Juvenile Chinook Salmon Weight Prediction Using Image‐Based Morphometrics

Abstract

We developed an empirical weight prediction model for juvenile Chinook Salmon, Oncorhynchus tshawytscha, ranging from 33-113 mm and 0.35-14.86 g using morphometrics analysis on images collected on live fish in the field. This method relies on consumer-grade hardware and free software and addresses an issue of a lack of valuable weight and condition data for life-history and survival modeling. A blind test revealed the method was relatively precise with mean absolute error of 0.075 g and mean absolute percent error of 2.86% and unbiased with a mean error of-0.001 g. Given inherent variability in field wet weight measurement estimated with an average standard deviation of 0.030 g and coefficient of variation of 1.02% from a blind repeated measurements experiment, the observed level of predictive error suggests the method is an adequate substitute for field weight measurement when conditions (i.e. boat sampling) preclude conventional scale taring and accurate weight measurement. Additionally, this method is useful in an automated data quality control workflow to identify erroneous length and/or weight measurements. The image landmark placement closely follows previous work on juvenile Chinook Salmon and can be incorporated with existing methodologies such as stock analysis and phenotypic differentiation to increase available biometric information.

The utility of non‐lethal morphometrics to evaluate fish condition

“…First, by integrating landmark‐based geometric morphometrics and prediction models, we can accurately estimate individual weight or volume without having to physically handle fish. Holmes and Jeffres (2021) digitized anatomical landmarks in images of juvenile Chinook salmon ( Oncorhynchus tshawytscha Walbaum) and applied predictive models to estimate the wet weight of each fish with a mean error of only 2.9%. Second, the use of artificial intelligence and deep learning to automatically detect and identify fish species from camera footage is rapidly accelerating (Ditria et al., 2020), and similar approaches are being investigated to automate sizing (Shi et al., 2020) and biomass estimation (Zhang et al., 2023).…”