Review of a model to assess stranding of juvenile salmon by ship wakes along the Lower Columbia River, Oregon and Washington

Long period wake waves from deep draft vessels have been shown to strand small fish, particularly juvenile Chinook salmon Oncorhynchus tschawytcha, in the lower Columbia River (LCR). The U.S. Army Corps of Engineers is responsible for maintaining the shipping channel in the LCR and recently conducted dredging operations to deepen the shipping channel from an authorized depth of 40 feet(ft) to an authorized depth of 43 ft (in areas where rapid shoaling was expected, dredging operations were used to increase the channel depth to 48 ft). A model was developed to estimate stranding probabilities for juvenile salmon under the 40- and 43-ft channel scenarios, to determine if channel deepening was going to affect wake stranding (Assessment of potential stranding of juvenile salmon by ship wakes along the Lower Columbia River under scenarios of ship traffic and channel depth: Report prepared for the Portland District U.S. Army Corps of Engineers, Portland, Oregon). The U.S. Army Corps of Engineers funded the U.S. Geological Survey to review this model. A total of 30 review questions were provided to guide the review process, and these questions are addressed in this report. In general, we determined that the analyses by Pearson (2011) were appropriate given the data available. We did identify two areas where additional information could have been provided: (1) a more thorough description of model diagnostics and model selection would have been useful for the reader to better understand the model framework; and (2) model uncertainty should have been explicitly described and reported in the document. Stranding probability estimates between the 40- and 43-ft channel depths were minimally different under most of the scenarios that were examined by Pearson (2011), and a discussion of the effects of uncertainty given these minimal differences would have been useful. Ultimately, however, a stochastic (or simulation) model would provide the best opportunity to illustrate uncertainty within a given set of model predictions, but such an approach would require a substantial amount of additional data collection. Several review questions focused on the accuracy and precision of the model estimates, but we were unable to address these questions because of the limited data that currently exists regarding wake stranding in the LCR. Additional field studies will be required to validate findings from Pearson (2011), if concerns regarding accuracy and precision remain a priority. Although the Pearson (2011) model provided a useful examination of stranding under pre-construction and post-construction conditions, future research will be required to better understand the effects of wake stranding on juvenile salmonids throughout the entire LCR. If additional information on wake stranding is desired in the future, the following topics may be of interest: (1) spatial examination of wake stranding throughout the entire LCR; (2) additional evaluation of juvenile salmonid behavior and population dynamics; (3) assessing and integrating predicted changes in ship development; and (4) assessing and integrating predicted changes in climate on environmental factors known to cause stranding.