Habitat Preferences of Natrix tessellata at Strofylia, Northwestern Peloponnese, and Comparison to Syntopic N. natrix

MERTENSIELLA 18

302-310

20 September 2011

ISBN 978-3-9812565-4-3

Habitat Preferences of Natrix tessellata at Strofylia, Northwestern Peloponnese, and Comparison to Syntopic N. natrix Yannis Ioannidis & Konrad Mebert Abstract. The habitat preferences of Natrix tessellata and N. natrix were studied in the wetlands of Strofylia in southern Greece. Both species used most available water habitats, but N. tessellata dominated in permanent lagoons with brackish water or saltwater. N. natrix uses equally temporary and permanent water bodies and prefered areas with large breeding concentrations of frogs and toads, such as ponds, wet meadows, marshes and a shallow lakes. Based on their surface frequency the habitat overlap between the two species is low. N. tessellata was found to be more aquatic, whereas N. natrix has been observed at large distances from water. Regarding the terrestrial habitats, N. natrix inhabited areas with mainly forests, whereas N. tessellata was common in more open situations, particularly rocky shores with scattered bushes. The hibernation period for N. tessellata is relatively short and lasted around three months. Some individuals left their hibernaculaum during warm winter days. The terrestrial activity pattern of N. tessellata is unimodal in spring with equal numbers observed throughout the day until dusk. In summer and autumn they switched to a bi-modal activity pattern with many animals active in the morning hours, followed by a sharp decrease during mid-day and a second higher peak around dusk and during the first hours of the night. High road mortality of N. tessellata was recorded along a stretch of 2.1 km on the north shore of Prokopos Lagoon We estimated that vehicles kill approximately 1100 dice snakes annually on this road. Key words. Peloponnese, Greece, dice snake, grass snake, habitat preferences, activity, road mortality

Introduction Natrix tessellata is widespread in the mainland of Greece, the Ionian Islands, some of the Aegean islands and Crete (Chondropoulos 1986). There are only a few data about its habitats in Greece, including wetlands such as lakes, ponds, canals, rivers and streams (Cyrén 1941, In den Bosch & Musters 1981, Brinsøe 1986, Keymar 1986, Bousbouras & Ioannidis 1994), but also brackish water lagoons in Corfu (Wütschert 1984) and Crete (Sowig 1985). However, information about the biology and ecology of this species in Greece remains scarce. Over the entire mainland and some of the islands N. tessellata is sympatric with N. natrix. According to the niche theory, some type of resource partitioning is necessary to avoid or limit interspecific competition in ecologically similar and sympatric species (Pianka 1981, Walter 1991). Most studies on reptiles focus on three main niche dimensions: food, space, and time, an approach introduced by Pianka (1973). The relatively few comparative studies on sympatric populations of N. tessellata and N. natrix emphasize the partitioning of food niche (Filippi et al. 1996, Luiselli & Rugiero 1991, Janev Hutinec & Mebert 2011). Generally, N. tessellata is a sit-and-wait predator, feeding mainly on fish, adding or switching locally to amphibian diet (Lanka 1978, Gruschwitz 1986, Lenz & Gruschwitz 1993, Filippi et al. 1996, Zimmermann & Fachbach 1996, Gruschwitz et al. 1999, Luiselli et al. 2007, Brecko et al. 2011), whereas N. natrix is predominantly an active forager on am-

phibians, but exhibiting a broader food spectrum that includes also birds, rodents, reptiles and fish (Madsen 1983, Hailey & Davies 1986, Luiselli & Rugiero 1991, Luiselli et al. 1997, Luiselli et al. 2005). In most vertebrates partitioning exists mainly at the habitat level (Schoener 1974). Snakes are considered as an exception by some authors (Toft 1985, Vitt 1987), as partitioning seems to exist mainly at the food level. A recent review by Luiselli (2006) supported this view, although it disagreed with the perception that most snakes are food specialists and highlighted the importance of interspecific competition in structuring the snake communities. Information on habitat partitioning between N. natrix and N. tessellata is only anecdotal with the exception of a recent investigation by Janev Hutinec & Mebert (2011). The activity of N. tessellata has been studied mainly on central European populations (see refs. in Gruschwitz et al. 1999). It is a diurnal species that seems more active in the morning and in the afternoon (Lenz & Gruschwitz 1993, Gruschwitz et al. 1999). Some degree of nocturnal behavior has been reported for northern Italy (Scali et al. 2001, Scali 2011) and variously documented by authors in Mebert (2011) and summarized in Mebert et al. (2011b). The activity period for the central European populations starts between late March to early May after a hibernation of around 6 months (Hecht 1930, Gruschwitz et al.1999). There are numerous reports of a strong decline of observations in the hot summer months (e.g. Lanka 1978, Lenz & Gruschwitz 1993, Gruschwitz et al.1999). The snakes

©  2011 Deutsche Gesellschaft für Herpetologie und Terrarienkunde e.V. (DGHT), Rheinbach, Germany

302

Yannis Ioannidis & Konrad Mebert

return to their hibernacula between mid-September to mid-October (Schweizer 1962, Gruschwitz et al.1999). In contrast, there are no equivalent data for populations from southern Europe. The aim of this study is to partially compensate for the lack of information from southern European populations by comparing the habitat preferences of N. tessellata and N. natrix in a wetland system with syntopic populations and to highlight some aspects of daily movement and terrestrial activity of N. tessellata based on road counts. Materials and Methods Study Area This study took place at the Natura 2000 site of Strofylia, northwestern Peloponnese, Greece (Fig. 1). The Stone pine (Pinus pinea) forest along the coast of Strofylia is probably the only remaining true forest of this tree species in Greece and represents almost 80% of the country’s total stone pine forests (Moussouris & Regato

1999). The land habitat types of this area in relation to reptiles have been described in Ioannidis et al. (2008). The flatlands of Strofylia are characterized by numerous wetlands, which are better known for their ornithological value, leading to their designation as aprotected area under the Ramsar Convention. The area includes large wetlands, such as the lagoon of Prokopos and the lake and marches of Lamia, as well as many smaller permanent or temporary ponds, ditches, wet meadows, and small rivers. Two more lagoons, Kalogria and Kotychi, are present in the north and south of the site (Fig. 1 insert). The climate is typical for the eastern Mediterranean with a mean annual rainfall of 706 mm. (range 405– 1044 mm.) concentrated in the period from October to March (Assimakopoulos 1999). A total of 18 sampling squares, representative of the various habitat types in the area, were installed around the Prokopos Lagoon to record the composition of local reptile species. The borders of each quadrat were defined with a combination of pieces of tape and landscape element. Nine squares included wetlands of different size, depth and salinity. The quadrats had a rela-

Fig. 1. Google Earth view of the Peloponnese Peninsula and the Strofylia area in the northwest. Insert: Aerial view of the Strofylia area with the localities mentioned in the text. 1. Kalogria Lagoon, 2. Prokopos Lagoon, 3. Lamia Lake and marshes, 4. Kotychi Lagoon.

303

Habitat Differences in Dice and Grass Snakes on the Peloponnese

tively large size (30 ha) which were sampled using general herpetological search techniques, including turning logs and all sorts of debris large enough to cover a resting snake, disturbing the leaf litter, and looking into the bundled stems of suitable plants. All individuals were counted, on land, in water, or in refugia. Each square was sampled for a total of approximately 20 hours, mainly in the spring, early summer and autumn during 2003–2004. The temporal sampling procedure for the various squares was almost identical during the seasons and day time. Road surveys were carried out in the same general region by driving a vehicle with a speed of maximally 30 km/hour. Every dead or alive reptile was recorded and it’s GPS position and some environmental data taken. Road surveys were conducted during the period of January 2003 – June 2004 for a total of 35 days. They were approximative equally spread among seasons and covered a total distance of 1135 km. To avoid inconsistencies, data for the habitat preferences and road counts were analysed separately. In addition, a more systematic road survey targeting dice snake mortality was initiated between January 2003 and December 2004 on a 2.1 km road segment north of Prokopos Lagoon (Figs. 2, 3). During the first passage on this road segment all visible road casualties were removed. Within the next 24 hours, the road was inspected again at a very slow speed up to three times in

Fig. 3. Photo of the shore road north of the Prokopos Lagoon. Photo: Benny Trapp.

order to avoid the loss of road casualties through scavengers. The time of each survey wasn’t always the same, but usually the first was conducted in the early morning, the second around noon and the third in the evening or early night. Additional counts were conducted on this road segment during the months of February, April, May, June, July and November in 2005 and 2006 to assess the start and end of the active period and to establish whether those mortality rates were temporary or show a similar pattern every year.

Fig. 2. Prokopos Lagoon and the shore road, where the road surveys took place.

304

Yannis Ioannidis & Konrad Mebert

Results Habitat Preferences In the study area Natrix tessellata is syntopic with N. natrix. Both species use most available aquatic habitats with the exception of a stream, where only N. natrix was observed. The total counts of both species from the nine wetland types are presented in Table 1. N. natrix was dominant in most wetlands with the exception of lagoons, where N. tessellata was more common. The value of Spearman’s correlation coefficient between the two species was negative (r = –0.244) but the correlation was very weak (Pp = 0.52), indicating that there is no strong habitat segregation between them. The niche breath, calculated with the Shannon-Wiener measure, was high for N. natrix (J = 1.934) with a maximum index value of 2.197, whereas that for N. tessellata was lower (J = 1.171). Niche overlap calculated with Pianka’s modification of MacArthur-Levin’s measure was Ojk = 0.166. Both species were observed mainly inside the water or on the banks, less than 5 m from the water line (Tab. 2). However 36% of the observations of N. natrix were at greater distances from the water, including 31 observa-

tions with a distance between 20–250 m from the nearest wetland. Most distant observations (84%) concerned inactive individuals that were found under logs and dead branches. N. tessellata was more frequently in close proximity to water and its farthest distance from water was around 60 meters. Regarding terrestrial habitat use for distances >20 meters from water, N. natrix was most common in forests (87%), followed by open areas with bushes (10%) and cultivated fields (3%). Among the forest observations, 55% were in stone pine, 30% in oak, and 15% in Aleppo pine forest. In contrast, terrestrial N. tessellata was found mainly in rocky areas scattered with bushes and only one observation came from a stone pine forest. Activity and Road Mortality A total of 236 road casualties, reptiles and amphibians, were found during the survey. They belonged to 21 different species and included 43 amphibians, 12 turtles and tortoises, 20 lizards and 161 snakes. Obviously this number is only a fraction of the animals that are killed every year. The smaller sized species, usually lizards and

Tab. 1. Number of Natrix natrix and N. tessellata in nine wetland types, observed during surveys of 18 sampling squares from 2003–2004 in the Strofylia area, Greece. Presence of anuran: Scarce = observed occasional single individuals during survey visits; Abundant = choruses of large number of individuals during the breeding season, but impossible to count individual males, and large number of individuals observed during survey visits; Temporary = large choruses early in the breeding season but scarce later. Salinity level: Freshwater < 0.5 ppm; Brackish 0.5–30.0 ppm; Saltwater > 30 ppm.   Natural stream Temporary pond Wet meadow Large channel by the sea Marsh Ditch: small artifical channel Saltwater lagoon Lake Brackish water lagoon

Natrix tessellata 0 1 2 4 4 7 19 29 120

Natrix natrix 13 43 34 6 33 35 1 35 7

Tab. 2. Distance of syntopic watersnakes species from nearest water body at Strofylia, Greece. The data include observations on 18 sampling quares and some additional observations made outside the squares. The surveys were conducted in Strofylia area during the years 2003–2004.   In the water 0–5 meters 5–20 meters > 20 meters

Natrix tessellata % (n = 192) 37.0 41.1 18.8 3.1

Natrix natrix % (n = 238) 27.3 36.6 23.1 13.0

Presence of fish No No No Yes No No Yes Yes Yes

Presence of anuran Scarce Abundant Abundant Scarce Abundant Temporary No Abundant Temporary

Salinity level Freshwater Freshwater Freshwater Brackish-Saltwater Freshwater Freshwater-brackish Saltwater Freshwater Brackish

frogs, probably are underrepresented in the samples due to the difficulty to recognize them from a moving car. The full list is presented in Table 3. Nearly 44% (n = 103) of all road casualties were N. tessellata. A 2.1 km segment of the road that passes between the Prokopos Lagoon and Mavra Vouna Hill was responsible for a disproportionably large percentage of road casualties, the majority of wich were N. tessellata (Fig. 4). From 3 to 7 systematic road surveys were conducted monthly from January 2003 until December 2004 to estimate the mortality rate on this road section. The results were pooled as daily mortality rates for N. tessellata per month and are pre-

305

Habitat Differences in Dice and Grass Snakes on the Peloponnese

Tab. 3: Taxa and numbers of individuals of reptiles and amphibians found dead on the network of paved and unpaved roads in the Strofylia area between January 2003 and June 2004. Those numbers do not include road casualties from the 2.1 km road segment north of Prokopos Lagoon where more systematic road surveys were conducted. Observations by B. Trapp (pers. comm.) add following species as casualties from that road, but not listed in the table: Ablepharus kitaibeli, Lissotriton graecus, and Bufo bufo. Amphibians Bufo (sensu lato) viridis Hyla arborea Pelobates syriacus Pelophylax or Rana sp. Reptiles Turtles and tortoises Emys orbicularis Mauremys rivulata Testudo hermanni Testudo marginata Lizards Pseudopus apodus Anguis cephallonicus Podarcis taurica Lacerta trilineata Snakes Hierophis gemonensis Platyceps najadum Elaphe quatuorlineata Zamenis situla Zamenis longissimus Malpolon insignitus Natrix natrix Natrix tessellata Vipera ammodytes Total

19 7 9 8

2 6 3 1 3 1 9 7 9 5 5 4 1 10 18 103 6 236

sented in Figure 5. The mean daily mortality rate from all counts was 3.25 individuals/day. There was a peak in spring (March to May) after which the numbers were relatively stable through the summer and most of the autumn. Road mortality sharply declined in winter. The additional road surveys in 2005–2006 showed no significant difference of mean daily mortality rate for the relevant months compared to the counts in 2003 and 2004 (x2 = 0.470, 0.25 < Pp < 0.50). A comparison of dead and alive individuals on the road during the survey days shows a strong relationship of 0.933 (F = 15.84, P < 0.0001). This means that road activity greatly increases the risk of being killed (coefficient of determination = 0.870). Active individuals were observed from February until November. Only one N. tessellata was found moving on land in January at a temperature of 14 °C (shadow) and water temperature of 10.5 °C. The beginning of the active season was around the mid-February. The highest frequency of snakes was

306

Fig. 4. Road casualities of Natrix tessellata and N. natrix. Photo: Benny Trapp.

Fig. 5. Mean daily mortality rate per month of Natrix tessellata on a 2.1 km road segment north of Prokopos Lagoon, Strofylia, Greece. Data collected during surveys in 2003 and 2004.

recorded from March until May with a decrease in summer and a second much lower peak in September and October, which involved mainly younger individuals of less than 45 cm including a few newborns under 25 cm (Fig. 6). Activity ceased after the first 10 days of November and by then, most of the snakes had retreated to their hibernacula where they remained for almost three months. From February until the mid of June most individuals were observed during the midday (43%) and morning (32%). This changed drastically in the summer and up to October, when most individuals were recorded in the evening and at night (59%). During two days (18 April 2003 and 29 October 2004), the search for active snakes was intensified by repeating road surveys every 2 hours for a 24 h period. The activity pattern that was observed in those two days was more or less similar with the previous results (Fig. 7). During the 24 h-survey in April there was a constant presence of snakes during the day with a sharp decline after dark, whereas in October the snakes were present in the morning hours

Yannis Ioannidis & Konrad Mebert

Discussion

Fig. 6. Daily mean number of active Natrix tessellata per month (bars) along a 2.1 km road segment north of Prokopos Lagoon, Strofylia, Greece, plotted against minimum and maximum water levels (triangles) and mean air temperatures (circles). Grey part of the bar represents adults and white part individuals under 45 cm total length. Data collected during surveys in 2003 and 2004.

Fig. 7. Number of active Natrix tessellata during road counts on 18 April 2003 (triangles) and 29 October 2004 (squares) along a 2.1 km road segment north of Prokopos Lagoon, Strofylia, Greece. The horizontal axis is time in GMT+2.

with a decline at noon and a significant increase around dusk and at the early night hours. The salinity of the Prokopos lagoon varies significantly during the year. In 2004 an environmental station 20 meters from the northern coast, where the road surveys were conducted and a connection to the sea exists, measured the salinity of the water. It ranged from 0.7 to 28 ppm (= psu) with the exception of the period from July to September, when the salinity increased to values between 32 to 43 ppm. A second station in the southern part of the lagoon where the majority of freshwater is collected, recorded similar fluctuations but the maximum salinity was with 29 ppm much lower. N. tessellata were active in the northern part of the lagoon even when the salinity reached saltwater levels.

The degree of habitat overlap between the two species is low in Strofylia, even though both species can be found in a similar variety of aquatic habitats and do not exclude each other (weak correlation of a negative Spearman’s coeffcient), but they inhabit the variety of aquatic habitats to different proportions. For example, Natrix tessellata prefers the larger water bodies that support large fish populations and avoids areas with temporary water, such as streams, ponds, and wet meadows. It seems to cope well with increased levels of salinity and the densest populations have been observed in a lagoon that yields brackish to saltwater during most of the year (Fig. 8). In contrast, N. natrix is distributed more evenly across the various water types. It appears to frequent similarly often temporary and permanent water bodies and prefers smaller open water bodies with abundance of shore- and floating-vegetation that have the largest breeding concentrations of frogs and toads such as ponds, wet meadows, marsh and shallow lake. A similar low overlap in the use of water habitats between N. tessellata and N. natrix has also been reported from southern Croatia (Janev Hutinec & Mebert 2011). The low interspecific overlap was attributed to their different availability and preference of prey, as N. tessellata was feeding exclusively on fish and N. natrix mostly on amphibians. In sympatric populations from Central Italy, N. tessellata is feeding exclusively on fully aquatic prey consisting of fish and tadpoles, whereas the more terrestrial N. natrix feeds to approximately 90% on amphibians, including migrating toads (Luiselli & Rugiero 1991, Filippi et al. 1996). An increased underwater vision ability of N. tessellata compared to N. natrix has been technically investigated and confirmed by Schaeffel & Mathis (1991). The availability of preferable prey items in correlation with distinct behavior was not studied in the Strofylia area, but could also explain the differences in habitat preferences (see Tab. 1 for availability and abundance of prey types in different habitats). For example, the absence of N. tessellata from the surveyed stream is not understandable at the first glance, as this species has been observed in many rivers and streams throughout Greece (Trapp 2007, Valakos et al. 2008). However, stream-lake segregation has also been observed in the region of Prespa Lake, northern Greece, where N. tessellata is considered more common in or near the lakes and N. natrix is found in drainage ditches and streams (Ioannidis & Bousbouras 1997). But elsewhere, N. tessellata also inhabits streams (see refs. in Gruschwitz et al. 1999) or simply occupies all available open aquatic habitats, lentic and lotic systems (Mebert et al. 2011a). For example, N. tessellata thrives in three Mediterranean streams of central Italy with permanent or temporary water, where it feeds up to 90% on fish (Luiselli et al. 2007). At our study site in Strofylia, the observed absence of fish and the scarcity of frogs in the studied stream probably render that aquatic habitat un-

307

Habitat Differences in Dice and Grass Snakes on the Peloponnese

Fig. 8. A Natrix tessellata hunting in Prokopos Lagoon, a large brackish water body with a high abundance of fish. Photo: Benny Trapp.

suitable for N. tessellata and supports only a low number of N. natrix. Although both species have been observed mainly near or in the water, N. tessellata seems to exhibit a more aquatic behavior than N. natrix as was often reported from other areas (e.g. Filippi et al. 1996, Gruschwitz et al. 1999). More than a third of N. tessellata (37%) were observed in the water. Other studies have reported much higher frequencies of observations in the water (e.g. Lenz & Gruschwitz 1993, Janev Hutinec & Mebert 2011), but this could be related to the sampling methods, as in both refered studies sampling was concentrated directly along the coast or in the water. In this study, the sampling time was spent proportionately to the size of each habitat and the water area represented a relatively small portion (