Terrestrial Ecology Projects

(a) = active project; (c) = completed project

Harris' Checkerspot

Movement, population dynamics, and evolution of Checkerspot Butterflies

Collaborators: Paul Severns, Mark Lewis, Felix Sperling, Devin Goodsman, Elizabeth Crone, Aubrie James,  and Andy Edwards

We have three ongoing projects on Euphydryas checkerspots:

1) Baltimore Checkerspots.(a) The tiny butterfly above is a uniquely marked (two green dots were added with a metallic gel pen) Harris' Checkerspot (Chlosyne harrisii, which we used in the spring of 2011 to refine marking methods. Like other checkerspots, this species forms small discrete colonies and is ideal for mark-recapture designs to understand dispersal and meta-population dynamics, and there is a healthy colony a 5 minute drive of Harvard Forest.

Baltimore CheckerspotIn 2011 we started pilot projects on the Baltimore checkerspot (Euphydryas phaeton - right), which developed into a much more extensive set of ongoing investigations related to movement, demography, evolution of host plant specializations, and landscape genetics.  This butterfly has been experiencing steady population increase in New England, and anecdotal reports suggest this might be due to the relatively recent evolution of the ability to utilize a weedy exotic host (Plantgo lancelota). Prior to adoption of this new host, this species was probably declining and relatively rare, with populations restricted by the patchy distribution of its single native host plant Chelone glabra (white turtlehead). In fact, this situation is still the case over most of its geographic range east of the Mississippi River. Many populations in New England still use the native host, but many others appear to persist using only the exotic Plantago. It is not clear if these populations are exchanging individuals, or if the movement behaviour of adults is affected by the host plant each population uses. All else being equal, Plantago using populations should develop a movement strategy the includes higher dispersal probabilities because and be more apt to disperse because Plantago is widespread, and thus suitable patches much closer together and easier to find.  We've published a number of papers on this already and several more are in later stages of preparation / submission.

Aubrie James, my 2012 REU student, discovered that host plant affects the morphology of males but not females. See her blog post on our summer 2012 field work! Aubrie is now a Ph. D. student in Monica Gerber's lab at Cornell.

Also, see Aubire's best David Attenborough impression as she captures a predation event during a flight map on her iPhone. Death by dragonfly.

Taylors Checkerspot 2) Editha Checkerspots.(a) In the Western US, the genus Euphydryas is highly diverse.  Many subspecies are endangered after becoming specialized on rare host plants or habitats.  The Taylor's Checkerspot (left) is one example, and we are developing meta-population and genetic models to understand why these rare species are doing so poorly.  We've already published two papers using Taylor's Checkerspots as model species and several more are in later stages of preparation. 

                Checkerspot3) Anicia Checkerspots.(a) In the far north (Alaska, Yukon), only one Ephydryas checkerspot is present.  In the southern and western parts of North America, this butterfly is widespread and has a number of described subspecies.   By contrast, its ecology in the Arctic is not well understood, and in the summer of 2015, we are planning a reconnaissance project.  Museum records and preliminary observations suggest it is an extreme habitat and hostplant specialist, and populations might be distributed as a dendritic metapopulation network along rivers, especially the Yukon River, which flows across both the Yukon Territory and the entire state of Alaska.  We hope to develop this species into a model to understand how metapopulations react to climate change at the northern edges of ranges.

Atlantis Fritillary MBC Report Bubble Map

Detecting Changes in Butterfly populations from Citizen Science data

Collaborators: Elizabeth Crone, Sharon Stichter, Massachusetts Butterfly Club, Leslie Reis, Elizbeth Zipkin, Karen Oberhauser, Derek Sikes and others.

1) Massachusetts Butterfly Club.(c) Using a new modeling approach known as list-length analysis (Szabo et al. 2010,Ecology) designed for analysis of Citizen Science sightings data, we have been able to rigorously estimate the population trends for nearly all New England butterfly species present in Massachusetts. The data analyzed were graciously provided by the Massachusetts Butterfly Club, who diligently collected, audited, and archived observations over the past 20 years. I cannot stress enough that this work could not have happened without the thousands of reports made by this dedicated group.

Above: The distribution of Atlantis fritillary (Speyeria atlantis) sightings in Massachusetts from 1992-2010. Size of circles indicates the number of individuals seen on a given day. Only two other species, the acadian hairstreak and aphrodite fritillary, are declining faster. Note that the bubble size scales as the log of the number of individuals seen, so as late at the 1990's hundreds of individuals were often reported. Now it is exceptional to see one or two.

The results so far are stunning. They suggest rapid climate induced population declines in nearly all cold adapted species and numerous invasions from southern species that were previously unrecorded or only vagrants. The overall impression is of a wholesale northward shift of butterfly population ranges in eastern North America. Some of our findings were recently published in Nature Climate Change (see publications).

These results garnered considerable media attention. Here are links to coverage in the Boston Globe which covered the work twice (linked is the longer Sunday Edition article), Smithsonian blog, CBCnews, one of 3 NPR interviews  (this one was broadcast nationally), and an interview I did for NBC Nightly News also broadcast nationally.  However, the results were reported by many other media outlets.

List-length results

Above: Population trends for all butterfly populations in the state of Massachusetts with enough reliable sightings data to estimate trends. Estimates of change in detection probability that are positive indicate increasing populations, negative indicate decreasing populations. We defined Northern species to have greater than 50% of their range north of the latitude of Boston, Southern species have ranges with greater then 50% of their range south of Boston. Cheat sheet for butterfly codes can be found here.

Ken Philips
                  Photo 2) Ken Philip collections.(a) Ken Philip was one of the most prolific collectors of Alaskan and Northern butterflies in history and collected up until his death in 2014.  His private collection, numbering over 100,000, is now being curated by the University of Alaska Museum.  With Derek Sikes and student Kathryn Daly we will use this collection, as well as the UAM lepidoptera collection, to understand range changes and extinctions under climate change and use what Ken started to launch a new citizen science effort to report and compile Alaskan butterfly sightings. (Ken is second from the left)

                  Brooks Photo of Monarch Larva 3) Monarch Larvae Monitoring Project.(a)  With Elise Zipkin and Karen Oberhauser, we're developing models of monarch larval survival using citizen science data collected by citizen scientists from the Monarch Larvae Monitoring Project.  We expect to test for links between survival and herbicide use as well as the overwintering survival and population size in Mexico.  This work developed as part of a SESYNC center project on modelling of citizen science insect monitoring data.

Caribou on Road

Predator - Prey Dynamics of Wolves and Caribou. (a)

Collaborators: Craig Demars, Jon Potts, Stan Boutin.

Working with Craig Demars, Jon Potts, and Stan Boutin, we are developing models to understand how effective different anti-predator strategies are at preventing the loss of caribou calves to wolves and bears.  This is of great concern, because many woodland caribou herds are declining in Western Canada.  The effectiveness of these strategies depends upon the type of predator and also the landscape and degree of disturbance.  The advantages and disadvantages of remaining in a group or dispersing to calf as individuals may be driven by these different factors.  The evolved strategy, which in forested areas is to be to disperse into as individuals in dense vegetation, may not be effective after roads or other linear features have been added to the landscape.  We are building models to understand how roads effect the predator-prey relationship and why these populations might be declining.

Optimal Movement Simulation

Movement Model Development. (a)

Model development to understand animal movement is the common thread between my marine and terrestrial research. Observations and data from our field work have sparked development of a number of lines of theoretical and analytical model development. These include dynamic behavioural programming approaches to understand how female butterflies (or females of a wide range of other species) should allocate reproductive effort spatially. In this case, where should females deposit eggs given a particular spatial arrangement of patches or resources and the disturbance intervals, and how should this be balanced against dispersal? We predict that in increasingly patchy habitat, females should increase allocation (lay more eggs) in their natal patch. However, this behaviour is likely evolved, and species that evolved in a less patchy, more connected habitat may disperse before allocating much reproductive effort into a natal colony. Not allocating to the natal patch, and dispersing into habitat where the odds of locating another suitable patch are small will cause such an individual's fitness to decrease or go to zero. However, allocating more eggs to the single, natal patch is also not ideal, as it decreases the stability of the metapopulation network as all patches will eventually experience a catastrophic disturbance and the local metapopulation will go extinct.

Last update May 2015