A comparison of methods for estimating juvenile salmon habitat capacity to assist with restoration planning and evaluation

Abstract

AbstractObjectiveA variety of analytical approaches have been developed in recent years to estimate salmon Oncorhynchus spp. freshwater habitat capacity to assist with planning and evaluating habitat restoration. We compared and contrasted seven different methods for estimating juvenile Chinook Salmon O. tshawytscha habitat rearing capacity (total number of fish that a given area of habitat can support), abundance, and habitat suitability in a 3.2‐km reach of the Grande Ronde River, Oregon.MethodsThis included approaches that used statistical methodologies applied to existing empirical data sets, including quantile random forests (QRFs) and spatial stream networks (i.e., Fish Data Analysis Tool [FDAT]), and deterministic models that included the habitat suitability index, the unit characteristic method (UCM), and three habitat expansion methods.ResultWe hypothesized that the methods would provide comparable estimates. However, the approaches provided very divergent estimates for the same stream reach, ranging from 1048 to 24,530 juvenile Chinook Salmon. The simple habitat expansion methods appeared to be the most similar to the actual parr (juvenile) abundance obtained from recent snorkel surveys (4179 juveniles). The fish density data (fish/m2) and habitat data used in the models were likely the main drivers of the differences between the UCM and habitat expansion methods.ConclusionThe habitat expansion approaches, which are based on detailed habitat surveys, appear to be particularly suitable for estimating capacity within a reach; thus, they may be suitable for evaluating changes due to habitat restoration. The model‐based approaches (QRF and FDAT) appear to yield coarser estimates, in part because they use various inputs at the reach and landscape scales. As such, they are more useful for relative comparisons among reaches in watersheds to assist with watershed‐scale restoration planning and prioritization. Because all of the methods rely on fish density in different types of habitats, regional or local data on fish densities would improve their accuracy.