Citizen-science data reveal similar fragmentation patterns for species throughout North America

We have often pointed out that data from citizen-science projects such as Birds in Forested Landscapes or its predecessor, Project Tanager, are unique and valuable resources. Just how valuable they are was brought home to me this past April when we were asked to present a paper at a symposium in Portland, Oregon, held in conjunction with the Cooper Ornithological Society’s annual meeting.

The symposium dealt with the effects of forest fragmentation on western bird populations. The emphasis on western populations is important, because much of our understanding about the effects of forest fragmentation comes from studies of birds in eastern or midwestern North America and from Europe. The main focus of the symposium was to contrast patterns of response shown by species in western versus species in eastern North America. Using data from Birds in Forested Landscapes and Project Tanager, we were able to show that these patterns were quite similar across the continent. Along the way, we also found that this story has an interesting twist, and that makes this a great example of how important scientific questions can be answered using citizen science.

To address the question of how forest fragmentation affects birds on both sides of North America, we first needed a way of collecting data across the continent using the same rigorous methods everywhere; our citizen-science projects provided that. Second, we needed a way to describe forest fragmentation wherever we found it. Project Tanager showed that we could describe fragmentation consistently for several species of birds across North America using an overall measure of fragmentation that combined several variables. We built on that base in Birds in Forested Landscapes, and for this analysis, we used data from both projects to calculate a remarkably similar measure of fragmentation that was usable with both tanagers and thrushes. Finally, we needed bird species whose ranges included both eastern and western North America or closely related species on opposite coasts. This, too, was provided by data from Birds in Forested Landscapes and Project Tanager.

Project Tanager showed that variation exists within the Scarlet Tanager’s range in its sensitivity to fragmentation. We also found that the overall negative response to fragmentation was quite similar for all three common North American tanagers. The Cooper Ornithological Society’s symposium gave us the impetus to broaden this comparison across genera. We chose the Western and Scarlet tanagers and the Swainson’s Thrush and the Veery from the thrushes as our species-pairs representing both the West and East. As a widespread species, we chose the Hermit Thrush. We then took fragmentation data from Birds in Forested Landscapes and Project Tanager for a total of 2,400 unique study sites across the United States and Canada. We put these data into a single, large database and used statistical techniques to combine a number of variables—such as forest-patch size and proportion of forest in the surrounding landscape—into one overall measure of fragmentation that would provide a valid description of fragmentation anywhere in North America. Using this new yardstick, we compared species’ responses to forest fragmentation using statistical models. This is what we found.

Both the Scarlet and Western tanagers were less likely to be found in sites with increasing fragmentation; this mirrored the results from the Project Tanager analyses. Both species showed an approximately five-fold decrease in the probability of attempted breeding as we progressed along the gradient from sites showing the least fragmentation to sites with the most fragmentation. In fact, the strength of this negative reaction for the two species was nearly identical.

When we analyzed data for the Wood Thrush, however, a surprising pattern emerged. This thrush, which was believed to require large contiguous blocks of forest to breed, and whose populations are steadily declining, did not show the same negative response to forest fragmentation as the other thrushes. In fact, Wood Thrushes were actually more likely to attempt breeding in more fragmented sites. This result actually agrees with recent studies that have shown that although Wood Thrushes frequently attempt to breed in highly fragmented forests, their nesting success there is lower than in less-fragmented forests. Hence, these birds may be susceptible to what has been termed an "ecological trap." The Hermit Thrush, on the other hand, whose populations are increasing over much of its range, is more likely to be found in the least fragmented forests. These very different patterns show that the relationship between sensitivity to fragmentation and population dynamics is still unclear and suggest directions for further research.

The results of these important citizen-science research programs continue to give us tools both to ask and to answer provocative questions and provide a basis for informed management decisions. Like all good science, these projects give us not only answers but new questions to investigate. We believe that with Birds in Forested Landscapes, we can build on the firm basis provided by Project Tanager and other citizen-science projects to understand better how populations of wild birds respond to human-caused stresses. Such an understanding is fundamental to our work for the conservation of wild birds.

Figure 1. This figure reveals that the probability of finding breeding Veeries is highest in low-elevation habitats with little fragmentation (rear-left portion of graph). The results were produced from a statistical technique known as logistic regression. For the Veery, both fragmentation and elevation produced statistically significant (p<0.10) effects. These data were collected during the 1997 Birds in Forested Landscapes field season at study sites east of the Great Plains.

Figure 2. This figure reveals that the probability of finding breeding Swainson's Thrushes is highest in habitats of moderate elevation with little fragmentation (rear-central portion of graph). The results wer produced from a statistical technique known as logistic regression. For the Swainson's Thrush, fragmentation produced statistically significant (p=0.004) effects, but elevation did not. These data were collected during the 1997 Birds in Forested Landscapes field season at study sites west of the Great Plains.

Figure 3.This figure reveals that the probability of finding breeding Hermit Thrushes is highest in high-elevation habitats with little fragmentation (rear-right portion of graph).  The results were produced from a statistical technique know as logistic regression. For the Hermit Thrush, both fragmentation and elevation produced statistically significant (p<0.002) effects. These data were collected during the 1997 Birds in Forested Landscapes field season at study sites east of the Great Plains. No significant difference exists between eastern and western Hermit Thrush populations in the strength of their response to fragmentation.

Figure 4.This figure reveals that the probability of finding breeding Wood Thrushes is highest in more fragmented habitats at any elevation (front portion of graph). The results were produced from a statistical technique know as logistic regression. For the Wood Thrush, fragmentation produced statistically significant (p<0.05) effects, but elevation did not. These data were collected during the 1997 Birds in Forested Landscapes field season at study sites east of the Great Plains.