Masters Thesis

A comparison of escapement estimate methods plus escapement-recruitment relationships for Chinook salmon and coho salmon in a coastal stream

This study compares methods for estimating escapement of Chinook salmon (Oncorhynchus tshawytscha) and coho salmon (Oncorhynchus mykiss) in a coastal stream without a counting fence or weir. Relationships between escapement estimates and juvenile recruitment were also analyzed. A six-year data set, collected from Prairie Creek, California, spanning the 1998/1990 through 2003/2004 spawning seasons, was used for this project. Five methods of escapement estimation were considered: area-under-the-curve of live spawner observations, carcass mark-recapture using a modified Jolly-Seber method, carcass mark-recapture using the Petersen method, a redd count based estimate assuming one redd per female, and a redd count based estimate using a redd-area method. A logistic regression model was developed to distinguish Chinook salmon from coho salmon redds for those redds which did not have a known associated species when observed. The model predicted species with 93.3% accuracy when time (water year day) was the only available parameter and 98.7% to 99.4% concordance when time and other redd measurements were available. Chinook salmon area-under-the-curve and one redd per female escapement estimates were stronger correlated in Upper Prairie Creek (Reaches 3 and 4, R-square = 0.9145) than the entire surveyed mainstem Prairie Creek (Reaches 1 through 4, R-square = 0.4789). Coho salmon area-under-the-curve escapement estimates were linearly correlated with one redd per female (R-square = 0.7492) and redd-area estimates (R-square = 0.6922) for the Prairie Creek Study Area. Carcass estimates were inconsistent with area-under-the-curve and redd-based estimates. Linear regression analyses of area-under-the-curve estimates and juvenile recruitment yielded R-square values of 0.3780 for Chinook salmon in Mainstem Prairie Creek and 0.9079, 0.3035, and 0.6681 for coho salmon in Upper Prairie Creek, Streelow Creek, and Boyes Creek, respectively. Linear regression analyses of redd-based escapement estimates and juvenile recruitment yielded R-square values less than 0.0258 for Chinook salmon in Mainstem Prairie Creek and between 0.6422 and 0.9169 for coho salmon in Upper Prairie Creek, Streelow Creek, and Boyes Creek.

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.