Tp preserve approximate stationarity, all time series were truncated to post-2000 due to an apparent regime shift. I was able to quantify previous patterns for predicting the Columbia run, giving a fairly precise answer. Patterns were still useful for the Fraser system, but were treated qualitatively. Patterns were present in the Bristol Bay system before 2000, but are no longer useful, so I used simple regression to extrapolate the strong temporal trend.