Introduction
Pollinators facilitate reproduction for over 80% of flowering plants
(Ollerton et al. 2011) and increase the yield, to varying extent, of
75% of crop species (Klein et al., 2007). Bees are the single most
important group of pollinators (Neff and Simpson, 1993) thus detecting
changes in bee biodiversity is important for developing pollinator
management plans to sustain wild plant communities while maximizing crop
yields (Garibaldi et al., 2013, Winfree et al., 2018). A variety of bee
monitoring efforts have found troubling declines among wild bees (Koh et
al., 2016). For example, some species have had substantial range
contractions and declines in abundance, especially bumble bees in North
America and Europe (Beiseijer et al., 2006, Williams and Osborne. 2009,
Cameron et al, 2011, Bartomeus et al., 2013). Overall wild bee abundance
is falling in over 23% of the United States land area (Koh et al.,
2016). And the number of bee species observed around the world in museum
collections and from community science observations has dropped by 25%
from 1990 to 2015 (Zattara and Aizen, 2021). Because of their importance
and growing evidence of declines, bee monitoring efforts that build a
better understanding of the dynamics of bee biodiversity are important
for developing plans that can lead to conserving and restoring wild bee
populations (Winfree, 2010, LeBuhn et al., 2013, Woodard et al., 2020).
Bee biodiversity can be measured in a variety of ways, all of which can
give unique insights into the dynamic of populations and communities
within and across years. The simplest component of biodiversity is the
total abundance (e.g. total number of bees) and the abundance of
individual species measured over a given period. Total abundance can
provide information on the times within years that are most favorable
for most species. Data on the abundance of individual species across
years is critical for understanding if species population trends are
stable, increasing, or declining over time. For bees, these types of
abundance data are often not available because of a lack of repeated and
standardized sampling over time (Portman et al., 2020). Richness, or the
number of species, is another metric of interest in biodiversity
studies, particularly from a restoration and conservation perspective.
However, richness can sometimes provide limited unique information
because the detection of species is highly dependent on sample sizes as
more individuals counted tend to lead to higher species detection.
Diversity metrics help solve these limitations by summarizing aspects of
richness and relative abundance among species (evenness) in a single
estimator. For example, measures like inverse Simpson’s and rarefied
richness represent the effective number of species and provide
biodiversity measures that are independent of abundance-driven changes
in richness (Jost, 2006). Diversity can also be measured in a way that
incorporates information about the evolutionary distance that is present
among all species in a community using tools from the field of community
phylogenetics (Webb et al., 2002). A community with many closely related
species is more clustered, while a community populated with distantly
related species is more even (also called overdispersed). Finally,
measures of community composition are among the most information-rich
measures of biodiversity because they can incorporate the relative
abundances of all species to determine how similar communities are. The
multivariate nature of these measures means it is possible to detect
changes among communities even if overall richness, abundance, and
diversity are the same. For this reason, community composition measures
can be particularly powerful in detecting changes over time or responses
of communities to environmental degradation or restoration (Nerlekar and
Veldman, 2020).
Adult bee communities are highly dynamic within years making
standardized and season-long sampling necessary to accurately
characterize seasonal variations in biodiversity (Leong et al, 2016). In
temperate climates, bees overwinter as larvae, pupae, or adults and then
emerge in spring or summer in response to a variety of environmental
cues (Cane, 2021). But the time of the year in which bees are active
(seasonality) and the duration of their period of activity (phenological
breadth) vary greatly among species, resulting in ever-shifting
communities within each year (Ogilvie and Forrest, 2017, Kammerer et
al., 2021). Some studies have investigated changes in bee community
composition using continuous standardized sampling across the entire
period of adult bee activity (e.g., Heithaus, 1979, Wilson et al., 2008,
Joshi et al., 2015, Kammerer et al., 2016, Leong et al., 2016, Naeve et
al., 2020, Stemkovski et al. 2020, Roubik et al., 2021) However, for
many bee communities, we still lack a detailed understanding of how
community biodiversity and composition change from month to month.
Relatively fewer studies have compared the phenological patterns, both
seasonality, and phenological breadth, for many co-occurring species
(but see: Stemkovski et al. 2020). One reason for this is that the focus
of many bee community studies are in agricultural settings where the
blooming period of crops is only over a small period of time (e.g. Russo
et al., 2015, Grab et al., 2019, Graham et al., 2021).
Bee abundance and richness can also change greatly across years (Ogilvie
et al., 2017, Graham et al., 2021). One method for studying changes in
bee species over time is to compare historical records with more recent
collections (Cameron et al, 2011, Bartomeus et al., 2013, Burkle et al.,
2013, Mathiasson et al., 2019, Wood et al., 2019). This approach is
typically only able to test for changes in a small number of species
that are popular among collectors (like bumble bees) and is most
informative to detect changes in species geographic distribution
(Cameron et al, 2011, Mathiasson et al., 2019, Wood et al., 2019). The
lack of standardized collection methods between historical and current
records means it is difficult to separate if the changes in abundance
are the result of changes in population size or changes in sampling
efforts (Portman et al., 2020). An alternative approach is to conduct
standardized sampling, often using passive traps, in the same locations
across multiple years (Iserbyt and Rasmont, 2012, Martins et al. 2013,
Gezon et al., 2015, Onuferko et al., 2018, Graham et al., 2021). These
types of long-term longitudinal studies of bee communities are rare but
the existing ones have reported a mix of species that are stable,
increasing, and decreasing over time (e.g. Graham et al., 2021, Roubik
et al., 2021). These sampling approaches can give more explicit insight
into the changes in relative abundance in a wide variety of species.
Jointly, these studies have increased the need and interest in formal
monitoring of bee biodiversity to assess if there are declines and
potential links to decreasing pollination services (Woodard et al.,
2020).
In this study, we conducted standardized bee collections across 6 years
in Southern Pennsylvania, USA to characterize changes in bee community
biodiversity and changes in abundance of specific species. Specifically,
we quantified abundance, richness, diversity, phylogenetic structure,
and composition of bee communities between months and years. We
collected bees continuously from April through October using passive
Blue Vane traps. With this, we asked the following specific research
questions:
- How does bee biodiversity change within-years?
- How does bee biodiversity change across years?
- How do seasonal patterns differ among bee families and species?
- Do bee families and species differ in their changes in abundance over
time?