Nonparametric Methods for Fisheries Data Analysis

Part I of this book contains nonparametric and alternative statistical procedures to analyze fisheries biology data as: non linear resistant smoothers; kernel density estimators (KDE) calculated by means of averaging shifted histograms (ASH) and weighted averaging of rounded points (WARP); computing intensive methods as cross-validation (CV) and bootstrapping (BS) for KDE bandwidth election. Other nonconventional methods such as k-nn and kernel regression are presented. Program listings (Turbo Pascal and Stata languages) are included in the several appendices. Part II contains a fisheries biology study case on Lateolabrax japonicus (Japanese sea bass) including nonparametric and two phase regressions, length frequency analysis by KDE's using CV and BS to choose the bandwidth, age and growth determination by scale and otolith, whole and sectioned, reading (catch composed mainly by immature, 1 and 2 years old; a female 14 years was the oldest). Finally, multivariate analysis of covariance (MANCOVA) was used to analyze fish condition (high in Summer), reproductive (high in Winter) and trophic (high in Spring) status in comparison with traditional morphophysiological indexes.