A cooperative effort gathered a large European length-at-age data set (N = 45,759, Lat. 36S61N Long. 10W27E) for Anguilla anguilla, covering one century. To assess the effect of global warming during the last century and habitat effects on growth, a model was fitted on the data representing the conditions met at the distribution area scale. Two GLMs were designed to predict eel log(GR): one model was fitted to the whole data and the other was fitted to the female data subset. A model selection procedure was applied to select the best predictors among sex, age class, five temperature parameters and six habitat parameters (depth, salinity and four variables related to the position in the catchment). The yearly sum of temperatures above 13 degrees C (TempSUP13), the relative distance within the catchment, sex, age class, salinity class and depth class were finally selected. The best model predicted eel log(GR) with a 64.46% accuracy for the whole data and 66.91% for the female eel data. Growth rate (GR) was greater in habitats close to the sea and in deep habitats. TempSUP13 variable had one of the greatest predictive powers in the model, showing that global warming had affected eel growth during the last century.
Abstract– AcooperativeeffortgatheredalargeEuropeanlength-at-agedataset(N =45,759,Lat.36S–61NLong.10W–27E)for Anguillaanguilla,coveringonecentury.Toassesstheeffectofglobalwarmingduringthelastcentury
andhabitateffectsongrowth,amodelwas?ttedonthedatarepresentingtheconditionsmetatthedistributionareascale.TwoGLMsweredesignedtopredicteellog(GR):onemodelwas?ttedtothewholedataandtheotherwas?ttedtothefemaledatasubset.Amodelselectionprocedurewasappliedtoselectthebestpredictorsamongsex,ageclass,?vetemperatureparametersandsixhabitatparameters(depth,salinityandfourvariablesrelatedtothepositioninthecatchment).Theyearlysumoftemperaturesabove13 C(TempSUP13),therelativedistancewithinthecatchment,sex,ageclass,salinityclassanddepthclasswere?nallyselected.Thebestmodelpredictedeellog(GR)witha64.46%accuracyforthewholedataand66.91%forthefemaleeeldata.Growthrate(GR)wasgreaterinhabitatsclosetotheseaandindeephabitats.TempSUP13variablehadoneofthegreatestpredictivepowersinthemodel,showingthatglobalwarminghadaffectedeelgrowthduringthelastcentury.
