Modeling Otolith Weight using Fish Age and Length: Applications to Age Determination

Abstract

Relationships among otolith weight and age have been explored widely as cost‐effective means to predict age. However, otolith weight is influenced by both fish age and somatic growth, making it necessary to partition these confounding effects to best use otolith weight to predict age. We used several hatchery strains of Lake Trout Salvelinus namaycush that varied in capture season, year, location, and mean size at stocking to develop a maximum likelihood model of otolith weight as complementary but independent functions of age (mg/year) and fish length (mg/mm). Once these relationships were established, we determined age probabilities for each fish based on their measurements. Our best model included age and somatic growth components, with cumulative mean temperature as a variable to scale for seasonal differences in otolith accretion. For fish between 3 and 19 years of age, otolith weight increased by 1.16 mg/year, with an additional somatic growth component of 0.62 × exp(fish length × 0.0038) mg/mm. We correctly assigned age to 60% of the age‐3 fish, but this decreased to about 10% among the oldest age‐classes; however, predicted age residuals approximated normal distributions within each age‐class. The model was robust to seasonal and annual variation, but somatic growth differences among some strain and stocking‐size groupings resulted in modal aging biases of ±1 year. Given the stability in otolith weight components arising from fish age and somatic growth, we suggest that modeling otolith weight using reader age estimates offers a new quality assurance/quality control tool to assess whether reader ages support the widely documented properties of otolith growth. When otolith weight and fish length are more informative of age, our model may also be useful in estimating age composition data similar to an age–length key. Received December 19, 2016; accepted March 20, 2017 Published online May 26, 2017

Mass‐effect: Understanding the relationship between age and otolith weight in fishes

“…Over the last two decades, the use of otolith weight has been used alternatively and supplementary to traditional age estimation (Boehlert, 1985; Morat et al., 2008; Nazir & Khan, 2019), mainly, based over the premise that long‐lived and slow‐growing fish tend to have relatively larger and heavier otoliths compared with those otoliths from fast‐growing and short‐lived fish (Radtke et al., 1985). Several authors have shown that there is a predictable relationship between the age and otolith weight (Britton & Blackburn, 2014; Francis et al., 2005; Ghanbarzadeh et al., 2014; Hanson & Stafford, 2017), which has generated debate within the scientific community about its real application in the processes of age estimation in fish. Nonetheless, this technique has been successfully implemented for age estimation in the Atlantic cod ( Gadus morhua , Gadidae), lake trout ( Salvelinus namaycush , Salmonidae), European eel ( Anguilla anguilla , Anguillidae) and Bata ( Labeo bata , Cyprinidae) (Bermejo, 2014; Hanson & Stafford, 2017; Kanjuh et al., 2018; Khan et al., 2018).…”

Age determination for whitespotted conger Conger myriaster through somatic and otolith morphometrics

“…Eels collected from coastal China seas were all young individuals up to 6 years, and older fish may prefer offshore areas. Otolith weight is more reliable to predict age compared with other morphological characteristics of otolith or body [ 17 – 19 ], which could be attributed to the fact that the growth of otolith and body length of C. myriaster are not synchronized [ 24 , 30 ]. The otolith weight increases continually over the life, while the growth of body and otolith length slows down with ages [ 11 , 31 ].…”

“…Many statistical methods were used to avoid subjectivity in age determination using otolith/somatic morphometrics, such as random forests (RF) [ 16 – 18 ], maximum likelihood-based mixture analysis [ 15 , 19 ], discriminant analysis [ 20 ], support vector machine [ 21 ], and ensemble of wrappers [ 22 ]. Random forest model is one of the most widely used approaches, which can model non-linear relationships and interactions [ 23 ].…”

Extrinsic Factors Influencing Somatic Growth of Lake Sturgeon

“…As has been done previously in relation to other fish species (e.g., Griffiths et al. ; Hanson and Stafford ), a decimal age correction factor was applied to account for seasonal growth. Hatch (birth) dates (start of growing season; ~10°C water temperature) and growth cease dates (~10°C; based on hatchery observations; C. Klassen, Manitoba Hydro, personal communication) were estimated for each location (Table ).…”

“…Lake Sturgeon gill netting occurred throughout the open-water season, meaning that an individual captured in June and another captured in October would have experienced different opportunities for growth, even if they were the same count age. As has been done previously in relation to other fish species (e.g., Griffiths et al 2009;Hanson and Stafford 2017), a decimal age correction factor was applied to account for seasonal growth. Hatch (birth) dates (start of growing season;~10°C water temperature) and growth cease dates (~10°C; based on hatchery observations; C. Klassen, Manitoba Hydro, personal communication) were estimated for each location (Table 2).…”