Assumptions and evidence for the existence of metapopulations in nature

The different types of metapopulations described above (mainland-island, classical, source-sink, patchy, and non-equilibrium) are all variations on the same themes. Local extinctions are commonplace, there is an equilibrium involving colonization and extinction rates, and so on. They differ in the levels of detail, whether they allow for patches to be of different quality, whether they allow for differing levels of connectivity between patches, and whether they include local population dynamics.

What are the general assumptions we are making in all of these models? How can we demonstrate that the long-term persistence of a species in a landscape is due to metapopulation, rather than local population, processes? One difficulty is that long-term population data on patch occupancy are hard to gather. For this reason, many studies of population persistence in a fragmented landscape, meeting the criteria for a metapopulation, come from short-lived, easily monitored organisms. Accordingly, there has been an emphasis on populations of butterflies such as the Glanville fritillary (Melitaea cinxia) and the bay checkerspot (Euphydryas editha) (Hanski 1999, Ehrlich and Hanski 2004). On the other hand, metapopulation theory was famously, though perhaps inappropriately, used in designing a conservation plan for the northern spotted owl (Strix occidentalis) (Boyce 2002), and is being implemented in the management of both black and golden lion tamarin (Leontopithecus chrysomelas and L. rosalia) populations in South America (J. Mickelberg, personal communication).

Assumptions that metapopulation dynamics are decisive to the structure of the regional population include (Hanski and Kuussaari 1995, Hanski 1999):

1 The species has local breeding populations in relatively discrete habitat patches. This condition stresses that the population is spatially structured and therefore most individuals interact with others only in the local habitat patch.

2 No single local population is large enough to have a longer expected lifetime than the expected lifetime of the metapopulation itself. This excludes mainland-island populations.

3 Empty habitat patches are common. In the Glanville fritillary butterfly study in Finland, for example, 70% of approximately 1600 habitat patches have been empty at a given time (Hanski et al. 1995).

4 The habitat patches are not too isolated to prevent re-colonization. Long-distance movements may be facilitated by habitat corridors or other mechanisms.

5 Local dynamics are sufficiently asynchronous to make simultaneous extinction of all local populations unlikely. With complete synchrony, the metapopulation only lasts as long as the local population with the lowest chance of extinction. The greater the asynchrony, the longer the metapopulation is likely to last. In a recent review of the literature, Elmhagen and Angerbjorn (2001) found eight studies (four insect species and four small mammal species) in which asynchrony of population dynamics among patches was confirmed.

6 Population turnover, local extinctions, and the establishment of new populations form the basis for metapopulation dynamics, and metapopulations persist despite population turnover. Elmhagen and Angerbjorn (2001), in a review of the literature, found 22 studies confirming turnover. See supportive data below.

7 Population size or density is significantly affected by migration. This is the basis for source-sink populations and the rescue effect (Pulliam 1988, 1996).

8 Population density, colonization rate, and extinction rates are all affected by patch size and isolation.

9 Metapopulations can affect competitive, predator-prey, and parasite-host interactions. These ideas were discussed at the beginning of this chapter and will be elaborated in later chapters.

Reviews of metapopulation studies by Elmhagen and Angerbjorn (2001) and Harrison (1991, 1994) found that many of these criteria are frequently not met in the published metapopulation literature. Nevertheless, there is an extensive literature, particularly from studies of butterflies, supporting most of these assumptions (Ehrlich and Hanski 2004).

Supportive field studies

1 Boycott (1930) studied freshwater mollusk populations in 84 ponds in England. Over a 10-year period he recorded 64 extinctions and 93 colonizations of 18 species.

2 In their study of arthropod populations on red mangrove islands, Simberloff and Wilson (1969, 1970) removed all arthropod species by fumigation. After re-colonization of the mangroves by the arthropods, they found an equilibrium of between 20 and 40 species per island, depending on island size. But the turnover rate of species was approximately 2% of the species pool per day.

3 As described at the beginning of this chapter, after the explosion on Krakatau the number of bird species on Rakata seemed to reach equilibrium at 30 by 1934. Yet a consistent turnover has continued.

4 Hanski et al. (2004) described an experiment with the Glanville fritillary in which 10 local populations, derived from 72 larval groups, were transported from the Finnish island of Aland to the island of Sottungia in August 1991. Sottungia is a 4 km by 2 km island containing 20 small meadows suitable for this butterfly, although there were none at the time of the introduction. An

Table 5.4 The number of surviving local populations, the number of extinctions, and the number of colonizations per year for the metapopulation of Melitaea cinxia on Sottungia island in the Baltic Sea off the coast of Finland. Extinctions + colonizations = turnover events. Larvae were transported in August 1991. Adapted from Hanski et al. (2004).

Table 5.4 The number of surviving local populations, the number of extinctions, and the number of colonizations per year for the metapopulation of Melitaea cinxia on Sottungia island in the Baltic Sea off the coast of Finland. Extinctions + colonizations = turnover events. Larvae were transported in August 1991. Adapted from Hanski et al. (2004).

Year

Number of

Number of

Number of

Total number of

local populations

extinctions

colonizations

turnover events

1991

10

-

-

-

1992

5

5

0

5

1993

5

1

1

2

1994

6

2

3

5

1995

3

3

0

3

1996

6

0

3

3

1997

10

0

4

4

1998

14

0

4

4

1999

2

12

0

12

2000

11

0

9

9

examination of Table 5.4 shows that this metapopulation has persisted, in spite of the fact that none of the original 10 populations lasted for the entire 11 years of the study. These data confirm the persistence of the metapopulation as a stochastic balance between local extinctions and re-colonizations of available habitat patches. Note that we are only keeping track of presence or absence of the population in a habitat; we are not making an assessment of local population numbers or local population dynamics, and there is no assessment of habitat quality.

5 Long-term work on the same species of butterfly on the Aland archipelago (Nieminen et al. 2004) over an eight-year period found: (i) the number of extinct populations varied from 131 to 234 per year; (ii) the number of colonizations ranged from 97 to 230 per year; (iii) the total number of extant populations varied from 303 to 496 per year; and (iv) the number of empty patches varied from 749 to 3507 per year.

6 Crone et al. (2001) examined the six-year data gathered by Pokki (1981) on vole (Microtus agrestis) populations found on the Tvarminne archipelago in Finland. Extinction and re-colonization of local island subpopulations were common, and a spatially explicit model such as the incidence function model provided reasonable predictions of the structure and function of this metapopulation. However, contrary to assumptions, the mainland did not prove to be a significant source of dispersing animals; in fact, an important source of immigrants to the larger islands came from tiny, ephemeral populations found on small islands. While these small populations were unlikely to persist, they were an important source of immigrants to larger islands on the archipelago. Crone et al. also found that the parameters fitted to the incidence function models varied dramatically from year to year. Most importantly, this study suggests that the "rescue effect" can be turned around such that a small ephemeral population may help stabilize a larger, more permanent population.

Was this article helpful?

0 0

Post a comment