resolution of museum records specifically for building improved SDMs. We used the high-elevation plant Saxifraga austromontana Wiegand (Saxifragaceae) as a case study to test the effect of using this protocol when developing an SDM. In MAXENT, we generated and compared SDMs using a comprehensive occurrence dataset that had undergone three different levels of georeferencing: (1) trained using all publicly available herbarium records of the species, minus outliers (2) trained using herbarium records claimed to be previously georeferenced, and (3) trained using herbarium records that we have manually georeferenced to a ≤ 1-km resolution using the SAGA protocol. Model predictions of suitable habitat for S. austromontana differed greatly depending on georeferencing level. The SDMs fitted with presence locations georeferenced using SAGA outperformed all others. Differences among models were exacerbated for future distribution predictions. Under rapid climate change, accurately forecasting the response of species becomes increasingly important. Failure to georeference location data and cull inaccurate samples leads to erroneous model output, limiting the utility of spatial analyses. We present a simple, standardized georeferencing method to be adopted by curators, ecologists, and modelers to improve the geographic accuracy of museum records and SDM predictions.
Full-text available
Floristic and faunal diversity fall within species assemblages that can be grouped into distinct biomes or ecoregions. Understanding the origins of such biogeographic assemblages helps illuminate the processes shaping present-day diversity patterns and identifies regions with unique or distinct histories. While the fossil record is often sparse, dated phylogenies can provide a window into the evolutionary past of these regions. Here, we present a novel phylogenetic approach to investigate the evolutionary origins of present-day biogeographic assemblages and highlight their conservation value.
Full-text available
Climate change affects not just where species are found, but also when species' key life-history events occur-their phenology. Measuring such changes in timing is often hampered by a reliance on biased survey data: surveys identify that an event has taken place (for example, the flower is in bloom), but not when that event happened (for example, the flower bloomed yesterday). Here, we show that this problem can be circumvented using statistical estimators, which can provide accurate and unbiased estimates from sparsely sampled observations. We demonstrate that such methods can resolve an ongoing debate about the relative timings of the onset and cessation of flowering, and allow us to place modern observations reliably within the context of the vast wealth of historical data that reside in herbaria, museum collections, and written records. We also analyse large-scale citizen science data from the United States National Phenology Network and reveal not just earlier but also potentially more variable flowering in recent years. Evidence for greater variability through time is important because increases in variation are characteristic of systems approaching a state change.
Full-text available
1.Many short-lived species complete their life cycles during brief seasonal windows of favorable environmental conditions. Such species may persist in the face of climate warming by migration to track their seasonal climate niche in space and/or by phenological shifts to track favorable conditions in time within the year. To describe the seasonal climate niche of the short-lived annual Mollugo verticillata in California, we used data from herbarium specimens and historic climate records to estimate environmental conditions at the location, month and year of each collection.
Full-text available
Changes in the life cycle of organisms (i.e. phenology) are one of the most widely used early-warning indicators of climate change, yet this remains poorly understood throughout the tropics. We exhaustively reviewed any published and unpublished study on fruiting phenology carried out at the community level in the American tropics and subtropics (latitudinal range: 26°N–26°S) to (1) provide a comprehensive overview of the current status of fruiting phenology research throughout the Neotropics; (2) unravel the climatic factors that have been widely reported as drivers of fruiting phenology; and (3) provide a preliminary assessment of the potential phenological responses of plants under future climatic scenarios. Despite the large number of phenological datasets uncovered (218), our review shows that their geographic distribution is very uneven and insufficient for the large surface of the Neotropics (~ 1 dataset per ~ 78,000 km²). Phenological research is concentrated in few areas with many studies (state of São Paulo, Brazil, and Costa Rica), whereas vast regions elsewhere entirely unstudied. Sampling effort in fruiting phenology studies was generally low: the majority of datasets targeted fewer than 100 plant species (71%), lasted 2 years or less (72%), and only 10.4% monitored > 15 individuals per species. We uncovered only 10 sites with ten or more years of phenological monitoring. The ratio of numbers of species sampled to overall estimates of plant species richness was wholly insufficient for highly diverse vegetation types such as tropical rainforests, seasonal forest and cerrado, and only slightly more robust for less diverse vegetation types, such as deserts, arid shrublands and open grassy savannas. Most plausible drivers of phenology extracted from these datasets were environmental (78.5%), whereas biotic drivers were rare (6%). Among climatic factors, rainfall was explicitly included in 73.4% of cases, followed by air temperature (19.3%). Other environmental cues such as water level (6%), solar radiation or photoperiod (3.2%), and ENSO events (1.4%) were rarely addressed. In addition, drivers were analyzed statistically in only 38% of datasets and techniques were basically correlative, with only 4.8% of studies including any consideration of the inherently autocorrelated character of phenological time series. Fruiting peaks were significantly more often reported during the rainy season both in rainforests and cerrado woodlands, which is at odds with the relatively aseasonal character of the former vegetation type. Given that climatic models predict harsh future conditions for the tropics, we urgently need to determine the magnitude of changes in plant reproductive phenology and distinguish those from cyclical oscillations. Long-term monitoring and herbarium data are therefore key for detecting these trends. Our review shows that the unevenness in geographic distribution of studies, and diversity of sampling methods, vegetation types, and research motivation hinder the emergence of clear general phenological patterns and drivers for the Neotropics. We therefore call for prioritizing research in unexplored areas, and improving the quantitative component and statistical design of reproductive phenology studies to enhance our predictions of climate change impacts on tropical plants and animals.
Full-text available
The capacity of grass species to alter their reproductive timing across space and through time can indicate their ability to cope with environmental variability and help predict their future performance under climate change. We determined the long-term (1895-2013) relationship between flowering times of grass species and climate in space and time using herbarium records across ecoregions of the western USA. There was widespread concordance of C3 grasses accelerating flowering time and general delays for C4 grasses with increasing mean annual temperature, with the largest changes for annuals and individuals occurring in more northerly, wetter ecoregions. Flowering time was delayed for most grass species with increasing mean annual precipitation across space, while phenology-precipitation relationships through time were more mixed. Our results suggest that the phenology of most grass species has the capacity to respond to increases in temperature and altered precipitation expected with climate change, but weak relationships for some species in time suggest that climate tracking via migration or adaptation may be required. Divergence in phenological responses among grass functional types, species, and ecoregions suggests that climate change will have unequal effects across the western USA.
The timing of phenological events, such as leaf-out and flowering, strongly influence plant success and their study is vital to understanding how plants will respond to climate change. Phenological research, however, is often limited by the temporal, geographic, or phylogenetic scope of available data. Hundreds of millions of plant specimens in herbaria worldwide offer a potential solution to this problem, especially as digitization efforts drastically improve access to collections. Herbarium specimens represent snapshots of phenological events and have been reliably used to characterize phenological responses to climate. We review the current state of herbarium-based phenological research, identify potential biases and limitations in the collection, digitization, and interpretation of specimen data, and discuss future opportunities for phenological investigations using herbarium specimens.