Comparing three methods to estimate the average size at first maturity: A case study on a Curimatid exhibiting polyphasic growth

Abstract

The average size at first |maturity (L50) is among the most important parameters for fisheries management and conservation. This paper aims to compare three different methods for its estimation. Considering a classical approach, a logistic model was used (a) by determining the gonadal stage macroscopically; and (b) by using the GSI as proxy of sexual maturity; and finally; (c) by using the length–weight relationship (LWR) in a theoretical approach. The proposed methods were applied using data of a detritivorous fish, Cyphocharax abramoides, monthly sampled using gill nets. Captured individuals were measured, weighed, sexed and the gonadal stage was classified macroscopically and weighed. Estimated L50 values using the macroscopic identification, GSI approach and LWR were not different from each other considering confidence intervals. Between the three different techniques, we concluded that the analysis of the LWR in fishes with polyphasic growth presented promising results as it only requires length and weight data to be performed and estimate a L50 within the range of both classical logistic models analysed.

Traditional approaches to estimate length at first maturity (L50) retrieve better results than alternative ones in a Neotropical heptapterid

“…On the other hand, macroscopic identification may display high accuracy as it considers the external morphological features of the gonads linked to the maturation stage, such as size and colour (Midway et al ., 2013; Peer et al ., 2012). Compared to the former, macroscopic identification requires less financial and time investment, but it is biased by individual perception (Hashiguti et al ., 2019) and exhibits high levels of misclassification ( e.g ., Ferreri et al ., 2009). Error in the classification of maturity status is primarily associated with classifying individuals in early maturation stages as mature or classifying resting individuals as immature (Midway et al ., 2013; Núñez & Duponchelle, 2009), which may bias L 50 estimates and the establishment of reproductive periods.…”

“…These approaches are the most accepted in fishery management studies (ICES, 2008) and are referred to hereafter as traditional approaches. Microscopic identification is expensive and time‐consuming, while macroscopic identification is usually biased by individual perception (Hashiguti et al ., 2019). An alternative approach is to define sexual maturity status by applying a cut‐off value in the gonadosomatic index ( I g , gonad weight relative to total weight), which represents the reproductive investment and may be used as a proxy to sexual maturity (Fontoura et al ., 2009).…”

“…An alternative approach is to define sexual maturity status by applying a cut‐off value in the gonadosomatic index ( I g , gonad weight relative to total weight), which represents the reproductive investment and may be used as a proxy to sexual maturity (Fontoura et al ., 2009). This method requires less financial and time investment than microscopic identification and it is not biased by individual perception (Flores et al ., 2015; Hashiguti et al ., 2019).…”

“…Different approaches to estimating L 50 retrieve estimates varying in precision, accuracy and demands of time, money, sampling effort and mortality of specimens (Hashiguti et al ., 2019). Regardless of the approach, estimates are more accurate the more precisely the identification of the specimen's maturation stage is performed, but nonlethal and less costly approaches are necessary when available resources are scarce.…”

“…Error in the classification of maturity status is primarily associated with classifying individuals in early maturation stages as mature or classifying resting individuals as immature (Midway et al, 2013; Núñez & Duponchelle, 2009), which may bias L 50 estimates and the establishment of reproductive periods. Sampling should be performed during the main spawning period to avoid such bias (ICES, 2008), which may not be possible for species that spawn continuously, such as R. transfasciatus.The I g approach underestimated and overestimated L 50 for females and males, respectively, and exhibited low precision in the estimates for males Hashiguti et al (2019). highlighted that the type of spawning of the target species is an important factor to determine the suitability of the model.…”