Australolacerta australis is a very poorly known species, discovered at the beginning of the 20th century on the Cederberg Mountains (Western Cape, South Africa). The few faunistic researches carried out in the area in the following years evidenced that A. australis is a rock dwelling lizard with a narrow range. The chance to modelling the habitat suitability of the species allows to partially overcome the scarcity of faunistic data predicting the potential distribution on the base of the ecological features of the study area. The eighteen georeferred records of the specimens preserved in the herpetological collection of the Stellenbosch University were randomly separated in four training (70% of the data) and four test (30%) subsets. The training subsets were utilized to elaborate the 19 habitat suitability models (DIVAGIS software; BIOCLIM algorithm; WorldClim databank). The obtained models were validated assigning the predicted habitat suitability values to the test subsets and to the random points. The assigned values were compared by AUC and Kappa values. The two best variables were jointly used to elaborate the four definitive maps of potential distribution, which were summarized in a single mean map and validated using the same procedure of the separated models. The distribution of the Southern Rock lizard is mainly affected by the “precipitation of coldest quarter” and the “mean temperature of wettest quarter” (AUC = 0.958 and 0.858; Kappa = 0.917 and 0.717 respectively). The final mean model (AUC = 0.929; Kappa = 0.857) shows that the presence of the species is related to the Mountain Fynbos biome and the elevation (>700 m asl). Therefore, A. australis could be potentially distributed continuatively on the Cape Fold Mountains, from Pakhuis Pass (Clanwilliam) to Kwadousberg (Worcester), but also on the Riviersonderendberge (Riviersonderend). The modelling of potential distribution of such poorly known species can be also helpful to increase the efficiency of future faunistic researches devoted to the conservation of the species, addressing the field efforts in the suitable areas
Atti del 6° Congresso Nazionale della Societas Herpetologica Italica (Roma, 27.IX-1. X.2006)
SALVI, Daniele;
2006-01-01
Abstract
Australolacerta australis is a very poorly known species, discovered at the beginning of the 20th century on the Cederberg Mountains (Western Cape, South Africa). The few faunistic researches carried out in the area in the following years evidenced that A. australis is a rock dwelling lizard with a narrow range. The chance to modelling the habitat suitability of the species allows to partially overcome the scarcity of faunistic data predicting the potential distribution on the base of the ecological features of the study area. The eighteen georeferred records of the specimens preserved in the herpetological collection of the Stellenbosch University were randomly separated in four training (70% of the data) and four test (30%) subsets. The training subsets were utilized to elaborate the 19 habitat suitability models (DIVAGIS software; BIOCLIM algorithm; WorldClim databank). The obtained models were validated assigning the predicted habitat suitability values to the test subsets and to the random points. The assigned values were compared by AUC and Kappa values. The two best variables were jointly used to elaborate the four definitive maps of potential distribution, which were summarized in a single mean map and validated using the same procedure of the separated models. The distribution of the Southern Rock lizard is mainly affected by the “precipitation of coldest quarter” and the “mean temperature of wettest quarter” (AUC = 0.958 and 0.858; Kappa = 0.917 and 0.717 respectively). The final mean model (AUC = 0.929; Kappa = 0.857) shows that the presence of the species is related to the Mountain Fynbos biome and the elevation (>700 m asl). Therefore, A. australis could be potentially distributed continuatively on the Cape Fold Mountains, from Pakhuis Pass (Clanwilliam) to Kwadousberg (Worcester), but also on the Riviersonderendberge (Riviersonderend). The modelling of potential distribution of such poorly known species can be also helpful to increase the efficiency of future faunistic researches devoted to the conservation of the species, addressing the field efforts in the suitable areasPubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
