Research Document - 2012/145
Pre-season run size forecasts for Fraser River Sockeye (Oncorhynchus nerka) and Pink (O. gorbuscha) salmon in 2013
By S.C.H. Grant and B.L. MacDonald
Abstract
Salmon forecasts remain highly uncertain, in large part due to wide variability in annual salmon productivity (recruits-per-spawner). For Fraser Sockeye, quantitative and qualitative indicators of productivity explored to date have not reduced forecast uncertainty and remain an active area of research. Fraser Sockeye forecasts have been particularly uncertain in recent years, due to the systematic declines in productivity exhibited by most stocks, which culminated in one of the lowest productivities on record in the 2005 brood year (2009 four year old and 2010 five year old returns). Subsequently (2010 to 2012 return years), productivity appears to have improved.
Similar to previous years, models selected for each stock in the 2013 forecast were chosen for their relative ability to predict true returns over the full stock-recruitment time series. Each model was compared to a suite of forecast models, which exclude the recent productivity model forms introduced in the 2010 forecast. Comparative rankings of model performance were obtained from the 2012 jackknife analysis output. A sensitivity analysis was also conducted to compare the 2013 forecasts with forecasts produced using top ranked models evaluated over the low productivity period (brood years 1997 to 2005) and includes 2010 recent productivity models.
To capture inter-annual random (stochastic) variability in Fraser Sockeye productivity, forecasts are presented as standardized cumulative probabilities (10%, 25%, 50%, 75%, and 90%). The 2013 forecast indicates a one in ten chance (10% probability) the total Fraser Sockeye return will be at or below 1,554,000 (lowest observed on this cycle) and a nine in ten chance (90% probability) it will be at or below 15,608,000, assuming productivity is similar to past observations. The mid-point of this distribution (50% probability) is 4,765,000 (there exists a one in two chance the return will be at or below this value). The four year old percentage of the total forecast is 91%, and ranges from 13% to 100%, depending on the stock. For the 2013 forecast, Raft, Harrison, and North Thompson miscellaneous stocks were moved into the Summer Run timing group, due to changes in the migration timing of these stocks, which is consistent with a recent decision by the Fraser Panel of the Pacific Salmon Treaty process.
Summer Run stocks, particularly Chilko & Quesnel, contribute 78% to the total return forecast, whereas Late Run (12%), Early Summer (5%) and Early Stuart Run stocks (4%) each contribute considerably less. The Harrison forecast for 2013 is particularly uncertain, and the return for this stock could fall outside the forecast range. The total forecasted 2013 Fraser Sockeye return largely falls (up to a three in four chance, based on past observations) below the cycle average (8.6 million), due to the below average 2009 and 2008 brood year escapements for most stocks. Conversely, there is a one out of four chance the return will be above the cycle average, if Fraser Sockeye productivity falls at the high end of past observations. If low productivity conditions resume, returns could be considerably lower than forecast, based on a sensitivity analysis forecast that ranges from 523,000 to 5,419,000 at the 10% to 90% probability level.
For Fraser Pink Salmon, the forecast ranges from 4,794,000 to 17,111,000 fish at the 10% to 90% probability levels (Table 1). The median (50% probability) forecast of 8,926,000 Pink Salmon is below the long-term (1959-2011) average return (12.6 million) (Table 1). The Fraser Pink forecast is highly uncertain given the changes in the methods used to estimate returns through time.
Accessibility Notice
This document is available in PDF format. If the document is not accessible to you, please contact the Secretariat to obtain another appropriate format, such as regular print, large print, Braille or audio version.
- Date modified: