Effects of variability and synchrony in assessing contributions of individual streams to habitat portfolios of river basins

Abstract

Despite growing recognition that complex population portfolios are important sources of ecosystem stability and resilience, the structure of such stock complexes is often not considered in monitoring schemes to inform environmental impact assessments. The use of “index” or “indicator” sites, and of short assessment windows, to make broad claims about the relative importance of specific habitats or populations within a portfolio, assume that regional populations are highly synchronous and that abundance is stable through time. Asynchrony among populations and changes in population productivity over time render these assumptions fundamentally flawed. We used 57 years of abundance data for sockeye salmon (Oncorhynchus nerka) spawning in a set of eight streams in the Wood River watershed, southwest Alaska, to demonstrate how natural patterns of variability affect the ability of fixed assessment windows to characterize the contribution of individual populations to the entire portfolio. Additionally, simulations were used to demonstrate how different levels of synchrony and autocorrelation affect the ability of monitoring schemes to estimate the contributions of individual populations to a portfolio over the long-term. We found that fixed assessment windows were distinctly limited in their ability to characterize a population’s contribution to a portfolio over the long-tem. Asynchronous dynamics among populations, and the presence of autocorrelation that creates slow changes in populations, weaken the ability to characterize a stream's potential contribution to a portfolio from short-term assessments. These results suggest that the structure of population portfolios, and the presence of directional changes in productivity within individual populations, need to be taken into account when carrying out environmental risk assessments that aim to measure the contribution of an individual population or habitat to system wide dynamics. Typical risk assessments that depend on short monitoring periods are likely to vastly underestimate the potential long-term value of any specific habitat, and the population it supports, as a component of a multi-population portfolio.