Assessment of river health: accounting for perturbation pathways in physical and ecological space

Summary 1. A full understanding of changes to river ecosystem structure and functioning along the continuum from relatively pristine to profoundly perturbed requires knowledge of physical, chemical and ecological properties at many spatial and temporal scales. Perturbations may span broad spatio‐temporal scales or be spatially and temporally discrete. 2. The relevant scale for study depends on the manner in which perturbations move through two kinds of pathways – through the physical space of the catchment hierarchy and through the ecological space of river food webs. Different kinds of perturbation (e.g. inputs of sediment, nitrogen or phosphorus, changes to shading or discharge) vary in the degree and manner in which they are propagated downstream and through food webs. 3. The fundamental importance of disturbance regimes and refugia to river health is now clear and managers need to take this into account when devising river management schemes. 4. A comparison of the physics, chemistry and ecology of streams in catchments of native tussock and exotic pasture grassland in New Zealand serves to emphasize (i) the extreme complexity of interacting factors resulting from land use change, (ii) the importance of disturbance regime (not encapsulated in biotic indices) and biogeographic patterns (usually unmeasured) in determining ecosystem structure and functioning, and (iii) the lack of concordance between any single index of health and various fundamental features of ecosystem functioning. 5. Practical considerations limit most evaluations of river health to a small suite of indices, though it is important that researchers continue to evaluate the spatial, temporal and biological limitations of these indices. The vagaries of history and geography (extinction and colonization trajectories in relation to natural disturbance regimes) and the complexity of interacting physical, chemical and ecological responses to perturbation suggest that multi‐scale, multi‐temporal studies of river function offer the best opportunity to evaluate river health.