Urban ecosystems are temporally dynamic systems, yet historical factors associated with land use legacy and time lags as a result of development have largely been overlooked. Because human-altered landscapes are relatively recent, it is particularly important to recognized that the observed biodiversity may be undergoing a process of change from the previous land use type to the new one when interpreting observations. Just as urban ecosystems are dynamic across years, they are also dynamic intra-annually, with resulting phenological shifts within urban landscapes compared to surrounding natural habitat. The urban heat island effect is a well-documented phenomenon where the city environment can be significantly warmer than the surrounding landscape as a result of impervious surface area that retains heat and higher energy usage, causing changes in the timing of ecological patterns. As a result, plants bloom earlier the more densely urban the surrounding habitat is and bird migration advances earlier in urban contexts. Another potential effect of land use change in urban and agricultural landscapes beyond climatic can be the variety and timing of floral resource availability. Urban areas, while having less green space, often grow many exotic plants which are supplemented with water and nutrient inputs that allow for an extended flowering season. As a result, urban areas are characterized by relatively low , but constant, floral resources throughout the year. Agricultural landscapes have large patches of dense, often homogenous, floral resources that will fluctuate greatly from early spring to the end of the summer due to mass-flowering monoculture crops. In contrast, pot with drainage holes many natural areas in California experience a large burst of diverse floral blooms in the spring, and by the end of the summer, there are very few floral resources available.

Bees provide the majority of animal mediated pollination services on which an estimated 87.5% of flowering plants depend . The value of pollination in agriculture is estimated at $200 billion worldwide, largely due to many foods that are essential for food security and a healthy human diet, including numerous fruits, vegetables, and nuts that require bee pollination. In addition, there has been growing interest in urban agriculture to ensure food security and access to healthy foods for urban populations. One study estimated the economic value of urban fruit trees in the one city of San Jose, California to be worth $10 million annually. Honey bee populations and many bumble bee species are declining worldwide, while many other bee species have not been closely documented enough to determine their status. One of the reasons proposed to be negatively affecting bee populations is land use change. A review of 265 papers studying the effect of land use change on pollinator populations found more negative than positive impacts with a wide window of variability. This could partially be attributed to the wide diversity of pollinators themselves, as well as the varying definitions that people use to constitute land use change. We propose that an additional cause of high variability can be explained by investigating how the seasonal patterns of bee communities may shift in different neighboring land use types which experience highly different availability and suites of floral resources. Although bee seasonality has been documented in urban and agricultural landscapes, no studies to our knowledge have specifically investigated the differences in seasonal patterns of population abundances of bees between human-altered landscapes and neighboring natural habitat, despite the established seasonality of bees and variability of floral resource availability.

Bees are often sampled throughout the season, but all of these data are typically lumped together, potentially obscuring subtleties in change . Here, we investigate how local bee communities shift over the course of the flowering season in urban, agricultural, and natural land use types. We make use of a “natural experimental design” in which urban, agricultural, and natural areas intersect in a peri-urban landscape on the outskirts of the San Francisco Bay Area in Contra Costa County, California. To study the impact of changing land use on local bee community population dynamics, we sampled the bee community flying through the landscape at four time points over the course of the season for three years at 24 sampling locations.At each site we laid out a standardized pan trapping transect of fifteen 12 ounce bowls spaced 5 meters apart in alternating colors of fluorescent blue, white, and fluorescent yellow. Bowls were filled to the brim with soapy water . In 2010, transects were set up for a 4 hour period between 10:30am to 2:30pm , with 4 sites sampled per day, and all sites sampled on consecutive days. These 2010 transects were run twice, once in the early summer, and once in the late summer. In 2011 and 2012, sampling was conducted over a 24 hour period, so that more sites could be run simultaneously and more samples could be collected per site per year. All 24 sites were sampled within two consecutive collecting windows , and were run four times each year: early spring, late spring, early summer, and late summer. Because we were interested in landscape level effects, we tried to control local variables as much as possible. All sites were selected in easily accessible, open areas that received full sun. Natural areas were in grassland habitat, so we selected agricultural sites that were either weedy field margin edges or fallow fields, and urban sites that were vacant lots or green ways.

The human-altered sites were deliberately selected to not be adjacent to any mass flowering plants of agricultural crops or gardens. The goal of collection was to sample the bee community that was flying through the site searching for resources. Bees were collected from the pan traps by using a metal strainer, rinsed with water, frozen overnight or longer, and then pinned and labeled. Specimens were sorted to the genus level, and then to the species level with the assistance of Dr. Robbin Thorp , Professor Emeritus, UC Davis. The only exception to identification at the species level were bees of the genus Lasioglossum, large pot with drainage due to their overwhelming abundance, limited availability of taxonomic expertise for this group, and lack of known ecological diversity. Voucher specimens and the majority of the total collection will be deposited at the Essig Museum of Entomology at UC Berkeley.For response variables including aggregate bee abundance, species richness, Shannon diversity, and number of rare species, we tested for the effect of land use type, seasonality, and their interaction with generalized linear mixed models using the R package lme4. We designated collecting method , land use type, seasonality, and the interaction of land use type and seasonality as fixed effects, and site and year as random effects. We analyzed the effect of time both categorically by collecting period and continuously by day of year. Day of year was normalized on a scale of 0 to 1 from the first collecting date to the last. Shannon diversity was fit with a Gaussian distribution while all other variables were fit with Poisson distributions.We found that the bee communities in human-altered landscapes experienced different phenological patterns than the neighboring natural areas. Increased temperature in urbanized areas as a result of the urban heat island effect is often cited as the driving force for changed ecological dynamics. We propose another driver of local phenological shifts: the timing and quality of floral resource availability, due to irrigation that extends the flowering season throughwater inputs and landscaping choices in urban residential, public, and commercial zones, as well as mass flowering crops in agricultural fields. For example, Eucera actuosa was collected most frequently in human altered sites in the early spring and very little in late spring, whereas in natural sites it was collected in lower numbers in early spring and peaked in late spring. Another bee, Melissodes lupina , is a later flying bee than E.actuosa and experienced the opposite pattern. Melissodes lupina was collected most frequently in the early summer for natural areas, but was collected more often in the late summer in human-altered landscapes. Both E.actuosa and M. lupina demonstrate a pattern of relative abundance of species being shifted in different land use types. Even further, the temporal direction of relative abundance in both examples areas skew towards higher abundance during the middle of the season for natural areas and higher abundance at the more extreme ends of the season in the human-altered landscapes.

In other words, these are not simple patterns where species in human-altered landscapes always are collected earlier, which is generally the result of the urban heat island effect. Instead, this is likely due to irrigation effects extending flowering times in human-altered landscapes that provide bees with necessary resources for extended flight periods. This supports our theory that patterns of change in bee community distribution throughout the year are a result of the different land use types offering variable seasonal floral resources. While ecologists have used time as an important variable in many different systems, only recently has time begun to be incorporated into urban ecology. These differences in relative bee abundance throughout the year could be the result of resource tracking or a shift in emergence timing between different land use types. Shifts in plant phenology have been well documented in temperate urban landscapes. Urban areas have been associated with earlier plant blooming closer to the city center, and the urban heat island effect also can directly affect animal populations. In addition, historical temperature and museum collection records show a link between climate change and advancing bee emergences. Bees could be responding to local climatic differences or floral availability respectively, with both emergence timing and length of the flight season of bee species being impacted differentially at a micro-scale between different land use types. Pollinator responses to land use change are generally more often negative than positive, although there is high variability of outcomes due to many different experimental design types, systems, and the use of simple community metrics rather than more species specific analyses. Using functional groups such as nesting type, generalized foraging , and sociality, more patterns have emerged about traits that may be most sensitive to anthropogenic disturbance, although the type of disturbance will affect bees differently. For example, ground nesting species may be more successful in intensified agricultural landscapes, while cavity nesting species may be more common in urban landscapes because of increased nesting resources. We had 7 species that were collected almost exclusively in human-altered sites . These species positively associated with anthropogenic change covered a range of functional groups for sociality, nesting type, foraging generalism, size, and distributional range. Two of these are non-native , and as a group they have a wide diversity of life history traits. For example, Ceratina dallatorreana, a species of small carpenter bee originally from theMediterranean region first collected in California in 1949, has the unusual life history trait of female parthenogenesis in its local population here. Megachile rotundata is another nonnative bee accidentally introduced to the United States from the Mediterranean , is a solitary leaf cutter species that have become an important managed pollinator in the Western United States. Andrena chlorogaster, is a native California bee, with a generalized life habitat and wide range, while the squash bee, Peponapis pruinosa, is a solitary specialist on cucurbits. These species that favor human-altered landscapes not only fail to share many life history traits, but they are also diverse phylogenetically, comprising several different bee families. These different patterns of bee distributions could be the result of two possibilities: either bee populations are tracking resources between the different land use types, or the bee communities in the different land use types are experiencing different emergence timing. Bee movement has been notoriously difficult to study because of the size, mobility, and life history of this group. Some species and system-specific conclusions about bee foraging distances and size based models of foraging distances have been made. However, these are all based on foraging movements anchored around a central nest that a female bee is provisioning, and in contrast almost nothing is known about dispersal movement—in other words, how far bees might travel from their emergence site to mate and select their own nest site.