**2. Material and methods**

*Habitats of the World - Biodiversity and Threats*

resulting in both African and Asian monsoons [7, 8].

variables spread over the survey area [20, 21].

patterns of the two mentioned species in Asia.

occur in Iran [30, 31].

Anderson [30].

During interglacial periods, the climate warmed, and forests returned to areas that

Analyzing species distribution models can help in conservation planning [9] and in understanding theoretical research [10] on ecological and evolutionary processes [1]. Species distribution models can be used to investigate the effect of climate changes on distributions and abundances of species [11], to determine biodiversity [12] and biogeographical patterns [13], to predict potential distribution [14], and to appraise possible future changes in the diversity [15]. Lizards, like other ectotherms [16], provide excellent models for analysis of species distribution under climate change [2]. MaxEnt is a general approach for characterizing probability distributions from small sample sizes [17–19]. MaxEnt estimates the probability distribution of maximum entropy (i.e., closest to uniform) based on environmental

The Scincidae family has more than 25% of all living genera and species of lizards [22]. The genus *Ablepharus* (Fitzinger, 1823) encompasses 10 valid species: *A. bivittatus* (Menetries, 1832), *A. budaki* (Göcmen, Kumlutas & Tosunoglu, 1996), *A. chernovi* (Darevsky, 1953), *A. darvazi* (Jeremčenko & Panfilov, 1990), *A. deserti* (Strauch, 1868), *A. grayanus* (Stoliczka, 1872), *A. kitaibelii* (Bibron & Bory, 1833), *A. lindbergi* (Wettstein, 1960), *A. pannonicus* (Fitzinger, 1824), and *A. rueppellii* (Gray, 1839) which are distributed in Europe, Turkey, Syria to Egypt, Azerbaijan, Armenia, Caucasus, Tajikistan, Kazakhstan, Kyrgyzstan, Uzbekistan, Turkmenistan, Afghanistan, Iran, Iraq, United Arab Emirates, Pakistan, and NW India [23–28]. The genus *Ablepharus* in the molecular phylogenic aspect is a sister taxon of the central and East Asian *Asymblepharus* [29]. *Ablepharus bivittatus* (Menetries, 1832), *A. grayanus* (Stoliczka, 1872), and *A. pannonicus* (Fitzinger, 1824)

*Ablepharus grayanus* was first described as *Blepharosteres grayanus* from Waggur District, northeast Kutch, India [26]. Later, Fühn [24] regarded it as a subspecies of *A. pannonicus* based on examination of a few specimens (three *A. grayanus*, four *A. pannonicus*). *Ablepharus grayanus* (Stoliczka, 1872) is now regarded as a distinct species. *Ablepharus grayanus* (Stoliczka, 1872) has a distribution range from northern and western India through Pakistan and Afghanistan to Eastern Iran [30, 31]. Researchers based on the morphological characters identified different species and subspecies—*A. brandtii* (Strauch, 1868) from Samarkand, Turkestan; *A. pusillus* (Blanford, 1874) from Basra, Iraq; *A. brandtii* vs. *brevipes* (Nikolsky, 1907) from Dech-i-Diz and Karun River, Iran; *A. persicus* (Nikolsky, 1907) from Shahrud, Iran; and *A. p. pannonicus* and *A. p. grayanus* [24]—in wide distribution range of *A. pannonicus*, that all species regarded to synonym *A. pannonicus* by

The general aim of this chapter is (1) to identify potential areas of distribution during three periods of the past, last interglacial (LIG: ∼120,000–140,000 years BP) and mid-Holocene (MH: ∼6000 years BP), (2) to describe current (~1950– 2000) distribution and suitable habitat, and to understand the biogeographical

once supported tundra vegetation [2]. During the last interglacial period (LIG: 150,000–120,000 years), temperature gradient increased in polar regions toward lower latitudes and caused sea level rise and reduction of ice sheets [4]. Briefly, the climate of the last interglacial had a relatively stable warm period [5]. Kerwin et al. [6] simulated terrestrial conditions at the mid-Holocene (6 ka) that indicated summer temperatures were warmer than at present in the high-latitude Northern Hemisphere. But during the mid-Holocene, northern Africa, Arabia, and southern Asia underwent conditions much wetter than at present, these conditions

**66**

## **2.1 Study area and records**

The study area encompasses the whole Iranian territory. We assembled the species occurrence data for each species based on a systematic biological survey by walking randomly through the habitat from 09:00 to 12:00 AM and 15.00 PM to evening (much of the activity time of species) during spring to summer 2010 and 2015. We used localities mentioned in previous studies (e.g., Anderson [30]; Vyas [28]). *Ablepharus grayanus* specimens were collected, and their distribution data were recorded (34 recorded) from Sistan and Baluchestan and Kerman Provinces, southeastern Iran. We gathered distribution data of *A. pannonicus* specimens collected under rocks or leaves on the floor of oak forest in the Zagros Mountains and in between the meadow grass in the Darvishab River Park (Baghmalek, Khuzestan Province) and recorded the exact location using the global positioning system (GPS). In other areas (Esfahan, Ilam, Kermanshah, Khorasan Razavi, Kurdistan, Lorestan, Mazandaran, Qum, Semnan, Zanjan, and Yasuj Provinces), we observed *A. pannonicus* in between the grasslands, shrubs, and steppes, and exact coordinates were marked with GPS (108 recorded).

#### **2.2 Data set and analysis**

We implemented maximum entropy modeling (MaxEnt, 3.3.3e http://www. cs.princeton.edu/~schapire/MaxEnt) of species geographic distributions with default parameters of the data to test samples. We examined 19 bioclimatic variables and 2 topographical variables with grids approximately 1 km2 precision (30 s × 30 s) for contemporary (~1950–2000) and 10 km2 precision (5 min × 5 min); we also examined 19 bioclimatic variables in the past (LIG and MH) in the related part of the world (Asia) [32, 33] (www.worldclim.org) (see the Appendix). To identify the correlation ratios between variables and presence records, openModeller (V. 1.0.7) [34] was used. Then we used SPSS IBM (version 22) for Pearson correlation coefficient [17]. We selected variables with a Pearson correlation lower than 0.75 to choose the variables that are ecologically important for species separation according to our observations and to describe habitat [35]. We conducted MaxEnt software with 10 replicates of the analysis that yield the best model for the studied species. MaxEnt provides state distribution models by the receiver operating characteristic (ROC) plots; ROC curves plot true-positive rate against false-positive rate [21, 36]. A value of the area under the curve (AUC) of 0.5–0.7 is taken to indicate that the result is a stochastic prediction [37, 38], values of 0.7–0.9 suggest useful models, and the values more than 0.9 indicate high accuracy [39]. We used DIVA-GIS 7.3.0.1 software for the mean predicted map and a logistic output of present records with suitability ranging from zero (unsuitable habitat) to one (the best suitable habitat) [40].

### **3. Results**

The final models in the present study showed good match and closely fitted the presence of the two species recorded in the study areas, as suggested by high AUC values (*A. grayanus* = 0.929 ± 0.087 and *A. pannonicus* = 0.979 ± 0.007). Moreover, two variables contributed for both species (BIO3 and slope), six variables for *A. grayanus*, and six variables for *A. pannonicus* were detected separately (**Table 1**). The last models in the mid-Holocene simulated high AUC values (*A. grayanus* = 0.975 ± 0.019 and *A. pannonicus* = 0.988 ± 0.006). In addition, three variables were important for both species, one variable for *A. grayanus*, and three variables for *A. pannonicus* were identified separately (**Table 2**). The last interglacial showed high AUC values (*A. grayanus* = 0.975 ± 0.019 and *A. pannonicus* = 0.988 ± 0.006) (**Table 3**). During this time, four variables for *A. grayanus* and six variables for *A. pannonicus* were recognized separately.

The model for *A. grayanus* predicted the distribution range presence of the species in the riparian and wet areas of northwest India, through Pakistan and Afghanistan, and oases and palm groves of the eastern and southeastern Iran. That distribution of the species was verified by using a comparison of environmental variables. Moreover, the climate variable model suggests that there are more suitable potential regions in the United Arab Emirates, Oman, Saudi Arabia, Iraq, Jordan, central Turkey, north Syria, south Turkmenistan and Uzbekistan, and west of China. The MH and the LIG simulated the distribution model for *A. grayanus*


#### **Table 1.**

*Relative of variables (in percentages) at the current period (1950–2000) used in MaxEnt model for the two studied species of the genus Ablepharus.*


#### **Table 2.**

*Relative of variables (in percentages) at the mid-Holocene, 6000 years ago (6 ka), used in MaxEnt model for the two studied species of the genus Ablepharus.*

**69**

**Figure 1.**

**Table 3.**

*Modeling the Past and Current Distribution and Habitat Suitability for Two Snake-eyed Skinks...*

**Variable Description of variables** *A. grayanus A. pannonicus*

BIO3 Isothermality [(BIO2/BIO7) × 100] 28.8 BIO4 Temperature seasonality (standard deviation × 100) 17 BIO7 Temperature annual range (BIO5–BIO6) 7.2

15.5

20.7

10.7

8.5

*DOI: http://dx.doi.org/10.5772/intechopen.82476*

BIO2 Annual daily temperature difference (minimal

BIO8 Average temperature of the wettest quarter of the

BIO9 Average temperature of the driest quarter of the

BIO15 Seasonality of precipitation (coefficient of

*model for two species of the genus Ablepharus.*

temperature maximal temperature)

year

year

BIO14 Precipitation of the driest month 56

variation)

BIO17 Precipitation of the driest quarter of the year 16.4 BIO19 Precipitation of the coldest quarter of the year 19.2

*Relative of variables (in percentages) at the last interglacial, 120,000 years ago (120 ka), used in MaxEnt* 

*Distribution map of Ablepharus grayanus in southwestern Asia and much of their potential distribution pattern in the region during: (A) current period (1950–2000); (B) the mid-Holocene, 6000 years ago (6 ka); and (C) the last interglacial, 120,000 years ago (120 ka). The four colored squares on the bottom left indicate the result of stochastic prediction of present species. The black circles refer to the collected specimens.*

*Modeling the Past and Current Distribution and Habitat Suitability for Two Snake-eyed Skinks... DOI: http://dx.doi.org/10.5772/intechopen.82476*


#### **Table 3.**

*Habitats of the World - Biodiversity and Threats*

recognized separately.

and *A. pannonicus* = 0.988 ± 0.006). In addition, three variables were important for both species, one variable for *A. grayanus*, and three variables for *A. pannonicus* were identified separately (**Table 2**). The last interglacial showed high AUC values (*A. grayanus* = 0.975 ± 0.019 and *A. pannonicus* = 0.988 ± 0.006) (**Table 3**). During this time, four variables for *A. grayanus* and six variables for *A. pannonicus* were

The model for *A. grayanus* predicted the distribution range presence of the species in the riparian and wet areas of northwest India, through Pakistan and Afghanistan, and oases and palm groves of the eastern and southeastern Iran. That distribution of the species was verified by using a comparison of environmental variables. Moreover, the climate variable model suggests that there are more suitable potential regions in the United Arab Emirates, Oman, Saudi Arabia, Iraq, Jordan, central Turkey, north Syria, south Turkmenistan and Uzbekistan, and west of China. The MH and the LIG simulated the distribution model for *A. grayanus*

**Variable Description of variables** *A. grayanus A. pannonicus*

BIO3 Isothermality [(BIO2/BIO7) × 100] 11.4 8.2 BIO4 Temperature seasonality (standard deviation × 100) 27

BIO8 Average temperature of the wettest quarter of the year 18.5

BIO11 Average temperature of the coldest quarter of the year 16

BIO15 Seasonality of precipitation (coefficient of variation) 10.5

Slope Slope 6.5 19.2

*Relative of variables (in percentages) at the current period (1950–2000) used in MaxEnt model for the two* 

**Variable Description of variables** *A. grayanus A. pannonicus*

BIO3 Isothermality [(BIO2/BIO7) × 100] 22.8 33.9

BIO7 Temperature annual range (BIO5–BIO6) 59.7 1

*Relative of variables (in percentages) at the mid-Holocene, 6000 years ago (6 ka), used in MaxEnt model for* 

2.1 0.6

27.5

0.5

16.6

20.3

BIO2 Annual daily temperature difference (minimal

BIO2 Annual daily temperature difference (minimal

temperature maximal temperature)

BIO5 Maximum temperature of the warmest month 1.1

BIO9 Average temperature of the driest quarter of the year 23.3

BIO14 Precipitation of the driest month 18.4

BIO17 Precipitation of the driest quarter of the year 24 BIO19 Precipitation of the coldest quarter of the year 15.4

BIO4 Temperature seasonality (standard deviation ×

BIO8 Average temperature of the wettest quarter of

BIO15 Seasonality of precipitation (coefficient of

*the two studied species of the genus Ablepharus.*

*studied species of the genus Ablepharus.*

temperature maximal temperature)

100)

the year

variation)

BIO9 Mean temperature of the driest quarter of the year 15.3

**68**

**Table 2.**

**Table 1.**

*Relative of variables (in percentages) at the last interglacial, 120,000 years ago (120 ka), used in MaxEnt model for two species of the genus Ablepharus.*

#### **Figure 1.**

*Distribution map of Ablepharus grayanus in southwestern Asia and much of their potential distribution pattern in the region during: (A) current period (1950–2000); (B) the mid-Holocene, 6000 years ago (6 ka); and (C) the last interglacial, 120,000 years ago (120 ka). The four colored squares on the bottom left indicate the result of stochastic prediction of present species. The black circles refer to the collected specimens.*

#### **Figure 2.**

*Distribution map of Ablepharus pannonicus in southwestern Asia and much of their potential distribution pattern in the region during: (A) current period (1950–2000); (B) the mid-Holocene, 6000 years ago (6 ka); and (C) the last interglacial, 120,000 years ago (120 ka). The four colored squares on the bottom left indicate the result of stochastic prediction of present species. The black circles refer to the collected specimens.*

that were more suitable areas than present in southwestern Asia today (**Figure 1**). The model for *A. pannonicus* predicted the occurrence of range of the species in steppe areas, grassy, rocky hills separated by oak forest of the Zagros Mountains in the west, and palm groves in southwestern Iran. In addition to the mentioned habitat, the distribution range model of the species predicted that *A. pannonicus* occurs in Iraq, Kuwait, Pakistan, Afghanistan, Tajikistan, Turkmenistan, Uzbekistan, and suitable potential northeast in Syria, Turkey, Kazakhstan, and patchwork areas of northern India. The simulated MH distribution range model for *A. pannonicus* had continuous restriction in east Syria, throughout Iraq, and north Saudi Arabia toward southeastern Turkmenistan. Also, simulated suitable potential fragmented areas of north India and central China were demonstrated. The LIG simulated distribution ranges were the same as the MH suitable potential habitat (**Figure 2**).
