|
Activity Number:
|
53
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
| Abstract - #306226 |
|
Title:
|
Forecasts of Salmon Returns
|
|
Author(s):
|
Saang-Yoon Hyun*+ and David H. Salinger
|
|
Companies:
|
Columbia River Inter-Tribal Fish Commission and University of Washington
|
|
Address:
|
729 NE Oregon Street, Suite 200, Portland, OR, 97232,
|
|
Keywords:
|
preseason and in-season forecasts ; salmon returns ; interannual variability ; probability density ; optimization ; Bayesian methods
|
|
Abstract:
|
Preseason and in-season forecasts of salmon returns are required for commercial fishermen and harvest managers. We make a preseason forecast with historical data on sibling returns and ocean conditions using ordinary regression and autoregressive models. Also, we make an in-season forecast with data on fish run proportions from the past years and daily fish returns collected during in-season. Interannual variability in the ecosystem and fish biotic metrics are the main obstacle to making an accurate forecast early during in-season. We develop a probability density and use a pattern-matching optimization to generate an in-season forecast. We also use Bayesian methods to express forecast uncertainty. Finally, we combine preseason and in-season forecasts to reduce forecast uncertainty. We illustrate our ideas using Columbia River Chinook salmon and Alaska Bristol Bay sockeye salmon.
|
