DRAFT Assessment of Two Methods for Estimating the Composition of Chinook Encounters Early in the Fishing Season

An important component of the in-season creel surveys conducted for Chinook mark-selective fisheries (MSF) is a test fishery (TF) and/or angler Voluntary Trip Report (VTR) survey that provides the data used to estimate the size (legal or sub-legal) and mark (marked or unmarked) composition of the Chinook being encountered by the fishery. Early in the fishing season, sample sizes for these programs are often small (< 20 fish) and estimates of the size|mark composition are imprecise and can vary greatly from day to day. This presents a challenge to fishery co-managers as the early-season assessments of the total encounters and mortalities of marked and unmarked Chinook that have occurred to date may fluctuate considerably (increasing or decreasing) until later in the season when sample sizes are larger and the estimated size|mark composition is more stable. The objective of the analyses conducted for this report is to determine if there are alternative methods of estimating the size|mark composition of Chinook encounters that can be used when the sample sizes for the TF or VTR programs are small early in a fishing season. Two alternative methods are examined: (1) the mean of the three prior years of size|mark composition estimates and (2) in-season information collected during dockside interviews of anglers exiting the MSF; from these interviews, Chinook harvested and Chinook released are tabulated by size|mark composition category.

Based on an examination of the size|mark composition estimates from each alternative method and retrospective analyses, the following recommendations are made:

• Dockside angler interview data collected in-season should be used to estimate the size|mark composition of Chinook encounters early in the season prior to the collection of an adequate sample size by the in-season TF/VTR programs.
• Adequate sample sizes for the in-season TF/VTR programs should be defined as a sample size that produces an estimate for the proportion of legal-size and marked encounters that has a coefficient of variation ≤20%.
• Whenever Method 2 is used to estimate the total number of Chinook encounters in a MSF, and dockside data are being used instead of TF/VTR data to estimate the size|mark composition of Chinook encounters, bias-correction methods should be applied to the estimate of the size|mark composition and the bias-corrected estimate of %LM should be used to estimate total encounters and apportion the encounters to size|mark categories.
• Size|mark composition estimates using the dockside angler interview data should replace the beta regression method currently used for post-season MSF analyses and reporting when sample sizes from an area’s TF/VTR programs are small and do not meet the 20% CV standard.