Power to detect trends in Missouri River fish populations within the Habitat Assessment Monitoring Program

Abstract

As with all large rivers in the United States, the Missouri River has been altered, with approximately one-third of the mainstem length impounded and one-third channelized. These physical alterations to the environment have affected the fish populations, but studies examining the effects of alterations have been localized and for short periods of time, thereby preventing generalization. In response to the U.S. Fish and Wildlife Service Biological Opinion, the U.S. Army Corps of Engineers (USACE) initiated monitoring of habitat improvements of the Missouri River in 2005. The goal of the Habitat Assessment Monitoring Program (HAMP) is to provide information on the response of target fish species to the USACE habitat creation on the Lower Missouri River. To determine the statistical power of the HAMP and in cooperation with USACE, a power analysis was conducted using a normal linear mixed model with variance component estimates based on the first complete year of data. At a level of 20/16 (20 bends with 16 subsamples in each bend), at least one species/month/gear model has the power to determine differences between treated and untreated bends. The trammel net in September had the most species models with adequate power at the 20/16 level and overall, the trammel net had the most species/month models with adequate power at the 20/16 level. However, using only one gear or gear/month combination would eliminate other species of interest, such as three chub species (Macrhybopsis meeki, Macrhybopsis aestivalis, and Macrhybopsis gelida), sand shiners (Notropis stramineus), pallid sturgeon (Scaphirhynchus albus), and juvenile sauger (Sander canadensis). Since gear types are selective in their species efficiency, the strength of the HAMP approach is using multiple gears that have statistical power to differentiate habitat treatment differences in different fish species within the Missouri River. As is often the case with sampling rare species like the pallid sturgeon, the data used to conduct the analyses exhibit some departures from the parametric model assumptions. However, preliminary simulations indicate that the results of this study are appropriate for application to the HAMP study design.