Research in population ecology

Current research and vision

Population ecology is currently the primary focus of my work, with two projects focusing on plants and bees, respectively. In each project, we are investigating the environmental factors that promote or limit population growth, particularly in the context of climate change. Long term, I aim to have a research program that continues to include both plants and bees. These taxa are intimately interconnected, and I am interested in how populations of each respond to each other and their shared environment.

Population consequences of phenological shifts

I am working with Dr. Amy Iler to investigate the effects of phenological shifts (driven by climate change) on plant population growth. While the effects of climate change have been studied before, they are often studied in isolation—that is, on individual vital rates. By following cohorts of plants through time, we will integrate the environmental effects across the whole life cycle while accounting for trade offs between vital rates. In doing so, we will be able to better understand whether, how, and why plant populations might respond to changing environments.

Effects of the biotic and abiotic environments on population viability

Together with Dr. Paul CaraDonna, I am co-PI on a recently funded NSF grant to study the effects climate change on solitary bees. Climate change can (potentially) affect bee populations in many ways, including direct effects of the abiotic environment (e.g., temperature), and indirect effects mediated by the biotic environment (i.e., food availability, predation rates). In this project, we will use an experimental demography approach to assess the role of multiple biotic and abiotic factors affecting population dynamics of a solitary bee (Osmia ribifloris), how their responses are shaped by plasticity versus adaptation, and how these environmental factors come together to affect population viability under different climate scenarios.

Past projects and collaborations

I am currently in a phase of transition, as I move out of my PhD into my postdoctoral work. Below are details from my doctoral work on bees, some of which is yet to be published.

In my work on wild bee populations, I used population genetics as a tool to better understand bumble bee populations. In one study, I used rates of genetic isolation-by-distance to jointly estimate effective population densities and dispersal distances of six bumble bee species. In another, I more closely examined the population genetics of Bombus impatiens, assessing colony densities and genetic diversity, and meausuring the effect of different habitats on rates of gene flow between subpopulations of Bombus impatiens (“isolation-by-resistance,” sensu McRae 2006).

Estimating effective population size and dispersal distances

Effective population size and intergenerational dispesral are two important aspects of any species’ population biology. Despite their importance, however, little is known about either of these demographic parameters for bees. This is likely due, in large part, to the difficulty of measuring them.

In this study, I use established but under-appreciated tools from population genetics to address these difficulties. Specifically, I use observed rates of genetic isolation-by-distance to jointly estimate effective population density and dispersal distances1. Using my own field data on Bombus impatiens, together with published data on B. impatiens and five other Bombus species, I demonstrate the utility of this method, assess its sensitivity to study design (namely spatial scale of sampling), and provide guidance on use and interpretation.

In short, I found that average effective population densities of Bombus are likely quite low—much lower than current estimates of census population densities. I also found, however, that these estimates are sensitive to the spatial extent of sampling; while average densities are quite low, local effective densities can be much higher. I thus recommend that any user of this method be intentional about the spatial extent of their study, and to be explicit about the scale of their study in their interpretation.

Population ecology of Bombus impatiens

Bombus impatiens is something of an enigma. On the one hand, it is ubiquitous, being found in our lawns, city parks, agricultural fields and roadside ditches. On the other hand, careful study of their colonies has found reproductive success of B. impatiens colonies to be higher in forests than in open habitats (Pugesek and Crone 2021), and correlational studies in the Winfree Lab have likewise shown B. impatiens to be forest-associated (Smith et al 2021; Simpson in prep). The contrast of these results makes me wonder what our eyes aren’t seeing. While we see B. impatiens workers everywhere, are there differences in colony size or abundance across different habitats? And if colonies are more succussful in certain habitats, how does landscape composition affect population connectivity?

In this study, I used microsatellites to genotype nearly 1000 B. impatiens workers from southern New Jersey. I use these to assess differences in colony abundance and genetic diversity between agricultural, forested, and sub/urban landscapes, and to assess the effect of these habitats on geneflow across the landscape.

There are two primary results from this study. First, I found that B. impatiens colony number increased with increasing amounts of deciduous forest and crop cover in the surrounding landscape, but that these effects did not carry through to affect gene flow. Second, I also found these estimates of colony number were completely uncorrelated with the abundance of workers I observed while collecting, which had distinct and sometimes contradictory associations with landcover. As a result, if I were to have inferred B. impatiens’s response to habitat from worker abundance alone, I would have come to very different conclusions.

This latter result is critical not just for this study, or even B. impatiens specifically, but the study of eusocial bees generally. It has been long recognized that colony number is the pertinent measure of population size for eusocial insects (Crozier et al 1979; Chapman and Bourke 2001), most studies of how bumble bees respond to their environment (including my own!) still rely on observed worker abundance. My results suggest this is a bad idea (see also: Herrmann et al 2007; Wood et al 2015).

There is a lot of hub-hub about bee declines, and the repurcussions of bee declines for agriculture. Yet long-term monitoring studies to demonstrate such declines are exceedingly rare. In this paper, led by Andrew Aldercotte, we utlized an eight-year study of bee visitation to watermelon crops to measure the presence and extent of bee declines (Aldercotte, Simpson and Winfree, 2019).

We demonstrated that there was a statistically significant decline in wild bee abundance on watermelon farms over eight years of the study, but that this result is not out of line with what we could expect given random population fluctations year to year. That is, yes, bee populations declined, but this was not necessarily evidence of a population trend, per se. Upshot: clearly identifying population trends requires many years of data—enough to parse actual trends from the rapid fluctations characteristic of insect populations.

  1. See Rousset 1997 and 2000, and Robledo-Arnuncio & Rousset 2010. ↩