Summary: | Native faunas and floras are especially susceptible to negative effects of invasive alien species in islands. The world´s fourth largest island, Madagascar, has very unique biota with high level of endemism. The black rat, Rattus rattus, is claimed to cause more extinctions of insular vertebrates than any other introduced rodent in the world. On Madagascar, R. rattus is suggested to competete with native rodents belonging to the endemic subfamily Nesomyinae. Extensive deforestation and fragmentation also threatens Malagasy forest-dwelling species. The aims of this thesis were to 1) study the occurence of native and introduced species and the different kind of factors determining occurence in southeastern Madagascar and 2) help to target future studies by estimating which native rodents are the most potential competitors with introduced rodents on Madagascar. Habitat use data of black rat and endemic rodents were collected both in fragmented and in unfragmented forest in Ranomafana National Park in southeastern Madagascar. A total of 698 rodent individuals were captured in 6204 trap nights. Logistic and Poisson regression models were used to determine the factors that influence the presence and abundance of rodent species and to investigate how sensitive one variable can be to other variables in regression models. This led to the introduction of a new approach called explanatory framework based regression analysis (EFRA) which rests on the social science based elaboration technique. EFRA enables systematization of the link between ecological knowledge and statistical analysis. From the point of it multicollinearity is more source of information than a problem for data analysis. The abundance of R. rattus increased with increasing forest disturbance. The spread of R. rattus was suggested to be associated with deforestation but not directly with fragmentation. It is not surprising when remembering that R. rattus utilizes open areas too. The measured value of the size of the fragment can be viewed as an index ...
|