Premium
SINGLE‐SPECIES DYNAMIC SITE SELECTION
Author(s) -
Moilanen Atte,
Cabeza Mar
Publication year - 2002
Publication title -
ecological applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.864
H-Index - 213
eISSN - 1939-5582
pISSN - 1051-0761
DOI - 10.1890/1051-0761(2002)012[0913:ssdss]2.0.co;2
Subject(s) - metapopulation , snapshot (computer storage) , selection (genetic algorithm) , computer science , butterfly , ecology , representation (politics) , endangered species , mathematical optimization , biology , machine learning , mathematics , population , habitat , biological dispersal , demography , sociology , politics , political science , law , operating system
Methods for designing regional reserve networks mostly concentrate on providing maximal representation of species occurring in the region. Representation‐based methods, however, typically consider a static snapshot of species incidences, and the spatial dynamics of the species are ignored. It has been empirically demonstrated that reserves designed using representation do not guarantee another important goal of reserve design: long‐term persistence. The question studied here is the following: Which subset of sites do you select to maximize the long‐term persistence of a species living in a metapopulation, given that each site has a cost and the amount of resource (e.g., money) available is limited? We present an optimization method, which uses a combination of evolutionary optimization (a genetic algorithm) and local search to find the optimal selection of sites. The quality of each candidate solution is evaluated using a spatially realistic metapopulation model, the incidence function model. The proposed method is applied to a metapopulation of the false heath fritillary butterfly, an endangered species in Finland. With this data set, the proposed estimation method produces intuitively acceptable and consistent results within minutes of computation time. Sites favored by the algorithm are located in three patch clusters, and they tend to be inexpensive and initially occupied. Expensive and/or very isolated patches are rarely selected into the optimal site selection.