oa Predicting The Distribution Of Lizard Species In Qatar Using Mathematical Models

Conducting biodiversity surveys is expensive and time consuming and we cannot always invest as much time as needed conducting field work. To compensate for such deficiencies, scientists can benefit from advanced techniques of species distribution modelling. Data on species distribution is essential for the correct conservation and management of the species and their habitats. The objective of this study has been to predict the distribution range of four lizard species in Qatar that were found in few locations during field surveys conducted in 2012-2013, and that we believe that could be present in a larger range. The species examined were: the Schmidt´s fringe-toed lizard, Acanthodactylus schmidti (with 18 field observations), the toad-headed agama, Phrynocephalus arabicus (with 27 field observations), the Arabian sand gecko, Stenodactylus arabicus (with 22 field observations) and the Eastern skink, Scincus mitranus (with 8 field observations). The four species have been only observed in the southern part of the Qatar country. To make predictions about the potential distribution of these lizard species we used climatic data obtained from the WorldClim database, and remote sensing data (Landsat 8 Image), from which we obtained high resolution data as surface temperature, and other features derived from the land surface reflectivity at different wavelenghts. We generated new maps for the four lizard species in Qatar based on different ecological niche models (ENMs). All these models appeared to be ‘good’, with a AUC value >0.8. There are not significant differences between bioclimatic and remote-sensing maps. However the bioclimatic maps were the ones that visually matched better with the observed distribution of the species. Furthermore, bioclimatic maps can be less reliable due to the low number of climatic observatories in the area used to build the databases. We found some differences between the predicted distribution maps depending on the environmental covariates used, being the relative rank between pairs of maps of any species always near 0.5. Despite the limitations of the models, they appear to be a good predictive tool for lizard distribution ranges. Future advances in the knowledge of the environment together with environmental and climatic maps of higher resolution will improve ecological niche modelling in Qatar.