Ecological Niche Modelling of the Genus Anatololacerta under Past, Historical and Future Bioclimatic Conditions

Ecological niche modelling of the genus Anatololacerta under past, historical and future bioclimatic conditions. Folia Biologica (Kraków) 70 : 45-53. Anatoliahasplayedanimportantroleinthespeciationofmanyspecies.Globalclimaticchangesaffectthedistributionofmanyreptilesindifferentways,includingtheirrangeexpansionorhabitatloss.In thisstudy,wefocusedonthegenus Anatololacerta whichinhabitsAnatoliaandsomeGreekislands.In total, 150 literature records and 20 items of observation data were analysed with the maximum entropy method (MaxEnt) for the last glacial maximum (LGM), as well as historical (1970-2000) and future (2081-2100; RCP 4.5) bioclimatic conditions. According to our model, the distribution ranges in the future conditions for A. danfordi , A. finikensis and A. ibrahimi were more extensive than for the historical conditions. However, A. anatolica and A. pelasgiana may experience a potential habitat loss in the future. Although physical barriers were obstacles in the LGM and historical records, these barriers may be overcome after the climatic changes taking place in the near future.

Reptiles have been widely distributed throughout the world in a time period ranging from the Palaeozoic era to the present day (REISZ et al. 2011;BÖHM et al. 2013). Although reptiles play important roles in nature, nearly one in five species of reptiles are currently at risk of extinction (BÖHM et al. 2013). The main risks for reptile extinction are habitat loss and degradation, pollution, invasive species, disease, climate change and harvesting from nature by the pet trade (GIBBONS et al. 2000;MARSHALL et al. 2020). Terrestrial reptiles have a narrower distributional range and are affected by ecological changes due to being ectotherms (BÖHM et al. 2013;CARRANZA et al. 2018;ALATAWI et al. 2020). The most dramatic scenario for reptiles related to climatic change is their extinction due to habitat changes, which are linked to precipitation and temperature changes (GIBBONS et al. 2000;ARAUJO et al. 2006;BÖHM et al. 2013). ARAUJO et al. (2006) reported that climate cooling would be more dangerous for reptile species than global warming. Since climate change also affects the sex determination of some species of turtles and crocodiles, male turtles may become extinct in certain future scenarios (JANZEN 1994;GONZALEZ et al. 2019). In the future climate scenario, the distributions of most species may change, along with reductions and increases in their range or shifts in the latitudes and elevations (GOMEZ-RUIZ & LACHER Jr. 2019).
Members of the Lacertidae family are widely distributed in Eurasia and Africa, and the family is represented by around 340 species (ARNOLD et al. 2007;GARCIA-PORTA et al. 2019). Lacertids live in numerous environmental conditions ranging from deserts to rainforests and from sea level to high mountains (GARCIA-PORTA et al. 2019). The Lacertidae family has been divided into two subfamilies -Lacertinae and Gallotiinae -as a result of DNA sequencing (ARNOLD et al. 2007). The Lacertinae subfamily is further split into two monophyletic tribes: Lacertini are distributed in Europe, Northwest Africa and Southwest and East Asia; while Eremiadini are present in Africa and in Southwest and Central Asia (ARNOLD et al. 2007).
Members of the genus Anatololacerta are distributed throughout Turkey and the Greek islands (Samos, Ikaria, Rhodes and Psomis) (ARNOLD et al. 2007;BELLATI et al. 2015;KARAKASI et al. 2021). ARNOLD et al. (2007) reported that the Western and Southern Anatolian Lacerta group should be defined as a new genus named Anatololacerta. This genus, which lives in Anatolia and the Aegean islands of Samos, Ikaria and Rhodes, was represented by three species at that time: A. anatolica (Werner, 1900), A. danfordi (Günther, 1876 and A. oertzeni. The subspecies taxonomy of this genus was controversial and included nine morphological subspecies (A. anatolica, A. a. aegaea, A. oertzeni, A. o. budaki, A. o. ibrahimi, A. o. finikensis, A. o. pelasgiana, A. o. quandtaylori, A. danfordi and A. d. bileki) (BELLATI et al. 2015;CANDAN et al. 2016). BELLATI et al. (2015) conducted a molecular study about this genus and reported that this genus was represented by four species in Anatolia (A. anatolica, A. pelasgiana (Mertens, 1959), A. budaki and A. danfordi). In this study, A. oertzeni was demoted to the subspecies level and two of the subspecies of A. oertzeni (A. o. pelasgiana and A. o. budaki) were recognised as having a species rank (BELLATI et al. 2015). CANDAN et al. (2016) reported that A. oertzeni and A. danfordi each have two well-supported lineages. The last revision of this genus was conducted by KARAKASI et al. (2021) who found five well-supported parapatric species in the genus Anatololacerta. However, A. anatolica, A. pelasgiana and A. danfordi still remain distinct species, as in BELLATI et al. (2015), while A. ibrahimi (Eiselt & Schmidtler, 1987) and A. finikensis (Eiselt & Schmidtler, 1987) were identified as separate species (KARAKASI et al. 2021).
The niche for one species includes both biotic (predations, competition, dispersal limitation, etc.) and abiotic (climatic, environmental, etc.) factors (STIGALL 2012;HEIDARA 2021). Ecological niche modelling (ENM) creates a predicted distribution map by using the climatic and geographical conditions based on known locality records (GÜL et al. 2015). Under the niche conservatism theory, a species preserves its ecological niche parameters following environmental changes, and such a species overcomes those environmental changes by shifting their preferred habitat laterally if they have the dispersal potential (STIGALL 2012;TOK et al. 2016). If the species does not have the ability for a dispersal potential, they will lose their range and may become extinct (ARAUJO et al. 2006;TOK et al. 2016). The prediction of distributions is more significant in terms of evolutionary biology due to the resulting gene flow (PRÖHL et al. 2010), and for ecology and conservation biology due to the shifting between habitats (TOK et al. 2016).
The aim of this study was to determine the possible distribution of the Anatololacerta genus and its related species by ecological niche modelling methods under the last glacial maximum, historical and future conditions.
For the past climate conditions, the LGM (nearly 21,000 years ago) at a spatial resolution of 2.5 arcminutes was used.  (GÜL et al. 2018). This program was also used to investigate the geographical distributions of species and the climatic limitations on their distribution (GÜL et al. 2018). The Maxent algorithm estimates (0 is the lowest and 1 is the highest probability) the habitat's suitability with a maximum entropy distribution by using locality point data (GÜL et al. 2015;ALATAWI et al. 2020). The features for this study were organised as 20 subsampled replicates, with 5000 iterations and cloglog outputs. We used Area Under the Curve (AUC) for an estimation of the significance of the model, calculated from the Receiver Operating Curve (ROC) (ALATAWI et al. 2020). The predicted distribution maps obtained from Maxent were imported into ArcGIS vers. 10.3.1 for a visualisation.

Anatololacerta anatolica
The areas predicted to be suitable for A. anatolica had a high AUC of 0.944±0.051 for the LGM, 0.945±0.044 for the historical and 0.943±0.038 for the future bioclimatic conditions (Figure 1, Table 1). There was only one suitable place, situated in the centre of the Taurus Mountains. Other parts of Anatolia were unsuitable during the LGM.
West of Amanos Mountain, east of the Antalya Gulf, Muðla Province, Gediz Basin, the Büyük Menderes basins, Gallipoli and south of the Marmara Sea were seen as suitable places for the historical bioclimatic conditions, while the rest of Anatolia was determined to be unsuitable.
As can be seen in the future climatic conditions, the distribution range was narrower than under the historical conditions. In the future, it is expected that the possible distribution of A. anatolica will include only the area south of the Marmara Region, especially around Uludað Mountain.

Anatololacerta danfordi
The areas predicted to be suitable for A. danfordi had a high AUC of 0.917±0.123 for the LGM, 0.912±0.145 for the historical and 0.906±0.141 for the future bioclimatic conditions ( Figure 2, Table 1). Southeast Anatolia east of Amanos Mountain, south of the Middle Taurus Mountain and Ýzmir Province were determined to be suitable places for this species, while Thrace, the Black Sea Region and west of Central Anatolia were identified to be unsuitable.
In the historical climatic conditions, the distribution model showed that the Adana and Mersin Region, Antalya and the Aegean Region were suitable places, while Eastern Anatolia, Thrace and the Black Sea Region were identified as unsuitable places. When the distributions between the LGM and historical predictions are compared, the suitable places drift from the east to the west of Anatolia for A. danfordi. historical; and c) future bioclimatic conditions. Warm colours refer to highly suitable places on the map. According to the future bioclimatic conditions, the Aegean Region, Central Anatolia and the Mediterranean Region were identified as suitable places, while Thrace, the Black Sea Region, Southeast Anatolia and Northeast Anatolia were determined to be unsuitable places. The distribution model for the future climatic conditions showed that suitable places will be more extensive than in the historical climatic conditions.

Anatololacerta finikensis
The mean AUC values for A. finikensis for the LGM, historical and future climatic conditions were calculated as 0.955±0.044, 0.958±0.040 and 0.959±0.041, respectively ( Figure 3, Table 1). While the Aegean Region, Adana and Sivas were identified as suitable places for this species, the rest of Anatolia was determined to be unsuitable.
According to the historical climatic conditions, the predicted distribution models showed that the LGM and historical climatic conditions were mainly similar, where the Cilicia Region and Antalya appeared to be suitable. The distribution of A. finikensis in the historical climatic conditions was more extensive than in the LGM.
The distribution model for the future conditions was mainly similar to that for the historical climatic conditions, but Central Anatolia appeared to be more suitable than in the historical climatic conditions. The Cilicia Region was also determined to be unsuitable under the future conditions.

Anatololacerta ibrahimi
The calculated mean AUC values for A. ibrahimi for the LGM, historical and future climatic conditions were 0.914±0.063, 0.927±0.052 and 0.921±0.043, respectively ( Figure 4, Table 1). All parts of Anatolia seemed to be unsuitable for A. ibrahimi under the LGM bioclimatic conditions. For the historical conditions, the Mediterranean Region, Hatay, Aegean Region, west of Central Anatolia, Yozgat and the localities between Gümüºhane and Amasya Provinces were determined to be suitable places, while the east and southeast areas of Anatolia and the coastline of the Black Sea appeared to be unsuitable places.
The distribution range of A. ibrahimi under the future conditions was mainly similar to the historical conditions, but it was more extensive than the historical climatic conditions. Also, Gallipoli and south of the mountains in Northeast Anatolia were determined to be suitable places.

Anatololacerta pelasgiana
The mean AUC values were calculated as 0.908±0.070, 0.902±0.071 and 0.921±0.072 for the LGM, historical and future bioclimatic conditions, respectively ( Figure 5, Table 1). Under the LGM bioclimatic conditions, Hatay, Cilicia Region, Antalya Basin and the Aegean Region were seen to be suitable places.   The distribution model in the historical conditions mainly fitted with that of the LGM, but the distribution was more suitable when compared with the LGM. It also widened towards the interior of Anatolia. East and Southeast Anatolia and the coastline of the Black Sea Region seem to be unsuitable places for A. pelasgiana.
The distribution range in the future mainly fitted with that of the LGM and historical bioclimatic conditions, but it was larger than in the LGM and smaller than in the historical conditions. While the Mediterranean and Aegean Regions and Hatay were suitable places for this species, the Cilicia Region, coastline of the Mediterranean Sea and south of the Marmara Sea were determined to be suitable places.

Discussion
In relation to speciation, Anatolia plays an important role and contains several refuges which have conserved species during ice ages and gave species the chance to migrate to suitable habitats during interglacial periods (GÜL et al. 2015). BELLATI et al. (2015) suggested that physical and climatic changes in the Pleistocene affected the distribution of the ancestral Anatololacerta, which allowed for an allopatric divergence of various lineages within Anatolia. Physical barriers could also have affected the divergence of species. While the Büyük Menderes River is considered to be a physical barrier between A. anatolica and A. pelasgiana, the Göksu River also isolated A. danfordi from the rest of Anatololacerta (BELLATI et al. 2015;KARAKASI et al. 2021). KARAKASI et al. (2021) suggested that there is no niche overlap for these five species belonging to Anatololacerta in the whole distribution region, except for one contact zone between A. finikensis and A. ibrahimi. In our study, possible distribution maps showed the niche overlaps for all species of Anatololacerta for both historical and future bioclimatic conditions, but physical barriers could restrict this possible niche overlapping.
The distribution of reptiles is influenced by anthropogenic factors including the human population size, trade, transportation and climatic changes (BICKFORD et al. 2010;SILVA-ROCHA et al. 2019). Global climatic changes also affect the distribution of many reptiles in different ways, leading to range expansion or habitat loss. One example of range expansion involves the Burmese python (Python bivittatus), which was introduced to North America by the pet trade. OSLAND et al. (2021) suggested that P. bivittatus could extend its range further north in response to winter warming. This assumption was tested not only in relation to invasive species but also to native species, and it has been predicted that some native species may expand or restrict their distribution range in the near future. SRINIVASULU et al. (2021) suggested that Eryx whitakeri, Eutropis clivicola and Uropeltis phipsonii may expand their ranges in Western Ghat with RCP 8.5, while the distribution range of Kaestlea laterimaculata may be restricted with RCP 4.5 in the same region. As a result of our study, we determined that while A. danfordi, A. finikensis and A. ibrahimi have the potential to expand their distribution range, there will be a potential narrowing in the distribution of A. pelasgiana and A. anatolica. However, the main question regarding this assumption is: how will they overcome the physical barriers? MARTINEZ-MEYER (2005) suggested that some barriers for one species may not be a barrier in the near future, and that ecological differences will make previously unsuitable areas suitable for certain species. This example provides an idea of how the physical barriers could be overcome. Our results also showed possible niche overlapping for members of the genus Anatololacerta. Thus, the interspecies competition that will emerge in the future will determine the species distribution, as well as the climatic suitability. LEGAULT et al. (2020) suggested that competition between species plays a large role during range expansions and that competition effects should be included in the models predicting future ranges.
As seen in our results, some species belonging to the genus Anatololacerta in Anatolia appear to have the potential to extend their range with the climatic changes occurring during the LGM and historical scenarios, as well as in the future RCP 4.5 scenarios. The range extension of the species determined by the models provides limited climatic information, but this assumption could be influenced by other factors such as species competition and physical barriers. Thus, our results suggest that the genus Anatololacerta should be monitored in the near future, with more locality records for niche overlapping and inter-or intraspecies competition.