| Since physical and biological responses to ecological restoration are influenced by processes acting at multiple spatial and temporal scales, attributing change to habitat restoration requires separating forcing functions imposed by restoration from those occurring naturally. A framework is proposed to separate factors influencing ecological response into four levels: external physical forcing functions, restoration-induced functions, and physical and biological response variables. The approach is illustrated through application to the Lower Red River Meadow Stream Restoration Project in north-central Idaho, USA, a soft-engineered restoration project designed to improve habitat for chinook salmon (Oncorhynchus tshawytscha) by restoring natural, physical processes to a 4.1 km reach. Field monitoring and hydrodynamic modeling were combined to quantify the magnitude and direction of physical and biological responses. While the direction of physical response to individual treatments was predictable in most cases, the magnitude of change and cumulative physical response, especially from treatments with potentially offsetting effects, were less predictable. Furthermore, reach-median values demonstrated overall trends, but the direction and magnitude of responses varied between types of restored channel. Even though many active stream restoration projects are funded and implemented on the premise that restoration of physical processes will result in improved biological conditions, biological responses due directly to restoration are difficult to quantify because of their inherent high variability. Power to detect response to habitat restoration is a function of effect size, variability, and sample size. From observed variability in physical and biological parameters monitored for 3–9 years at the project reach, magnitude of detectable response was estimated as a function of years of monitoring. With five years of post-restoration monitoring, biological parameters (e.g., age 0 chinook density) would require an order of magnitude larger response than physical parameters (e.g., median particle size) to be statistically detectable. By combining the proposed framework with the Before-After, Control-Impact (BACI) sampling design, physical parameters can be used to detect initial responses to restoration and complement biological parameters in the long-term evaluation of stream restoration. |
