Many techniques for predicting species potential distribution were recently developed. Despite the international interest for these procedures, applications of predictive approaches to the study of Italian fauna distribution are exceptionally rare. This paper aimed at: (a) detecting climatic exigencies of A. bedriagae in Sardinia; (b) predicting the Archaeolacerta bedriagae Sardinian potential distribution; (c) identifying the most vulnerable Italian populations of the species. Literature and field data were utilized as presence records. Six modelling procedures (BIOCLIM, DOMAIN, ENFA, GAM, GLM, and MAXENT) were adopted. The species climatic requirements were defined using the WorldClim databank for deriving the environmental predictors. AUC and Kappa values were calculated for models validation. AUC values were compared by using Anova Monte Carlo. The best four models were combined through the weighted average consensus method for producing a univocal output. GAM and MAXENT had the best performances (respectively: AUC = 0.93 ± 0.03, Kappa = 0.77 ± 0.08; AUC = 0.93 ± 0.03, Kappa = 0.78 ± 0.07). Good results were also obtained by GLM and DOMAIN (respectively: AUC = 0.89 ± 0.04, Kappa = 0.72 ± 0.05; AUC = 0.88 ± 0.04, Kappa = 0.69 ± 0.07). BIOCLIM and ENFA gained relatively low performances (respectively: AUC = 0.78 ± 0.07, Kappa = 0.57 ± 0.14; AUC = 0.75 ± 0.06; Kappa = 0.49 ± 0.10). In Sardinia A. bedriagae is mainly influenced by seasonality, which causes the evidenced range fragmentation. Moreover, the general importance of multi-methods approaches and consensus techniques in predicting species distribution was highlighted. © 2009 Brill Academic Publishers.

Modelling Bedriaga's rock lizard distribution in Sardinia: An ensemble approach

Salvi, Daniele;
2009-01-01

Abstract

Many techniques for predicting species potential distribution were recently developed. Despite the international interest for these procedures, applications of predictive approaches to the study of Italian fauna distribution are exceptionally rare. This paper aimed at: (a) detecting climatic exigencies of A. bedriagae in Sardinia; (b) predicting the Archaeolacerta bedriagae Sardinian potential distribution; (c) identifying the most vulnerable Italian populations of the species. Literature and field data were utilized as presence records. Six modelling procedures (BIOCLIM, DOMAIN, ENFA, GAM, GLM, and MAXENT) were adopted. The species climatic requirements were defined using the WorldClim databank for deriving the environmental predictors. AUC and Kappa values were calculated for models validation. AUC values were compared by using Anova Monte Carlo. The best four models were combined through the weighted average consensus method for producing a univocal output. GAM and MAXENT had the best performances (respectively: AUC = 0.93 ± 0.03, Kappa = 0.77 ± 0.08; AUC = 0.93 ± 0.03, Kappa = 0.78 ± 0.07). Good results were also obtained by GLM and DOMAIN (respectively: AUC = 0.89 ± 0.04, Kappa = 0.72 ± 0.05; AUC = 0.88 ± 0.04, Kappa = 0.69 ± 0.07). BIOCLIM and ENFA gained relatively low performances (respectively: AUC = 0.78 ± 0.07, Kappa = 0.57 ± 0.14; AUC = 0.75 ± 0.06; Kappa = 0.49 ± 0.10). In Sardinia A. bedriagae is mainly influenced by seasonality, which causes the evidenced range fragmentation. Moreover, the general importance of multi-methods approaches and consensus techniques in predicting species distribution was highlighted. © 2009 Brill Academic Publishers.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11697/142328
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 23
  • ???jsp.display-item.citation.isi??? 18
social impact