Research in landscape ecology

Recent research and vision

A recurring theme in my research is how populations and communities are shaped by the landscapes they inhabit. That is, how are populations or communities affected by the quality, availability, and arrangement of habitat in their landscape? Given the rapid pace of global land use and landcover change, such questions are clearly pertinent to conservation. Beyond landscape management, however, such questions are also key to understanding population dynamics in real-world landscapes.

In my doctoral research, I examine the habitat associations of wild bees, and assess the spatial scale at which they respond to habitat in their landscape (i.e., their “scale of response,” or “scale of effect”). In my previous work, I have studied the effects of forest cover and fragmentation on tick-borne pathogen prevalence.

In my future work, I would like to more concretely connect landscape context to population dynamics. That could mean i) more clearly defining and measuring habitat availability or quality, such that “landscape context” can be described more directly in terms of specific resources or conditions that affect the focal organisms, and ii) connecting those resources and conditions to individual- or population-level processes like survival, reproduction, foraging, or dispersal. For example, how does an organism’s movement abilities and foraging behavior shape its landscape-scale resource requirements? And how does the spatial distribution of that organism’s food resources affect population size and connectivity?

Bee habitat associations and scales of response

Habitat loss is considered one of the primary threats to wild bees, yet surprisingly little is known about the habitat needs of most bee species. And even less—next to nothing!—is known about the spatial scales at which bee populations respond to their environment. This “scale of response” (or “scale of effect”) is an important aspect of a species’ ecology, and is critical for effective conservation and management. It is not well understood, though, what determines a species’ scale of response—not for bees or any other taxa.

In this study, I assess the habitat associations and scales of response of 84 bee species. To do so, I use an amalgamated dataset of 30k specimens collected across 165 sites in the eastern US. I found (unsurprisingly) that bees have diverse and distinct habitat associations. There are broad trends, though, driven by phenology. Spring flying bees are typically forest-associated, while summer-flying bees are more often synanthropic, being associated with agriculture or sub/urban development. I also found that bees have (surprisingly) idiosyncratic scales of response. Importantly, scales of response were not associated with body size, which is a very common assumption in the literature.

Past projects and collaborations

Geometric effects of habitat fragmentation on biodiversity

It has long been thought that habitat fragmentation is bad for biodiversity, an idea that’s been affirmed by many experiments. Yet reviews of large scale, real-world landscapes have found fragmented landcapes to be no less diverse – or even more diverse – than contiguous landscapes.

In this study, led by Colleen Smith, we examined “geometric effects” of fragmentation (sensu May et al 2019) as a potential explanation for this disagreement between experimental and observational studies (Smith et al 2024). To do so, we simulated forest-specialist bee communities that were parameterized after real-world bee communities (Smith et al 2021), and used actual landscapes from the northeastern USA to simulate deforestation, then measured changes in biodiversity after forest loss. We found that more species were retained in more fragmented landscapes. This suggests that, at least in the immediate aftermath of habitat loss, fragmentation can indeed mitigate the effects of habitat loss on biodiversity.

Landscape context and tick-borne pathogen prevalence

Landscape context, namely forest loss and fragmentation, is hypothesized to affect the abundance of ticks and the prevalence of tick-borne pathogens by way of its effects on the ticks’ and pathogens’ vertebrate hosts. While some evidence supports this hypothesis in the northeast, with respect to deer ticks and Lyme disease, it had not been explored in other regions that are dominated by other ticks and pathogens.

In these studies, I examined the spatiotemporal distribution of the bacterial pathogen Ehrlichia chaffeensis and its vector, the lone star tick (Amblyomma americanum), in relation to variation in weather and landscape context (Simpson et al 2019 a and b). I found pathgoen prevalence in the tick population to vary by an order of magnitude across years (from < 1 to 4%), and to vary with forest cover, type, and fragmentation, and with temperature in the previous year. Similarly, tick turnover was affected by surrounding forest type and an interaction between forest amount and fragmentation, but in a way that appeared de-coupled from pathogen prevalence.

In subsequent years, however, my former mentee Olivia Spencer (a superstar, now PhD student at Penn State) continued this study for her honors thesis and found that results were very idiosyncratic. In particular, we found that the apparent drivers of tick occupancy and pathogen prevalence differed between adult and nymph ticks, and k-fold cross-validation showed results to depend greatly on which data were included in the analysis. Ultimately, tick abundance and pathogen prevalence both appear highly stochastic. This, and the relative rarity of E. chaffeensis (relative to the bacterial agent of Lyme disease) make predicting the distribution of E. chaffeensis or its drivers very difficult.