Widespread consumption-dependent systematic error in fish bioenergetics models and its implications

Abstract

Data from laboratory evaluations of seven fish bioenergetics models (BEMs) were used to investigate possible associations between BEM prediction error in relative growth rate (RGR error ) and levels of model input variables: mean daily food-consumption rate and fish body weight. Correlation between RGR error and fish body weight was found in three BEMs applied under submaintenance feeding conditions. A strong correlation between RGR error and mean daily consumption level was observed in all models over full consumption ranges; consumption level explained 70%-96% of variation in RGR error . All BEMs underestimated (by 2-to 5-fold) growth at lower consumption levels and overestimated (by 2-to 3-fold) growth at higher consumption levels. RGR error values associated with higher consumption levels were greater (up to 22 cal·g -1 ·day -1 ) than those at lower consumption levels (up to 10 cal·g -1 ·day -1 ). Correlation between consumption rate and RGR error in all seven models indicates widespread systematic error among BEMs that likely arises from deficiencies in consumption-dependent model parameters. Results indicate that many BEMs are substantially inaccurate when predicting fish growth from higher feeding rates or estimating consumption from higher growth rates, even when higher consumption levels or growth episodes are of short duration. Findings obtained under submaintenance feeding conditions indicate that additional body-weight-and consumption-dependent terms should be added to BEM subequations for routine metabolism to account for metabolic reduction.Résumé : Les données des évaluations en laboratoire de sept modèles bioénergétiques (BEM) de poissons nous ont servi à étudier les associations possibles entre l'erreur de prédiction par les BEM du taux relatif de croissance (RGR erreur ), d'une part, et la valeur des variables d'entrée des modèles, soit le taux journalier moyen de consommation de nourriture et la masse corporelle des poissons, d'autre part. Il existe une corrélation entre RGR erreur et la masse corporelle des poissons dans trois des BEM utilisés dans des conditions d'alimentation inférieures au niveau de maintien (SMFC). Il y a aussi une forte corrélation entre RGR erreur et le taux journalier moyen de consommation dans tous les modèles sur toute l'étendue des taux de consommation; le taux de consommation explique 70 -96 % de la variation de RGR erreur . Tous les BEM sous-estiment la croissance (par un facteur de 2-5) aux faibles taux de consommation et la surestiment (par un facteur de 2-3) aux taux de consommation plus élevés. Les RGR erreur associées aux taux plus élevés de consommation sont plus grandes (pouvant atteindre 22 cal·g -1 ·jour -1 ) que celles liées aux taux de consommation plus faibles (atteignant 10 cal·g -1 ·jour -1 ). La corrélation entre le taux de consommation et RGR erreur dans les sept modèles indique qu'il y a une erreur systématique générale dans les BEM qui provient vraisemblablement d'imprécisions des paramètres des modèles reliés à la consommation. Nos rés...

Simultaneous identification and correction of systematic error in bioenergetics models: demonstration with a white crappie (Pomoxis annularis) model

“…Correction of systematic error in BEMs is possible, owing to the identification of significant relationships between model prediction error and levels of certain PIO variables (consumption rate, temperature, and fish body weight) for several BEMs (Bajer et al 2003(Bajer et al , 2004; present study). The correction approach that we present is considered efficient because model error evaluation and correction can be accomplished for a broad array of application conditions from a single laboratory data set that can be generated as a result of a reasonable level of effort.…”

“…Recent laboratory evaluations (Madenjian and O'Connor 1999;Bajer et al 2003) and the present study have identified at least two forms of systematic error within BEMs. Evidence that this systematic error is widespread among BEMs and can cause substantial over-and under-estimation of fish growth and consumption rates (Bajer et al 2004) heightens the need to improve the predictive accuracy of BEMs, particularly given that rates of application of these models are rapidly increasing.…”

“…That many BEMs apparently contain important systematic error, while being applied at increasing rates, heightens the need for actions to improve them. Bajer et al (2003), and to a further extent Bajer et al (2004), discuss likely sources of the systematic error found within BEMs. These insights should facilitate efforts to fundamentally improve BEMs by identifying specific internal subequations that warrant reevaluation and possible improvement.…”

“…Recent studies have identified the presence of strong systematic error in BEMs (Madenjian and O'Connor 1999;Bajer et al 2003), while another study has shown this type of error to be widespread among these models and to have potentially important consequences relating to predictions of consumption and growth rates (Bajer et al 2004). That many BEMs apparently contain important systematic error, while being applied at increasing rates, heightens the need for actions to improve them.…”

Estimation of Bioenergetics Parameters for Rainbow Trout (Oncorhynchus mykiss) Using Capture-Recapture Data with Comparison to Estimates from a Laboratory-Based Model

“…However, as with all *Address correspondence to this author at the Fisheries Centre, University of British Columbia, Vancouver, B.C. V6T 1Z4, Canada; Tel: 604-822-0046; Fax: 604-822-8934; Email: b.vanpoorten@fisheries.ubc.ca models, there are misgivings about bioenergetics models related to parameter uncertainty [2,19] and model structure [12,20,21]. One drawback of these models is the use of laboratory estimates from other populations or closely related species as parameter inputs for use in estimating growth and consumption of a population of interest [2].…”