Premium
Assessing Accuracy and Bias of Protocols to Estimate Age of Pacific Salmon Using Scales
Author(s) -
Harris Julianne E.
Publication year - 2020
Publication title -
north american journal of fisheries management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.587
H-Index - 72
eISSN - 1548-8675
pISSN - 0275-5947
DOI - 10.1002/nafm.10468
Subject(s) - oncorhynchus , chinook wind , fish measurement , hatchery , affect (linguistics) , fishery , stock assessment , stock (firearms) , biology , age groups , fork (system call) , fish stock , fish <actinopterygii> , scale (ratio) , statistics , demography , geography , psychology , fishing , mathematics , computer science , cartography , communication , archaeology , sociology , operating system
Accurate estimates of age composition are essential for fisheries management, but evaluation of protocols to estimate age data are rarely conducted. Herein, known‐age Pacific salmon Oncorhynchus spp. from six hatchery stocks were used to examine accuracy of multiple scale‐aging protocols and to assess how errors in age assignment may affect age composition estimated from scale data. Accuracy reached 1.0 for one stock of Coho Salmon Oncorhynchus kisutch , over 0.9 for three stocks of spring Chinook Salmon Oncorhynchus tshawytscha , and was no higher than 0.8 for two stocks of fall Chinook Salmon. For all hatchery stocks, accuracy increased and bias decreased when readers had auxiliary information about a fish's fork length and sex. Results suggest that a protocol that collects scales from above the lateral line, provides auxiliary information on fork length and sex, allows readers to exclude individuals for which age appears unclear, and uses an arbitrator to assign age when two initial readers disagree may produce the most accurate and unbiased age data from scales. This study highlights the importance of assessing patterns in age errors both qualitatively and quantitatively. A general pattern of overestimating age for younger age‐classes and underestimating age for older age‐classes was evident. This pattern in age errors could affect age composition estimates for adult Pacific salmon and other species with few (e.g., two to six) age‐classes. In addition, individual variation in demographic rates may affect individual patterns identified on scales; thus, for stocks with more extensive individual demographic variability, like fall Chinook Salmon stocks, accuracy may be lower and bias higher, resulting in less accurate estimates of age composition.