Mesopredator release and avifaunal extinctions in a fragmented system

1. What they did

In coastal southern California, a century of building has left only steep canyons of native sage scrub as islands in a sea of houses. Scrub-breeding birds have been disappearing quickly from these canyons. One explanation is “mesopredator release”: when a large predator (here the coyote) declines, smaller carnivores (striped skunk, raccoon, grey fox, domestic cat and opossum) become more numerous and eat more birds. The idea had received little critical testing, yet it was already being used to justify protecting carnivores.

The authors used the natural differences in coyote numbers between and within 28 habitat fragments in San Diego to test it:

• Carnivore surveys (September 1995 to August 1997). They set track detection stations every 250 m along transects in each fragment and checked them for five days in a row in autumn, winter, spring and summer. The number of visits by each species, divided by survey effort, gave a relative abundance. Scats and remotely triggered cameras confirmed which species were present. In five fragments where coyote numbers changed in the first year, they surveyed for a second year.
• Bird surveys (April to June 1997). Teams of trained observers did point counts and walked transects to count how many of eight scrub-specialist bird species were still present in each fragment.
• Fragment features. They measured each fragment’s area, age (years since it was cut off by development) and isolation.
• Cats and people. They collected coyote scats, followed radio-collared cats, and sent questionnaires to residents living next to the fragments about their cats, what the cats brought home, and coyotes.

2. What they found

Coyotes and smaller carnivores

• Coyotes were more abundant in larger fragments.
• Where there were fewer coyotes, there were more mesopredators. The link was strongest for grey foxes, cats and opossums, and held even after allowing for fragment area, age and isolation.
• In fragments that coyotes never visited, total mesopredator abundance was more than twice as high as in fragments where coyotes were detected at least once.
• Cats were more common in small fragments, which have more urban edge and so more neighbouring pet cats.

Avoiding coyotes in time and space

• In 13 fragments that coyotes visited only some of the time, mesopredators were more abundant in the seasons when coyotes were absent. This was significant for foxes, cats and skunks, but not for opossums or raccoons.
• In the five fragments surveyed for two years, mesopredator visits rose as coyote visits fell, driven mostly by cats.
• In 11 fragments where coyotes were always present, coyotes and mesopredators visited the same track station on the same night less often than expected by chance. This was clear for foxes, weaker for opossums, and not significant for cats, skunks or raccoons.

Coyotes kill cats

• Cat remains were found in most fragments with coyotes, and in 21% of 219 coyote scats.
• 25% of radio-collared cats were killed by coyotes.
• Of 636 residents who answered the questionnaire, 71% knew coyotes were a threat to cats. 42% of cat owners in areas with coyotes said coyotes had attacked or killed their cats, and 46% kept their cats in more when they believed coyotes were around.

Cats and birds

• 32% of residents owned cats (1.7 cats per owner on average); 77% of owners let their cats outdoors, and 84% of outdoor cats brought kills home.
• Each hunting cat brought home on average 24 rodents, 15 birds and 17 lizards a year. The authors estimate that about 35 outdoor cats around a fragment of about 20 hectares with 100 neighbouring homes bring back about 840 rodents, 525 birds and 595 lizards a year, probably an underestimate. Such a fragment may hold only one or two pairs of native predators such as foxes or coyotes.
• Of 68 prey items identified, 67% of 26 rodents, 95% of 21 birds and 100% of 11 lizards were native species.

Birds

• Fragment area (positive) and fragment age (negative) were the strongest predictors of how many scrub bird species remained.
• Even so, fragments with more mesopredators, especially cats and raccoons, had fewer bird species, and fragments where coyotes were present or more abundant had more bird species.

3. What they concluded

• The disappearance of a dominant carnivore leads to more numerous and more active mesopredators, which put strong predation pressure on native prey.
• The changes in mesopredator activity as coyotes came and went, coyotes killing cats, and cats killing birds all strengthen this conclusion.
• The interactions between coyotes, cats and birds probably have the strongest impact on the decline and extinction of scrub-breeding birds. Perhaps the strongest effect of coyotes on cats is indirect: owners keep cats inside when coyotes are about.
• Pet cats are “recreational hunters”, fed by their owners far above what the land could support, so they keep killing prey even when prey are scarce.
• This level of bird predation appears unsustainable: some bird populations in small fragments number 10 or fewer, and at least 75 local bird extinctions may have occurred in these fragments over the past century.
• These predator interactions combine with fragmentation effects (area and age) to help shape the community.

4. How tracking was or was not used

Tracks were the main source of carnivore data. Track detection stations were set every 250 m along transects, and each species’ relative abundance was the number of visits to the stations divided by effort. The paper does not describe the stations or the tracking surface. Tracks at the same station on the same night were used to test whether mesopredators avoided coyotes. Scats confirmed which species were present, and coyote scats were checked for cat remains.

5. Tracker or observer certification

Not reported. The paper does not say who identified the tracks or how their identifications were checked, although all carnivore abundance results rest on telling coyote, fox, cat, skunk, raccoon and opossum tracks apart. Presence was also checked with scats and cameras. The bird surveys were done by “trained observers”.

6. Other technologies, skills or innovations

• Remotely triggered cameras to confirm species presence.
• Radio-collars on domestic cats.
• Questionnaires to residents bordering the fragments.
• Fragment areas measured from digitized, scaled aerial photographs (1995).

7. Implications for conservation (from the paper)

• Large carnivores are especially vulnerable to extinction in fragmented landscapes, and their loss can cascade down to birds and other small animals.
• The study supports the mesopredator release hypothesis, which underpins conservation programmes that justify protecting carnivores.
• In small habitat fragments, even modest increases in predation by smaller carnivores, together with other fragmentation effects, may quickly drive native prey, especially rare species, to local extinction.
• Pet cats around fragments kill large numbers of native rodents, birds and lizards.