Human-induced alteration of biogenic habitats at the landscape scale is increasing and directly influences local diversity and system functioning (Vitousek et al. 1997, Duarte 2002, Lozte et al. 2006 and Airoldi & Beck 2007). Alteration of natural landscapes can result in the loss of entire habitats or the transition from more complex to less complex habitats, for example from shell reefs or seagrass beds to mudflats (Short and Wyllie-Echeverria 1996, Hughes et al. , Thrush et al. 2006). Habitat modifications can generate unsuitable habitat between isolated patches of contiguous habitat (MacArthur and Wilson 1967, Hanski 1994). These changes, in turn, have a direct effect on the dynamics of a community (total abundance, diversity and species richness), particularly on the design and influence of the functional roles of the species occupying the system (Gray 1997, Tilman et al. 1994, Loureau et al. 2001). For example, much of the temperate continental shelf is being homogenized through bottom trawling and dredging, resulting in a loss of individuals and functional groups (Thrush and Dayton 2002, Gray et al. 2006 ). The synergistic effects of habitat loss in quantity and quality, and the timing and pattern of habitat alteration can result in threshold levels of habitat loss below which diversity, abundance and wildlife survival can be resilient. Habitat loss has been indicated as one of the major threats to marine biodiversity (Gray, 1997), and temperate biogenic reefs are likely among the most threatened habitats in the world (Coleman and William 2002, Barbera et al. 2003, Airoldi and Beck 2007, Airoldi et al. These structurally complex habitats are declining in marine environments, in many spa... middle of paper ......does the loss of kelp forest cover and the spatial scale over which it occurs influence community dynamics associated in the canopy? (2) How do alternative undisturbed habitats (undisturbed canopy or subcanopy kelp habitats) buffer the invertebrate community after disturbance? (3) Does the timing of habitat loss influence how community dynamics respond? (Moved to discussion) (3) How do community dynamics recover after habitat loss, and if so, (4) how long does recovery take? To answer these questions, we experimentally removed canopy habitat from replicated kelp forests. The manipulations were carried out at approximately the same spatial scale at which kelp forest covers are typically fragmented by human-induced alterations (e.g. mechanical harvesting activities; Bodkin 1988, Foster & Donnellan 1999).