Avian persistence in fragmented rainforest

Science

Volumn 298, Issue 5596, 2002, Pages 1236-1238

  Lens, Luc ^a   Van Dongen, Stefan ^a   Norris, Ken ^a   Githiru, Mwangi ^a   Matthysen, Erik ^a  

^a GHENT UNIVERSITY   (Belgium)

Author keywords

[No Author keywords available]

Indexed keywords

BIODIVERSITY; DATA ACQUISITION; LANDFORMS; MUSEUMS; RAIN;

AVIAN PERSISTENCE; PHENOTYPIC DEVELOPMENT;

FORESTRY;

AVIFAUNA; DISPERSAL; HABITAT FRAGMENTATION; RAINFOREST; SPECIES DIVERSITY; SPECIES RICHNESS;

ARTICLE; BIODIVERSITY; BIRD; HABITAT FRAGMENTATION; LANDSCAPE; MEASUREMENT; NONHUMAN; PRIORITY JOURNAL; RAIN FOREST; SPECIES DIFFERENTIATION;

ANIMALS; BIRDS; CONSERVATION OF NATURAL RESOURCES; ECOSYSTEM; ENVIRONMENT; KENYA; LOGISTIC MODELS; MODELS, BIOLOGICAL; MODELS, STATISTICAL; PHENOTYPE; POPULATION DYNAMICS; PROBABILITY; SPECIES SPECIFICITY; TARSUS, ANIMAL; TREES;

KENYA;

AVES;

EID: 0037044772     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.1075664     Document Type: Article

References (26)

1. G. R. Robinson, F. R. Thompson III, T. M. Donovan, D. R. Whitehead, J. Faaborg, Science 257, 524 (1992).
2. M. L. Hale et al., Science 293, 2246 (2001).
3. R. S. Etienne, J. A. P. Heesterbeek, Am. Nat. 158, 389 (2001).
4. N. Myers, R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, J. Kent, Nature 403, 853 (2000).
5. A selection of maps of the study area, illustrating its global position, its elevational structure, and spatial relationships of the forest fragments, can be found at http://bio-www.uia.ac.be/bio/deco/topics.html#study areas.
6. P. Dias, Trends Ecol. Evol. 11, 326 (1996).
7. tot = 9) with confirmed breeding activity during one or more breeding seasons between 1996 and 2001.
8. G. White, K. P. Burnham, Bird Study 546, 120 (1999).
9. Individual recapture histories were generated using time intervals of 1 month. Such a short time interval was needed because over longer intervals (3, 6, or 12 months), a number of individuals were recaptured in multiple forest fragments within a single interval A total of 889 individuals were captured-recaptured in two or more months, and among these, 47 individuals were captured-recaptured in two or more fragments.
10. We used Akaike's Information Criteria (AIC) to determine the most parsimonious model among a candidate set and computed species-specific dispersal probabilities for further analysis from this model. Recapture data were too sparse to include fully time-dependent models of survival and recapture probabilities in the candidate set of models, and sample size did not allow us to calculate complex multistrata models for Apalis thoracica. For this species, no single individual was captured in more than one fragment, hence Ψ was estimated as 0. There is no formal goodness-of-fit test for multistrata models, yet our results were robust for a range of 1 to 1.5 in the overdispersion parameter ĉ.
11. More complex dispersal models could not be fitted to all species in MARK because models were overparameterized. For the more mobile species with adequate numbers of recaptures after dispersal between fragments, a model assuming direction-dependent dispersal rates between fragment pairs fitted the recapture data significantly better than the direction-independent model described in Table 1 for two species, but equally as well as the direction-independent model for two other species (nested models were compared using likelihood ratio tests).
12. L. Van Valen, Evolution 16, 125 (1962).
13. L. Lens, S. Van Dongen, S. Kark, E. Matthysen, Biol. Rev. 77, 27 (2002).
14. S. Van Dongen, G. Molenberghs, E. Matthysen, J. Evol. Biol. 12, 94 (1999).
15. L. Lens, S. Van Dongen, C. M. Wilder, T. M. Brooks, E. Matthysen, Proc. R. Soc. London Ser. B 266, 1241 (1999).
16. The validity of this relationship depends on the assumption that tarsi develop in the patch where they are measured. Violation of this assumption, however, can only bias results against the hypothesis tested in this paper.
17. No museum samples were available for A. thoracica (rare species) and Nectarinia olivacea (abundant species yet not collected in the Taita Hills).
18. 1 = 4.1, P < 0.05) in a three-way mixed analysis of variance with unsigned FA as the dependent variable; fragment as a fixed factor; and species, time, and all relevant two- and three-way interaction terms as random factors [model specifications in (15)].
19. L. Lens, S. Van Dongen, K. Norris, M. Githiru, E. Matthysen, data not shown.
20. Logistic regression analysis was performed in WINBUGS version 1.3 to allow for missing values and use of all available data [W. R. Gilks, A. Thomas, D. J. Spiegelhalter, The Statistician 43, 169 (1994)]. The occupancy-dispersal association was based on eight species and the occupancy-asymmetry association on six species. An extra dispersion variance parameter [parameter estimate 0.22 ± 0.71, 95% confidence interval (Cl) 0.001 to 1.63] was added to correct for overdispersion (25).
21. S. L. Lima, P. A. Zollner, Trends Ecol. Evol. 11, 131 (1996).
22. C. Carbone, J. L. Gittleman, Science 295, 2273 (2002).
23. G. M. Clarke, Conserv. Biol. 9, 18 (1995).
24. L. Lens, S. Van Dongen, E. Matthysen, Conserv. Biol. 16, 479 (2002).
25. P. Congdon, Bayesian Statistical Modelling (Wiley, Chichester, UK, 2001).
26. We thank B. Amakobe, J. Barnes, R. Barnes, T. Brooks, A. Cooke, D. Gitau, T. Imboma, C. Jackson, J. Kageche, S. Karimi, J. Lindsell, G. Mwachala, D. Samba, E. Waiyaki, C. Wilder, and the staff of the Taita-Taveta District Forest Department, the Ornithology Department of the National Museums of Kenya, and the East African Herbarium for help with fieldwork. B. Bytebier and M. De Meyer of the Taita Hills Biodiversity Project provided logistic support. L. Bennun, P. Sweet, J. Bates, D. Willard, R. Paynter, and F. Sibley allowed access to the historical specimens. A. Notenbaert and S. Vanlishout provided help with the compilation of maps of the study area. Supported by the Fund for Scientific Research Flanders (L.L., S.V.D.) and Flemish Interuniversity Council project 02/6/7-338-607 (E.M.).