Patterns and processes in complex landscapes: testing alternative biogeographical hypotheses through integrated analysis of phylogeography and community ecology in Hawai’i JON ELDON,* † JONATHAN P. PRICE,* ‡ KARL MAGNACCA* § and DONALD K. PRICE* § *Tropical Conservation Biology and Environmental Science Graduate Program, University of Hawai’i, Hilo, HI 96720, USA, †Department of Environmental Studies, University of California, 1156 High Street, Santa Cruz, CA 95064, USA, ‡Department of Geography and Environmental Sciences, University of Hawai’i, Hilo, HI 96720, USA, §Department of Biology, University of Hawai’i, Hilo, HI 96720, USA Abstract The Island of Hawai’i is a dynamic assemblage of five volcanoes with wet forest habi- tat currently existing in four distinct natural regions that vary in area, age and geo- graphical isolation. In this complex landscape, alternative assumptions of the relative importance of specific habitat characteristics on evolutionary and ecological processes predict strikingly different general patterns of local diversity and regional similarity. In this study, we compare alternative a priori hypotheses against observed patterns within two distinct biological systems and scales: community composition of wet for- est vascular plant species and mitochondrial and nuclear genes of Drosophila sproati, a wet-forest-restricted endemic. All observed patterns display strong and similar regio- nal structuring, with the greatest local diversity found in Kohala and the windward side of Mauna Loa, the least in Ka’  u and Kona, and a distinctive pattern of regional similarity that probably reflects the historical development of this habitat on the island. These observations largely corroborate a biogeographical model that integrates multiple lines of evidence, including climatic reconstruction, over those relying on sin- gle measures, such as current habitat configuration or substrate age. This method of testing alternative hypotheses across biological systems and scales is an innovative approach for understanding complex landscapes and should prove valuable in diverse biogeographical systems. Keywords: biogeography, community ecology, Drosophila, Hawai’i, hypothesis testing, phyloge- ography Received 9 May 2012; revision received 12 March 2013; accepted 14 March 2013 Introduction Observable patterns in biogeographical systems are the result of ecological and evolutionary processes that are heavily influenced by the spatial and temporal charac- teristics of the associated habitat. These characteristics are often described through simple proxies, such as current habitat patch size and intervening geographical distance, which are then used to infer relevant processes from observed patterns. However, in complex land- scapes, these habitat characteristics may be dynamic over space and time, with the associated processes dri- ven by the interaction of a variety of factors. In this case, simple proxies of habitat characteristics may be mislead- ing. This study applies an innovative method for understanding biogeographically complex landscapes, in which alternative general hypotheses of relevant habi- tat characteristics are tested by integrating observable patterns across multiple biological systems and scales. The resulting biogeographical model of the landscape provides a corroborated general framework for Correspondence: Jon Eldon, Fax: 831-459-4015; E-mail: jeldon@ucsc.edu © 2013 Blackwell Publishing Ltd Molecular Ecology (2013) doi: 10.1111/mec.12326