2006 Bristol Bay Sockeye Salmon Fishery Forecast

Issued Tuesday, November 15, 2005

FORECAST AREA: Bristol Bay

SPECIES: Sockeye Salmon

FORECAST OF 2006 RUN:

METHODS

The forecast for the sockeye salmon run to Bristol Bay in 2006 is the sum of individual predictions for nine river systems (Kvichak, Alagnak, Naknek, Egegik, Ugashik, Wood, Igushik, Nushagak-Mulchatna, and Togiak) and four age classes (ages 1.2, 1.3, 2.2, and 2.3, plus ages 0.3 and 1.4 for Nushagak River). Adult escapement and return data from brood years 1976-2002 were used in the analyses.

Predictions for each age class returning to a river system were calculated from models based on the relationship between adult returns and spawners or siblings from previous years. Tested models included simple linear regression, multiple regressions, and 5-year averages. In addition, univariate and multivariate time series analysis models were examined. The models chosen were those with statistically significant parameters having the greatest past reliability (accuracy and precision) based on mean absolute deviation, mean absolute percent error, and mean percent error between forecasts and actual returns for the years 2003 through 2005.

The forecast range was the upper and lower values of the 80% confidence bounds for the total run forecast. The confidence bounds were calculated using deviations of actual runs from published run predictions for the 2001 through 2005 runs.

RESULTS

A total of 32.7 million sockeye salmon are expected to return to Bristol Bay in 2006. This prediction is similar (2 percent higher) to the previous 10-year mean of total runs (32.1 million; range of 19.7 million to 44.7 million). The 80% confidence bounds for the 2006 forecasted run ranged from 27.0 million to 38.5 million. All systems are expected to exceed their minimum spawning escapement goals except the Kvichak River. There is a forecasted shortfall of 60 thousand sockeye salmon to the Kvichak River based on the escapement goal of 2.0 million fish and the forecasted run of 1.94 million fish. The Department does not actively manage for an escapement goal for the Alagnak River stock of sockeye salmon. The expected escapement to the Alagnak River was forecast to be 2.0 million sockeye salmon based on the Alagnak River run developing as forecast and the commercial fisheries in Bristol Bay being managed similar to recent years.

A run of 32.7 million sockeye salmon can potentially produce a total harvest of 23.7 million fish if escapement goals are met for managed stocks at their mid range and industry is capable of taking the surplus fish. A harvest of this size would be 26 percent higher than the previous 10-year mean harvest (18.8 million; range was 9.9 million to 29.5 million).

The forecasted run to each district and river system was as follows: 11.98 million to Naknek-Kvichak District (1.94 million to Kvichak River; 2.86 million to Alagnak River; 7.18 million to Naknek River); 9.30 million to Egegik District; 3.34 million to Ugashik District; 7.52 million to Nushagak District (4.67 million to Wood River; 2.01 million to Nushagak River; 0.84 million to Igushik River) and 0.59 million to Togiak District.

The forecasted total run of 32.7 sockeye salmon is expected to be comprised of 13.76 million age 1.3 fish (42%) followed by 8.86 million age 1.2 fish (27%), 5.86 million age 2.2 fish (18%), 3.96 million age 2.3 fish (12%) and 0.29 million age 0.3 and age 1.4 fish (1%).

DISCUSSION

There is always uncertainty in forecasting returns of sockeye salmon to Bristol Bay. The 2006 forecast is no different than previous years. The total run forecast this year may be biased low. This is based upon the uncertainty in predicting the returns of age 1.2 sockeye salmon and the recent tendency to under forecast total run and expected harvests. We had large returns of age 1.1 fish to the Naknek, Egegik and Wood river systems during 2005. In addition, we had age 1.1 fish return to Kvichak, Alagnak and Ugashik river systems. The large presence of age 1.1 fish in the 2005 run suggests a large return of age-1.2 fish in 2006. We used both age 1.1 sibling and spawner-return data to forecast age 1.2 fish in 2006. However, if we had only used age 1.1 sibling data in the forecast, we would have forecast a larger return of age 1.2 fish in 2006 to the Egegik, Naknek and Wood river systems and to a lesser extent to the Kvichak, Alagnak and Ugashik river systems.

Similar methods have been used to produce the Bristol Bay sockeye salmon forecast since 2001. The forecast methods during this recent time have performed fairly well. There is still a tendency for the forecasts and expected harvests to be biased low. The five previous total run forecasts (2001-2005) have averaged 4% below the total run. The forecast run differences have ranged from 18% below in 2005 to 10% above in 2001. The expected harvests have averaged 10% below the actual harvest for the last five years. The expected harvest differences have ranged from 9% below in 2002 to 33% above in 2004. Even though the differences between the overall forecasts and expected harvests are improvements over previous forecasting methods, we will continue to work on improving our ability to forecast sockeye salmon returning to Bristol Bay.

Tim Baker
Area Fishery Research Biologist
Bristol Bay Salmon
Anchorage

2006 Bristol Bay Sockeye Salmon Forecast by District