ON SOME FACTORS DRIVING THE PRESENCE OF AMPHIBIANS IN WATER BODIES OF THE UPPER OKA BASIN (CENTRAL RUSSIA)

A better understanding of the factors influencing the distribution of amphibians is needed to conserve amphibian species in regions highly populated by people. A total of 71 water bodies were examined in the Upper Oka Basin (the Central Part of the Russian Plain, Russia, Kaluga Oblast and adjacent regions). For each site, the presence and absence of 11 amphibian species were determined. We use the logistic regression within the glm function from the R Stats Package to estimate the influence of predictors on the probability of the species presence. The species with the highest occurrence were Pelophylax lessonae, P. ridibundus, Rana temporaria, and Bufo bufo. The type of terrestrial vegetation that surrounded the water body was a significant factor for three amphibian species. We found R. temporaria and B. bufo avoiding open habitats; and P. ridibundus avoiding wooded habitats. The degree of water moving was a significant factor for two amphibian species. Lissotriton vulgaris was present more frequent at lentic waters, and P. ridibundus preferred lotic waters. The percentage of the vegetation covering the water surface was a significant factor for Rana arvalis, which was more often present in water with 50% or more coverage. Acidity was a significant factor for Pseudepidalea viridis, which was detected only in neutral and alkaline waters. Total dissolved salts and the area of water body were significant predictors for no amphibian species. The presence of the Chinese sleeper (Perccottus glenii), an invasive fish species, was not significantly important in predicting detection or non-detection for any species. Many water bodies in the Upper Oka Basin that were likely once suitable for amphibians may not be occupied by amphibians due to barriers to dispersal from other sites and due to stochastic extinction. To estimate the capability of amphibian immigration from other sources, we identified the presence of lotic and lentic permanent water bodies within 1 km of a surveyed site. These factors were not significant for any species. Further investigations may achieve the best measure of connectivity of amphibian habitats. To conserve amphibians we need to keep terrestrial habitats surrounding the water bodies, especially wooded habitats.


Introduction
Amphibians have been the object of numerous ecological investigations in both field and in laboratory settings (Beebee, 1981(Beebee, , 1983Kaufman, 1989;Fayzulin, 2010;Smirnov, 2013). Many investigations in amphibian ecology are concerned with the quality of breeding and residential water bodies, which has great value for the continued persistence of amphibian populations. The quality of amphibian habitats is influenced by the type of vegetation in the water body and surrounding terrestrial habitat, the hydroperiod and water quality, the presence of predators and competitors, the prevalence of diseases, and the nature and frequency of human disturbances (Hamer & Mc-Donnell, 2008;Collins et al., 2009). Specifically, aquatic vegetation provides shelter for larval and adult amphibians and oviposition sites (Hartel et al., 2007). Terrestrial vegetation provides opportunities for dispersal, food, shelter and overwintering sites once individuals have metamorphosed (Hamer & McDonnell, 2008). The terrestrial vegetation influences the water temperature, and the water temperature determines the development of eggs of amphibians (Nikolaev, 2007). The most important factor of water quality is water acidity, which can affect the success of reproduction and causes various morphological anomalies in some frogs (Flax, 1986;Nikolaev, 2007;Fayzulin, 2010). Amphibians are highly sensitive to environmental pollutants in the water; particularly, the presence of dissolved metals and salts in water (i.e. high conductivity) and high nutrient loads negatively affect amphibian populations in urban and suburban areas (Hamer & McDonnell, 2008). Water current and water transparency are also important to amphibians (Semenov et al., 2000).
Populations of amphibians across the globe have been declining for the last few decades due to climate change, habitat loss and diseases (Alford & Richards, 1999;Lips et al., 2003;Stuart et al., 2004;Collins et al., 2009). To conserve threatened amphibian species we need to identify species' tolerance and optimal environmental conditions especially under increasing urbanisation and the development of recreational activities. Understanding why threatened species and biological communities persist in urbanised areas can help us create successful conservation management plans for preventing and reversing future declines (Valdez et al., 2015).
The vulnerability of amphibians to disturbance relates to their limited dispersal ability (Araújo et al., 2006). Many ponds may be not inhabited by am-Many ponds may be not inhabited by amphibians due to the isolation of ponds and the limited dispersal distance of amphibians (Semenov et al., 2000;Smith & Green, 2005). For example, the number of permanent water bodies within a given distance of a surveyed water body has been shown to be the most important predictor for occupancy, colonisation, and abundance of amphibian species (Valdez et al., 2015). However, it has been shown that dispersal distance might be much larger than expected for amphibians (Smith & Green, 2005).
There is significant geographic variation in life-history characteristics of amphibians and in tolerance to some environmental factors (Vershinin, 1995;Morrison & Hero, 2003;Kuzmin, 2012). The different abilities of amphibians to cope with the effects of urbanisation are also likely to generate regionally contrasting long-term trends in their community dynamics (Hamer & McDonnell, 2008). So it is necessary to survey ecological features of species to conserve amphibians in a given region.
The Upper Oka Basin is one of the most populated and anthropogenically disturbed regions of Russia. Previous research focused on the species composition, distribution and abundance of amphibians in terrestrial habitats, bionomics, and feeding of amphibian species in this region (Alekseev & Sionova, 2002;Ruchin & Alekseev, 2008;Korzikov et al., 2014;Korzikov, 2016); but a better understanding of the factors influencing distributions of amphibians is needed to plan further investigations. Therefore, the aim of the study was to evaluate the parameters of water bodies, which determine the presence or absence of different amphibian species in the Upper Oka Basin.

Study area
The Upper Oka Basin is located in the centre of the Russian Plain and contains the Basin of the River Oka from its source to the mouth of the River Nara near Serpukhov in Moscow Oblast. This region encompasses primary areas of the Kaluga oblast excluding western territories.
We surveyed a total of 71 water bodies from different districts of the Kaluga and Tula Oblast. We observed water bodies such as puddles, quarries, ponds, rivers, streams, and streams with one or several dams, backwaters (parts of a river in which there is little or no current), bogs, oxbows, and waterworks (Table 1). Sixty water bodies were permanent, and eleven water bodies were temporal. Forty-one sites were located in rural areas, and twenty-nine sites in urban areas.
We analysed four quantitative and five qualitative parameters of water bodies (Table 1).

Field surveys
Amphibian surveys were conducted from March to August in 2005-2016. The presence or absence of amphibian species was registered in every water body. Amphibians were detected in the daytime visually by hand and using a net with 0.6 m diameter of mouth, 1.7 m length of handle, and 2 mm diameter mesh. Most of the water bodies were surveyed at least three times during the season. A few ponds were surveyed one time during the season because they were located in hard-to-reach areas, and we assumed that single surveys are sufficient for the detection of amphibians based on expert experience (Alekseev, personal communication).
Species were identified using sounds and morphological characteristics. Species of the genus Pelophylax, which are not distinguishable by morphological characteristics, were identified using analysis of intron-1 of the serum albumin gene (SAI-1) from tissue of frogs through polymerase chain reaction (Plötner et al., 2009).
The Chinese sleeper, Perccottus glenii (Dybowski, 1877), is an introduced fish species, which is known to negatively impact populations of amphibians (Semenov et al., 2000;Reshetnikov, 2003). This fish was detected using dipnetting. The pH and total dissolved salts (TDS) were measured in 37 water bodies using portative electronic readers pH-009 and TDS-3 (equipment consisted 0.1 for pH and 1 mg to litre for TDS). Areas of water bodies were de-tected using satellite images on Google Maps Calculator (https://3planeta.com/googlemaps/googlemaps-calculator-ploschadei.html).

Data analysis
Data were analysed using R (R Core Team, 2014). The logistic regression within glm function from the R Stats Package was applied to estimate the influence of predictors on the probability of presence of every species using two different models. First, we computed the glm model for six factors which were measured in all 71 water bodies: type of terrestrial vegetation, degree of water moving, presence of lentic and lotic water bodies within 1 km, presence of Perccottus glenii, and percentage of the vegetation covering the water surface. Secondly, we computed the glm model for three factors, which were measured in the subset of 37 waterbodies: surface area, pH, TDS. Independent variables were uncorrelated (p > 0.2 for Pearson's correlation).

Results
There were 11 amphibian species in water bodies of the Upper Poochye (Table 2). No species occupied more than half of all examined sites. Species with the highest frequency of occurrence were the Pool Frog (Pelophylax lessonae Camerano, 1882), the Common Frog (Rana temporaria Linnaeus, 1758), the Common Toad (Bufo bufo Linnaeus, 1758), and the Marsh Frog (Pelophylax ridibundus Pallas, 1771). Table 2. Number of water bodies of different kinds, which were inhabited by different species of amphibians. Sample size in brackets after water body type
The degree of water flow was a significant factor for two amphibian species (Appendix 1, Table 4). The Smooth Newt (Lissotriton vulgaris Linnaeus, 1758) was present more frequently in lentic waters (p = 0.016), and Pelophylax ridibundus preferred lotic waters (p = 0.040).
The percentage of the vegetation covering the water surface was a significant factor (p=0.019) for Rana arvalis (Nilsson, 1842). This frog was more often present in waters with 50% or more coverage.
The presence of the Chinese sleeper and the presence of lotic or lentic water bodies within 1 km were not significant in predicting detection or non-detection for any species (Appendix 1).
The acidity was a significant factor (p = 0.048) for Pseudepidalea viridis (Laurenti, 1768), which was detected only in neutral and alkaline waters, with pH = 7-10 (Appendix 2). Total dissolved salts and the area of water body were significant predictors for no amphibian species (Appendix 2).

Discussion
The terrestrial vegetation was the best predictor of the presence of three amphibian species (Rana temporaria, Bufo bufo, and Pelophylax ridibundus). This result agrees with previous research from other regions: terrestrial habitat factors were found to be especially important in determining the suitability of breeding sites for Rana temporaria in Ireland (Marnell, 1998) and Sweden (Loman & Lardner, 2006). Previous studies have also found the importance of terrestrial habitats for amphibians within conservation programmes (Valdez et al., 2017) and for females of the green and golden bell frog Litoria aurea (Lesson, 1830) (Valdez et al., 2016). The terrestrial vegetation can determine the temperature of water, which plays an important role in amphibian breeding and development (Moore, 1939). Wooded sites can provide superior refugia and feeding for adult amphibians, and some species require forests for part of their life cycle (Guerry & Hunter, 2002;Kuzmin, 2012). Rana temporaria is sensitive to air humidity (Dinesman, 1948), so it is a forest species to minimize the exposure to drier conditions (Kutenkov, 2017). Bufo bufo is also characterised as a forest species although it is less sensitive to humidity than the Common Frog (Kuzmin, 2012). Pelophylax ridibundus prefers open habitats because open habitats are often due to overflowing large rivers and ponds, and this species is connected with large rivers and ponds (Ruchin et al., 2009;Svinin, 2013).
The water quality significantly influenced the presence of Pseudepidalea viridis, a species that is not abundant in the Upper Oka Basin (Korzikov, 2016). The range of pH in which P. viridis was present in the Upper Oka Basin was close to those observed in the Middle Volga (6.6-10.0; Fayzulin, 2010). This species is qualified as synanthropic, e.g. it lives near, and benefit from, an association with humans and the somewhat artificial habitats that humans create around them (Alekseev & Sionova, 2002), and artificial water bodies tend to be more often alkaline than natural ones.  Generally, the range of acidity and total dissolved salts sets were not limiting factors to the occurrence of amphibians in the Upper Oka Basin: the focal species in this study can tolerate a wide range of variation in pH and TDS (Kuzmin, 2012). Besides, the non-significance might be explained by the absence of extremely low or high values at the water bodies surveyed. Our data agrees with the results of research in Sweden (Loman & Lardner, 2006) that found the scarcity of amphibians in some areas could not be explained by water quality alone but the quality of the terrestrial habitat surrounding the ponds and the metapopulation structure.
Our results indicated that many water bodies are suitable for amphibians but are not occupied by amphibians now. It is known that some ponds may not be inhabited by amphibians owing to their isolation and barriers to migration of amphibians (Semenov et al., 2000). To estimate the capability of amphibians' immigrations from other sources we identified the presence of lotic and lentic permanent water bodies within 1 km of a surveyed site. These factors were not significant for any species. This fact contrasts with Valdez et al. (2015) who found that the number of permanent water bodies within a kilometre of surveyed sites was the best predictor of the occupancy of the green and golden bell frog (Litoria aurea), and also with the paper of Pellet et al. (2004) who found that the presence of calling males of the tree frog (Hyla arborea (Linnaeus, 1758)) is influenced by urbanisation around the pond and closeness of roads. This difference may be due to the fact that Valdez et al. (2015) used the number of water bodies rather than our study that simply used the presence or absence of water bodies. Furthermore, forest landscapes of the Up-per Oka Basin due to relative soft microclimatic conditions in wooded habitats and numerous puddles of water may be less stressful for amphibians than industrial landscapes of Australia and landscapes of Switzerland, so the «ponds-as-patches» view (Marsh & Trenham, 2001) is less applicable to our data. Lastly, investigations of this subject require more rigorous assessment of amphibian populations based on the metapopulation approach (Marsh & Trenham, 2001;Smith & Green, 2005;Griffiths et al., 2010).