From metapopulation theory to conservation recommendations: Lessons from spatial occurrence and abundance patterns of Maculinea butterflies

Abstract

Classic metapopulation theory assumes relatively frequent population extinctions and colonisations on local habitat patches and consequently its interest is focused on spatial presence–absence patterns rather than on local population densities and their dynamics. However, the latter may also be important for metapopulation functioning, especially in the case of low turnover of local populations. We investigated spatial occurrence and abundance patterns of three species of endangered Maculinea butterflies in the Kraków region, southern Poland, in relation to various habitat parameters. For all three species investigated we found almost complete occupancy of their foodplant patches. The few patches lacking Maculinea populations were significantly smaller and more isolated. Foodplant availability proved to be the main factor limiting population densities for M. alcon, but not for M. teleius and M. nausithous, for which patch size and shape mattered the most. Small and highly internally fragmented patches supported higher densities of these two species. We hypothesise that the negative density–area relationship as well as positive impact of patch fragmentation on butterfly densities derive from differences in relative abundance of Myrmica ant hosts, which are a vital resource for myrmecophilous Maculinea butterflies. Since ants are under strong parasite pressure from Maculinea within their foodplant patches, and spread mainly through nest budding from surrounding refuge areas, their densities can be expected to be higher on small and fragmented patches. This underlines the importance of not only foodplant patches, but also their surroundings for the conservation of Maculinea butterflies.

Introduction

With many species living nowadays in increasingly fragmented landscapes metapopulation theory has become a paradigm in biodiversity conservation (Hanski and Gaggiotti, 2004). In its classic form, exemplified by Levins’ (1970) model or Hanski’s (1994) incidence function model, this theory assumes high frequency of extinction and colonisation events in local habitat patches. Consequently its main interest is focused on spatial presence–absence patterns rather than on abundance patterns. Only recently there have been attempts to incorporate local population sizes and their dynamics into metapopulation models (Keeling, 2002, Ovaskainen and Hanski, 2004). On the other hand, many authors have argued that classic metapopulations are in fact rare in the real world (Harrison and Taylor, 1997, Szacki, 1999, Smith and Green, 2005). Butterflies are believed to be one of the groups that best fit the theoretical framework of the classic metapopulation theory, but even among them there are exceptions (Thomas and Hanski, 1997, Thomas and Hanski, 2004, Baguette, 2004).

Large Blue butterflies of the genus Maculinea, which over the last two decades have become the ‘flagships’ of biodiversity conservation in Europe (Thomas and Settele, 2004), obviously experience severe fragmentation of their habitats (Munguira and Martin, 1999). Nevertheless, based on their biology one may expect Maculinea not to be particularly prone to exist in classic metapopulations. They are characterised by unique life cycles requiring specific foodplants and host ants in the larval period (Elmes and Thomas, 1992, Thomas, 1995). Due to high density-dependant mortality experienced by larvae (Hochberg et al., 1992, Hochberg et al., 1994) populations of Maculinea are typically small, but remarkably demographically stable (Thomas et al., 1998a, Nowicki et al., 2005a, Nowicki et al., 2005b). Apart from this, mobility of Maculinea is very low compared to most other butterflies (Nowicki et al., 2005b, Nowicki et al., 2005c). Both features reduce the turnover of local populations and thus should increase the importance of their sizes and dynamics for the functioning of metapopulations.

The aim of the present study was to test (i) how well in reality Maculinea butterflies conform to the classic metapopulation theory, and (ii) which factors affect their occurrence and abundance patterns. Identifying such factors is per se crucial for Maculinea conservation, but it is also important to know whether the probability of occupancy of local habitat patches and population densities on them are determined by the same or different habitat parameters. The former situation would allow monitoring the success of conservation activities with presence–absence data, which are much easier to obtain than abundance estimates. It would also make possible to apply relatively simple classic metapopulation models, e.g. for testing future management scenarios. The opposite situation, in which occupancy probabilities and population densities are shaped by different habitat parameters, would obviously be less desirable for conservation, because it may involve conflicting management recommendations and the need to choose between different priorities, e.g. high occupancy rates vs. high local densities. To achieve our goals we surveyed occurrence and local densities of M. teleius, M. nausithous, and M. alcon on habitat patches in the Kraków region, southern Poland, where the three species live sympatrically in very large (except for M. alcon) metapopulations (Nowicki et al., 2005c, Nowicki et al., 2005d). Subsequently we analysed this information against a wide range of habitat parameters derived from a GIS analysis. In particular we were interested in assessing the relative importance of factors reflecting habitat quality of Maculinea foodplant patches versus spatial characteristics of these patches.