NESTED HIERARCHIES AND SCALE-DEPENDENCE OF MECHANISMS OF FLOW REFUGIUM USE

We outline a hierarchical framework of classes of mechanisms, operatin g at different spatial and temporal scales, by which populations may e xploit environmental heterogeneity to ensure persistence in environmen ts subject to physical disturbance. Examples are provided for stream s ystems subject to high flow disturbances. Genotypic heterogeneity, enc ompassing morphological and physiological adaptations, operates at evo lutionary and biogeographic scales, and ultimately constrains mechanis ms at smaller scales. On ecological scales, organisms may exploit temp oral or spatial heterogeneity, or both, to maintain population sizes l arger than possible by morphological adaptations alone. For 2 classes of mechanism operating at large scales (>1 generation and >1 habitat p atch), individuals do not survive disturbances and population persiste nce depends on recruitment from external sources. Where generations ov erlap and populations exploit temporal refugia in complex life cycles, dispersal occurs between habitat patches of different types. Alternat ively dispersal occurs between habitat patches of similar type. For 2 other classes of mechanism, operating at smaller scales (<1 generation and less than or equal to 1 habitat patch), individuals survive distu rbances by moving between microhabitat patches of different types, or by changes in habitude which reduce the negative effects of disturbanc e. A simple mathematical model is used to explore the efficacy of 4 me chanisms of refugium use operating through microhabitat heterogeneity within a single habitat patch. Simulations of population changes over a series of disturbance events showed that each of 4 mechanisms could maintain a viable population under certain conditions. Total refugium area and proportion of the population lost at each disturbance event h ad a strong effect on final population size for all mechanisms. Doubli ng disturbance frequency had little effect on final population size if movements of individuals into refugia were facilitated by small acces sible refugium patches, or if individuals remained in refugia during b enign inter-disturbance periods. Empirical evidence supports the effic acy of these mechanisms: organisms with short life cycles and, perhaps , poor mobility may be lost from erosive patches but persist in refugi a; macroinvertebrates may accumulate in refugia during disturbance eve nts and redistribute throughout the stream after the disturbance. More empirical data on the spatial attributes of refugium patches and the mobility of stream organisms are required to test and to add realism t o these models.