Nonparametric quantile regression captures regional variability and scaling deviations in Atlantic surfclam...

Area of Science:

• Marine Biology
• Quantitative Ecology
• Fisheries Science

Background:

• The allometric model's universality for marine species length-weight relationships is debated, especially for invertebrates.
• Nonparametric regression offers flexibility for complex growth patterns, like inflection points, missed by standard models.
• Identifying biologically relevant thresholds is crucial for fisheries management and size-dependent yield.

Purpose of the Study:

• To compare parametric (allometric) and nonparametric regression models for Atlantic surfclam (Spisula solidissima) length-weight relationships.
• To assess model performance in mean regression and quantile regression contexts across different coastal regions.
• To investigate size-dependent growth variations and condition-dependent deviations using advanced statistical approaches.

Main Methods:

• Compared a classic allometric model with a kernel-based nonparametric mean regression model using bootstrap procedures.
• Applied parametric and nonparametric quantile regression to analyze size-dependent growth patterns.
• Utilized hypothesis testing for mean regression model selection and goodness-of-fit tests for quantile regression evaluation.

Main Results:

• The allometric model demonstrated a superior fit for mean regression analysis of Atlantic surfclam data.
• Nonparametric models were more effective for quantile regression, revealing condition-dependent deviations and regional growth variability.
• The study highlights the utility of distinguishing central from marginal populations in growth modeling for bivalves.

Conclusions:

• While the allometric model suffices for average length-weight trends, nonparametric methods are essential for detailed analysis of growth variability in marine bivalves.
• Nonparametric quantile regression provides deeper insights into factors influencing growth and condition across different populations.
• Findings suggest broader applicability for these advanced modeling techniques in fisheries science for species like Arctica islandica and Mercenaria mercenaria.