Scale-Dependent Effects of a Fragmented Agro-Ecosystem on a Reptile Community

Thesis submitted in partial fulfillment of the requirements for the degree of

“DOCTOR OF PHILOSOPHY”

by Guy Rotem

Submitted to the Senate of Ben-Gurion University

of the Negev

29-October-2012

Beer-Sheva

Scale-Dependent Effects of a Fragmented Agro-Ecosystem

on a Reptile Community

Thesis submitted in partial fulfillment of the requirements for the degree of

“DOCTOR OF PHILOSOPHY”

by Guy Rotem

Submitted to the Senate of Ben-Gurion University

of the Negev

Approved by the advisor

Date: 26 March 2014

Approved by the advisor

Date: 26 March 2014

Approved by the Dean of the Kreitman School of Advanced Graduate Studies

29-October-2012

Beer-Sheva

This work was carried out under the supervision of

Prof. Yaron Ziv

The Life Sciences Department, Faculty of Natural Science

Ben-Gurion University of the Negev

Prof. Amos Bouskila

The Life Sciences Department, Faculty of Natural Science

Ben-Gurion University of the Negev

Research-Student's Affidavit when Submitting the Doctoral Thesis for Judgment

I Guy Rotem, whose signature appears below, hereby declare that

(please mark the appropriate statements):

___ I have written this thesis by myself, except for the help and guidance offered by my thesis advisors.

___ The scientific materials included in this thesis are products of my own research, culled from the period during which I was a research student.

___ This thesis incorporates research materials produced in cooperation with others, excluding the technical help commonly received during experimental work. Therefore, I am attaching another affidavit stating the contributions made by myself and the other participants in this research, which has been approved by them and submitted with their approval.

Date: 26 March 2014 Student's name: Guy Rotem

Signature:______

Table of Contents Abstract ...... I General Introduction ...... 1 Chapter 1: Scale-Dependent Variables Affect Reptile Communites on Natural Patches Within a Fragmented Agro-Ecosystem ...... 6 Introduction ...... 6 Methods ...... 8 Results ...... 14 Discussion ...... 26 Appendices ...... 30 Chapter 2: Combined Effects of Climatic Gradient and Domestic Grazing on Reptile Community Structure in a Heterogeneous Agro-Ecosystem ...... 42 Introduction ...... 42 Methods ...... 45 Results ...... 50 Discussion ...... 57 Appendices ...... 61 Chapter 3: Wheat Fields as an Ecological Trap: Trachylepis vittata as a Case Study ...... 62 Introduction ...... 62 Methods ...... 64 Results ...... 66 Discussion ...... 70 General Discussion ...... 74 References ...... 80

Abstract Processes that affect ecological community measurements, such as abundance, species richness or species diversity, form the framework of ecology. For years these processes were investigated by examining local factors (explanatory variables) that may affect the relationship between individuals or species on a local scale. Over the last two decades, due to the rise of advanced scientific theory and practice, new computational and statistical approaches have been developed, allowing ecologists to examine community structure and measurements using several scale-dependent variables and processes. Despite much research conducted in this field, reptiles as a biological group have been somewhat neglected compared to other groups (e.g. birds, insects, rockpools).

Over the years, various studies have been conducted on the structure of ecological communities within agricultural systems, using diverse approaches, including advanced methods of spatial ecology. However, contiguous farmland is rarely spread over a sharp climatic gradient, so the effect of such a gradient on community structure within an agricultural system is currently understudied. In addition, in cases where a climatic gradient exists, it is usually accompanied by changes in elevation, ground composition, and geology, and, as expected, in the agricultural activity itself.

The study area for this thesis, the Southern Judea Lowlands, is characterized by a very sharp climatic gradient over a short distance, with no significant change in elevation, soil, geology, human history or agricultural activities. This PhD thesis examines scale- dependent variables that affect the reptile community within natural patches in the fragmented agricultural system of the Southern Judea Lowlands.

The first chapter addresses the effect of different spatial scale variables on the reptile community located in natural patches within an agricultural matrix. Three 3.2×4 km land units were chosen, located from north to south – Galon, Lachish and Dvir. These land- units reflect the north-south climatic gradient that exists at the landscape scale. Patches of varying size, shape and spatial configuration were identified within the land-units. Within these patches, I marked 100×50 m equal-sized plots which were used for sampling reptiles. This sampling method allowed me to examine how a series of

I variables (e.g. plot heterogeneity, patch size and spatial configuration), which operate at different spatial scales and may be related to different ecological processes, affect the reptile community. By using an AICc-based (Akaike Information Criterion with correction for finite sample sizes) model selection approach, I examined which variables most affect community structure. The models that offered the best explanation for the three community measurements – abundance, species richness and species diversity – were all multiple scale models. However, all three community measures were strongly affected by local scale variables which suggested a strong influence of local hetrogeneity on reptile communities. Moreover, for all three community measures, both grazing and climatic gradients together were found to be an important variable, which indicates the significance of this combined environmental phenomenon.

The second chapter deals with one of the most common agricultural activities within my research area -- seasonal grazing of sheep (Ovis aries) and cattle (Bos primigenius). Grazing takes place mainly in the stubble after harvest, but herds also find natural patches located within the agricultural matrix. Previous studies have examined the effect of domestic animal grazing on reptile communities, but few of these have examined the integrated effects of grazing and climate, especially within a similar area. In Chapter 2, I examine the combined effect of grazing and climatic gradient on the reptile community of the Southern Judea Lowlands with an additional site to the south – Rahat. I list the proportion of species according to their biogeographical origin. The results indicate a decrease in the percentage and relative abundance of Mediterranean species and an increase in the percentage and abundance of desert species with decreasing precipitation. The effect of grazing itself also changed according to the location of the plot along the climatic gradient. Within the Galon land-unit, which belongs to the Mediterranean climate, grazing was found to increase plot heterogeneity and species richness. In contrast, in the southern area, near Dvir, a negative effect of grazing on plot heterogeneity and species richness was found. At Dvir, grazing was also found to affect community composition. Mediterranean-oriented species richness was negatively affected by grazing intensity. In contrast, arid-oriented species richness was positively related to grazing intensity at Dvir.

Apart from grazing, the study area is also characterized by the presence of vast wheat fields. In the third chapter of the work I address the effect of wheat fields on the reptile communities. I chose to focus on plots located in the center of wheat fields which were all concentrated in the northern land-unit Galon, in order to avoid in this analysis the impact of the climatic gradient. In addition, due to the paucity of observations of other species, this part of the study focused on the lizard Bridled Mabuya (Trachylepis vittata) only. Arthropod abundance found in the early spring within the wheat fields was significantly higher than in natural patches. In addition, I found a significant movement of reptiles from natural patches to the wheat fields, but very little in the opposite direction. The physical condition of the individuals who left the natural patches for the fields was significantly better than that of the individuals who remained in the natural patches. Finally, no individuals were found within the wheat fields after the harvest in contrast to the natural patches. These results indicate that wheat fields act as an ecological trap for T. vittata. This third chapter has been accepted for publication in the journal Biological Conservation.

In conclusion, the results of this work suggest that the reptile community within a fragmented agro-ecosystem is affected by many scale-dependent ecological processes. The results also suggest that the presence of a reptile community within an agricultural area is affected by the type of agricultural crop and other agricultural practices. My research highlights the need to consider various scale-dependent ecological variables, as well as the type of agricultural activity, when investigating ecological communities within an agricultural area.

Keywords: Abundance, Agro-ecosystem, AIC, Climate, Community, Ecological trap, Ecology, Ecotone, GIS, Grazing, Habitat fragmentation, Habitat selection, Israel, Reptile, Scale-dependence, Southern Judea Lowlands, Spatial scale, Species diversity, Species richness.

III

General Introduction Ecologists have been concerned with questions about species diversity within natural patches or islands for more than forty years. While the classical approach to these questions dealt with the study of processes occurring within a local area or patch (Cornell and Lawton 1992), a new research approach, examining a variety of biological processes related to different spatio-temporal scales, has become widely accepted over the last two decades (Palmer and Dixon 1990, Wu and Loucks 1995, Harrison and Cornell 2008). This new approach can also be applied to systems of natural habitats located within agricultural fields. Such systems, termed ‘agro-ecosystems’, are now frequently used to study ecological patterns and processes in semi-natural agricultural environments (Vandermeer 2011).

Scale is the foundation of spatial ecology (Turner et al. 2001, Farina 2006). Spatial scale refers to ecological or biological processes that uniquely determine a given biological pattern at a particular defined area. A change in the research area that does not affect the ecological processes does not reflect a real change in the spatial scale. Therefore, a change in the spatial scale means a change in the relative contribution of ecological processes. In addition, scales are usually hierarchical, i.e., "a system of interconnections wherein the higher levels constrain the lower levels to various degrees” (Turner et al. 2001).

In order to analyze community structure and diversity patterns, ecologists measure different community characteristics (hereafter, measures), such as abundance (i.e., total and proportional abundance), species richness (i.e., raw number and/or corrected number of species) and species diversity (e.g., Fisher’s α; Fisher et al. 1943). These community measures may be affected by both local scale processes, such as local competition (Hardin 1960), higher scale processes, such as migration or dispersal (Hamback and Englund 2005), and even climatic change or biogeographic origin (Urbina-Cardona et al. 2006, Harrison and Cornell 2008). These processes do not necessarily operate solely but rather may interact to affect any of the above community characteristics (Ziv 1998).

My research area, the Southern Judea Lowlands (SJL), is located between Galon in the north and Rahat in the south. This area is characterized by a sharp climatic gradient, from an average annual rainfall of 450 mm (Mediterranean climate) in the north to 300 mm (semi-arid climate) in the south (Gvirtzman 2002), over a distance of just 30 km. This sharp climatic gradient confers a research advantage because there is little change in soil type, geology, agricultural crops or human history over the entire range. This allowed me to look at the effects of climate, grazing and fragmentation on the reptile community. The unique position of the SJL in a transition zone between the Mediterranean and arid ecosystems (Kark et al. 1999) brings together biota from different biogeographical zones: Mediterranean, Irano-Turanian and Saharo-Arabic (Giladi et al. 2011). Thousands of years of human development have amplified the patchy nature of the landscape (Ben-Yosef 1980).

Within the SJL, I chose three 4×3.2 km ‘land-units’ ranging from north to south: Galon, Lachish and Dvir (with the addition of Rahat in the south for the grazing study). Within those land-units I chose natural or semi-natural patches differing in size, shape, isolation and physiognomy. Within those patches I used 100×50 m sampling units to sample the reptile community, in order to analyze the effect of different variables related to different spatial scales on reptile community measurements.

Within the sampling plots, I measured various local-scale explanatory variables, such as percentage rock cover and presence of shrubs and trees. These local-scale variables may affect local heterogeneity, habitat variability, and, consequently, local competition (Wright et al. 2002, Weibull et al. 2003). Patch size may also affect heterogeneity and habitat variability (Rosenzweig 1995). Patch size can also affect the chances of migration from other patches (MacArthur and Wilson 1967) and thus affect community structure via a higher scale. Patch spatial configuration and patch fragmentation may also affect a species' chances of dispersal or migration between patches (Hanski and Gilpin 1997), as well as its chances of population or community survival within the isolated patches (e.g. the rescue effect; Hanski and Gaggiotti 2004). Finally, the location of patches along the climatic gradient can affect the overall species pool (James and Shine

2000), the balance of biogeographic origin, and thus the community structure via higher spatial scales.

Israeli herpetofauna includes 87 terrestrial species (Bouskila and Amitai 2001), all of which are protected by the Israeli law, but feature in different categories of extinction risk in the Red Book (Bouskila 2002). The Red Book also reveals that many reptile species are threatened by the processes of habitat loss (Bouskila 2002). This finding is consistent with other studies showing that species diversity all over the world is declining (Andren 1994, Stuart et al. 2004). In addition, it is generally assumed that the major contributors to loss of species diversity are reduction and destruction of natural habitats, habitat fragmentation and entry of invasive species (Sodhi and Ehrlich 2010). One of the most common land use practices that has led to the destruction of natural habitats is modern agriculture (Tscharntke et al. 2005).

About 80% of the terrestrial area on Earth is affected by agricultural activities (FAO 2007), while a third of this area experiences intensive agricultural activity. Much of this intensive agriculture involves production of the same crop over a huge area, i.e., monoculture (Vandermeer 2011), which consequently causes a reduction in heterogeneity, damage and loss of natural habitats, and fragmentation of the remaining natural habitats within the agricultural area.

One possible solution to reduce the negative impact of farmland on biodiversity is to keep natural patches in-between agricultural fields, in order to increase the heterogeneity within the agricultural matrix (Benton et al. 2003). Natural or semi-natural patches can affect biodiversity through many biological and ecological mechanisms (Kleijn et al. 2006, Tschantke et al. 2012). Natural patches that conserve the natural habitat contain native species and may help to increase spatial heterogeneity. It is generally assumed that spatial heterogeneity correlates with niche diversity. An increased number of niches may enable a greater variety of organisms to exist in the landscape (Cumming 2007, Cunningham et al. 2007). The need to maintain habitat patches within unnatural environment is also supported by other well-established theories that are specifically related to patches, such as the Theory of Island Biogeography (MacArthur and Wilson 1967), which assumes a positive connection between patch size and number of species, or

3 the Metapopulation Dynamics (Levin 1970), which assumes that movement of organisms between the patches is necessary to maintain viable populations. These theories highlight the importance of preserving natural patches within the agricultural matrix for maintaining biodiversity.

Agricultural activity may also affect community measures within the natural patches (Vandermeer 2011). The main agricultural activities in my research area include wheat cropping and domestic grazing. Grazing of sheep and cattle began some seven thousand years ago with the beginning of the process of domestication of these species (Perevolotsky and Seligman 1998). Today, approximately 30% of the Earth’s terrestrial area is used for grazing of domestic animals (Steinfeld et al. 2006). The ecological literature reports different effects of livestock grazing on biodiversity. Several studies have found that grazing can cause biodiversity to decrease, while other studies have found opposite trends (Adler et al. 2001).

Field crops (e.g. wheat, vineyards, orchards, etc.) can also affect the populations living within natural patches. Agricultural crops with nutritional potential (such as fruit or roots) can attract animals to move from the natural area into the agricultural system (Bino et al. 2010). Situations in which the agricultural system offers richer resources than the preserved natural areas can lead to an increase in abundance, species richness and diversity within the agricultural lands (Rotem et al. 2011). Such high-quality agricultural land relative to the adjacent natural areas at certain times of the year can lead animals to choose the agricultural area as their preferred habitat. Sometimes this choice is successful and the fitness of the individuals which choose to live in an agricultural area is positive, maintaining a long-term persistent population. However, in other cases, the agricultural activity may result in a high mortality rate and lower fitness. In extreme cases, agricultural fields may become an ecological trap (Best 1986).

My research examines several processes, through particular explanatory variables (factors), that affect the reptile community in the SJL agro-ecosystem. In the first chapter I explore the effect of different spatial variables on the reptile community within this fragmented agro-ecosystem. In the second chapter I explore the combined effect of domestic animal grazing and climatic gradient on the reptile community. Finally, in the

4 third chapter I explore the effect of wheat cropping on populations of Trachylepis vittata (Bauer 2003) in natural patches adjacent to wheat fields, by exploring movement dynamics and survival between the natural patches and the wheat fields.

Chapter 1: Scale-Dependent Variables Affect Reptile Communites on Natural Patches Within a Fragmented Agro-Ecosystem

(This chapter is prepared for publication with the authorship of Guy Rotem, Amos Bouskila, Itamar Giladi and Yaron Ziv)

Introduction Large areas of the globe are affected by agricultural activity. Modern agriculture is characterized by high homogeneity, natural habitat loss and fragmentation (Robinson and Sutherland 2002, Green et al. 2005, FAO 2007). Decrease in spatial heterogeneity, reduction and fragmentation of natural habitats are key factors in biodiversity loss around the world (Sodhi and Ehrlich 2010).

Major ecological theories such as Island Biogeography (MacArthur and Wilson 1967), Niche Theory (Hutchinson 1959) and Metapopulation Dynamics (Hanski and Gilpin 1997), rely on a positive relationship between heterogeneity and connectivity between remaining natural habitats (“patches”) and biodiversity. It is therefore important to understand which features of such patches facilitate the existence of a diverse ecological community at different heterogeneity levels. Consequently, contemporary ecology focuses on natural patch community structure as a product of many processes operating at different scales (Ziv 1998, Turner et al. 2001, Farina 2006). Given that processes at different scales can simultaneously influence community properties and determine differences among communities (Wiens 1989, Turner et al. 2001, Farina 2006), a comprehensive understanding of community diversity and composition patterns requires a multi-scale study.

In order to analyze community structure and diversity patterns, ecologists apply different measures, such as abundance (i.e., total and proportional abundance), species richness (i.e., raw number and/or corrected number of species) and species diversity (e.g., Fisher’s α). Each of these measures may reflect different responses to environmental conditions and/or ecological constraints including climatic gradients. For example, while total abundance may reflect the biomass and food availability for all individuals regardless of their species identity (Shine and Madsen 1997, Smart et al. 2000), species richness may correlate to the means by which different organisms utilize resources within a community

(i.e., their niche; Rocha et al. 2008, Soininen et al. 2011). Species diversity, on the other hand, may reflect the proportional use and subdivision of resources among the existing species, providing additional information about realized niches and interactions among a set of existing species (Hiltunen et al. 2006). In addition, each of these community measures might be scale-dependent in a different way, given that determinants of community structure and composition change with scale (Dumbrell et al. 2008).

My research area -- the Southern Judea Lowlands (SJL) agro-ecosystem -- is characterized by a very sharp climatic gradient from an average annual rainfall of 450 mm in the north (Galon) to an average annual rainfall of 300 mm in the south (Dvir) over a distance of only 30 km, with no change in elevation. However, over the study period the actual precipitation was lower than the annual average (Appendix 1; Israeli Meteorological Service http://www.ims.gov.il). This sharp climatic gradient confers a scientific advantage since the agricultural activity within it is not accompanied by any large geological, lithological or human history gradients. In more moderate climatic gradients those variables may also change and present confounding effects. This sharp climatic gradient confers an additional scientific advantage since, due to artificial irrigation, the agricultural matrix and products can be maintained with little change along the climatic gradient. This allowed me to examine the characteristics of patches which support reptile communities under different climatic conditions with no change in the agricultural matrix.

Specifically, in this chapter I ask: (1) What are the scale-dependent ecological variables that correlate to reptile community measures on natural patches within an agro-ecosystem along a sharp climatic gradient?; and (2) Are particular community measures (i.e., abundance, species richness and species diversity) differentially affected by variables (determinants) of different scales?

With respect to the first question, I distinguish between two non-mutually exclusive hypotheses -- the biotic hypothesis and the abiotic hypothesis. The biotic hypothesis states that reptile communities are affected by biotic variables that operate both at a local scale (e.g., food availability) and at a regional scale (i.e., evolutionary origin). The abiotic hypothesis argues that different abiotic variables of different scales (e.g., soil type,

7 humidity and stoniness at the local scale, spatial configuration at the land-unit scale, and climatic conditions at the regional scale) determine reptile communities.

With respect to the second question I distinguish between two mutually exclusive hypotheses. The first hypothesis suggests that all the community measures are affected in the same way by the same variables. The second hypothesis suggests that different community measures are differentially affected by variables (determinants) of different scales. Specifically, I predict that in my study abundance will be associated with variables that might affect reptiles’ food availability (Smart et al. 2000), expressed by local-scale heterogeneity (e.g., percentage of exposed rocks). I also predict that species richness and species diversity will be associated with niche-related heterogeneity (Soininen et al. 2011), such as patch size and pach heterogeneity, movement ability and connectivity (Dumbrell et al. 2008, Krewenka et al. 2011).

Methods Study Area

The SJL landscape is located in central Israel (31°30’52’’N 34°52’36’’E), approximately 50 km southwest of Jerusalem (Figure 1). The area is dominated by soft limestone and chalk low-lying hills surrounded by loessal valleys that are used for agriculture. At the regional scale, the sharp climatic gradient changes from an average annual rainfall of 450 mm (Mediterranean climate) to 300 mm (semi-arid climate) (Gvirtzman 2002) over a distance of 30 km. The unique position of the SJL in a transition zone between the Mediterranean and arid ecosystems (Kark et al. 1999) brings together biota from different biogeographical zones: Mediterranean, Irano-Turanian and Saharo-Arabic (Giladi et al. 2011). Thousands of years of human activity have amplified the patchy nature of the landscape (Ben-Yosef 1980) and affected the fauna and flora. Today, within the SJL, there are no untouched “natural” areas with “original” fauna.

Figure 1. Hierarchical scales of the study system. (a) The sharp precipitation gradient of Israel and location of Southern Judea Lowlands (SJL) along this gradient. (b) Galon, Lachish and Dvir land-units within the SJL. Note the sharp decline in mean annual precipitation along a short distance of 30 km. (c) the Galon land-unit - the distribution of remnant natural patches within the agricultural matrix. (d) A patch, with its specific attributes of area, shape, and internal heterogeneity. (e) Perennial shrub, (bottom right corner) surrounded by annual plants (<15 cm tall), and the contrast with the agricultural matrix during the sampling period.

Study Design and Survey Protocol

Within the SJL landscape, I chose three 4×3.2 km land-units (from north to south: Galon, Lachish and Dvir; Figure 2), characterized by remnant patches of natural vegetation of differing size and configuration, surrounded by a matrix of agricultural fields. I surveyed reptiles in 63 100×50 m sampling plots located within 30 natural patches that were chosen to represent patches of different size and isolation in each land-unit. I established one sampling plot in each of the smaller patches (0.5 ha50 ha). Nine very small patches (1000 m2< area <0.5 ha ), three in each land-unit ,were sampled exhaustively.

I actively searched for reptiles in springtime (March to June) of 2009 and 2010 on clear days with a daily maximum temperature above 25 °C. I used three complementary active methods for sampling the reptile community within each plot: (i) line transects, (ii) turning over rocks, and (iii) pitfall traps. I conducted line transect surveys during the late morning, after commencement of reptile activity, by walking in a straight line along the long axis of the sampling plot. Turning-over stones was performed along the long axis of the sampling unit during the late morning, immediately after completion of the line- transect survey. I made an effort to turn over any stone with a diameter exceeding 10 cm, and in any case at least 100 stones per plot. All the stones turned were returned to their original position in order to avoid damage to the habitat. I used 20 one-liter dry pitfall traps in each plot (total of 1260 traps). The traps were arranged in two lines parallel to the long axis of the plot. The lines were 25 m apart and the traps along each line were positioned at 10 m intervals (see Figure 2). The traps were opened before sunset and left open for two entire days. To prevent injury to animals and to reduce their discomfort a tiled roof was built over each trap. This roof protected the trap from solar radiation and reduced the exposure of trapped animals to predation risk. In the first year of the study captured individuals were marked individually. Since there were no repeated captures during the first year, marking of individuals was discontinued in all subsequent years.

Figure 2. The sampling design, based on three complementary active methods for sampling the reptile community within each plot: (i) line transects, (ii) over-turning stones, and (iii) pitfall traps

Measurements of Variables at Different Scales

I focused on three community measures: abundance, local species richness and species diversity (using Fisher’s alpha index of diversity; Fisher et al. 1943, Rosenzweig 1995). I recorded a range of categorical and continuous variables at multiple spatial scales -- plot scale, patch scale, land-unit scale, and landscape scale to serve as explanatory variables (hereafter, determinants) for the above measures (Table 1).

I used rectified aerial photographs (Ofek, 2005, pixel size = 1 m2) to identify all the patches of natural vegetation within each of the land units (Galon, Lachish and Dvir ). I then demarcated their boundaries on a digitized map and stored the information as a vector-based coverage in a Geographical Information Systems (GIS) platform (ArcInfo TM; ESRI). The data were converted to a raster-based layer (grid cells size = 5×5 m) and

11 exported to FRAGSTATS© (McGarigal and Cushman 2002) for calculating patch shape and patch proximity index.

Using line transects along the long axis of each sampling plot we measured plot-level explanatory variables: percentage stone cover, percentage vegetation cover and percentage exposed soil. Using a compass we measured the plot aspect: 270°- 89° were defined as north-facing and 90°-269° were defined as south-facing. South-facing slopes receive higher solar radiation, and consequently have lower water availability and higher ground temperatures than north-facing slopes. As a result of ancient agricultural activity, some plots contained ancient cisterns and/or stone walls. The presence of each of these structures was considered as a categorical explanatory variable. When plots included stone heaps we measured their total area and used it as an explanatory variable.

I quantified domestic animal grazing levels in Dvir and Lachish by measuring grazer feces density in ten 1×1 m quadrats in each plot. I averaged the ten densities to a single- plot average feces density. In Galon, due to the presence of a different grazing species (i.e. cows), we measured grazing feces density in five 2×2 m quadrats in each plot. The five densities were then averaged to a single-plot density.

I used patch area and patch shape as explanatory variables at the patch-scale level. Patch shape was calculated by the shoreline fractal-dimension index :

2 ln(.25p) FRAC  , (1) ln a where a is patch area and p is patch perimeter. Proximity index is a measure of patch isolation, which quantifies the spatial context of a patch in relation to neighboring patches. (equation source: FRAGSTATS©; see also Farina 2006, P 340.) For each patch i, the proximity index (PIi) is given by:

n a j PI i   2 , (2) j1 hij

where aj is the area of patch j, hij is the distance (edge-to-edge) between patch i and patch j, and the summation is done for all n patches that are within a certain search radius, rs of

12 the focal patch i (i.e. hij < rs) (Gustafson and Parker 1992). I reported the data analysis conducted using a proximity index calculated using a radius of 1000 m. Preliminary data analysis conducted using a proximity index with smaller rs (500 m and 100 m) yielded very similar results. The three chosen land-units -- Galon, Lachish and Dvir (Figure 2) -- reflect positions along a north-south climatic gradient, which we considered as an explanatory variable at the landscape scale.

Table 1. Explanatory variables. A list of independent explanatory variables, their description and the scale at which they were measured. Explanatory variable Variable type Description Landscape Scale North-South gradient Categorical Land-unit Scale Proximity index Continuous Patch Scale Patch size Continuous Shoreline fractal dimension Continuous Plot Scale Grazing Continuous Based on average feces density per plot as grazing pressure measure Aspect Categorical Measured using a compass Rock cover Continuous Based on line transects along the long axis of the plot Exposed soil Continuous Based on line transects along the long axis of the plot Cistern Categorical Present / absent Historical wall Categorical Present / absent Heap area Continuous Total area and stone size Plot heterogeneity Continuous Shannon diversity, based on line transects along the long axis of the plot

Statistical Analysis

Based on the prospective explanatory variables I constructed a series of 68 biologically meaningful General Linear Models (GLM) for each community measure (i.e., abundance, species richness and species diversity). Some models included different variables related to one scale only, but most models included variables related to several scales and interactions. I then applied a model selection procedure using the Akaike information criterion corrected for small sample size (AICc):

2k(k 1) AIC  2log L  2k  (3) c (n  k 1) where L is the maximum likelihood for the candidate model and K is the number of independently estimated parameters in the model where the lowest AICc is the better fit (Anderson 2008).

I also considered parameter estimates averaged over the entire model set, proportional to the support that each model receives (Yasuhara et al. 2012). This approach accounts for uncertainty in model selection and thus leads to appropriately broader confidence intervals than would be obtained by relying only on the single, best-supported model. The influences of the various predictor variables were measured as relative importance, which is the sum of the Akaike weights of models that include the variable in question (Burnham & Anderson, 2002). Analyses reported here were implemented in the R programming language (R Development Core Team, 2011), using functions from the R packages: MuMIn (Barton , 2013) and AICcmodavg (Mazerolle 2013) to perform the model averaging.

Results I observed a total of 781 individuals from 29 species and 12 families in 63 plots (see Appendix 5 for species list). There were 349, 226 and 206 individuals, representing 20, 19 and 25 species in the northern (Galon), central (Lachish) and southern land-units (Dvir), respectively. Three species (Trachylepis vittata, Chalcides guentheri and Ableharus rueppelli) accounted for over 53% of all observations (Table 2). The mean

14 number of reptile species per plot was 5.01 (range = 1-11), and mean abundance was 10.2 individuals per plot (2- 24). The mean ‘Fisher’s alpha index of diversity’ per plot was 5.726 (0.34-26.78).

Table 2. Summary of reptile sampling. General summary of reptile sampling results with the species composition, site occupancy (number of occupied plots) and species abundance of reptiles recorded. Family Species Occupancy Total no. of (# (%) plots individuals (%) occupied) Agamidae Roughtailed Rock Agama (Laudakia stellio) 7 (11.1) 20 (2.5) European Glass Lizard (Ophisaurus apodus) 4 (6.3) 4 (0.5) Atractaspididae Muller's Two-headed Snake (Micrelaps muelleri) 2 (3.1) 2 (0.2) Boidae Javelin Sand Boa (Eryx Jaculus) 4 (6.3) 5(0.6) Chamaeleonidae Mediterranean Chameleon (Chamaeleo chamaeleon 1 (0.1) 5(0.6) recticrista) Colubridae Dice Snake (Natrix tessellata) 1 (0.1) 1(0.1) Mediterranean Cat Snake (Telescopus fallax 2 (3.1) 2 (0.2) syriacus) Large Whip Snake (Coluber jugularis asianus) 14 (22.2) 26 (3.3) Collared Dwarf Racer (Coluber rubriceps) 17 (26.9) 30 (3.8) Montpellier Snake 3 (4.7) 5 (0.6) (Malpolon monspessulanus insignitus) Palestine Kukri Snake 4 (6.3) 6 (0.7) (Rhynchocalamus melanocephalus) Roth's Dwarf Racer (Eirenis rothi) 21 (33.3) 25 (3.2) Gekkonidae Lichtenstein's Short-Fingered Gecko 9 (14.2) 29 (3.7) (Stenodactylus sthenodactylus) Mediterranean House Gecko (Hemidactylus 9 (14.8) 26 (3.3) turcicus) Sinai Fan-Fingered Gecko (Ptyodactylus guttatus) 4 (6.4) 27 (3.4) Lacertidae Olivier's Sand Lizard (Mesalina olivieri) 1 (1.8) 1 (0.1) Small Spotted Lizard (Mesalina guttulata) 2 (3.7) 2 (0.2) Snake-Eyed Lizard (Ophisops elegans) 15 (23.8) 30 (3.8) Leptotyphlopidae Hook-snouted Worm Snake (Leptotyphlops 5 (7.3) 7 (0.8) macrorhynchus) Scincidae Bridled Mabuya (Trachylepis vittata) 53 (84.2) 253 (32.3) Ocellated Bronze Skink (Chalcides ocellatus) 11 (17.6) 25 (3.2) Gunther's Cylindrical Skink (Chalcides guentheri) 36 (57.4) 79 (10.1) Rüppell's Snake-eyed Skink (Ablepharus 28 (44.4) 87 (11.1) rueppellii) Algerian Skink (Eumeces schneideri pavimentatus) 21 (33.3) 22 (2.8) (Eumeces schneideri schneideri) 2 (3.7) 2 (0.2) Testudinidae Mediterranean Spur-thighed Tortoise 15 (23.8) 31 (3.9) (Testudo graeca) Typhlopidae Eurasian Blind Snake (Typhlops vermicularis) 12 (19.0) 18 (2.3) Simon Worm Snake (Typhlops simoni) 4 (6.3) 6 (0.7) Viperidae Palestine Viper (Vipera palaestinae) 3 (4.7) 5 (0.6) Total abundance 63 781 (Source of English common names: www.catalogueoflife.org and www.reptile- database.org)

The abundance model set (Tables 3-5 and Appendix 2) indicates that models that included explanatory variables from multiple scales had lower AICc values (i.e., better fit the data) than models with explanatory variables from a single scale only.

Table 3. Reptile abundance AICc based model selection. The table presents the four best models with accumulative probability of 0.87 (~0.9) and the “null” model as well.

Model K AICc Delta AIC Model Accumulative Relative R- ID probability probability squared Wi (Wi)

1 16 -449.203 0 0.437041 0.437041 1 0.6369

2 16 -447.638 1.565227 0.199819 0.636861 0.45721 0.635

3 17 -447.363 1.839705 0.174195 0.811056 0.398578 0.637

4 16 -445.299 3.904038 0.062054 0.87311 0.141987 0.634

Null 2 39.36843 488.5714 8.09E-107 1 3.54E-107

Table 4. Reptile abundance AICc based model selection. The table presents the parameter estimations of the four best models with accumulative probability of 0.87 (~0.9).

Parameters Model 1 Model 2 Model 3 Model 4

Landscape Scale GA -0.015 -0.018 -0.02 -0.015

LA 0.124 0.123 0.129 0.124

Land-unit Scale PR -2E-06 -1.9E-06 -2E-06 -1.9E-06

Patch scale PS -0.034 -0.028 -0.032 -0.034

F 0.442 0.435 0.435 0.442

Plot Scale GR -0.079 -0.084 -0.078 -0.079

C -0.065 -0.071 -0.059 -0.065

W 0.096 0.104 0.089 0.096

HA 0.0009 - 0.001 0.009

PD -0.604 0.648 -0.608 -0.604

EX - 2.91E-04 8.80E-04 -

RO 0.021 0.02 0.021 -

NAS 0.398 0.451 0.518 0.021

Interactions GA×PD 0.18 0.177 0.181 0.18

LA×PD -0.018 -0.016 -0.023 0.018

Values R2 0.636 0.635 0.637 0.696

AICc -449.203 -447.638 -447.636 -445.299

AW 0.437041 0.199819 0.174195 0.06205

GA, Galon; LA, Lachish; PR, Proximity; PS, Patch size; F, Fractal; GR, grazing; C, Cistern; W, Wall; HA, Heap area; PD, plot diversity; EX, Exposed soil; RO, Rock cover percent; NAS, Northern aspect.

The table shows the parameter estimates, R2, the Akaike information criterion corrected for small sample size (AICc), and the Akaike weight (AW).

Table 5. Model-averaged parameter estimates and CIs of reptile abundance

Parameter RI Estimates Upper CI Lower CI

GR 1 -0.081 -0.036 -0.125

GA 1 -0.009 0.121 -0.139

LA 1 0.122 0.246 -0.002

RO 1 0.021 0.025 0.016

GA×GR 1 0.177 0.228 0.125

LA×GR 1 -0.078 0.043 -0.078

PD 1 -0.613 -0.408 -0.818

F 1 0.446 0.611 0.281

W 1 0.093 0.146 0.04

C 0.95 -0.065 -0.011 -0.119

PS 0.94 -0.031 -0.001 -0.062

PR 0.92 -2E-06 -9.1E-07 -3.1E-06

NAS 0.91 0.457 0.885 0.03

HA 0.8 0.0009 0.002 -0.0005

EX 0.55 0.0004 0.004 -0.003

GR, grazing; GA, Galon; LA, Lachish; PR, Proximity; PS, Patch size; F, Fractal; C, Cistern; W, Wall; HA, Heap area; PD, plot diversity; EX, Exposed soil; RO, Rock cover percent; NAS, Northern aspect; CIs, confidence intervals; RI, relative importance (the sum of the Akaike weights of models that include the variable, see statistical analysis).

Hence, my analyses suggest that grazing, the location along the climatic gradient and the interaction between them both are very important determinants of reptile abundance. The

18 analysis also suggests that several plot scale variables (e.g. plot diversity, rock cover percent etc.), patch proximity and patch shape are also important determinants of reptile abundance. Surprisingly, I found percentage of exposed soil to be a relatively unimportant explanatory variable of reptile abundance.

The species richness model set (Tables 6-8 and Appendix 3) showed that multiple-scale models had lower AICc values than single-scale models. All three best models included variables belonging to all four scales. Similar to the result obtained for reptile abundance my analyses suggest that grazing, location along the climatic gradient and interaction between grazing and climatic gradient are important variables in determining reptile species richness. I also found that patch scale variables (e.g. plot diversity, walls, rocks cover percent, exposed soil etc. ) are important variables as well. Surprisingly, I found that proximity and patch size are relatively less important for species richness.

Table 6. Reptile species richness AICc based model selection. The table presents the three best models with accumulative probability of 0.93 and the “null” model as well.

Model K AICc Delta AIC Model Accumulative Relative R- ID probability probability squared Wi (Wi)

1 16 -584.802 0 0.541936246 0.541936 1 0.519

2 17 -583.636 1.165331 0.302621466 0.844558 0.558408 0.52

3 16 -581.302 3.499523 0.094196858 0.938755 0.173815 0.516

Null 2 -239.455 345.346 5.53312E-76 1 1.02099E-75

Table 7. Reptile species richness AICc-based model selection. The table presents the parameter estimations of the three best models with accumulative probability of 0.93

Parameters Model 1 Model 2 Model 3

Landscape Scale GA -0.4125 0.392 -0.3946

LA 0.4973 -0.512 -0.505

Land-unit Scale PR -1.17E-06 -1.17E-06 -1.3E-06

Patch scale PS - 0.012 0.007885

F 0.2696 0.29 0.29

Plot Scale GR -0.2021 0.208 -0.2002

C -0.102 -0.096 -0.08193

W 0.1304 1.218 0.101

HA -1.32E-03 -1.44E-03 -

PD -0.5366 0.565 -0.5112

EX -0.006 -0.005 -0.00519

RO 0.012 0.011 0.01155

NAS -0.718 0.011 -0.6198

Interactions GA×PD -0.718 0.252 0.2578

LA×PD 0.263 2.887 0.2796

Values R2 0.519 0.52 0.516

AICc -584.802 -583.636 -581.302

AW 0.541 0.302 0.094

Table 8. Model-averaged parameter estimates and CIs of reptile species richness

Parameter RI Estimates Upper CI Lower CI

GR 1 -0.204 -0.166 0.241

-0.404 GA 1 -0.296 -0.513

LA 1 -502 -0.395 -0.609

GA×GR 1 0.259 0.3 0.217

LA×GR 1 0.284 0.337 0.231

PD 1 -0.709 -0.374 0.719

W 1 0.124 0.171 0.077

RO 1 0.011 0.015 0.007

F 1 0.277 0.418 0.137

EX 1 -0.006 -0.003 -0.009

C 1 -0.098 -0.049 -0.147

NAS 0.99 -0.709 -0.265 -1.153

PR 0.93 -1.18E-06 -2.231E-07 -2.139E-06

HA 0.91 0.001 -1.5168E-05 -0.002

PS 0.45 0.011 0.038 -0.014

Hence, my analyses suggest that grazing, location along the climatic gradient and the interaction between them both are very important determinants of reptile species richness. My analysis also suggests that many plot scale variables (e.g. plot diversity, rock cover percentage etc.), are also important determinants of reptile species richness. Surprisingly I found patch size to be a relatively unimportant explanatory variable of reptile species richness.

Similar to the other two community measures, the model set for species diversity (Tables 9-11 and Appendix 4) showed that integrating explanatory variables of different scales provided the models with the best fit to the data. Similar to the results obtained so far I found that grazing, location along the climatic gradient ,the interaction between the two and patch hetrogeneity are relatively important variables. I also found that patch size, shape (i.e., fractal) and proximity are (in contrast to my hypotheses) relatively less important variables in determining species diversity.

Table 9. Reptile species diversity AICc based model selection. The table presents the 14 best models with accumulative probability of 0.95 and the “null” model as well.

Model K AICc Delta AIC Model Accumulative Relative R- ID probability probability squared (Wi) Wi

1 14 13.67171 0 0.152877 0.152877 1 0.373

2 14 13.99228 0.320573 0.130236 0.283113 0.8519 0.372

3 14 14.29947 0.627762 0.111693 0.394805 0.730606 0.372

4 15 14.94603 1.274317 0.08084 0.475646 0.528793 0.374

5 16 15.11995 1.448238 0.074107 0.549753 0.484752 0.376

6 16 15.48797 1.816261 0.061652 0.611405 0.403277 0.376

7 14 15.65366 1.981949 0.05675 0.668155 0.371215 0.242

8 15 15.75023 2.078524 0.054075 0.72223 0.353716 0.373

9 16 15.92346 2.251748 0.049588 0.771818 0.324369 0.375

10 16 16.29839 2.626681 0.041112 0.81293 0.26892 0.375

11 15 16.3799 2.708191 0.03947 0.8524 0.258181 0.372

12 16 16.45378 2.782066 0.038039 0.890438 0.248818 0.375

13 16 16.50841 2.836697 0.037014 0.927452 0.242114 0.375

14 17 17.10146 3.42975 0.027516 0.954968 0.179986 0.376

Null 2 227.5835 2.14E+02 5.42E-48 1 3.55E-47

Table 10. Reptile species diversity AICc based model selection. The table presents the parameter estimations of the 14 best models with accumulative probability of 0.95 Model Model Model Model Model Model Model Model Model Model Model Model Model Model Parameters 1 2 3 4 5 6 7 8 9 10 11 12 13 14

GA -1.097 -1.135 -1.122 -1.141 -1.191 -1.186 -1.151 -1.091 -1.187 -1.139 -1.121 -1.139 -1.202 -1.184 Landscape Scale LA -1.233 -1.226 -1.210 -1.226 -1.192 -1.173 -1.196 -1.227 -1.181 -1.191 -1.201 -1.213 -1.159 -1.178 Land-unit PR >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 >0.000 Scale PS -0.021 -0.018 -0.018 -0.024 -0.032 -0.023 -0.023 -0.030 -0.021 -0.022 -0.029 Patch Scale F -0.293 -0.315 -0.318 -0.326 -0.386 -0.387 -0.344 -0.292 -0.387 -0.342 -0.320 -0.340 -0.411 -0.387

GR -0.350 -0.341 -0.342 -0.347 -0.337 -0.326 -0.339 -0.348 -0.334 -0.333 -0.339 -0.345 -0.327 -0.331

C -0.044 -0.030 -0.044 -0.033 -0.033 -0.061 -0.048 -0.060 -0.029 -0.021 -0.043

W 0.123 0.120 0.115 0.138 0.125 0.115 0.146 0.122 0.107 0.141 0.114 0.109 0.104 0.128

HA -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 -0.001 0.000 -0.001 Plot Scale PD -0.646 -0.564 -0.576 -0.620 -0.662 -0.602 -0.653 -0.640 -0.646 -0.576 -0.563 -0.700 -0.650 -0.636

EX -0.005 -0.005 -0.005 -0.006 -0.003 0.003 -0.006 -0.005 -0.002 -0.006 -0.005 -0.002 -0.003

RO 0.002 0.001 0.000 0.003 0.001 -0.001 0.000 0.001

NAS 0.518 0.516 0.599 0.476 0.887 0.460

GA×PD 0.439 0.458 0.451 0.455 0.470 0.477 0.461 0.438 0.468 0.465 0.452 0.448 0.476 0.473 Interaction LA×PD 0.627 0.617 0.609 0.621 0.597 0.589 0.615 0.625 0.587 0.611 0.606 0.608 0.577 0.595

2 R 0.373 0.372 0.372 0.374 0.376 0.376 0.242 0.373 0.375 0.375 0.372 0.375 0.375 0.376

AIC Values c 13.672 13.992 14.299 14.946 15.120 15.488 15.654 15.750 15.923 16.298 16.380 16.454 16.508 17.101

AW 0.153 0.130 0.112 0.081 0.074 0.062 0.057 0.054 0.050 0.041 0.039 0.038 0.037 0.028 GR, grazing; GA, Galon; LA, Lachish; PR, Proximity; PS, Patch size; F, Fractal; C, Cistern; W, Wall; HA, Heap area; PD, plot diversity; EX, Exposed soil; RO, Rock cover percent; NAS, Northern aspect; CIs, confidence intervals; RI, relative importance (the sum of the Akaike weights of models that include the variable, see statistical analysis).

Table 11. Model-averaged parameter estimates and CIs of reptile species diversity

Parameter RI Estimates Upper CI Lower CI

GR 1 -0.3402 -0.27456 -0.40584

GA 1 -1.14 -0.94716 -1.33284

LA 1 -1.207 -1.01874 -1.39526

GA×GR 1 0.4566 0.534138 0.379062

LA×GR 1 0.61 0.704413 0.515587

PD 1 -0.6195 -0.32864 -0.91036

W 1 0.1231 0.20152 0.04468

F 0.99 -0.3361 -0.08444 -0.58776

EX 0.96 -0.0047 -0.00058 -0.00877

C 0.91 -2E-06 7.07E-08 -3.3E-06

HA 0.81 -0.0011 0.001341 -0.00353

PR 0.8 -0.0412 0.036946 -0.11927

PS 0.75 -0.0228 0.020262 -0.06582

RO 0.39 0.00081 0.007849 -0.00624

NAS 0.3 0.5592 1.333596 -0.2152

Discussion In this chapter, I examined scale-dependent variables that affect reptile communities on natural patches in an agricultural system. I asked two main questions: (1) What are the scale-dependent ecological variables that affect reptile communities on natural patches within a sharply gradiented agro-ecosystem and (2) Are particular community measures differently affected by variables (determinants) of different scales?

The results obtained can be explained simoultaneousely by two non-mutually exclusive hypotheses: the biotic and the abiotic. With respect to scale effects, my results supported the hypothesis that the structure of a natural patch's reptile community is affected by different ecological processes related to different spatial scales (Michael et al. 2008). Consistently, models that included several variables from different scales attained lower AICc values than models that included several variables from the same scale.

The results of the GLMs also strengthened my prediction that each of the three community measures -- abundance, species richness and diversity -- is affected by different variables operating at different scales. With regard to plot scale variables, the results clearly indicate that all the community measures were strongly affected by variables representing plot variability and associated with plot heterogeneity. Those variables (e.g. ancient wall and exposed soil) affected plot heterogeneity and habitat availability within each patch. Based on niche theory, an increase in plot scale heterogeneity could explain the increases in richness and species diversity (Vandermeer 1972). In addition, the relatively high importance of grazing for each one of the community measures may also relate to the effect of grazing on plot heterogeneity and niche availability. Grazing can change plant cover percentage, and by doing so may change habitat quality and availability.

With respect to patch scale, in contrast to my hypothesis, patch size had a relatively low effect on reptile community measures. However, it is important to note that reptiles were sampled within equal-sized plots, hence measuring “species density” rather than a patch- corrected species richness (Holt 1992, Gotelli and Graves 1996, Connor et al. 2000).

Therefore, the relationship between patch size and species richness in my study does not represent a classic species-area relationship. However, the results indicate that large and rich reptile populations can also be found within small patches.

With regard to land-unit scale, surprisingly, the variable that describes patch isolation (proximity value) was found to be of a relatively little importance to reptile species diversity. Since this was not tested in a controlled experiment I cannot point to the exact mechanism, but I suggest two alternative explanations for this pattern. One possibility is that the agricultural matrix, within which the patches are located, is completely hostile to reptiles and no matter what the isolation is, reptiles cannot move between patches. Conversely, the agricultural matrix may not constitute any barrier at all to the movement of reptiles, rendering the degree of isolation meaningless; in other words, none of the patches is really isolated. In order to be able to accept or reject one of these prospective hypotheses, further controlled experiments are needed. Such information is important in planning an agricultural system that supports the existence of reptile communities.

At the landscape scale, in accordance with my hypothesis, I found that climatic gradient is a major factor in the three community measures. Since I have not tested this in a controlled experiment I cannot point to the exact mechanism, but I suggest two explanations for this pattern. The climatic change from Mediterranean climate to semi- arid climate is reflected, among other things, in reduced plant biomass. It is reasonable to assume that this decrease leads to changes in abundance, species richness and species diversity of higher trophic levels, which can explain the change in community measures according to the climatic gradient (Wadie et al. 1999, Mittelbach et al. 2001). It is also possible that the reptile community, like other biological groups (e.g. plants, Giladi unpublished data; arthropods, Gavish unpublished data) changed from a Mediterranean community to a desert-oriented community along the SJL climatic gradient. If this scenario is true, then this process can lead to changes in the community measures according to the climatic gradient. In order to be able to accept or reject one of these prospective explanations, further controlled experiments are needed.

Based on their ecological requirements, (e.g. being ectothermic predators), reptiles depend on many particular environmental characteristics. For example, they need the

27 option of exposing themselves to solar radiation in order to achieve active temperature as early as possible; however they also need protection from high temperatures at other times of the day / night. They also need ambush and chasing sites of particular nature in order to meet their specific predatory strategies and behaviors, yet they need also shelters from their own predators. In addition, the overall population persistence depends on the movement of individuals between remote patches for mating and demographic-related effects (e.g., rescue effect; Brown and Kodric- Brown 1977). Consequently, only a heterogeneous habitat and habitat clusters at different scales can successfully provide all the conditions for maintaining a reptile population. This is supported by the relatively high importance we found for the local scale and the variables related to habitat- heterogeneity. Since modern agricultural areas reduce habitat heterogeneity, they affect the ability of reptiles to find suitable habitats. One option for solving this negative effect is to protect natural patches within the agricultural system that retain the local and spatial heterogeneity. This approach is known as ‘Wildlife Friendly Agriculture’ (WFA) and is a relatively cheap and simple method to protect biodiversity within an agricultural system (Green et al. 2005, Phalan et al. 2011).

My results support the approach that protecting natural patches within an agro-ecosystem using WFA is an effective tool to protect biodiversity. Based on previous literature (Bouskila and Amitai 2001) and expert information (A. Bouskila, B. Shacham, personal communication), over the whole research area 38 reptile species should exist (out of the 87 native terrestrial reptile species in Israel). From this potential species pool I found 29 species within 29 ha (the total size of the patches where sampling was carried out). This strongly verifies that the herpetological richness in this region heavily benefits from those natural patches.

Natural patches in the agricultural system of the Southern Judea Lowlands have remained untouched due to their physically rocky structure that has not allowed heavy agricultural machinery to invade. Therefore, these patches are relicts rather than products of a-priori planning. Early planning of an agricultural system which includes maintaining heterogeneous natural patches can help to protect reptile communities within intensive modern agricultural systems. The fact that a large and diverse reptile community can

28 exist within a relatively small patch collection is important to include in any discussion about the feasibility of gaining farmers' cooperation and their willingness to keep natural patches within the agricultural area, since the loss of agricultural land is relatively minor (Herzon and Mikk 2007).

There are still many open questions which need to be considered in order to be able to plan sustainable agricultural systems that will maintain reptile species diversity over a long time. The impact of the agricultural matrix on the possibilities for reptile movement between natural patches needs further research. The influence of agricultural matrix heterogeneity on reptile community needs additional research since different crops probably have different impacts on reptile populations in a patch. In addition, other agricultural practices, such as grazing, should also be studied in the context of maintaining biodiversity in general and reptile communities in particular.

Appendices

Appendix 1: Precipitation (mm) over the research area during the winter of 2008-9 and 2009-2010*

Land unit 2008-9 2009-10

Galon 283 317

Lachish 241 264

Dvir 197 231

*Data source: Israeli Meteorological Service

Appendix 2: Abundance Model Set

AICc Delta Wi ID Model AICc

1 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+RO+NAS -449.203 0.000 0.437

2 AB=LU+LU*GR+PS+F+PR+GR+C+W++PD+EX+RO+NAS -447.638 1.565 0.200

3 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+RO+NAS -447.363 1.840 0.174

4 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+RO -445.299 3.904 0.062

5 AB =LU+LU*GR+F+PR+GR+C+W+HA+PD+EX+RO+NAS -444.845 4.358 0.049

6 AB=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX+RO+NAS -444.784 4.419 0.048

7 AB=LU+LU*GR+F+PR+GR+C+W+HA+PD+EX+RO -442.516 6.687 0.015

8 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+RO -441.568 7.635 0.010

9 AB=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX+RO+NAS -438.241 10.962 0.002

10 AB=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX+RO -437.907 11.296 0.002

11 AB=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX+RO+NAS -435.877 13.326 0.001

12 AB=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX+RO -435.475 13.728 0.000

13 AB=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX+RO -431.052 18.151 0.000

14 AB=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX+RO+NAS -421.428 27.775 0.000

15 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX+RO -415.299 33.904 0.000

16 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX+RO+NAS -413.168 36.035 0.000

17 AB=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX+RO -409.575 39.628 0.000

18 AB=LU+PS+F+PR+GR+C+W+HA+PD+EX+RO -397.150 52.053 0.000

19 AB=LU+PS+F+PR+GR+C+W+HA+PD+EX+RO+NAS -395.074 54.129 0.000

20 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+NAS -375.831 73.372 0.000

21 AB=LU*GR+PS+F+PR+GR+C+W+PD+EX -352.491 96.712 0.000

22 AB=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX -351.340 97.863 0.000

23 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX -350.422 98.781 0.000

24 AB=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX -349.254 99.949 0.000

25 AB=LU+LU*GR+PS+F+PR+GR+C+W+PD -347.316 101.887 0.000

26 AB=LU+LU*GR+PS+F+GR+C+W+HA+PD -345.739 103.464 0.000

27 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD -345.294 103.909 0.000

28 AB=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX -344.673 104.530 0.000

29 AB=LU+LU*GR+PS+F+PR+GR+C+HA+PD -344.594 104.609 0.000

30 AB=LU+LU*GR+PS+F+PR+GR+W+HA+PD -341.119 108.084 0.000

31 AB= PS+F+PR+GR+C+W+HA+PD+EX -339.142 110.061 0.000

32 AB= PS+F+PR+GR+C+W+HA+PD -337.730 111.473 0.000

33 AB=LU+LU*GR+F+PR+GR+C+W+HA+PD+EX -337.151 112.052 0.000

34 AB=LU+LU*GR+F+PR+GR+C+W+HA+PD -311.611 137.592 0.000

35 AB=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX -303.052 146.151 0.000

36 AB=LU+LU*GR+PS+PR+GR+C+W+HA+PD -286.379 162.824 0.000

37 AB=LU+LU*GR+PS+F+PR+GR+C+HA -278.441 170.762 0.000

38 AB=LU+LU*GR+F+PR+GR+C+W+HA -277.767 171.436 0.000

39 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA -277.762 171.441 0.000

40 AB=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX -277.722 171.481 0.000

41 AB=LU+LU*GR+PS+F+GR+C+W+HA -275.133 174.070 0.000

42 AB=LU+LU*GR+PS+F+PR+GR+C+W -270.483 178.720 0.000

43 AB=LU+LU*GR+F+PR+GR+C+W -268.951 180.252 0.000

44 AB= LU*GR+PS+F+GR+C+W -268.311 180.892 0.000

45 AB=LU+LU*GR+PS+F+PR+GR+C -266.428 182.775 0.000

46 AB=LU+LU*GR+PS+F+PR+GR+W+HA -265.300 183.903 0.000

47 AB=LU+LU*GR+PS+F+GR+C -263.391 185.812 0.000

48 AB=LU+LU*GR+PS+F+PR+GR+W -263.082 186.121 0.000

49 AB=LU+LU*GR+F+PR+GR+C -259.848 189.355 0.000

50 AB=LU+PS+F+PR+GR+C+W+HA -257.442 191.761 0.000

51 AB=LU+PS+F+PR+GR+C+W -252.039 197.164 0.000

52 AB=LU+LU*GR+PS+F+GR -251.594 197.609 0.000

53 AB=LU*GR+PS+F+PR+GR -251.078 198.125 0.000

54 AB=LU+PS+F+PR+GR+C -248.174 201.029 0.000

55 AB=LU+LU*GR+PS+PR+GR+C+W+HA -245.093 204.110 0.000

56 AB=LU+LU*GR+PS+PR+GR+C+W -241.603 207.600 0.000

57 AB= LU*GR+F -241.433 207.770 0.000

58 AB=LU+LU*GR+F+PR+GR -239.377 209.826 0.000

59 AB=LU*GR+PS+PR+GR+C -235.872 213.331 0.000

60 AB=LU+PS+F+PR+GR -235.392 213.811 0.000

61 AB=LU*GR+PS -222.299 226.904 0.000

62 AB=LU+LU*GR+PS+PR+GR -220.649 228.554 0.000

63 AB=LU+LU*GR -215.717 233.486 0.000

64 AB=LU+PS+F -213.047 236.156 0.000

65 AB=LU+PS+F+PR -211.149 238.054 0.000

66 AB=LU+PS -185.911 263.292 0.000

67 AB=LU -157.210 291.993 0.000

68 Null 39.368 488.571 0.000

AB-Abundance, LU-Land-unit, GR-Grazing, PR-Proximity, PS-Patch size, F-Fractal, , C-Cistern,W-Wall, HA-Heap area, PD-Plot diversity, EX-Exposed soil, RO-Rock cover percent, NAS –Northern aspect

Appendix 3: Species Richness Model Set

AICc Delta Wi ID Model AICc

1 SR =LU+LU*GR+F+PR+GR+C+W+HA+PD+EX+RO+NAS -584.802 0 0.542

2 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+RO+NAS -583.637 1.165 0.303

3 SR=LU+LU*GR+PS+F+PR+GR+C+W+PD+EX+RO+NAS -581.302 3.500 0.094

4 SR=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX+RO+NAS -579.881 4.921 0.046

5 SR=LU+LU*GR+F+PR+GR+C+W+HA+PD+EX+RO -576.462 8.340 0.008

6 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+RO -575.501 9.301 0.005

7 SR=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX+RO+NAS -569.860 14.942 0.000

8 SR=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX+RO -569.746 15.056 0.000

9 SR=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX+RO+NAS -569.246 15.556 0.000

10 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+RO -569.208 15.594 0.000

11 SR=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX+RO -568.447 16.355 0.000

12 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+RO+NAS -567.897 16.905 0.000

13 SR=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX+RO -567.228 17.574 0.000

14 SR=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX+RO+NAS -558.805 25.997 0.000

15 SR=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX -558.070 26.732 0.000

16 SR=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX+RO -557.710 27.092 0.000

17 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX -556.901 27.901 0.000

18 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+NAS -556.348 28.454 0.000

19 SR=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX -555.779 29.023 0.000

20 SR=LU+LU*GR+PS+F+PR+GR+C+W+PD+EX -554.934 29.868 0.000

21 SR=LU+LU*GR+PS+F+PR+GR+W+HA+PD -551.587 33.215 0.000

22 SR=LU+LU*GR+PS+F+PR+GR+C+W+PD -549.601 35.201 0.000

23 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD -549.532 35.270 0.000

24 SR=LU+LU*GR+PS+F+GR+C+W+HA+PD -547.577 37.225 0.000

25 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX+RO+NAS -545.004 39.798 0.000

26 SR=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX -543.996 40.806 0.000

27 SR=LU+LU*GR+PS+F+PR+GR+C+HA+PD -543.832 40.970 0.000

28 SR=LU*GR+PS+F+PR+GR+C+HA+PD+EX -543.265 41.537 0.000

29 SR=LU+LU*GR+F+PR+GR+C+W+HA+PD+EX\ -540.435 44.367 0.000

30 SR=LU+LU*GR+PS+PR+GR+C+W+HA+PD -529.504 55.298 0.000

31 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX+RO -517.975 66.827 0.000

32 SR=LU+LU*GR+F+PR+GR+C+W+HA+PD -507.915 76.887 0.000

33 SR=LU+LU*GR+PS+F+PR+GR+W -469.673 115.129 0.000

34 SR=LU*GR+PS+F+PR+GR+W+HA -469.239 115.563 0.000

35 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX -468.958 115.844 0.000

36 SR=LU+LU*GR+PS+F+PR+GR+C+W+HA -468.946 115.856 0.000

37 SR=LU*GR+PS+F+PR+GR+C+W -468.463 116.339 0.000

38 SR=LU+LU*GR+F+PR+GR+C+W+HA -468.312 116.490 0.000

39 SR=LU+LU*GR+F+PR+GR+C+W -466.920 117.882 0.000

40 SR=LU+LU*GR+PS+F+PR+GR+C+HA -466.304 118.498 0.000

41 SR=LU+LU*GR+PS+PR+GR+C+W -464.377 120.425 0.000

42 SR=PS+PR+GR+C+W+HA -464.036 120.766 0.000

43 SR=LU+LU*GR+PS+F+GR+C+W+HA -462.587 122.215 0.000

44 SR=LU+LU*GR+PS+F+GR+C+W -462.404 122.398 0.000

45 SR=LU+LU*GR+PS+F+PR+GR+C -461.733 123.069 0.000

46 SR=LU*GR+PS+F+PR+GR -459.410 125.392 0.000

47 SR=LU+LU*GR+PS+PR+GR+C -456.648 128.154 0.000

48 SR=LU+LU*GR+PS+F+GR -454.519 130.283 0.000

49 SR=LU+LU*GR+PS+F+GR+C -454.321 130.481 0.000

50 SR=LU+LU*GR+PS+PR+GR -454.147 130.655 0.000

51 SR=LU+F+PR+GR+C -454.076 130.725 0.000

52 SR=LU*GR+PS -450.631 134.171 0.000

53 SR=LU+LU*GR+F+PR+GR -448.883 135.919 0.000

54 SR=LU+LU*GR+F -448.304 136.498 0.000

55 SR=LU+LU*GR -445.585 139.217 0.000

56 SR=LU+PS+F+PR+GR+C+W+HA+PD+EX+RO+NAS -421.634 163.168 0.000

57 SR=PS+F+PR+GR+C+W+HA+PD+EX+RO -418.620 166.182 0.000

58 SR=LU+PS+F+PR+GR+C+W+HA+PD -410.206 174.596 0.000

59 SR=LU+PS+F+PR+GR+C+W+HA+PD+EX -408.474 176.328 0.000

60 SR~LU+PS+F+PR+GR+C -372.480 212.322 0.000

61 SR=LU+PS+F+PR+GR+C+W -371.362 213.440 0.000

62 SR=PS+F+PR+GR+C+W+HA -369.543 215.259 0.000

63 SR=LU+PS+F+PR -368.714 216.088 0.000

64 SR=LU+PS+F+PR+GR -367.726 217.076 0.000

65 SR=LU+PS+F -366.124 218.678 0.000

66 SR=LU+PS -360.020 224.782 0.000

67 SR=LU -297.466 287.336 0.000

68 Null -239.456 345.346 0.000

SR-Species Richness, LU-Land-unit, GR-Grazing, PR-Proximity, PS-Patch size, F-Fractal, , C-Cistern,W-Wall, HA-Heap area, PD- Plot diversity, EX-Exposed soil, RO-Rock cover percent, NAS –Northern aspect

Appendix 4: Species Diversity Model Set

AICc Delta Wi ID Model AICc

1 SD=LU+LU*GR+F+PR+GR+C+W+HA+PD+EX 13.672 0 0.153

2 SD=LU+LU*GR+PS+F+PR+GR+C+W+PD+EX 13.992 0.321 0.130

3 SD=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX 14.299 0.628 0.112

4 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX 14.946 1.274 0.081

5 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+NAS 15.120 1.448 0.074

6 SD=LU+LU*GR+PS+F+PR+GR+C+W++PD+EX+RO+NAS 15.488 1.816 0.062

7 SD=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX 15.654 1.982 0.057

8 SD=LU+LU*GR+F+PR+GR+C+W+HA+PD+EX+RO 15.750 2.079 0.054

9 SD=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX+RO+NAS 15.923 2.252 0.050

10 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+RO 16.298 2.627 0.041

11 SD=LU+LU*GR+PS+F+PR+GR+W+HA+PD+EX+RO 16.380 2.708 0.039

12 SD =LU+LU*GR+F+PR+GR+C+W+HA+PD+EX+RO+NAS 16.454 2.782 0.038

13 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+RO+NAS 16.508 2.837 0.037

14 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+EX+RO+NAS 17.101 3.430 0.028

15 SD=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX+RO 17.542 3.871 0.022

16 SD=LU+LU*GR+PS+F+GR+C+W+HA+PD+EX+RO+NAS 18.826 5.154 0.012

17 SD=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX 20.079 6.407 0.006

18 SD=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX+RO 21.991 8.319 0.002

19 SD=LU+LU*GR+PS+PR+GR+C+W+HA+PD+EX+RO+NAS 24.062 10.390 0.001

20 SD=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX+RO+NAS 24.341 10.670 0.001

21 SD=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX 24.617 10.945 0.001

22 SD=LU+LU*GR+PS+F+PR+GR+C+HA+PD+EX+RO 26.450 12.778 0.000

23 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX+RO 29.152 15.480 0.000

24 SD=LU+LU*GR+PS+F+PR+GR+W+HA+PD 30.610 16.939 0.000

25 SD=LU+LU*GR+PS+F+PR+GR+C+W+PD 30.627 16.956 0.000

26 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+EX+RO+NAS 31.244 17.572 0.000

27 SD=LU*GR+PS+F+PR+GR+C+W+HA+EX 32.374 18.702 0.000

28 SD=LU+LU*GR+PS+F+PR+GR+C+HA+PD 32.442 18.770 0.000

29 SD=LU*GR+PS+F+PR+GR+C+W+HA+PD 32.704 19.032 0.000

30 SD=LU+LU*GR+F+PR+GR+C+W+HA+PD 32.872 19.200 0.000

31 SD= PS+PR+GR+C+W+HA+PD 33.229 19.558 0.000

32 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA+PD+RO 33.627 19.956 0.000

33 SD=LU+LU*GR+F+PR+GR 33.711 20.039 0.000

34 SD=LU+LU*GR+PS+F+PR+GR+W 33.933 20.262 0.000

35 SD=LU+LU*GR+F+PR+GR+C+W 33.934 20.262 0.000

36 SD=LU+LU*GR+PS+F+GR+C+W+HA+PD 34.519 20.847 0.000

37 SD= LU*GR+PS+F+PR+GR 34.959 21.287 0.000

38 SD=LU+LU*GR+F+PR+GR+C 34.967 21.295 0.000

39 SD=LU+LU*GR+F+PR+GR+C+W+HA 35.515 21.844 0.000

40 SD=LU+LU*GR+PS+F+PR+GR+W+HA 35.623 21.951 0.000

41 SD=LU+LU*GR+PS+F+PR+GR+C+W 36.010 22.338 0.000

42 SD=LU+LU*GR+F 36.119 22.447 0.000

43 SD=LU+LU*GR+PS+F+PR+GR+C 36.481 22.809 0.000

44 SD=LU+LU*GR+PS+F+PR+GR+C+HA 37.028 23.356 0.000

45 SD=LU+LU*GR+PS+PR+GR 37.522 23.850 0.000

46 SD=LU+LU*GR+PS+F+PR+GR+C+W+HA 37.620 23.948 0.000

47 SD= LU*GR+PS+F+GR 38.139 24.467 0.000

48 SD=LU+LU*GR+PS+F+GR+C+W 39.089 25.418 0.000

49 SD=LU+LU*GR+PS+PR+GR+C 39.105 25.433 0.000

50 SD= PS+PR+GR+C+W 39.126 25.454 0.000

51 SD=LU+LU*GR 39.877 26.205 0.000

52 SD=LU+LU*GR+PS+F+GR+C 40.193 26.521 0.000

53 SD=LU+LU*GR+PS+PR+GR+C+W+HA 40.340 26.668 0.000

54 SD=LU+LU*GR+PS+F+GR+C+W+HA 40.787 27.115 0.000

55 SD=LU*GR+PS 41.753 28.082 0.000

56 SD=LU+PS+F+PR+GR+C 203.135 189.464 0.000

57 SD=LU+PS+F+PR+GR+C+W 204.006 190.334 0.000

58 SD=LU+PS+F+PR+GR 204.673 191.001 0.000

59 SD=LU+PS+F+PR+GR+C+W+HA+PD+EX+RO+NAS 205.580 191.908 0.000

60 SD=LU+PS+F+PR+GR+C+W+HA 205.875 192.203 0.000

61 SD=LU+PS+F+PR+GR+C+W+HA+PD+EX 207.070 193.398 0.000

62 SD= PS+F+PR+GR+C+W+HA+PD 207.434 193.762 0.000

63 SD=LU+PS+F+PR+GR+C+W+HA+PD+EX+RO 207.931 194.259 0.000

64 SD=LU+PS+F+PR 207.995 194.324 0.000

65 SD=LU+PS 208.221 194.549 0.000

66 SD=LU+PS+F 208.597 194.926 0.000

67 Null 227.584 213.912 0.000

68 SD=LU 228.720 215.049 0.000

SD-Species Diversity, LU-Land-unit, GR-Grazing, PR-Proximity, PS-Patch size, F-Fractal, , C-Cistern,W-Wall, HA-Heap area, PD- Plot diversity, EX-Exposed soil, RO-Rock cover percent, NAS –Northern aspect

Appendix 5: Species list according the three land-units.

Species Galon Lachish Dvir

Ablepharus rueppellii 1 1 1

Chalcides guentheri 1 1 1

Chalcides ocellatus 1 1 1

Chamaeleo chamaeleon recticrista 1 1 1

Coluber jugularis asianus 1 1 1

Coluber rubriceps 1 1 1

Eirenis rothi 1 1 1

Eryx jaculus 1 1 1

Eumeces schneideri pavimentatus 1 1 1

Eumeces schneideri schneideri 0 0 1

Hemidactylus turcicus 0 1 1

Laudakia stellio spp 1 1 1

Leptotyphlops macrorhynchus 0 1 1

Mabuya vittata 1 1 1

Malpolon monspessulanus insignitus 1 1 0

Mesalina olivieri 0 0 1

Mesaling guttulata 0 0 1

Micrelaps muelleri 1 0 0 natrix tessellata 1 0 0

Ophisaurus apodus 1 0 0

Ophisops elegans 1 1 1

Ptyodactylus guttatus 0 0 1

Rhynchocalamus melanocephalus 1 0 1

Stenodactylus sthenodactylus 0 0 1

Telescopus fallax syriacus 0 1 1

Testudo graeca 1 1 1

Typhlops simoni 0 1 1

Typhlops vermicularis 1 0 1

Vipera palaestinae 1 1 1

1 – Species was found in the land-unit

0- Species was not found in the land-unit

Chapter 2: Combined Effects of Climatic Gradient and Domestic Grazing on Reptile Community Structure in a Heterogeneous Agro-Ecosystem (This chapter is prepared for publication with the authorship of Guy Rotem, Amos Bouskila, Itamar Giladi and Yaron Ziv)

Introduction

Ecological communities are affected by many biotic and abiotic factors. These include climate conditions and livestock grazing. Climate is a basic environmental factor. Climate conditions within a specific area can affect community structure via their influence on many biotic and abiotic attributes (De Bello et al. 2005, Rowe 2007, Jiguet et al. 2011). Such attributes may include physiological state, fitness and survival rate. Grazing is a common agricultural activity in many parts of the world: 30% of ice-free land area is used for pastoral activity worldwide (Steinfeld et al. 2006). Also many studies have examined the combined effect of these two factors – climate and grazing – on plant community structure (for example see: Pakeman 2004, De Bello et al. 2005, Diaz et al. 2007), only few studies examined this combination on reptile community structure.

Studies on grazing effects show that sheep and cattle grazing may influence several common biotic and abiotic factors, including plant diversity and community structure (Adler et al. 2001). However, the effects of grazing on plant diversity are not uniform (Olff and Ritchie 1998). Adler et al. (2001) found that the effects of grazing on spatial heterogeneity are also not uniform. They found that an increase or decrease in spatial heterogeneity is the product of different interactions between the spatial patterns of grazing and vegetation.

The influence of grazing on plants itself also depends on climatic conditions (Osem et al. 2002, Olff and Ritchie 1998). Different plant traits, such as plant identity, life form, and anti-grazing mechanisms, may differ between plant communities based on the climatic conditions (Perevolotsky and Seligman 1998, De Bello et al. 2005). Grazing within a dense vegetation environment (e.g., Mediterranean forest) may decrease the percentage of vegetation cover, which, consequently, can increase the plant species diversity and landscape heterogeneity (Perevolotsky and Seligman 1998, Germano et al. 2012), as

42 predicted by the Intermediate Disturbance Hypothesis (Huston 1979). In contrast, grazing in semi-arid environments may damage the sporadically scattered vegetation, reducing landscape heterogeneity and resulting in a decline in biodiversity (Walker et al. 1981). In extreme cases, intensive grazing in semi-arid environments can lead to desertification (Fleischner 1994, Van Auken 2000).

Grazing effects on habitat heterogeneity may also affect reptile communities (Castellano and Valone 2006). Previous studies have found a positive effect of grazing on reptile abundance and species richness in cases where grazing increased habitat heterogeneity (Fabricius et al. 2003, Germano et al. 2012). In cases where grazing reduced the landscape heterogeneity, negative effects on reptile abundance and species richness were observed (Attum et al. 2006, Brown et al. 2011).

Israel is located between latitudes 29.5° and 33.3° north, and serves as a transition zone for flora and fauna between the deserts of North Africa in the south to temperate areas in the north. This unique positioning marks out the country as a meeting point for biota from different biogeographic zones such as the Mediterranean, Irano-Turanian, Saharo- Arabic and Sudanese (Alon and Waisel 1984, Svoray et al. 2005). As a result, the numbers of plant and animal species in Israel are relatively high, making Israel a Palearctic provincial hotspot (Gavish 2011). Even in an article that claims that Israel is not exceptional as a global hotspot, the authors point out that the richness of reptiles is indeed higher in Israel than expected (Roll et al. 2009).

My study area – the Southern Judea Lowlands (SJL) - largely represents the ecotone existing between the Desert and Mediterranean biomes over a limited area. Average annual rainfall ranges from 450 mm in Galon (Mediterranean climate, northern part of SJL) to 250 mm in Rahat (semi-arid climate, southern SJL) (Gvirtzman 2002), over 43 km and with no significant change in the topographic elevation. This position of the SJL in a transition zone between Mediterranean and arid ecosystems (Kark et al. 1999) uniquely brings together biota from the Mediterranean, Irano-Turanian and Saharo- Arabic biogeographic zones within a single ecotone (Giladi et al. 2011).

The extremely sharp climatic gradient of the SJL confers a scientific advantage since it provides a study area without the confounding effects of geological, lithological and human historical variation. One characteristic of human presence over the entire study area is seasonal grazing of sheep and cattle. Given the extremely sharp climatic gradient and presence of grazing throughout the area, I considered the SJL as an ideal landscape to study the unexplored aspects of combined effects of domestic animal grazing and climatic gradient on reptile community structure.

This study examines the combined effect of a sharp climatic ecotone and pastoral livestock on reptile communities in the SJL. Specifically, the study addresses three main questions: (1) What is the effect of domestic grazing and climatic gradient on reptile abundance, species richness and species diversity? (2) How do climatic gradient and grazing together affect the structure of species from different biogeographic origins? (3) What is the impact of grazing and climatic gradient on reptiles’ physical state?

With respect to the first question, I hypothesize that grazing of sheep and cattle affect reptile community measures according to the location along the climatic gradient. Specifically, I predict that in the southern area, due to low vegetation biomass, scarce precipitation and slow vegetation regeneration rates, spatial heterogeneity will decrease as a result of intensive grazing. This will lead to a decline in reptile abundance and species richness. Conversely, in the northern research area I predict that pastoral activity will increase the landscape heterogeneity, resulting in increased reptile abundance and species richness.

With respect to the second question, I hypothesize that grazing of sheep and cattle affect reptile community structure of species from different biogeographic origins differently according to the climatic conditions. Specifically, I predict changes in reptile species composition according to the climatic gradient. A high proportion of Mediterranean species is predicted to be found in the northern area. This proportion will decrease with respect to the climatic gradient. Since grazing can affect percentage plant cover, I also predict a negative effect of grazing on Mediterranean reptile species frequency. However, since the impact of grazing on plant and spatial heterogeneity depends on

44 climatic conditions, I predict that the negative effect of grazing on Mediterranean species composition will be stronger in the southern area than in the north.

With respect to the third question, I hypothesize that grazing affect reptiles’ physical state differently according to the location long the climatic gradient and to the biogeographic origins. Specifically, I predict that the physical state of Mediterranean species will decline from north to south according to the climatic gradient. I also predict that grazing will have similar effect and that the physical state of individuals will decline according to grazing pressure.

Methods Study Area

The geology of the Southern Judea Lowlands (SJL; 31°30’52’’N 34°52’36’’E; Figure 1) is dominated by soft limestone and chalk, covered by “Nari” (a thin, strong, calcium- based layer; (Kloner and Tepper 1987). Thousands of years of human development have amplified the patchy nature of the landscape (Ben-Yosef 1980).

Figure 1. Study area location within Israel on the left. The Southern Judea Lowlands with the four land-units are listed on the right. Note the sharp decline in mean annual precipitation along a short gradient of 43 km.

Domestic animals have been grazed in Israel and in the SJL for about seven thousand years (Perevolotsky and Seligman 1998). In the northern area (Galon), the main grazing animals are cattle. In the rest of the area, sheep and goats are the most common grazers. The grazing season extends from May to September. In May, after the end of the wheat harvest, shepherds are granted permission to bring their herds into the agricultural area. Grazing is partly in stubble fields and partly in the natural areas surrounding the agricultural fields. Once most of the stubble has been consumed, the grazing is

46 concentrated on natural areas only. In September, as part of preparation for wheat seeding, the grazing period ends, and the herds leave the agricultural area. In most cases, at the end of the grazing season herds are put into closed pens and fed on bought food.

Study Design and Survey Protocol

Within the SJL landscape, I chose four 4×3.2 km ‘land-units’ (from north to south: Galon, Lachish, Dvir and Rahat; Figure 1), each characterized by remnant patches of natural vegetation of different size and configuration, surrounded by a matrix of agricultural fields. In three of the land-units (Galon, Lachish and Dvir) I surveyed reptiles in 63 sampling plots 100×50 m in size, located within 21 natural patches. In Rahat, due to very heavy grazing and new planting, twenty survey sites were chosen where human impact was relatively minimal. I used rectified aerial photographs (Ofek, 2005; pixel size = 1m2) to identify all the patches of natural vegetation within each of these land-units. Their boundaries were then demarcated on a digitized map and the information stored as vector-based coverage in a Geographical Information System (GIS) platform (ArcInfo TM; ESRI).

I actively searched for reptiles in springtime (March to June) of 2009 and 2010 on clear days with a daily maximum temperature above 25 °C. In Galon, Lachish and Dvir three complementary active methods were used for sampling the reptile community within each plot: (i) line transects, (ii) over-turning stones, and (iii) pitfall traps. In Rahat, due to accelerated human activity, pitfall traps were not used. Line transect surveys were conducted during the late morning, after the onset of reptile activity, by walking in a straight line along the long axis of the sampling plot. Over-turning of stones was performed along the long axis of the sampling unit during the late morning, immediately after completion of the line transect surveys. I made an effort to turn over any stone with a diameter exceeding 10 cm, and in any case at least 100 stones per plot. All the stones were replaced in their original position to avoid habitat change. For pitfall trapping, twenty one-liter dry pitfall traps were dug in each plot (total of 1260 traps). The traps were arranged at 10 m intervals in two rows 25 m apart, parallel to the long axis of the plot. The traps were opened before sunset and left open for two entire days (Fisher et al. 2008). To prevent injury to animals and to reduce their discomfort a tiled roof was built

47 over each trap. This roof protected the trap from solar radiation and reduced the exposure of trapped animals to predation risk. In the first year of the study captured individuals were marked individually. Since no repeated captures were made during the first year, this practice was stopped.

Since I monitored reptiles in springtime, the observed data were influenced from the end of last year’s grazing season. Grazing levels were indirectly measured from feces density levels (Nicholson et al. 2006, Abaturov et al. 2008).

Measuring the Variables

I focused my analyses on the combined effects of grazing and climatic gradient on three community measures: abundance (i.e., the total number of individuals of all species present), local species richness (i.e., the number of species at a defined sampling unit) and species diversity, using ‘Fisher's alpha index of diversity’(Fisher et al. 1943). I was also interested in the combined effects of grazing and climatic gradient on the relative representation of species from different biogeographic zones. The assignment of species to biogeographic zones was done a-priori and was based on a previous knowledge of each species I recorded (Bouskila and Amitai 2001, Bouskila 2002).

Using line transects along the long axis of each sampling plot I measured plot-level explanatory variables including percentage cover of Thorny burnt (Sarcopterium spinosum),Thatching grass (Hyparrhenia hirta), all the other shrubs, total vegetation cover, stone cover and exposed soil.

I quantified grazing levels in Dvir and Lachish by measuring grazer feces density in ten 1×1m quadrats in each plot. The ten densities were then averaged to a single-plot average feces density (i.e. feces patches). In Galon, due to the difference in grazing species (i.e. cows), I measured grazing feces density in five 2×2 m quadrats in each plot. The five densities were then averaged to a single-plot density. In Rahat these data were missing. For most analyses, I refer to feces density data as a continuous variable, but for some analyses (e.g. MDS), feces density was divided into categories and referred to as a categorical variable. In order to categorize these data, the average feces values obtained were classified into three density categories: (I) from 0 to 50 feces patches (low grazing

48 level); (II) from 51 to 100 feces patches (moderate grazing level); and (III) more than100 feces patches (high grazing level) for Dvir and Lachish. In Galon, to suit the different grazing species, three slightly different density categories were applied: (I) from 0 to 5 feces patches (low grazing level); (II) from 5 to 10 feces patches (moderate grazing level); and (III) more than 10 feces patches (high grazing level). In Rahat, these data were missing. I therefore did not include Rahat in any analysis of grazing effect.

In order to analyze climatic and grazing effects on reptile body condition I used the “Index of Body Condition” (IC)(Andrews and Wright 1994):

푚푎푠푠0.3 퐼퐶 = ( ) ∗ 100 (1) 푆푉퐿 where mass is the reptile weight in grams and SVL is reptile length from snout to vent in millimeters. Trachylepis vittata, which is a Mediterranean species, was the only species for which there were enough data to calculate IC in all land-units and grazing level combinations. In Dvir, enough data were also gathered to calculate IC for Stenodactylus sthenodactylus, a desert-oriented species; thus in this land-unit I could compare IC values of the two species.

Statistical Analysis

Using Multi-Dimensional Scaling (MDS) analysis (Primer 6 software) and based on the plot-level explanatory variables, I calculated the habitat-type similarity for Galon, Lachish and Dvir plots. I analyzed the effect of grazing on percentage vegetation cover with a linear regression approach. I also used linear regression to test the effect of grazing on each of the three community measures (abundance, species richness and diversity). To do so I took the data collected in 2009 and 2010 and performed an analysis separately for each year. I also analyzed the effect of grazing on the proportion and abundance of reptile species of different biogeographic origins using a linear regression approach.

I used a one-way ANOVA and linear regression to test the effect of climatic gradient and grazing levels, respectively, on the body condition of T. vittata (Bauer 2003) and S. sthenodactylus, as represented by the IC. The analysis of body condition was restricted

49 to individuals that had intact tails and were adults at the time of capture (Dickinson and Fa 2000).

In order to avoid uncontrolled variables effect on the results, I analyzed the effect of grazing on the proportion and abundance of reptile species of different biogeographic origins and the effect of grazing and climatic gradient on body condition for each year separately (i.e. 2009 and 2010). Since the results were similar for the two years, in the ‘results’ section I present the 2009’s results and the results of 2010 I enclose in appendix A.

All of the above-mentioned analyses were conducted using R statistical platform (R development Core Team 2006).

Results In total, 1061 reptile observations were made (349, 226, 206 and 280 in Galon, Lachish, Dvir and Rahat, respectively), representing 34 reptile species (20, 19, 25 and 20 in Galon, Lachish, Dvir and Rahat, respectively). Fisher’s Alpha of diversity for each land-unit was 4.8, 4.5, 8.2 and 4.9 from north to south, respectively.

In Galon, Lachish and Dvir I found a negative association between grazing level and 2 percentage vegetation cover (Galon: Linear regression, N = 19, p = 0.023, R = 0.23; 2 Lachish: Linear regression, N = 19, p = 0.005, R = 0.3; Dvir: Linear regression, N = 19, p = 0.012, R2 = 0.25; data for Rahat was not obtained) and a positive association with 2 exposed soil (Galon: Linear regression, N = 19, p = 0.015, R = 0.24; Lachish: Linear 2 2 regression, N = 19, p = 0.03, R = 0.21; Dvir: Linear regression, N = 19, p = 0.018, R = 0.26).

Examining the degree of habitat-type similarity (MDS analysis, stress = 0.12; Figure 2) between the plot habitats indicates high similarity between Galon and Lachish with little effect of grazing. However, in Dvir plot habitats were clearly affected by grazing: plots with low grazing levels were situated closer towards Galon and Lachish plots, whereas plots with high grazing levels were further away.

Stress level =0.12

Figure 2. Non-metric multi-dimensional scaling (MDS) of habitat-type similarity. The symbols represent plots at three land-units (Galon, Lachish and Dvir) and the numbers represent grazing levels (from 1 to 3, where 1 is lowest grazing level and 3 is the highest grazing level). The data were square-root transformed and similarity was calculated using Bray-Curtis similarity. The results indicate high similarity between the plots of Galon and Lachish. The results also show similarity between the plots in Dvir which were subjected to low grazing levels and plots in Galon and Lachish. However, plots in Dvir which were subjected to high grazing levels show very little similarity to any other plots.

Reptile community composition was affected by the north-south climatic gradient. The proportion of observations of Mediterranean reptiles decreased from 95% of all observations in Galon to 5% in Rahat (Figure 3a). With respect to desert-oriented species, I found increasing numbers from Galon to Rahat. I also found a decrease in the frequency of Mediterranean species from 90% in Galon to 33% in Rahat (Figure 3b).

Figure 3. Change in reptile abundance (a) and species richness (b) over the SJL climatic gradient. Galon (northern land-unit) is characterized by a high frequency of Mediterranean reptiles with no arid species at all. The other land-units have fewer Mediterranean species and more generalist and arid species.

The sharp change in community composition over the climatic gradient is reflected by the percentage of species found in all four land-units -- only 20% of the species were recorded in all four land-units (Figure 4). Eighteen percent of the species were recorded in the three northern land-units only (Galon, Lachish and Dvir). The two southern land-

52 units (Rahat and Dvir) shared exclusively 15% of the species, and 12% of the species were recorded exclusively in the most southern land-unit (Rahat). The more distant units shared fewer species than the closer units (e.g. the proportion of species shared by Galon and Rahat only was lower than that shared uniquely by Dvir and Rahat) as presented in Table 1.

Figure 4. Reptile species uniqueness (from the total poll) and biogeographic origin along the sharp SJL north-south climatic gradient. The position of each circle on the graph represents the species

composition found in the same land-unit, two land-units, etc. Colored circles represent the proportion of species according to their biogeographic origin. The number in each circle represents the percentage of unique species from the total species pool in the same land-unit, two land-units, etc .

Table 1. Pair-wise table, The proportion of species shared by each pair of units decreases with distance.

Land-unit 1 Land-unit 2 Distance (km) Unique Shared

Galon Lachish 15 0.05 0.41

Galon Dvir 30 0.03 0.41

Galon Rahat 43 0 0.23

Lachish Dvir 14 0.03 0.5

Lachish Rahat 28 0 0.32

Dvir Rahat 14 0.15 0.32

Grazing affected reptile community measures in different ways according to location along the climatic gradient. In Galon, I found a significant positive effect (Figure 5a) of grazing on reptile species richness (Linear regression, N=19, p = 0.006, R2 = 0.46) and species diversity (i.e. 'Fisher's alpha index of diversity'; Linear regression, N=19, p = 0.005, R2 = 0.33). The impact of grazing on abundance was not significant (Linear regression, N=19, p = 0.08, R2 = 0.14 ). In Lachish, I found a significant positive effect of grazing on reptile species diversity (Linear regression, N=16, p = 0.018, R2 = 0.30). No significant effect of grazing on abundance (Linear regression, N=19, p = 0.286, R2 = 0.05) or species richness (Linear regression, N=19, p = 0.262, R2 = 0.06) was found. In Dvir, I found a significant negative effect of grazing on reptile species richness (Figure 5b; Linear regression, N=19, p = 0.004, R2 = 0.34), reptile abundance (Linear regression, N=19, p = 0.023, R2 = 0.20) and 'Fisher's alpha index of diversity’ values (Linear regression, N=15, p = 0.052, R2 = 0.22). All above analyses are based on 2009 data; the results for 2010 data were similar and are presented in the appendix.

(a) (b)

Figure 5. Effect of grazing on species richness varies with location along the climatic gradient. Galon (a) reveals a positive effect while Dvir (b) reveals negative effect of grazing on species richness.

I also found different effects of grazing on reptile community composition according to specific location along the climatic gradient. In Galon and Lachish no effects of grazing levels on proportions of Mediterranean, generalist or desert-oriented species from the total species poll were found (Galon, Mediterranean species proportion: Linear regression, N=19, p = 0.234, R2 = 0.07; Lachish, Mediterranean species proportion: Linear regression, N=19, p = 0.812, R2 = 0.003; Lachish, generalist species proportion, Linear regression, N=19, p = 0.19, R2 = 0.08; Lachish, desert-oriented species proportion, Linear regression, N=19, p = 0.325, R2 = 0.05) In Dvir, a significant effect of grazing on the balance of biogeographic origin was found (Figure 6a, Mediterranean species, negative effect, Linear regression, N=19, p = 0.04, R2 = 0.2; desert-oriented species, positive effect, Linear regression, N=19, p = 0.009, R2 = 0.3). Similarly the effect was found for individuals (Figure 6b, Mediterranean species, negative effect, Linear regression, N=19, p = 0.042, R2 = 0.19; desert-oriented species, positive effect, Linear regression, N=19, p = 0.011, R2 = 0.29). In Dvir, at low grazing levels the most common species were Mediterranean species (e.g. Trachylepis vittata). The proportion of Mediterranean species from the entire species pool decreased in correlation with an increase in the grazing level. Correspondingly, desert-oriented species (e.g. Stenodactylus sthenodactylus) were relatively uncommon at low grazing levels. The

55 proportion of desert-oriented species from the entire species pool increased with grazing level. Data for Rahat are lacking for this comparison.

(a) (b)

Figure 6. Effect of grazing (i.e. feces density) on relative proportion of Mediterranean and southern species (a) and Mediterranean and southern individuals (b) in Dvir.

Physical condition can indicate the extent of habitat matching to focal species. I found a significant effect (one-way ANOVA F(2,159) = 15.1, P<0.05) of the climatic gradient on body condition (IC) of T. vittata, with the highest values found in Galon and Lachish (there was no significant difference between them) and the lowest value found in Dvir (Rahat did not provide enough data). In Dvir (Figure 8a) I also found a significantly negative effect of grazing on T. vittata’s IC values (Linear regression, N=45, p = 0.00001, R2 = 0.19). This effect was not seen in Galon and Lachish (P>>0.005 for both). In contrast, a positive effect (Figure 8b) of grazing on S. sthenodactylus’ IC values in

Dvir was found (one way ANOVA, F(2, 24) = 20.073, P<0.05).

Figure 7. A significant effect of grazing intensity on reptile physical condition (CI) in Dvir. Trachylepis vittata (a), a Mediterranean species, showed a decline in physical state at high grazing levels. In contrast, Stenodactylus sthenodactylus (b), an arid species, showed an improvement in physical state at high grazing levels.

Discussion At the onset of this chapter I raised three questions: (1) What is the effect of domestic grazing and climatic gradient on reptile abundance, species richness and species diversity? (2) How do climatic gradient and grazing together affect the structure of species from different biogeographic origins? (3) What is the impact of grazing and climatic gradient on reptiles’ physical state?

With respect to the first question and as predicted, I found different effects of grazing on reptile community measures according to patch location along the climatic gradient. For example, in the two northern land-units, a significant positive effect of grazing was found on reptile species diversity whereas in the southern land-unit a negative effect of grazing on reptile species diversity was found. In the northern land-unit, a significant positive effect of grazing was found on reptile species richness whereas in the middle land-unit (Lachish) no significant result was found and in the southern land-unit (Dvir) a significant negative result was found. Similar results were obtained for reptile abundance.

With respect to the second question, I found that Mediterranean-oriented reptile species are extremely abundant in the northern land-unit and become less common as we move southward with the climatic gradient. In the southern land-unit many desert-oriented species were found which were completely absent from the northern units. Consequently, I found a low similarity between the community structure in the southern and northern parts of the research area. I also found different effects of grazing on reptile community structure according to the location along the climatic gradient. In the northern land-unit no effect of grazing on reptile community composition was found. However, in the other land-units more desert-oriented species were found with increasing grazing pressure.

With regard to the third question, a negative affect of the north-south climatic gradient was found on the physical state of the Mediterranean skink Trachylepis vittata. Grazing was also found to exert different effects on the physical state of T. vittata. In the two northern land-units no effect of grazing on T. vittata physical state were found. However, in the southern land-unit there was a negative association between grazing pressure and T. vittata physical state.

All the above results suggest that the north-south climatic gradient and domestic animal grazing have a combined effect on many aspects of the reptile community. As expected (see Olff and Ritchie 1998), a negative association was found between grazing and percentage plant cover. However, the effects of a decreased plant cover on the reptile community measures were different according to the location along the climatic gradient.

In Galon, which enjoys a Mediterranean climate, grazing decreased the percentage of plant cover but increased plot-scale heterogeneity. Effects of grazing on species diversity are often attributed to the Intermediate Disturbance Hypothesis (Huston 1979), which predicts that under conditions of moderate interference, high biodiversity will be found. This may explain the positive effect found between grazing and reptile species richness and diversity in this land-unit and in Lachish. This hypothesis can also be the explanation for similar results that reported in the literature.

In contrast, under the semi-arid conditions at Dvir, a negative association was found between grazing and percentage plant cover leading to a decline in plot-scale heterogeneity (i.e. a negative effect of grazing on plot-scale heterogeneity). The niche theory (Hutchinson 1959, Vandermeer 1972) predicts a positive association between variety of available niches and species diversity. Since one of the consequences of a reduction in heterogeneity is a decrease in available niches, this may explain the negative association found between grazing and species richness and diversity in Dvir (Soininen et al. 2011).

In Dvir, due to grazing and reduction in percentage of plant cover, large areas are exposed to erosion processes. I suggest that high grazing levels in Dvir caused soil and habitat degradation, with ground becoming more exposed at lower plant cover. These processes of soil and habitat degradation (which could lead to desertification) may explain the change in reptile community structure (e.g. decreasing frequency of Mediterranean species and increasing frequency of desert-oriented species under high grazing pressures). Under high grazing levels the habitat becomes hostile to Mediterranean species and suitable for desert species, as reflected in the results.

The negative association I found between grazing and the physical state of T. vittata (Mediterranean species), and the positive association I found between grazing and the physical state S. sthenodactylus (desert gecko) can be also explained in the light of soil and habitat degradation due to heavy grazing in a semi-arid climate in Dvir. Those results support my prediction that under strong grazing pressure the semi-arid area of Dvir becomes hostile to Mediterranean species and more suitable for desert species.

The plant community in Dvir consists primarily of annual vegetation with natural adjustments to grazing since grazing of large domestic herbivores has been common in this area for thousands of years (Osem et al. 2002). Despite this, various studies that have been conducted in this area found different effects of grazing on the plant community, including annual plant density, seed-bank density and species richness (Osem et al. 2002, 2006). The plant community did not show a sharp or clear change indicative of the beginning of desertification processes in the research area. In contrast, the reptile community did show a directional change in community composition from a

Mediterranean community to a more desert-oriented community associated with grazing intensity. This suggests that reptile communities can be used as a highly sensitive indicator for the effects of grazing on the overall biological community in semi-arid and desert borderland environments.

My findings are important in the light of climate change and its anticipated impact on the SJL area. In accordance with the conventional scenarios, climate change will cause expansion of the desert further north to the areas now experiencing a Mediterranean climate (Gabbay 2000). According to this scenario many desert-oriented species will extend their range to the north, forcing the area of overlap between Mediterranean and arid species to shift north accordingly. In addition, and dependent upon climate change, it is possible that areas in which grazing at present increases heterogeneity and consequently reptile species richness and diversity, will shift to become semi-desert areas in which, as I have shown for Dvir, grazing reduces the landscape heterogeneity and consequently reptile species richness and diversity.

These possible changes in reptile community structure could have a wider influence on the entire ecosystem. For example, as I have found, reptile abundance within the semi- arid zone is relatively low. If this pattern shifts north as a result of desertification, other species (e.g. Circaetus spp.), which currently live in a Mediterranean climate and rely on reptiles as a source of food, may suffer from lack of resources.

I found that reptile community structure is affected by climate conditions, grazing of domestic animals and the interactions between them. Understanding these effects is important because grazing is one of the most common agricultural activities in many parts of the globe. Moreover, various studies indicate different impacts of grazing on a reptile community. As far as I know, this study is the first to offer a mechanism that could explain this difference.

Appendices Correlation between feces density and community measures based on 2010 data.

Trend Galon, total observation of 203 individuals

0 Reptile abundance; Linear regression, N=19, p = 0.3, R2 = 0.04

+ Species richness; Linear regression, N=19, p = 0.002, R2 = 0.39

2 + Fisher’s alpha index of diversity; Linear regression, N=16, p = 0.01, R = 0.34

Lachish, total observation of 115 individuals

0 Reptile abundance; Linear regression, N=19, p = 0.827, R2 = 0.002

+ Species richness; Linear regression, N=19, p = 0.09, R2 = 0.14

2 + Fisher’s alpha index of diversity; Linear regression, N=13, p = 0.015, R = 0.37

Dvir , total observation of 72 individuals

- Reptile abundance; Linear regression, N=19, p = 0.000006, R2 = 0.57

- Species richness; Linear regression, N=19, p = 0.0005, R2 = 0.47

2 - Fisher’s alpha index of diversity; Linear regression, N=11, p = 0.02, R = 0.39

Chapter 3: Wheat Fields as an Ecological Trap: Trachylepis vittata as a Case Study (This chapter is already published in Biological Conservation with the authorship of Guy Rotem, Yaron Ziv, Itamar Giladi and Amos Bouskila. The chapter also includes the data required by one of the reviewers)

Introduction A rapidly growing global human population coupled with an increase in per-capita consumption challenge modern agriculture to increase productivity in order to meet the increasing demand. This challenge is being tackled by both an expansion of farming area and an intensification of agricultural practices. The vast terrestrial areas affected by agriculture (about 80% globally; MEA 2005), agricultural intensification, and the cultivation of monocultures are all expected to cause biodiversity loss (FAO 2007; Green et al. 2005). One recent approach to alleviate the negative effects of agriculture on biodiversity is ‘Wildlife Friendly Agriculture’, which apparently promotes a balance between food production and conservation by, among others, leaving natural habitat patches within a heterogeneous agricultural landscape (Green et al. 2005). Accordingly, preservation of natural or semi-natural patches within the agricultural matrix is considered an effective and relatively cheap way to preserve biodiversity (Aarssen ad Schamp 2002; Benton et al. 2003; Duelli and Obrist 2003). In addition to biodiversity conservation, this approach may be beneficial also for farmers because of the positive ecosystem services that natural habitats provide for agriculture (Rosenzweig, 2003a, b, Tscharntke et al. 2005, Bommarco et al. 2013).

However, the proximity of natural habitat patches to agricultural matrix may also affect animal behavior, in general, and habitat selection, in particular (Tscharntke et al. 2012). The selection of habitats in which to shelter, feed and reproduce can dramatically impact organism fitness. Consequently, most animals have evolved abilities to sense reliable cues regarding habitat quality and to move to a better habitat whenever possible (Abramsky et al. 1985; Pulliam 1988).

However, the ability to reliably assess habitat quality is often compromised in human- made environments (Kristan 2003; Battin 2004). Cultivation-related fluctuation in habitat quality may attract individuals at certain times and be detrimental at other times

(Best 1986; Bollinger et al. 1990). The case where an organism prefers low-quality habitats over other available better habitats is called an ‘ecological trap’ (Dwernych and Boag 1972; Donovan and Thompson 2001; Hawlena et al. 2010), which might be considered a special case of source-sink dynamics (Pulliam 1988; Battin 2004). Such ecological traps may have far-reaching consequences for the populations in both the low and the high quality habitats. Robertson and Hutto (2006) offer three criteria that define an ‘ecological trap’:” 1) individuals should have exhibited a preference for one habitat over another; 2) a reasonable surrogate measure of individual fitness should have differed among habitats; and 3) the fitness outcome for individuals settling in the preferred habitat must have been lower than the fitness attained in other available habitat”.

Our study area, the Beit-Nir agroecosystem, is located at the northern part of Southern Judea Lowlands (SJL), central Israel (31°30’52’’N 34°52’36’’E), approximately 50 km southwest of Jerusalem (Figure 1a). Thousands of years of human inhabitance (Ben- Yosef 1980) and recent intensive agricultural practice formed a landscape consisting of natural habitat patches at different degrees of isolation, surrounded by agricultural fields (mainly wheat) vineyards and olive groves. The presence of semi-natural patches within this agricultural landscape can potentially host a high diversity of reptiles. However, these patches are positioned within wheat fields, a habitat with potentially highly fluctuating quality due to seasonal cultivation.

Using reptiles, we examine the main hypothesis that the agricultural system serves as an ecological trap, as defined by Robertson and Hutto (2006), where many individuals move to and permanently occupy the agricultural fields, eliminated by the agricultural machinery before or during the reproduction season. We contrast this hypothesis with an alternative one, stating that the agricultural system is used for a daily foraging ground by individuals reptiles that mainly occupy the adjacent natural habitats.

Our model species Trachylepis vittata [Scincidae] is common along the eastern Mediterranean basin and in North Africa (Van der Winden et al. 1995). It is frequently found under stones in the early morning until the ambient temperature rises above 14° C. This species also uses rocks as shelters to escape rain and other extreme weather (Clark

63 and Clark 1973). It measures 225 mm from snout to tail and feeds on arthropods (Schleich et al. 1996). Females give birth to live offspring between July and August (Disi et al. 2001, P 226).

Methods Study Design and Survey Protocol

We surveyed reptiles in thirteen sampling sites each including a natural patch, an adjacent wheat field and the patch-field edge (Figure 1b). At each site we installed 40 traps, ordered in two trapping arrays, each comprised of 20 one-liter dry pitfall traps arranged in two parallel lines that were marked at distances of 10 m and ~15 m on either side of the patch-field edge (Figure 1b). Additionally, on the patch-field edge we used a polypropylene multiwall sheet to build a 100 m-long and 40 cm-high fence (Figure 1b) that directed movement of all reptiles between the natural patch and the agricultural field to passageways located every 20 meters along the fence (Fisher et al. 2008). At those passageways we placed two one-liter dry pitfall traps, one at each side (total of 10 traps along each fence). These sampling methods enabled us simultaneously to assess the community structure and monitor the physical condition of reptiles in the natural patch, in the field, while crossing from the natural patch to the field and while crossing in the opposite direction (Jenkins and McGarigal 2003).

Figure 1 :(a) Map of the research area and the study site. White polygons represent natural and semi-natural patches surrounded by agricultural fields, mainly wheat. (b) A diagram of a trapping array showing the fence (black line) and the traps in each habitat and along the separating fence. (c) A picture of the patch-field edge and the separating fence.

We trapped reptiles during six trapping sessions throughout the spring (March to June) - four times before the wheat harvest, immediately after the harvest and one week later. In each session, the traps were left open for 72 hours. Trapped animals were measured (i.e. weight, snout-to-vent-length, tail length) and identified to species (and sex when possible). Individuals’ physical condition was assessed by an index of body condition (IC; Andrews and Wright 1994). Marking of individuals during the four first sampling sessions resulted in no recapture at all and therefore this method was not used further on. We released all captured individuals back to the habitat where they were captured (in the natural patch or agricultural field) or to the habitat they were aiming for (in the patch- field edge). We averaged all the observations from each combination of ‘habitat ×

65 session × site’ prior to any statistical analysis and used these summarized data as our replicates, thus avoiding any pseudo-replication.

Incidentally, the pitfall traps also collected arthropods that were later identified in the laboratory to family level. Previous studies have found a positive correlation between insect abundance and reptile abundance (Rocha et al. 2008). As all the studied reptile species were predators, with insects forming a dominant component of their diet, we assumed that arthropod abundance could serve as a good indicator for habitat quality.

Results Throughout the study, we trapped 352 reptiles from 9 species. Most of the trapped individuals (271) belonged to one species, Trachylepis vittata. Thus, we focus our analysis of population density and body condition on this particular species. Of the 271 individuals of T. vittata, 244 were identified as adults throughout the season and in all habitat types, 16 were identified as sub-adults (mainly in the pre-harvest sessions only, but in all habitat types) and only 11 were juveniles, all of which were captured in the natural patch habitat in the post-harvest session. Although it was sometimes possible to determine the sex of trapped individuals, in most cases it could not reliably be done. Therefore, our analysis was not stratified by sex or age.

We found a significant effect of both sampling time and habitat (repeated-measures

ANOVA, F(5, 240) = 10.43, p < 0.001, and F(3, 48) = 72.46, p < 0.001, respectively) as well as their interaction (F(15,240) = 9.0643, p < 0.0001) on abundance of T. vittata(Figure 2).

Trachylepis vittata abundance (Figure 2) in natural patches remained relatively constant throughout the entire study period. In contrast, the number of T. vittata individuals found in the wheat field varied. Early in the season only a few individuals occurred within the field habitat, but their number increased throughout the spring until the harvest. After the wheat harvest, not a single individual was found within the field habitat. The reptiles’ movement across habitats was unidirectional with intensive movement from the natural patches into the wheat fields in early spring (38 individuals observed). Only two individuals attempted crossing in the opposite direction throughout the entire season. The

66 very low densities of other reptile species precluded us from conducting meaningful analyses at the species level.

Nevertheless, the general pattern for all the rest of the reptile community combined was similar to that found for T. vittata. The number of reptiles (excluding T. vittata) captured per trapping array per session remained constant in the natural patch habitat throughout the season (0.69 and 0.77 for pre-harvest and post-harvest, respectively). It dropped sharply in the field habitat (from 0.25 individuals in the pre-harvest to 0 in the post- harvest). Prior to the harvest, twice as many individuals crossed from the patch to the field than in the opposite direction (0.15 and 0.08 individuals per trapping array per session, respectively) and no movement was observed in either direction after the harvest.

Figure 2: Mean number of Trachylepis vittata individuals per trapping array.

Habitat type significantly affected body condition of T. vittata (Figure 3; one-way

ANOVA, F(2, 61) = 33.7, p < 0.05) and we found that individuals in the natural patches were in poorer physical condition than those captured in the field or those crossing from the natural habitat to the agricultural field (p = 0.0001, Tukey’s Honestly Significant Difference post hoc test).

All individuals in the field and those crossing from the natural patch to the field were adults, whereas juveniles and newborns were found in the natural patches only.

Figure 3: Index of body condition (IC index) of individuals in fields, natural patches and along the fence separating patch and field. The analysis is based on adults trapped in early spring with intact tails.

Arthropod abundance within the wheat fields was significant higher in early spring (see complete family list and abundance in Table 1) compared to that in the natural patches (in early spring: T-test, t(1, 17) = 3.791, p = 0.001,SD = 77.173; arthropod abundance within the wheat fields was not significantly higher after the harvest compared to that in the natural patches: T-test , T(1, 17) = 1.912, p = 0.07, SD=72.115).

Table 1. List of arthropod families and their abundance within the wheat fields and the natural patches before the harvest.

Family Field Patch Family Field Patch

Anobiidae 1 0 Myrmeleontidae 3 2

Armadillidiidae 86 0 Phalangiidae 14 40

Carabidae 429 154 Polyphagidae 3 2

Chrysomelidae 8 0 Porcellionidae 179 123

Coccinellidae 1 3 Salticidae 1 0

Collembola 1 0 Scarabaeidae 316 26

Curculionidae 0 1 Scolopendra 5 6

Dermestidae 2 1 Scorpionidea 0 1

Elateridae 77 14 Scutigeridae 1 1

Gryllidae 1 0 Silphidae 0 5

Histeridae 0 5 Spirostreptidae 0 11

Japigidae 1 0 Staphylinidae 18 24

Lasiocampidae 0 15 Tenebreonidae 195 121

Lithobidae 11 17 Tenebrionidae 64 79

Lycosidae 1 6 Thysanura 9 37

Meloidae 1 3 unknown 85 106

Total 1513 803

Discussion Robertson and Hutto’s (2006) criteria for the existence of an ecological trap include preference for one habitat over another and lower fitness (measured directly or using a surrogate) in the preferred relative to the other habitat (see Introduction). The relatively

70 higher insect abundance in the field in early spring may explain the extreme asymmetrical movement of individuals of the insectivorous skink, T. vittata, from the natural patches to the field. Furthermore, the superior body condition of adults crossing to, or already in the field, relative to those remaining in the natural patches, clearly indicates that the movement to the habitat with high food abundance is mainly executed by the individuals in a better condition, with high potential for reproduction. Meylan et al. (2002) showed that when movement between patches (i.e., dispersal) incurs a high energetic cost, only individuals of better body condition make an attempt for such movement. Whether or not this mechanism drove our results, the superior body conditioned individuals clearly showed preference to the agricultural field,

Combined our observations suggest that in early spring, individuals behave according to the expectation of ideal density-dependent habitat selection (Fretwell and Lucas 1969), i.e., moving to a higher quality habitat to increase fitness. However, this seemingly optimal habitat selection eventually led individuals to be trapped in a very poor habitat following the wheat harvest. We have not found even a single live T. vittata in the field following the harvest activity. Furthermore, we have not found evidence of movement of any individual from the field into the patch during or after the harvest, nor immediately prior to the harvest, despite a buildup of substantial population in the field at this time (Fig. 2). Some of the reptiles in the field were presumably killed by the agricultural machinery; others, exposed to predators like Corvus monedula , Falco tinnunculus or Circaetus gallicus that accompanied the harvest activity, were likely consumed. We indeed counted more than 20 individuals of each of these predatory birds following the harvester, apparently collecting prey uncovered by the harvester (personal observations). This phenomenal scene is typical to harvesting activity throughout SJL and Israel, and, as far as we know, also throughout the world.

Much of the results that we obtained, at least early in the season, could have been generated by a daily movement of individuals that reside and reproduce in the natural patch and conduct daily foraging forays into the fields where they benefit from the high arthropod abundance. If that was the case, we would expect that at least some individuals that arrived during a trapping session will be captured on their return to the natural patch.

The almost complete lack of such movement, especially in the last session prior to the harvest, when substantial population was found in the field, clearly rejects the daily pattern hypothesis. Furthermore, throughout the spring, the T. vittata population within the agricultural field grew, which could indicate that individuals that moved from the natural patches remained in the field and have not used it just for diurnal foraging. Finally, if daily foraging individuals benefited from high quality food in the fields, one could expect a negative correlation between the physical state of individuals in natural patches and the distance from the high quality field habitat. Using auxiliary data, where traps were located in different distances from the patch-field edge , we found no such 2 correlation (Linear regression, F 1,34 = 1.10, p = 0.30, R = 0.03).

Clearly, our results and observations indicate a large difference between the fitness provided by the two habitats – while reproduction of T. vittata occurs in the natural patch (as evident by the observation of a few newborns late in the season after the harvest), the fitness in the fields equals zero (not even a single live T. vittata, adult or juvenile, in the field following the harvest activity). Following the criteria set by Robertson and Hutto (2006), we affirm that the agricultural fields serve as an ecological trap for T. vittata -- better-conditioned individuals show preference for the field, as indicated by their directionality of movement; the field offers higher resource quantity indicating a potential difference in prospective fitness; and the preferred habitat has eventually a lower fitness. Although the data enabled us to conduct detailed analysis for only one common species, the similar patterns observed for the rest of the community suggest that the implications of the results may be pertinent for many species, including rare species for which data is always hard to obtain.

Organisms make decisions regarding their future success based on currently available information. Most of these decisions are based on the long process of evolutionary promotion of optimal habitat selection (Schlaepfer et al. 2002). However, adaptations for optimal habitat selection that have been shaped by long-term evolutionary processes may be out of context in cases of anthropogenic intervention with the natural environment (Hawlena et al. 2010). Such intervention, which is usually much faster than almost any evolutionary process, leads to situations in which organisms select habitats according to

72 their “evolutionary knowledge” (Battin 2004), leading, on occasions, to ecological traps (e.g., Hawlena et al. 2010). In our case, the wheat field serves as an ecological trap by attracting individuals of better physical condition in the population to migrate to the seemingly better habitat. These individuals, of high prospective fitness, find themselves in a very poor habitat after the harvest, leading to no fitness at all.

The passage of individuals in better physical condition from the natural patches into the wheat fields, where their fitness is very low, may further decrease both population size and the quality of the natural patches’ populations (Schlaepfer et al. 2002). Small fragmented populations are exposed to inbreeding depression and genetic drift, which further decrease the population’s genetic diversity and weaken its ability to cope with both short-term stochasticity (e.g., drought period) and long-term environmental change (Porlier et al. 2009). The effects of genetic isolation and the negative qualitative and quantitative effects of ecological traps on isolated populations in natural patches may be additive, or even synergistic, increasing the probability of extinction for those populations. The asymmetric, almost unidirectional, movement across habitats and the functioning of the wheat fields as an ecological trap pose a risk, particularly to small patches that share a long border with the fields relative to their patch area. In such small patches, the loss of individuals that do not reproduce in the patch may be detrimental, due to the reduction in the number of individuals available to replace natural mortality in the patch and due to potentially gradual loss in the quality of the remaining individuals in the patch.

We believe that the phenomenon of an ecological trap in agroecosystems is not unique to our study area or to our study species, but may represent an example of a broad phenomenon, probably found in agricultural areas in many places worldwide (see Bollinger et al. 1990, Shochat et al. 2005). Consequently, we think that possible risks of ecological traps should be incorporated in the ‘Wildlife Friendly Agriculture’ approach (see Introduction) that is currently proposed to promote conservation.

General Discussion This thesis consists of three complementary parts, each addressing a different aspect of the effects of scale-dependent variables and agricultural activity on the reptile community within a fragmented agro-ecosystem. The spatial scale over which I have been exploring these effects encompassed local, patch, land-unit and landscape scales. These scales enabled me to examine how is the reptile community affected by a long list of variables influencing several biological or ecological mechanisms such as competition, migration, and biogeographical orientation. I also examined prevailing agricultural activities common throughout the research area – domestic animal grazing and wheat cultivation – and their effect on the reptile community.

This thesis brings together scientific tools and knowledge of the spatial ecology approach and ideas and principles of nature conservation in agro-ecosystem landscape. Ecology has a long history of interest in the spatial patterning and geographic distribution of organisms which are the basis for the modern spatial ecology. The niche theory (Hutchinson 1959) suggested that local-scale ecological processes shape the community structure. More recent theories, such as Island Biogeography (MacArthur and Wilson 1967), Metapopulation Dynamics (Levin 1970), Metacommunity Dynamics (Leibold et al. 2004), suggest that the diversity, composition and structure of a local community is the result of ecological processes operating at different spatial scales. Those theoretical developments and advances in computational capabilities have led to the contemporary spatial ecology approach that reviews any community as a product of many scale- dependent ecological processes.

Not only the field of spatial ecology has undergone significant developments during the last decades; the field of nature conservation has rapidly developed as well. A few years ago it was common to assume that nature conservation and biodiversity protection should be done only in protected areas such as a nature reserves, national parks, forests and others. In recent years the understanding that protected areas are not sufficient for the protection of biodiversity became common. As a result, a new approach gives higher importance to open unprotected areas for the maintenance of biodiversity (Rosenzweig 2003b, Rosenzweig 2003a). Based on this new approach, agricultural fields that

74 constitute a significant percentage of Earth are important for the maintenance of biodiversity and nature conservation.

The global increase in food consumption and accompanying increase in land used for agriculture (Robinson and Sutherland 2002, MEA 2005, FAO 2007) emphasize the importance of understanding the ecological processes affecting biodiversity in intensive modern agricultural systems. Two main approaches are now widely accepted by the scientific community as tools for maintaining biodiversity in agricultural systems (Green et al. 2005). The first one, ‘land sparing’, argues that biodiversity cannot be protected within an agricultural system. Based on this approach, biodiversity protection should be carried out in protected areas separated from agricultural fields (Balmford et al. 2005, Fischer et al. 2008). The second approach, ‘land sharing’, which is also known as ‘wildlife friendly agriculture’, suggests that biodiversity protection can be maintained in natural or semi-natural patches nested within the agricultural system (Green et al. 2005, Mattison and Norris 2005, Phalan et al. 2011). In order for this approach to be applied, a great deal of information is necessary. This includes, among other things, information about the organisms within the relevant area, the patch size required, patches' spatial configuration and connectivity between patches. In this thesis, I measured and gathered information on various landscape-oriented and scale-dependent variables (factors) in an agro-ecosystem over three land units, enabling me to evaluate the suitability of the current agricultural practices as ‘wildlife friendly agriculture’.

In the first chapter, I explored the effect of different spatial scale-dependent variables on reptile community measures -- abundance, species richness and species diversity -- in natural patches using the AICc-based model selection approach (Anderson 2008). Consistently, models that included several variables from different scales attained lower AICc values than models that included several variables from the same scale. This means that the structure of the reptile community in natural patches is affected by different ecological processes related to different spatial scales (Michael et al. 2008). Among others, I found that the location along the climatic gradient and the domestic grazing activity are both important variables that strongly affect the community measures.

With relation to the importance of natural patches in the wildlife friendly agricultural approach, the results of this study indicate that reptile species richness within the natural patches remains high in relation to the entire potential species pool. Thirty-four reptile species were recorded from the regional inventory of about 40 species (Bouskila and Amitai 2001, Bouskila 2002), and out of a total of 87 native terrestrial reptile species in Israel (Bouskila and Amitai 2001). This information reinforces the approach that sees maintaining natural patches within the agricultural system as an effective tool in maintaining biodiversity (Balmford et al. 2005, Green et al. 2005, Fischer et al. 2008, Phalan et al. 2011).

In the second chapter, I explored the combined effect of domestic animal grazing and climatic gradient on the reptile communities residing in the natural patches of this heterogeneous landscape. The relatively high reptile species richness, the varied biogeographical orientation of those species and the negative correlation along the gradient between different land-units and reptile community similarity suggest that the SJL area serves as an ecotone along the changing environments, with Dvir located at its center.

The effect of grazing on the reptile community was correlated to the location along the north-south climatic gradient. In the northern land-unit – Galon – which enjoys a Mediterranean climate, grazing increases local plot heterogeneity and, consequently, increases reptile species richness. However, the southern land-unit – Dvir – showed an opposite trend. In Dvir, which has a semi-arid climate, grazing had a negative effect on local (within-plot) heterogeneity and a negative effect on species richness. Moreover, in the semi-arid conditions of Dvir I found an association between grazing and arid-oriented reptile species, suggesting that this area suffers from over-grazing and desertification (Van Auken 2000). This suggestion is supported by the fact that the body condition of the Mediterranean species (T. vittata) decreased under strong grazing pressure at Dvir, whereas the body condition of the arid-oriented species (Stenodactylus sthenodactylus) increased under the same grazing pressure at the same location.

The first and the second chapters provide evidence for a clear effect of domestic animal grazing on reptile community structure and community measures, such as abundance,

76 species richness and species diversity. Since grazing of domestic animals is a common agricultural activity (26 percent of the ice-free terrestrial surface of the Earth is used for grazing; FAO 2006), it is likely that many reptile communities are under pressure from the effects of grazing.

One of the preconditions for a long-term persistence of a community within a given patch is the ability to migrate (disperse) between patches (e.g., Hanski 1999). In the SJL, grazing is not limited only to the natural patches but is also permitted on the stubble fields. Since grazing removes the vegetation soil cover and simplifies the vegetation structure, it may increase the matrix hostility and decrease the predation risk for reptiles. This may reduce a reptile's chances of migrating successfully between patches and indirectly, this process can affect the within-patch reptile population. In the current study I have not check this issue. Therefore, I suggest that a following research should focus on the effect of grazing on dispersal probabilities, through changing the surrounding habitat’s physical conditions, particularly in relation to the wildlife friendly agriculture approach.

In the third chapter, I explored a particular effect of wheat fields on reptile communities in natural patches. I found that wheat fields are ecological traps for T. vittata. It is likely that this phenomenon is not limited only to T. vittata, but may be quite common within this agricultural area, affecting a number of other species. Since wheat is a major agricultural activity in the SJL and is the most widely grown cereal crop in the world, with an ever-increasing demand (FAO 2002), this phenomenon may have wide implications on many reptile communities worldwide.

Contrary to my expectations and with respect to the importance of connectivity between patches, I found in the first chapter the variable that describes patch isolation (proximity value) was found to be of a relatively little importance to reptile species diversity. In the third chapter I found an asymmetrical reptile movement between the natural patches and agricultural fields, with massive movement from the patch to the field and negligible movement in the other direction. Based on this, I conclude that wheat fields are an almost insurmountable barrier between patches. Hence, there were no differences in the

77 community structure under different isolation values, because even a small wheat field is impassable for reptiles.

Small and isolated populations are expected to suffer from inbreeding depression and other negative density-dependent processes, which can lead to local extinctions (Whitlock 2004). However, as already mentioned, I found a diverse community at each patch I sampled. A possible explanation for this gap between the expectation according to this theory and the result I obtained may be the spatio-temporal heterogeneity which exists within the agricultural system of the Southern Judea Lowlands and its origin in agricultural crop rotation

Agricultural crop rotation, which includes the planting of nitrogen-fixing plants in agroecosystems, has been shown to sustain ecosystem functions and land productivity (Simms and Taylor 2002, Sileshi et al. 2008). The agricultural crop rotation practiced in the research area includes annual alternation between wheat and legumes. Wheat is grown every year on about 60% of the area and legumes (meanly peas, alfalfa and hummus) on about 40% of it. In the following year, the entire area utilized for growing legumes is used for growing wheat, about half that utilized for growing wheat is used for growing legumes, and half of it is used for growing wheat for a second year in a row. Areas where legumes are grown enjoy a different agricultural schedule to the wheat. Alfalfa is harvested early (March to April), while it is still green. It then stays on the ground for a few weeks until it dries up (Y. Klein, director of field crops, personal communication). So, it is reasonable that reptiles can move through alfalfa fields after the harvest (which is a period of peak activity), thus allowing the passage of individuals between patches. Since there is no need for massive movement each year to maintain genetic diversity (and thus maintain healthy populations to prevent inbreeding depression and reduce the probability of local extinction; Hedrick and Gilpin 1997), it might be one of the mechanisms that allows the continued existence of reptile populations within the natural relict patches.

If indeed movement between patches is a mandatory mechanism for maintaining diversity within the landscape, then conserving natural patches is not enough to maintain biodiversity. It is also necessary to maintain crossing options for individuals between

78 patches. Switching between different crops (i.e., crop rotation) with different agricultural management needs can be a good solution that meets both the needs of farmers and the biological, ecological and conservation needs of the ecosystem as well.

In general, the results of this thesis suggest that the reptile community inhabiting the fragmented agro-ecosystem studied is strongly affected by many scale-dependent ecological processes, as well as by multi-scale processes. The results also provide evidence that the reptile communities inhabiting agricultural areas are affected by agricultural crop choices and practices. Although there are still many gaps of knowledge in our understadning of biodiversirtty conservation in agricultural matrix, my results support the approach that maintaining natural patches within an agricultural area can support species diversity. Based on the above findings I suggest that preserving natural patches, while taking into account various spatial scale-dependent ecological processes on the one hand and the effect of the agricultural matrix itself on the other, can be an efficient tool to protect species diversity and stabilize ecological systems within an agricultural area.

References

Abaturov, B. D., I. A. Dmitriev, L. Jargalsaikhan, and K. Z. Omarov. 2008. Phytomass utilization and deposition of feces by ungulates on steppe pastures of eastern Mongolia. Biology Bulletin 35:302-309. Adler, P. B., D. A. Raff, and W. K. Lauenroth. 2001. The effect of grazing on the spatial heterogeneity of vegetation. Oecologia 128:465-479. Alon, A. and Y. Waisel. 1984. Israel Plants.in A. Alon, editor. Plants and Animals of the Land of Israel: An Illustrated Encyclopedia. Ministry of Defence and The Society for the Protection of Nature in Israel (In Hebrew). Anderson, D. R. 2008. Model Based Inference in the Life Sciences: a Primer on Evidence. Springer, New York, NY. Andrews, R. M. and S. J. Wright. 1994. Long-Term Population Fluctuations of a Tropical Lizard: A Test of Causality.in J. L. Vitt and R. E. Pianka, editors. Lizard Ecology . Historical and Experimental Perspectives. Princeton University Press, Princeton, New Jersey. Attum, O., P. Eason, G. Cobbs, and S. M. B. El Din. 2006. Response of a desert lizard community to habitat degradation: Do ideas about habitat specialists/generalists hold? Biological Conservation 133:52-62. Balmford, A., R. E. Green, and J. P. W. Scharlemann. 2005. Sparing land for nature: exploring the potential impact of changes in agricultural yield on the area needed for crop production. Global Change Biology 11:1594-1605. Bauer, A. M. 2003. On the identity of Lacerta punctata Linnaeus 1758, the type species of the genus Euprepis Wagler 1830, and the generic assignment of Afro-Malagasy skinks. African Journal of Herpetology 52:1-7. Ben-Yosef, S. 1980. Israel Guide (vol. Judea). Keter Publishing House, Jerusalem. (In Hebrew). Benton, T. G., J. A. Vickery, and J. D. Wilson. 2003. Farmland biodiversity: is habitat heterogeneity the key? Trends in Ecology & Evolution 18:182-188.

Best, L. B. 1986. Conservation tillage: ecological traps for nesting birds? Wildlife Society Bulletin 14:308-317. Bino, G., A. Dolev, D. Yosha, A. Guter, R. King, D. Saltz, and S. Kark. 2010. Abrupt spatial and numerical responses of overabundant foxes to a reduction in anthropogenic resources. Journal of Applied Ecology 47:1262-1271. Bollinger, E.K., Bollinger, P.B., Gavin, T.A., 1990. Effects of hay-cropping on eastern populations of the Bobolink. Wildlife Society Bulletin 18, 142-150. Bouskila, A. 2002. Reptiles. Pages 71-121 in A. Dolev and A. Perevolotsky, editors. The Red Book: Vertebrates in Israel. Israel Nature and Parks Authority and The Society for the Protection of Nature in Israel. Bouskila, A. and P. Amitai. 2001. Handbook of Amphibians and Reptiles of Israel. Keter Publishing House, Jerusalem. Brown, G. W., J. W. Dorrough, and D. S. L. Ramsey. 2011. Landscape and local influences on patterns of reptile occurrence in grazed temperate woodlands of southern Australia. Landscape and Urban Planning 103:277-288. Castellano, M. J. and T. J. Valone. 2006. Effects of livestock removal and perennial grass recovery on the lizards of a desertified arid grassland. Journal of Arid Environments 66:87-95. Connor, E. F., A. C. Courtney, and J. M. Yoder. 2000. Individuals-area relationships: The relationship between animal population density and area. Ecology 81:734-748. Cornell, H. V. and J. H. Lawton. 1992. Species interactions, local and regional processes, and limits to the richness of ecological communities - a theoretical perspective. Journal of Animal Ecology 61:1-12. Cumming, G. S. 2007. Global biodiversity scenarios and landscape ecology. Landscape Ecology 22:671-685. Cunningham, R. B., D. B. Lindenmayer, M. Crane, D. Michael, and C. MacGregor. 2007. Reptile and arboreal marsupial response to replanted vegetation in agricultural landscapes. Ecological Applications 12:609-619. De Bello, F., J. Leps, and M. T. Sebastia. 2005. Predictive value of plant traits to grazing along a climatic gradient in the Mediterranean. Journal of Applied Ecology 42:824-833.

Diaz, S., Lavorel, S., McIntyre, S., Falczuk, V., Casanoves, F., Milchunas, D.G., Skarpe, C., Rusch, G., Sternberg, M., Noy-Meir, I., Landsberg, J., Zhang, W., Clark, H., Campbell, B.D., 2007. Plant trait responses to grazing - a global synthesis. Global Change Biology 13, 313-341. Dickinson, H. C. and J. E. Fa. 2000. Abundance, demographics and body condition of a translocated population of St Lucia whiptail lizards (Cnemidophorus vanzoi). Journal of Zoology 251:187-197. Dumbrell, A. J., E. J. Clark, G. A. Frost, T. E. Randell, J. W. Pitchford, and J. K. Hill. 2008. Changes in species diversity following habitat disturbance are dependent on spatial scale: theoretical and empirical evidence. Journal of Applied Ecology 45:1531-1539. Fabricius, C., M. Burger, and P. A. R. Hockey. 2003. Comparing biodiversity between protected areas and adjacent rangeland in xeric succulent thicket, South Africa: arthropods and reptiles. Journal of Applied Ecology 40:392-403. FAO. 2007. The state of food and agriculture paying farmers for environmental services. Food and Agriculture Organization of the United Nations. FAO Agriculture. Series No. 38. Farina, A. 2006. Principles and Methods in Landscape Ecology. Towards a Science of Landscape. Springer. Fischer, J., B. Brosi, G. C. Daily, P. R. Ehrlich, R. Goldman, J. Goldstein, D. B. Lindenmayer, A. D. Manning, H. A. Mooney, L. Pejchar, J. Ranganathan, and H. Tallis. 2008. Should agricultural policies encourage land sparing or wildlife- friendly farming? Frontiers in Ecology and the Environment 6:380-385. Fisher, R., D. Stokes, C. Rochester, C. Brehme, S. Hathaway, and T. Case. 2008. Herpetological Monitoring Using a Pitfall Trapping Design in Southern California. U.S. Geological Survey Fisher, R. A., A. S. Corbet, and C. B. Williams. 1943. The relation between the number of species and the number of Individuals in a random sample of an snimal population. The Journal of Animal Ecology 12:42-58. Fleischner, T. L. 1994. Ecological cost of livestick grazing in Western North-America. Conservation Biology 8:629-644.

Gavish, Y. 2011. Holy or not, Israel is a Palaearctic biodiversity hotspot. Israel Journal of Ecology & Evolution 57:207-211. Germano, D. J., G. B. Rathbun, and L. R. Saslaw. 2012. Effects of grazing and invasive grasses on desert vertebrates in California. Journal of Wildlife Management 76:670-682. Giladi, I., Y. Ziv, F. May, and F. Jeltsch. 2011. Scale-dependent determinants of plant species richness in a semi-arid fragmented agro-ecosystem. Journal of Vegetation Science 22:983-996. Gotelli, N. J. and G. R. Graves. 1996. Null Models In Ecology. Smithsonian Institution Press, Washington, D.C. Green, R. E., S. J. Cornell, J. P. W. Scharlemann, and A. Balmford. 2005. Farming and the fate of wild nature. Science 307:550-555. Gvirtzman, H. 2002. Israel Water Resources. . Yad Ben-Zvi, Jerusalem. (In Hebrew) Hamback, P. A. and G. Englund. 2005. Patch area, population density and the scaling of migration rates: the resource concentration hypothesis revisited. Ecology Letters 8:1057-1065. Hanski, I. and O. E. Gaggiotti. 2004. Ecology, Genetics and Evolution of Metapopulations. Elsevier Academic Press, San Diego, CA, USA. Hanski, I. and M. E. Gilpin, editors. 1997. Metapopulation Biology. Ecology, Genetics, and Evolution. Academic Press, San Diego, CA, USA. Hardin, G. 1960. The competitive exclusion principle. Science 131:1292-1297. Harrison, S. and H. Cornell. 2008. Toward a better understanding of the regional causes of local community richness. Ecology Letters 11:969-979. Hiltunen, T., J. Laakso, and V. Kaitala. 2006. Interactions between environmental variability and immigration rate control patterns of species diversity. Ecological Modelling 194:125-131. Holt, R. D. 1992. A neglected facet of island biogeography - the role of Internal spatial dynamics in area effects. Theoretical Population Biology 41:354-371. Huston, M. 1979. A general hypothesis of species diversity. American Naturalist 113:81- 101.

Hutchinson, G. E. 1959. Homage to Santa Rosalia or why are there so many kinds of animals? American Naturalist 93:145-159. James, C. D. and R. Shine. 2000. Why are there so many coexisting species of lizards in Australian deserts? Oecologia 125:127-141. Jiguet, F., L. Brotons, and V. Devictor. 2011. Community responses to extreme climatic conditions. Current Zoology 57:406-413. Kark, S., P. U. Alkon, U. N. Safriel, and E. Randi. 1999. Conservation priorities for chukar partridge in Israel based on genetic diversity across an ecological gradient. Conservation Biology 13:542-552. Kloner, A. and Y. Tepper. 1987. The Hidden Complexes in the Judean Shephelah. Hakibbutz Hameuchad Publishing House, Tel- Aviv. (In Hebrew.) Krewenka, K. M., A. Holzschuh, T. Tscharntke, and C. F. Dormann. 2011. Landscape elements as potential barriers and corridors for bees, wasps and parasitoids. Biological Conservation 144:1816-1825. Leibold, M. A., M. Holyoak, N. Mouquet, P. Amarasekare, J. M. Chase, M. F. Hoopes, R. D. Holt, J. B. Shurin, R. Law, D. Tilman, M. Loreau, and A. Gonzalez. 2004. The metacommunity concept: a framework for multi-scale community ecology. Ecology Letters 7:601-613. Levin, S. A. 1970. Community equilibria and stability, snd An extension of the competitive exclusion principle. American Naturalist 104:413-423. MacArthur, R. H. and E. H. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press. Mattison, E. H. A. and K. Norris. 2005. Bridging the gaps between agricultural policy, land-use and biodiversity. Trends in Ecology & Evolution 20:610-616. McGarigal, K. and S. A. Cushman. 2002. Comparative evaluation of experimental approaches to the study of habitat fragmentation effects. Ecological Applications 12:335-345. Nicholson, M. C., R. T. Bowyer, and J. G. Kie. 2006. Forage selection by mule deer: does niche breadth increase with population density? Journal of Zoology 269:39- 49.

Palmer, M. W. and P. M. Dixon. 1990. Small-scale environmental heterogeneity and the analysis of species distributions along gradients. Journal of Vegetation Science 1:57-65. Pakeman, R.J., 2004. Consistency of plant species and trait responses to grazing along a productivity gradient: a multi-site analysis. Journal of Ecology 92, 893-905. Perevolotsky, A. and N. G. Seligman. 1998. Role of grazing in Mediterranean rangeland ecosystems - Inversion of a paradigm. Bioscience 48:1007-1017. Phalan, B., A. Balmford, R. E. Green, and J. P. W. Scharlemann. 2011. Minimising the harm to biodiversity of producing more food globally. Food Policy 36:S62-S71. Robinson, R. A. and W. J. Sutherland. 2002. Post-war changes in arable farming and biodiversity in Great Britain. Journal of Applied Ecology 39:157-176. Rocha, C. F. D., H. G. Bergallo, C. F. Vera y Conde, E. B. Bittencourt, and H. d. C. Santos. 2008. Richness, abundance, and mass in snake assemblages from two Atlantic Rainforest sites (Ilha do Cardoso, Sao Paulo) with differences in environmental productivity. Biota Neotropica 8:117-122. Roll, U., L. Stone, and S. Meiri. 2009. Hot-spot facts and artifacts questioning Israel's great biodiversity. Israel Journal of Ecology & Evolution 55:263-279. Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge University Press, Cambridge, United Kingdom. Rosenzweig, M. L. 2003a. Reconciliation ecology and the future of species diversity. Oryx 37:194-205. Rosenzweig, M. L. 2003b. Win-Win Ecology. How the Earth's Species Can Survive in the Midst of Human Enterprise. Oxford University Press, Okford. Rotem, G., H. Berger, R. King, P. Bar, and D. Saltz. 2011. The effect of anthropogenic resources on the space-use patterns of Golden Jackals. Journal of Wildlife Management 75:132-136. Rowe, R. J. 2007. Legacies of land use and recent climatic change: The small mammal fauna in the mountains of Utah. American Naturalist 170:242-257. Shine, R. and T. Madsen. 1997. Prey abundance and predator reproduction: Rats and pythons on a tropical Australian floodplain. Ecology 78:1078-1086.

Smart, S. M., L. G. Firbank, R. G. H. Bunce, and J. W. Watkins. 2000. Quantifying changes in abundance of food plants for butterfly larvae and farmland birds. Journal of Applied Ecology 37:398-414. Sodhi, N. S. and P. R. Ehrlich. 2010. Conservation Biology for All. Oxford University Press, New York, USA. Soininen, J., J. Heino, J. Lappalainen, and R. Virtanen. 2011. Expanding the ecological niche approach: Relationships between variability in niche position and species richness. Ecological Complexity 8:130-137. Steinfeld, H., P. Gerber, T. Wassenaar, V. Castel, M. Rosales, and C. De Hann. 2006. Livestock’s Long Shadow. Environmental Issues and Options., Food and Agriculture Organization of the United Nations, Rome. Svoray, T., P. Bar, and T. Bannet. 2005. Urban land-use allocation in a Mediterranean ecotone: Habitat Heterogeneity Model incorporated in a GIS using a multi-criteria mechanism. Landscape and Urban Planning 72:337-351. Turner, G. M., H. R. Gardner, and R. V. O'nill. 2001. Landscape Ecology In Theory And Practice. Springer Press, USA. Urbina-Cardona, J. N., M. Olivares-Perez, and V. H. Reynoso. 2006. Herpetofauna diversity and microenvironment correlates across a pasture-edge-interior ecotone in tropical rainforest fragments in the Los Tuxtlas Biosphere Reserve of Veracruz, Mexico. Biological Conservation 132:61-75. Van Auken, O. W. 2000. Shrub invasions of North American semiarid grasslands. Annual Review of Ecology and Systematics 31:197-215. Vandermeer, H. J. 2011. the Ecology of Agroecosystems. Jones and Bartlett Publishers. Walker, B. H., D. Ludwig, C. S. Holling, and R. M. Peterman. 1981. Stability of semi- arid savanna grazing systems. Journal of Ecology 69:473-498. Weibull, A. C., O. Ostman, and A. Granqvist. 2003. Species richness in agroecosystems: the effect of landscape, habitat and farm management. Biodiversity and Conservation 12:1335-1355. Wiens, J. A. 1989. Spatial scaling in ecology. Functional Ecology 3:385-397. Wright, J. P., C. G. Jones, and A. S. Flecker. 2002. An ecosystem engineer, the beaver, increases species richness at the landscape scale. Oecologia 132:96-101.

Wu, J. G. and O. L. Loucks. 1995. From balance of nature to hierarchical patch dynamics: A paradigm shift in ecology. Quarterly Review of Biology 70:439-466. Ziv, Y. 1998. The effect of habitat heterogeneity on species diversity patterns: a community-level approach using an object-oriented landscape simulation model (SHALOM). Ecological Modelling 111:135-170.

תקציר

מאפייני חברה אקולוגית כמו: שפע פרטים, עושר מינים ומגוון מינים והתהליכים המשפיעים עליהם- מהווים חלק משמעותי מאבני היסוד של האקולוגיה. במשך שנים מקובל היה לבחון תהליכים אלו דרך גורמים )משתנים מסבירים( מקומיים המשפיעים על מערכת יחסי הגומלין בין פרטים או מינים בסקאלה המקומית. בעשורים האחרונים, הודות להתפתחות תיאורטית מדעית, חישובית וסטטיסטית התגבשה גישה הבוחנת את מאפייני ומבנה החברה באמצעות שימוש במגוון משתנים ותהליכים תלויי-סקאלה. אולם, למרות המחקר הרב שנעשה בתחום הרי שקבוצת הזוחלים הוזנחה באופן יחסי לקבוצות ביולוגיות אחרות והמחקרים הבוחנים השפעה מרובת סקאלות על טקסה זו מעטים.

במהלך השנים נעשו מחקרים לא מעטים על מבנה חברה בתוך מערכות חקלאיות תוך שימוש בגישות שונות, כולל שיטות מתקדמות של אקולוגיה מרחבית. אולם, לעיתים רחוקות שטחים חקלאים פרושים לאורך מפלים אקלימיים חדים, כך שהשפעה של מפל אקלימי חד על מבנה החברה האקולוגית בתוך השטח החקלאי הינה תחום בו חסר כיום ידע מחקרי. בנוסף, במקרים בהם ישנו מפל אקלימי לרוב הוא מלווה בשינוי בגובה הטופוגרפי או בהרכב הקרקע, בגיאולוגיה וכמובן בפעילות החקלאית עצמה. שטח המחקר שלי, שפלת יהודה הדרומית, מתאפיין במפל אקלימי חד ביותר על פני מרחק קצר וזאת ללא שינוי מהותי בגובה הטופוגרפי, בקרקע, במסלע, בהיסטוריה האנושית ובפעילות החקלאית. עבודת הדוקטורט שלי בחנה בסקאלות מרחביות שונות מהם המשתנים המשפיעים על חברת הזוחלים המתקיימת בכתמים טבעיים בתוך המערכת החקלאית המקוטעת של שפלת יהודה הדרומית.

בפרק הראשון בחנתי כיצד משתנים מרחביים מסקאלות שונות משפיעים על חברת הזוחלים המתקיימת בכתמים הטבעיים המצויים במערכת החקלאית. לשם כך בחרתי שלושה שטחים )מצפון לדרום – גלאון, לכיש ודביר( בגודל 4×2.3 קילומטר כל אחד. ארבעת השטחים הללו מיצגים את המפל האקלימי הפועל בסקאלת הנוף. בתוך שטחים אלו בחרתי כתמים השונים זה מזה בגודלם, בצורתם ובקונפיגורציה המרחבית שלהם. בתוך כתמים אלו סימנתי חלקות קבועות בגודל 011×01 מטרים בהם בצעתי את דיגום הזוחלים. בחירה זו של מערך דיגום אפשרה לי לבחון כיצד סדרה ארוכה של משתנים )דוגמת: הטרוגניות הכתם, גודל הכתם והקונפיגורציה המרחבית של הכתמים( הפועלים בסקאלות מרחביות שונות וקשורים לתהליכים ביולוגים / אקולוגים שונים משפיעים על מבנה חברת הזוחלים.

באמצעות שימוש בתהליך בחירת מודלים )Model selection method( המבוססים AICc בחנתי מהם המשתנים המשפיעים ביותר על מבנה החברה. המודלים שהציעו את ההסבר הטוב ביותר עבור שלושת המאפיינים של מבנה חברה - שפע פרטים עושר ומגוון מינים - היו כולם מרובי סקאלות.

אולם בחינה של מידת ההשפעה שיש למשתנים השונים על כל אחד משלושת מאפייניי החברה מצביעה על כך שכל אחד מהם מושפע בצורה חזקה ממשתנים מרחביים אחרים.

המודלים הציעו ששלות המרכיבים של חברת הזוחלים ) כלומר שפע, עושר ומגוון( מושפעים ממשתנים הפועלים בסקאלה המקומית כדוגמת: הטרוגניות החלקה, אבניות ועוד. עוד מציעים המודלים שחברת הזוחלים מושפעת מהמיקום לאורך הגרדינט האקלימי מרעיית צאן ובקר ומהאניטראקציה בין הגורמים הללו.

אחת הפעילויות החקלאיות הנפוצות ביותר בשטח המחקר הינה רעייה עונתית של צאן ובקר. הרעייה מתבצעת בשטחי השלף לאחר הקציר. במהלך תקופת הרעייה רועים העדרים גם בכתמים הטבעיים המצויים בלב השטח החקלאי. מחקרים קודמים כבר בדקו בעבר את השפעת הרעייה על חברת הזוחלים, אולם בפרק השני של העבודה, בחנתי את ההשפעה המשולבת של מפל אקלימי ורעייה על מבנה חברת הזוחלים. מעטים המחקרים הקודמים שבחנו השפעה משולבת כזו בתאי שטח דומים.

לצורך בחינת השפעת המפל האקלימי על הרכב החברה בחנתי את פרופורציית המינים בהתאם למוצא הביוגיאוגרפי שלהם. התוצאות העידו על ירידה באחוז המינים )והשפע היחסי( של מינים ים- תיכוניים עם הירידה בכמות המשקעים. במקביל לכך נמצאה עליה באחוז המינים )ובשפע היחסי( של מינים ממוצא מדברי.

גם השפעת הרעייה עצמה השתנתה בהתאם למיקום החלקות לאורך המפל האקלימי. בשטחים הסמוכים לגלאון, השייכים לאקלים ים תיכוני, מצאתי שהרעייה מעלה את הטרוגניות החלקות ואת עושר המינים. בניגוד לכך, בשטחים הדרומיים, בקרבת דביר, מצאתי השפעה שלילית של הרעייה על הטרוגניות החלקות ועושר המינים. בדביר מצאתי גם שהרעייה משפיעה על הרכב החברה. עושר המינים ממוצא ים תיכוני נמצא בדביר בהתאמה שלילית עם עוצמת הרעייה. בניגוד לכך עושר המינים ממוצא מדברי נמצא בדביר בהתאמה חיובית עם עוצמת הרעייה.

מלבד רעייה, שטח המחקר מתאפיין גם בשדות חיטה נרחבים. בפרק השלישי של העבודה בחנתי את השפעת שדות החיטה על חברת הזוחלים. לשם כך, בחרתי להתמקד בחלקות המצויות במרכז שדות החיטה ואשר היו מרוכזות כולן בשטח הצפוני, גלאון, זאת בכדי להימנע מההשפעה האקלימית. בנוסף, עקב מיעוט תצפיות על מינים אחרים המחקר התמקד במין 'חומט פסים'.

תוצאות פרק זה של העבודה הצביעו על כך ששפע פרוקי הרגליים המצוי בשדות החיטה בראשית האביב גבוה מאשר בכתמים הטבעיים. בנוסף, התוצאות מצביעות על תנועה משמעותית של זוחלים מתוך הכתמים הטבעיים אל שדות החיטה, אך מועטה מאוד בכיוון ההפוך. בנוסף, המצב הגופני של הפרטים שנעו מהכתמים לשדות היה, באופן מובהק, טוב יותר משל פרטים שנשארו בכתמים הטבעיים. לבסוף, לאחר הקציר לא נמצאו יותר פרטים בתוך שדות החיטה וזאת בניגוד לכתמים

הטבעיים. תוצאות אלו מצביעות על כך ששדות החיטה מהוות מלכודת אקולוגית עבור המין 'חומט פסים'.

תוצאות העבודה מצביעות שחברת הזוחלים החיה במערכת חקלאית מקוטעת מושפעת מתהליכים אקולוגים שונים המתרחשים בסקאלות מרחביות שונות. בנוסף, העבודה תומכת בכך שחברת הזוחלים המתקיימת בכתמים הטבעיים המצויים במערכת החקלאית מושפעת מהפעילות החקלאית המרחשת במטריקס. גורמים אלו תורמים להבנה שישנה השפעה הדדית של השדה על הכתם ולהפך. מכאן נובע שמחקר העוסק בחברות אקולוגיות במערכות אגרו-אקולוגיות צריך להתייחס למשתנים אקולוגים מסקאלות שונות ולבחון את השפעת הפעילות החקלאית עצמה.

מילות מפתח: AIC, אגרו-אקולוגיה, אקוטון, אקולוגיה, אקלים, בחירת בית גידול, חברה, ישראל זוחלים , מגוון מינים, מלכודת אקולוגית, ממ"ג, סקאלות מרחביות, עושר מינים, קיטוע בתי גידול, רעייה, שפלת יהודה הדרומית, שפע פרטים, תהליכים תלויי סקאלה

תודות

ברצוני להודות לשני מנחיי ירון ועמוס. תודה על החופש לעסוק בשאלות שעניינו והעסיקו אותי, על הזמינות והנכונות להתפנות ולעזור כאשר נזקקתי לכך. תודה על האמון, העזרה והתמיכה לאורך הדרך ותודה מעל לכל על אשר הראתם לי שגם באקדמיה חשוב להיות קודם כל אדם.

תודה לחברי במעבדה על רגעים יפים, על עבודה משותפת, על נכונות לעזור ולהתפנות בשעת הצורך. תודה על החשיבה המשותפת ועל האווירה הנעימה והחברית. תודה לאנשי השפלה. לבוקרים של בית ניר איתם אני בקשר עוד מימי עבודת המסטר ולאנשי הגד"ש אשר אפשרו לי לעבוד בשטחם, להיכנס לשדות ואשר עדכנו אותי בפעילות החקלאית אותה הם מבצעים. לאישתי ולילדי, עומר, יהל ופלג אהוביי. תודה לכם על התמיכה, העידוד והפרגון. תודה על הסבלנות וההקשבה. ותודה על כך שלמרות הקושי והמעמסה תמיד מצאנו את הביחד, ראינו את המצחיק ויחד צעדנו צעד נוסף קדימה.

העבודה נעשתה בהדרכת:

פרופ. ירון זיו.

המחלקה למדעי החיים, הפקולטה למדעי הטבע

אוניברסיטת בן גוריון בנגב

פרופ. עמוס בוסקילה.

המחלקה למדעי החיים, הפקולטה למדעי הטבע

אוניברסיטת בן גוריון בנגב

הצהרת תלמיד המחקר עם הגשת עבודת הדוקטור לשיפוט

אני החתום מטה מצהיר/ה בזאת: )אנא סמן(:

___ חיברתי את חיבורי בעצמי, להוציא עזרת ההדרכה שקיבלתי מאת מנחה/ים.

___ החומר המדעי הנכלל בעבודה זו הינו פרי מחקרי מתקופת היותי תלמיד/ת מחקר.

___ בעבודה נכלל חומר מחקרי שהוא פרי שיתוף עם אחרים, למעט עזרה טכנית הנהוגה בעבודה ניסיונית. לפי כך מצורפת בזאת הצהרה על תרומתי ותרומת שותפי למחקר, שאושרה על ידם ומוגשת בהסכמתם.

תאריך: 26/03/2014 שם התלמיד/ה: גיא רותם

חתימה: ______

השפעה תלוית סקאלה של מערכת אקולוגית חקלאית מקוטעת

על חברת הזוחלים

מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה"

מאת גיא רותם

הוגש לסינאט אוניברסיטת בן גוריון בנגב

אישור המנחה______תאריך: 31/12/04

אישור דיקן בית הספר ללימודי מחקר מתקדמים ע"ש קרייטמן______

יג' בחשוון תשע"ג 32 באוקטובר 3103

באר שבע

השפעה תלוית סקאלה של מערכת אקולוגית חקלאית מקוטעת על חברת הזוחלים

מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה"

מאת גיא רותם

הוגש לסינאט אוניברסיטת בן גוריון בנגב

יג' בחשוון תשע"ג 32 באוקטובר 3103

באר שבע