The Work Speaks—Part 7: Leaky Sink

In April, Conservation Biology published a comment authored by Christopher A. Lepczyk, Nico Dauphiné, David M. Bird, Sheila Conant, Robert J. Cooper, David C. Duffy, Pamela Jo Hatley, Peter P. Marra, Elizabeth Stone, and Stanley A. Temple. In it, the authors “applaud the recent essay by Longcore et al. (2009) in raising the awareness about trap-neuter-return (TNR) to the conservation community,” [1] and puzzle at the lack of TNR opposition among the larger scientific community:

“…it may be that conservation biologists and wildlife ecologists believe the issue of feral cats has already been studied enough and that the work speaks for itself, suggesting that no further research is needed.”

In fact, “the work”—taken as a whole—is neither as rigorous nor as conclusive as Lepczyk et al. suggest. And far too much of it is plagued by exaggeration, misrepresentations, errors, and obvious bias. In Part 6 of this series, I critiqued Christopher Lepczyk’s paper Landowners and cat predation across rural-to-urban landscapes, published in 2003. Here, I’m going to examine two studies conducted by Philip J. Baker and various collaborators.

The Studies
In the first study, Baker et al. distributed questionnaires to 3,494 households across a 4.2 km2 area of northwest Bristol (UK), and used responses to estimate cat ownership and predation levels (via prey returned home). [2] This work served as a pilot study for the subsequent study.

The second study, conducted August 2005–July 2006, was also conducted in Bristol. Added to the original 4.2 km2 site were nine 1 km2 sites. The researchers used very similar sampling methods, but, based on results of their pilot study, had somewhat more specific objectives:

  1. To quantify cat density
  2. To quantify the various species of birds killed by cats.
  3. To estimate the impact of cat predation by species and site.
  4. To determine whether the predation observed was compensatory or additive. [3]

Sources and Sinks
Among the authors’ conclusions from the pilot study was that, at least for three of the ten bird species surveyed:

“…it is possible that cat predation was significantly affecting levels of recruitment and creating a dispersal sink for more productive neighboring areas.” [2]

Dispersal sinks or habitat sinks, are patches of low-quality habitat that are unable to sustain a population of a particular species were it not for immigration from higher quality habitat patches—called sources—nearby. So, what Baker et al. are suggesting is that predation by cats may be extensive enough to deplete populations of certain bird species at their study site, such that at least some of the birds observed there were immigrants from nearby habitat.

But the authors also point out that, “despite occurring at very high densities, the summed effects on prey populations appeared unlikely to affect population size for the majority of prey species.” [2] And even for House sparrows, which were among the three species of concern (and, apparently, in decline throughout the UK’s urban areas), Baker et al. note that their “numbers appear to be stable in Bristol as a whole.”

So, is the area a habitat sink or not?

A cursory look at the theory and empirical measurement of source-sink dynamics reveals great complexity. Variations across time and geography must be taken into account—the ebb and flow of local populations might easily be overlooked or misunderstood by applying a short time horizon (i.e., 12 months) and arbitrary boundaries (i.e., those that define the study site). Annual rainfall, for example, can dramatically influence yearly population levels on a local scale. And it’s been shown that source-sink dynamics can occur over distances of 60–80 km. [4] In fact, the determination of sinks and sources in the field can be problematic enough that sources sometimes appear to be sinks and vice-versa. [5]

Given the complex nature of source-sink dynamics, the suggestion by Baker et al. that cat predation may be creating a habitat sink seems rather premature. Such assertions—despite the requisite disclaimers (the authors note only that “it is possible”)—tend to attract attention and gain traction. Longcore et al., for example, cited the pilot study in their 2009 essay, “Critical Assessment of Claims Regarding Management of Feral Cats by Trap-Neuter-Return.” [6]

Of greater interest to me, though, are the assumptions Baker et al. used to estimate the impact of cat predation.

Counting Cats and Counting Birds
In both studies, the authors quantified the impact of cat predation on bird populations by comparing different levels of predation with different bird densities. Their maximum impacts, for example, assumed that all cats were hunters—despite the fact that 51–74% of the cats included in the two studies brought home no prey at all—and that bird productivity was zero (i.e., no young birds survive to adulthood). As the authors admit:

“This was clearly not realistic, as the estimated maximum numbers of birds killed typically exceeded breeding density and productivity combined, such that the prey populations studied would probably have gone extinct rapidly at a local level or acted as a major sink for birds immigrating from neighbouring areas.” [3]

But how realistic are their other estimates?

A detailed examination of a single species at a one site (taken from the second study, for which such information is available) illustrates some flaws. I looked at House sparrows for the 1 km2 site designated as ST5277. Here, 18 participants reported that their 22 cats returned a total of 30 prey items, nine of which were birds (two of them “unidentified”). Of the birds returned home, two were House sparrows.

When it comes to estimating impacts, though, Baker et al. use figures of 332–1,245 House sparrows killed by the cats of ST5277. The maximum, we already know, is “not realistic,” but even the minimum seems awfully high. So, where are these birds coming from?

To start with, two adjustments have to be made to the original predation figure. First, the two unidentified birds are “distributed” across the categories of bird species that were identified. Then, we have to account for participant drop-out; not all of the 22 cats were surveyed for the entire year of the study. Now we’re up to an average of 8.7 House sparrows brought home annually by the cats at this site.

But of course there are more than 22 cats at ST5277. Baker et al. estimate that there are 314 of them (although we know very little about the factors that affect their hunting ability and success—for example, their access to the outdoors, age, etc.). We also know that only seven of the 22 cats included in the study brought home prey. In other words, 32% of the cats surveyed were documented hunters. Based on these numbers, then, we can estimate the yearly predation rate of House sparrows at ST5277 to be roughly 125—well short of the minimum proposed by Baker et al. (and just a quarter of their intermediate rate).

There are some minor differences between their method for estimating predation rates and mine. For the most part, though, the “missing” sparrows can be found in the authors’ use of a correction factor (3.3) proposed by Kays and DeWan to account for prey killed but not returned home. [7] Undoubtedly, cats fail to bring home all the prey they catch (though they also undoubtedly bring home prey they didn’t kill), but there is good reason to doubt Kays and DeWan’s “correction.” Among the flaws in their analysis were small, dissimilar samples of cats, and a failure to account for highly skewed data sets.

So, even setting aside the complexities of source-sink dynamics, these inflated predation rates, combined with the fact that “the estimates of breeding density presented in this manuscript should be regarded as minima,” [3] raise serious doubts about whether the site is in fact a habitat sink (or, if so, to what extent).

Compensatory and Additive Predation
As I’ve discussed previously, even accurately predicted levels of predation can be deceptive. There’s compensatory predation (in which prey would have died even in the absence of a particular predator, due to illness, starvation, other predators, etc.) and additive predation (in which healthy prey are killed). It’s the difference between, as Beckerman et al. put it, the “doomed surplus hypothesis” and the “hapless survivor hypothesis.” [8]

When it comes to relating predation to population levels, it’s critical to understand the difference, and know the extent to which each type is occurring.

To get at this critical issue, Baker et al. compared the physical attributes (e.g., muscle mass score, mean fat score, etc.) of 86 birds killed by collisions (e.g., with cars, windows, etc.) to those of 48 birds killed by cats. Although the authors point out, “the relationship between body mass and quality (i.e., likelihood of long-term survival and therefore reproductive potential) in passerines is complex,” they nevertheless conclude that the birds killed by cats “were likely to have had poor long-term survival prospects.” [3] (An earlier study comparing spleen mass arrived at essentially the same conclusion: that birds killed by cats “often have a poor health status.” [9])

Still, Baker et al. express caution about their findings:

“The distinction between compensatory and additive mortality does, however, become increasingly redundant as the number of birds killed in a given area increases: where large numbers of prey are killed, predators would probably be killing a combination of individuals with poor and good long-term survival chances. The predation rates estimated in this study would suggest that this was likely to have been the case for some species on some sites.”

But their inflated predation rates and low estimates of breeding density combine to diminish the apparent level of compensatory predation. Were these estimates adjusted to better reflect the conditions at the site, the “redundancy” the authors refer to would be reduced considerably.

*     *     *

It’s not clear why Longcore et al. cited the pilot study their essay, but left out any mention of the much larger subsequent study. Perhaps it was just a matter of timing—“Cats About Town” was published in August of 2008, while “Critical Assessment” was published in August of 2009. A year is not much time in the world of scientific journals, and it’s possible that the two manuscripts more or less crossed in the mail. On the other hand, the pilot study fits more neatly into the argument put forward by Longcore et al.—an argument that doesn’t even recognize the distinction between compensatory and additive predation.

Of course, Baker et al. did themselves no favors, either. By using inflated predation rates—the result of some peculiar, unjustified assumptions—they virtually buried the most important findings of their study.

References
1. Lepczyk, C.A., Mertig, A.G., and Liu, J., “Landowners and cat predation across rural-to-urban landscapes.” Biological Conservation. 2003. 115(2): p. 191-201.

2. Baker, P.J., et al., “Impact of predation by domestic cats Felis catus in an urban area.” Mammal Review. 2005. 35(3/4): p. 302-312.

3. Baker, P.J., et al., “Cats about town: is predation by free-ranging pet cats Felis catus likely to affect urban bird populations? Ibis. 2008. 150: p. 86-99.

4. Tittler, R., Fahrig, L., and Villard, M.-A., “Evidence of Large-Scale Source-Sink Dynamics and Long-Distance Dispersal among Wood Thrush Populations.” Ecology. 2006. 87(12): p. 3029-3036.

5. Runge, J.P., Runge, M.C., and Nichols, J.D., “The Role of Local Populations within a Landscape Context: Defining and Classifying Sources and Sinks.” The American Naturalist. 2006. 167(6): p. 925-938.

6. Longcore, T., Rich, C., and Sullivan, L.M., “Critical Assessment of Claims Regarding Management of Feral Cats by Trap–Neuter–Return.” Conservation Biology. 2009. 23(4): p. 887–894.

7. Kays, R.W. and DeWan, A.A., “Ecological impact of inside/outside house cats around a suburban nature preserve.” Animal Conservation. 2004. 7(3): p. 273-283.

8. Beckerman, A.P., Boots, M., and Gaston, K.J., “Urban bird declines and the fear of cats.” Animal Conservation. 2007. 10(3): p. 320-325.

9. Møller, A.P. and Erritzøe, J., “Predation against birds with low immunocompetence.” Oecologia. 2000. 122(4): p. 500-504.