Ben-Gurion University of the Negev

The Jacob Blaustein Institutes for Desert Research

The Albert Katz International School for Desert Studies

Reducing the competitive load on desert bat communities by

hampering the drinking ability of an invasive bat species,

Kuhl’s pipistrelle (Pipistrellus kuhlii)

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

"Master of Science"

By: Asael Greenfeld

April 2012

Ben-Gurion University of the Negev

The Jacob Blaustein Institutes for Desert Research

The Albert Katz International School for Desert Studies

Reducing the competitive load on desert bat communities by

hampering the drinking ability of an invasive bat species,

Kuhl’s pipistrelle (Pipistrellus kuhlii)

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

"Master of Science"

By: Asael Greenfeld

Under the Supervision of Prof. David Saltz and Prof. Carmi Korine

Mitrani Department of Desert Ecology

Author's Signature …………….……………………… Date …………….

Approved by the Supervisor…………….…………….. Date …………….

Approved by the Director of the School …………… Date ………….…

I

Reducing the competitive load on desert bat communities by

hampering the drinking ability of an invasive bat species, Kuhl’s

pipistrelle (Pipistrellus kuhlii)

By: Asael Greenfeld

This thesis is in partial fulfillment for the degree of Master of Science, Ben-Gurion

University of the Negev, Jacob Blaustein Institutes for Desert Research, Albert Katz

International School for Desert Studies, 2011

Abstract

Among the species comprising the bat community of the Negev desert in

Israel, three have expanded their range into the desert from Mediterranean habitats,

probably due to anthropogenic developments such as the addition of bodies of open

water and lights. The abundance and activity of bats in the desert is affected by the

distribution of bodies of open water, which are used by bats for drinking and foraging

sites. One of the species that has penetrated the Negev from Mediterranean habitats is

Kuhl’s pipistrelle (Pipistrellus kuhlii), the most common bat in Israel. Kuhl’s

pipistrelle competes for foraging habitats and food with three other species of bats

(P.rueppellii, Hypsugo bodenheimeri and Eptesicus bottae), comprising the

“background cluttered space” foraging guild. Kuhl’s pipistrelle drinks more

frequently than the other species of its guild and requires a clear “swoop zone” to

drink. I hypothesized that Kuhl’s pipistrelle uses newly established water bodies of

II

open to expand into desert habitats. To reduce the competitive load that Kuhl’s

pipistrelle has on desert dwelling species, I proposed and tested a management tool to

reduce the ability of Kuhl’s pipistrelle to drink from open bodies of water. I predicted

that, by installing obstructions above the water surface, I could reduce the drinking

ability of Kuhl’s pipistrelle, thus reducing their activity.

First, using acoustic methods, I surveyed natural and artificial bodies of open

water in the Ramon region in order to learn the current limits of Kuhl’s pipistrelle’s

range expansion and to study the mechanisms of its expansion. To study the effect of

obstructions on the water surface, I did field experiments. In each experiment, the

manipulation and control treatments were alternated, and the activity levels of the

different species present were monitored by acoustic methods.

I found in the survey that Kuhl’s pipistrelle are abundant in the Central Negev

Highlands, but absent from Makhtesh Ramon, and found activity levels of Kuhl’s

pipistrelle high in natural sites as well as in artificial ones. In the first field experiment

that tuck place in a swimming pool, I managed to prevent Kuhl’s pipistrelle's drinking

by obstructing the water surface and leaving just a 1x1m of “swoop zone”. In the

second manipulation experiment, in natural ponds in the Zin valley, I found that the

obstructions had no effect on activity levels or species composition. This might have

been the result of the proximity of untreated ponds to manipulated ponds in the study

area. In the third experiment, I chose isolated natural ponds throughout the Negev. I

found that, in some of the ponds, the manipulation significantly reduced Kuhl’s

pipistrelle activity. This reduction was observed in ponds that initially had higher

activity of Kuhl’s pipistrelle then of desert dwelling species.

III

I concluded that artificial bodies of open water may be part of the mechanism by

which Kuhl’s pipistrelle is expanding its range into the Negev. The existence of bodies of

open water pools in Makhtesh Ramon and the Southern Negev poses a risk of further

expansion of Kuhl’s pipistrelle into desert habitats from which they are currently absent.

With some further investigation of my predictions, manipulating artificial bodies of open

water could be used as a management tool to reduce competitive load from desert

dwelling species. Such a tool would be especially useful in the troughs established by the

National Parks Authority in the Negev, which were intended to facilitate biodiversity by

supporting wildlife, but are simultaneously threatening desert dwelling bat species.

IV

Acknowledgments

First and foremost I would like to thank my supervisors, Prof. Carmi Korine who

introduced me to the world of desert bats and invited me to continue my work at the

Ramon Science Centre and turn it into an M.S. research project, and Prof. David Saltz, for

giving me the opportunity to do this study under his supervision. I thank them both for

their advice, help and guidance throughout the planning, executing and writing of my

thesis.

I would have not been able to carry out my field work and experiments without

the help of many volunteers, neighbourhood friends and friends from the Libratory,

family members, and especially my students from the Mizpe Ramon Environmental

Studies Yeshiva high school. The pool experiment was performed with the help of two of

the course students: Keira Edwards and Ben Evans from the University of Bristol, UK.

I thank Noam Werner and Uri Shanas for lending me equipment, and Atlantis

Aquariums and the Alpaca Farm for their help with pool construction. I am also grateful

to the Israel Nature and National Parks Protection Authority for allowing me to carry out

my research in nature reserves, and individually, for all their help and advice, to: Ben

Drori, Nadav Tauba, Amram Tsabari, Gal Vine, and Assaf Tzoar.

Finally, nothing would come to be without the love and support of my beloved

wife Lea or without her smiles and those of our dear children, Rotem, Adi and Boaz.

My research was funded through grants from Ben-Gurion University Seed Money

to Carmi Korine and David Saltz, and from the Ministry of Technology and Science to

Carmi Korine. This work was supported through a scholarship from the Albert Katz

International School for Desert Studies.

V

Table of Contents

1. Chapter One - General Introduction.............................................................. 1

1.1 - Anthropogenic bodies of open water in desert environments and

their effect on local wildlife ....................................................... 1

1.2 - Diversity of insectivorous bats in association with water sources

.................................................................................................... 3

1.3 - The association between distribution of bodies of water and bats’

community structure in the Negev highlands ........................... 5

2. Chapter Two - Species composition of bat communities in different

habitats in the Ramon region, and the range expansion

mechanism of the Kuhl’s pipistrelle ...................................... 10

2.1 - Introduction ............................................................................... 10

2.2 - Methods ..................................................................................... 13

2.2.1 – Study area .................................................................... 13

2.2.2 – Survey species .............................................................. 15

2.2.3 – Field sampling methods .............................................. 17

2.2.4 – Data analysis of field sampling ................................... 20

2.3 - Results ...................................................................................... 21

2.3.1 – Total bat activity and Kuhl’s pipistrelle activity levels in

the different areas of the Ramon region ...................... 22

2.3.2 – Activity levels in the different site types .................... 25

2.4 - Discussion ................................................................................ 26

2.4.1 – Species composition of the “background-cluttered space”

guild in the Ramon region ........................................... 26

2.4.2 – Range expansion of Kuhl’s pipistrelle in the Ramon

region............................................................................ 28

3. Chapter Three - Managing bat community species composition by placing

obstructions to hamper the Kuhl’s pipistrelle ability to drink

from desert ponds ................................................................. 30

3.1 - Introduction ............................................................................... 30

3.1.1 – Management of alien invasive bat species .................. 30

VI

3.1.2 – Interspecific differences in the use of bodies of water as

the key for the management of desert bat communities 33

3.2 - Methods .................................................................................... 37

3.2.1 – Field experiment I: manipulation of the swimming pool

...................................................................................... 37

3.2.2 – Field experiment II: manipulation of natural ponds in the

Zin valley .................................................................... 40

3.2.3 – Field experiment III: manipulation of isolated natural

ponds ........................................................................... 44

3.3 – Field experiment results .......................................................... 48

3.3.1 – Manipulation of the swimming pool ........................... 48

3.3.2 – Manipulation of natural ponds in the Zin valley ......... 49

3.3.3 – Manipulation of isolated natural ponds ....................... 52

3.4 - Discussion ................................................................................ 55

3.4.1 – Manipulation of the swimming pool ........................... 55

3.4.2 – Manipulation of natural ponds in the Zin valley ......... 57

3.4.3 – Manipulation of isolated natural ponds ....................... 58

4. Chapter Four - General Discussion .............................................................. 60

4.1 - The current distribution of Kuhl’s pipistrelle, and its impact on

local bat communities ............................................................. 60

4.2 - The effect of pond manipulation on species composition ........ 61

4.3 - The potential use of the grid as a management tool for the Negev

bats ........................................................................................... 62

4.4 - Mitigation and Reconciliation rather than aggressive eradications

as the future of invasive species management ........................ 63

5. Chapter Five - Bibliography ..................................................................................... 65

VII

List of Figures

Chapter 2

Figure 2.1 - The known distribution of Pipistrellus kuhlii in the Negev (2009)…...…

Figure 2.2 – Map of field sampling sites in the Ramon region…………………….…

Figure 2.3 – Activity of Pipistrellus kuhlii in the Ramon region……………………..

Figure 2.4 – Species composition in each area of the Ramon region…………………

Figure 2.5 – Genaral bat activity levels at different site types ……………………….

Chapter 3

Figure 3.1 - Diagram of the grid covering of the swimming pool experiment………..

Figure 3.2 - Field experiment II – temporary pond sites in the Zin valley…………....

Figure 3.3 – An example of the network that was placed on pond surfaces…………..

Figure 3.4 – Map of sites for field experiment III in isolated ponds in the Negev........

Figure 3.5 – Activity and drinking passes in the swimming pool experiment ……..…

Figure 3.6 –Average bat activity levels in the different sites used for experiment III..

Figure 3.7 – Pipistrellus kuhlii proportion in different treatments in experiment III ..

11

20

23

24

25

38

41

43

45

49

53

54

VIII

List of Tables

Chapter 2

Table 2.1 - Field sampling sites in the Ramon region…………………………………

Table 2.2 – Number of recordings from each survey site in the Ramon region………

Table 2.3 – Activity of bat species at the survey sites in the Ramon region………….

Chapter 3

Table 3.1 – Sites for experiment II in the Zin valley…………………………………

Table 3.2- Staggered use of ponds for experiment II in the Zin valley……….……....

Table 3.3 – Sites for experiment III in isolated bodies of water………………...…….

Table 3.4 - Design of experiment III based on alternating the treatments……...…….

Table 3.6 – Bat activity at ponds in experiment II in the Zin valley…………………..

Table 3.7 – Regression slopes of proportions in experiment III………………………

Table 3.8 – Activity of Pipistrellus kuhlii in experiment III.........................................

19

21

22

40

42

45

47

51

51

52

1

Chapter One - General Introduction

1.1 Anthropogenic bodies of open water in desert environments and their

effect on local wildlife

Man-made elements that influence water availability are having increasing

impact on arid ecosystems worldwide (Ragab and Prudhomme 2002). Arid and

semiarid landscapes are unique ecosystems where continual water availability

results in hotspots of biological diversity in an otherwise water-limited environment

(Noy-Meir 1973, Stromberg et al. 1996, Stromberg 2001). This results sharp special

differences between the vast xeric uplands and narrow mesic riparian environments

and by sharp temporal transitions (Stromberg et al. 1996).

Human development of arid and semiarid environments has, on the one hand,

lowered the water levels that support the base flow in some deserts by ground water

pumping which reduces the water availability to wildlife (Pool and Coes 1999), and

on the other hand, has created new water sources in many places, increasing the

water supply to desert fauna (O'Brien et al. 2006) and flora (Brooks et al. 2006). An

important type of anthropogenic water source is “water developments” used by

conservation authorities as an integral component in the maintenance and

enhancement of wildlife habitats in arid regions. In the deserts of the USA, there are

thousands of water developments (over 800 in Arizona alone), which go back as far

as the 1940s, intended to benefit game species and other wildlife (Rosenstock et al.

1999). Contemporary water developments are now constructed in the USA to

accommodate use by a variety of wildlife and to mitigate the loss of natural water

sources (Rosenstock et al. 2004).

2

The importance of bodies of open water to wildlife increases during the dry

season, and visits to anthropogenic bodies of open water by birds and mammals

intensify with the rise in Ta (O'Brien et al. 2006). There has long been a

controversy regarding the effects of water developments on local wildlife, in which

critics argue that these developments do not yield expected benefits to game species

and may have adverse impacts, such as affecting prey species by localizing

predation (O'Brien et al. 2006). Additionally, the use of water for agriculture,

sewage treatment, and recreational reservoirs affects the distribution of local

wildlife and enhances the successful introduction and spread of invasive species

(Griffin 1998, Dukes and Mooney 2004, Lobos et al. 2005).

In recent years, the increase in human population and the concomitant increase

of inland agricultural development in Israel have had strong negative effects on

many vertebrate species (Yom-Tov and Mendelssohn 1988). In the Negev Desert,

agricultural development has brought about an increase in the number of bodies of

open water, mostly by collecting ground water in reservoirs (Portnov and Safriel

2004). An important type of water development in the Negev is construction of

artificial water troughs for wildlife, first used in the 1990s to support the

reintroduced population of Asiatic wild ass (Equus hemionus) (Saltz and

Rubenstein 1995). Water developments in the Negev affect the distribution of

wildlife and promote invasion by invasive species and the range expansion of others

(Hawlena and Bouskila 2006). Invasive alien species have been found harmful to

local native communities by means of resource competition with the local species,

increased predation on the local prey communities, and the introduction of

pathogens and parasites that may threaten local fauna (Simberloff 2010). Thus, the

3

invasion of alien Mediterranean species into the desert wildlife community may

have a negative impact on local populations.

1.2 Diversity of insectivorous bats in association with water sources

Among terrestrial mammals, bats play key roles in temperate and tropical

ecosystems. This is due to their large number of species (species richness),

enormous numbers of individuals, mobility, geographical distribution, and their

important functions as seed dispersers, pollinators, and arthropod predators (Hill

and Smith 1984). The activity level of insectivorous bats varies in time and space

and is associated with several factors that have been well studied. Of these factors,

those influencing the abundance of insects are of primary importance because the

choice of foraging site is directly linked to food availability (Rydell 1992), the

length of foraging bouts (Swift 1980), and the time of parturition (Kunz 1973,

Arlettaz et al. 2000). Many studies report that insect activity is directly proportional

to ambient temperature (Ta), and consequently, at low Ta, bat activity is reduced

(Williams 1961, Bradley 1970, Rydell 1989, Wai-Ping and Fenton 1989, O'Farrell

and Hayes 1997) and it is higher during the summer (O'Farrell and Bradley 1970,

Seidman and Zabel 2001, Korine and Pinshow 2004, Adams and Thibault 2006).

Bodies of open water also play a key role in bat distribution and habitat

preference since they are used by bats as a source of drinking water and as foraging

areas (Vaughan 1996, Grindal et al. 1999, Ciechanowski 2002, Razgour et al.

2010). The importance of drinking is demonstrated by Webb et al. (1995), who

report immense water loss, up to 30% of a bat’s body mass in Plecotus auritus and

Myotis daubentoni, two non-desert species, during one day while resting. Since

4

bats drink in flight, their drinking behavior depends on their flight performance

(Norberg and Rayner 1987). Flight performance, in turn, is a combined product of

wing morphology and mechanical properties of the wing membrane and bones.

These attributes vary considerably among bats (Swartz et al. 2004). Bats that can

hover can drink from almost any size of water body, while high speed flyers require

an unobstructed ‘swoop zone’ (Harvey et al. 1999), and consequently, can drink

from smaller and from more obstructed bodies of water, than larger bats (Tuttle et

al. 2006). Insectivorous bats likely prefer hunting over open water because the

echoes of insect bodies are reflected from the water surface and provide a stronger

target than similar size prey in the open air (Siemers et al.2001). In a recent study

(Jackrel and Matlack 2010) demonstrated how the drinking preferences of bats from

artificial water tanks follow the physical properties of the water source. The authors

found that bats drink more frequently from larger tanks (diameter = 3 m) than from

smaller tanks (d = 1.2 m), from full tanks more than from half-full tanks, and that

bats prefer tanks with less surrounding vegetation. All three criteria affected the

drinking frequency, but not the total flight activity, suggesting that all tanks were

equally suitable as foraging areas. Bat activity and species richness increase with

pond size, and temporary ponds do not significantly differ from permanent ones in

activity levels (Razgour et al. 2010). High levels of bat activity are consistently

found over open water (Rydell et al. 1994, Walsh and Harris 1996, Vaughan et al.

1997, Russo and Jones 2003), including in the Negev desert (Korine and Pinshow

2004). The importance of open water bodies was further demonstrated by Greif and

Siemers (2010), who showed that the recognition of bodies of water by the bats is

innate. Their experiments may indicate that bats have evolved the adaptation to

5

innately locate water in complete darkness, allowing even juveniles, which have

never before encountered an open body of water, to identify it as a source of

drinking water by echolocating its smooth surface.

1.3 The association between distribution of bodies of water and bats’

community structure in the Negev highlands

The availability of open bodies of water plays a key role in determining the

abundance and composition of bat communities in specific habitats. The proximity

to water was found to be an important factor in the choice of diurnal roosts by

species of bats in semi-arid northern Arizona (Rabe et al. 1998), while similar

studies done in wetter climates in Oregon (Betts 1998, Waldien 1998) found no

such effect. However, other studies found no effect of water availability on roost

preference, although night activity levels were higher near bodies of water (Waldien

1998). Since different species vary in their water requirements (Razgour et al.

2010), the abundance and nature of open bodies of water are expected to influence

species composition, especially in desert bat communities. The link between water

distribution and community structure has been studied in Poland (Ciechanowski

2002, Mysłajek et al. 2007) and, recently, in the Negev desert by Razgour et al.

(2010). In the mesic Polish forest, smaller ponds have higher species richness than

larger ponds or rivers. This was explained by the fact that, in ponds, the habitat

requirements of both water-obligated foragers and of dense vegetation foragers are

met (Ciechanowski 2002, Mysłajek et al. 2007). In the Negev desert, on the other

hand, bat species richness and activity are inversely proportional to pond size

(Razgour et al. 2010). Razgour et al. (2010) also report that, while total bat activity

6

is not affected by the hydro-period (the time that ponds hold water) of a pond, the

bat community composition differs between permanent and temporary ponds. Non-

desert bat species were found to drink more frequently at larger ponds either

permanent, or with lengthy hydro-periods. Pond size also affects bat community

structure based on species maneuverability, with larger species avoiding smaller

ponds (Ciechanowski 2002, Razgour et al. 2010).

Insectivorous bats are among the species most affected by anthropogenic

disturbances to natural habitats (Yom-Tov and Mendelssohn 1988), and it seems

that, in some cases, modification of the natural environment causes changes in bat

community structure, resulting in competitive exclusion of some species. Arlettaz et

al. (2000) suggest that artificial foraging sites facilitated the expansion of common

pipistrelle (Pipistrellus pipistrellus) populations in the Swiss Alps. This apparently

resulted in competition with the lesser horseshoe bat (Rhinolophus hipposideros),

contributing to a decline in the population size of the latter. High levels of bat

activity in artificial sites in the Negev Desert are linked with lower bat species

richness (Korine and Pinshow 2004). Similarly, urban parks with lakes were found

to support fewer bats and less diverse bat communities than those close to rural

lakes (Kurta and Teramino 1992).

Of the 34 insectivorous bat species found in Israel, 12 inhabit the Negev Desert

(Korine and Pinshow 2004). Nine of the 12 Negev bat species are associated with

arid areas, and the other three -- Kuhl's pipistrelle, the European free-tailed bat

(Tadarida teniotis), and the rare lesser horseshoe bat (Rhinolophus hipposideros) --

are invasive Mediterranean species that have expanded their distribution into the

Negev in the 20th

century (Yom-Tov and Mendelssohn 1988, Yom-Tov and

7

Kadmon 1998). Kuhl's pipistrelle, one of these Mediterranean species, shares the

same guild with three other local species: Ruppell's pipistrelle (P. rueppellii),

Bodenheimer's pipistrelle (Hypsugo bodenheimeri) and Botta's serotine (Eptesicus

bottae). All four have similar diets of flying insects, mainly Lepidoptera (Whitaker

et al. 1994, Feldman et al. 2000), have similar wing morphology and forage in

similar habitats (Korine and Pinshow 2004). The four species belong to the same

guild as defined by wing and ear morphology (Yom-Tov 1993), and also by habitat,

as defined by Korine and Pinshow (2004), who named them the “background-

cluttered space” guild, after their preferred foraging habitat. Kuhl's pipistrelle,

Ruppell's pipistrelle, and Bodenheimer's pipistrelle evidently compete over the use

of ponds for drinking and foraging, resulting in temporal and spatial partitioning

between these species (Razgour et al. 2011). These authors also report that Kuhl's

pipistrelle prefers smaller ponds and, when competitors are present, its activity

peaks several hours after sunset; Ruppell's pipistrelle prefers medium-sized ponds

and its activity peaks after sunset and before sunrise, in the presence of competition;

and that Bodenheimer's pipistrelle prefers larger ponds, where its activity peaks

three hours after sunset. On a seasonal scale, the increase in the size of the

competing populations of Ruppell’s pipistrelle and Bodenheimer’s pipistrelle, in

autumn and spring, was correlated with a decrease in the presence of Kuhl’s

pipistrelle at ponds frequented by the two other species (Razgour et al. 2010). While

these findings do not prove competitive displacement, they do suggest incomplete

resource partitioning that leads to competition. Accessibility to drinking water is

probably the limiting factor for the use of smaller ponds, and Mediterranean species

8

that drink most frequently were found to be associated with larger, permanent ponds

(Razgour et al. 2010).

Kuhl’s pipistrelle has expanded its range following human development (Yom-

Tov and Mendelssohn 1988). Artificial bodies of open water in the Negev highlands

were found to be used mainly by two species of bats: Kuhl’s pipistrelle and the

European free-tailed bat. These species are the two most common insectivorous bat

species in Israel (Yom-Tov and Mendelssohn 1988) and are closely associated with

human habitats in the Negev (Korine and Pinshow 2004).

The documented competition of Kuhl’s pipistrelle with the desert-dwelling

Ruppell’s pipistrelle and Bodenheimer’s pipistrelle that are restricted to only a few

foraging areas (Razgour et. al 2010), combined with the increasing development of

new bodies of open water in the Negev, raises two conservation-related concerns

regarding local bat biodiversity:

1) Mediterranean bat species, such as Kuhl’s pipistrelle, may further expand

their distribution and penetrate into even the more rural habitats, bringing with them

possible resource competition, introduction of pathogens and parasites and other

problems associated with invasive species.

2) In areas where Kuhl’s pipistrelle populations are already established, they

may change the local species composition by means of competitive exclusion of

other members of the guild. Ruppell’s pipistrelle and Bodenheimer’s pipistrelle are

classified as regionally endangered (EN), and Botta's serotine as vulnerable (VU)

(Dolev and Perevolotsky 2004), implying that any decrease in populations of these

species would be harmful to regional biodiversity.

9

The main goal of my research was to address these two concerns by:

1) Exploring the more Southern parts of the Negev highlands, which were

unexplored in terms of bat communities. Accordingly, I made a survey in which I

determining the possible boundaries of the current range of Kuhl’s pipistrelle in this

region.

2) Experimentally assessing the feasibility of a management tool which I

propose, namely hampering the drinking ability of Kuhl’s pipistrelle from bodies of

water by limiting the available “swoop zone.” This was done in an attempt to create

“protected habitats” for desert-dwelling bats, in order to reduce the competitive use

of bodies of water by Kuhl’s pipistrelle and the desert-dwelling species of its

feeding guild.

10

2 Chapter Two – Species composition of bat communities in different habitats

in the Ramon region, and the range expansion mechanism of the Kuhl’s

pipistrelle

2.1 Introduction

The dispersal of Kuhl’s pipistrelle into the Negev was first observed by Yom-

Tov and Mendelssohn (1988) and supported by the results of a cluster analysis of

their distribution by Yom-Tov and Kadmon (1998), yet the limits of this range

expansion are not clear. Most of the available data regarding species distribution in

desert habitats in Israel were, until the late nineties, collected from the Dead Sea

area and the Arava rift valley (Yom-Tov and Kadmon 1998).Yom Tov and Kadmon

(1998) created a model predicting the regional distribution of bat species. The

model is based on temperature and annual precipitation, and they predicted the

presence of Kuhl’s pipistrelle throughout the entire Negev. The model does not

take into account other local scale elements, such as water availability or inter-

species interactions. In a later study of the insectivorous bat community of the

northern Negev highlands, Kuhl’s pipistrelle was found to be highly abundant in all

natural and man-modified sites surveyed as far South as the Zin Valley (Korine and

Pinshow 2004). Their data, together with accumulated unpublished data (Korine

personal communication and the SPNI database of 2011 that reports observations

collected by the Society for the Protection of Nature in Israel, on the abundance of

Kuhl’s pipistrelle in the Arava, covered most of the more developed regions of the

Negev (see Figure 2.1), but left an information gap comprising the whole southern

11

Negev, including the Ramon ridge, with only a vague indication of the extent to

which Kuhl’s pipistrelle has penetrated this rural area. Incidental observations

reported in the SPNI database 1122 report Kuhl’s pipistrelle in the Paran valley

(1991), in Mitzpe Ramon (1991), at the Hemet and Lotz cisterns and Ein Hava

(1994, 1996), and once at Ein Saharonim (1994). The first published report of the

capture of a Kuhl’s pipistrelle in Eilat was in 2002 (Zelenova and Yosef 2003).

More recently reports of Kuhl's pipistrelle in Eilat have become common in the

International Birding and Research Centre, Eilat area (Shalmon, personal

communication). Benda et al. (2008) surveyed the Sinai Peninsula and found no

evidence of further expansion into the Sinai. The importance of the Ramon region,

compared with the northern Negev and the Arava, is that it comprises the largest

portions of undeveloped wilderness in Israel, with some of the largest nature

reserves in the country, such as the “Tzinim Cliffs” and the “Har Hanegev”

reserves. Many of the areas, in addition to being

declared nature reserves, are further protected from

development by serving as military training zones (Tal

2008). This makes them important as wildlife refuges,

protecting desert species from the habitat modification

caused by the residential and agricultural development

occurring in the rest of the Negev (Mazor 2001).

Figure 2.1 – The known distribution of Pipistrellus kuhlii in

the Negev prior the survey (dots) was limited to the Northern

Negev and the Arava, and anecdotal observations in the

Ramon region and southern Negev. Map based on BioGIS (2012)

Northern Negev

Southern Negev

12

Korine and Pinshow (2004) suggested that bat populations are affected by the

creation of artificial feeding and drinking sites in desert habitats. They based this

contention on the fact that they recorded the highest activity levels of bats at

artificial sites, and that Kuhl’s pipistrelle and the European free-tailed bat were the

only species that were abundant at these sites.

Artificial sites in the Ramon region are of two sources: 1) In towns and

agricultural settlements (limited to the town of Mitzpe Ramon, and some small

farms along Road 40 leading to the town). Mitzpe Ramon and the small farms have

water treatment, swimming pools and artificial lights, all of which cause insects to

congregate. Several army camps are located within the region and produce similar

modifications of their surroundings. 2) Artificial watering troughs, built to support

biodiversity of large animals, are also a potential threat to bat biodiversity since

they may support the invasion of alien bat species that may eventually reduce bat

biodiversity. Yet, there has been no monitoring of these troughs, as far as bats are

concerned (NPA rangers, personal communication).

One goal of my research was to achieve a better understanding of the role that

artificial ponds may play in the mechanism by which Kuhl’s pipistrelle expand their

range into desert habitats. This was done by studying the distribution of the bats at

the edge of their range along their suspected expansion front, the Ramon region.

Along with the abundance of Kuhl’s pipistrelle, I also monitored the abundance of

the other members of the “background-cluttered space” guild, Botta's serotine,

Ruppell’s pipistrelle and Bodenheimer’s pipistrelle, to explore understand possible

competition or coexistence within this guild (Korine and Pinshow 2004). I also

studied the effect of water availability on the activity levels of the abovementioned

13

species. I hypothesized that the distribution of Mediterranean insectivorous bats in

desert habitats is influenced by anthropogenic activities. This led to the following

predictions:

1) On the regional scale, Kuhl’s pipistrelle is more abundant in areas with

anthropogenic disturbances, such as residential and military development, than in

the rural areas, leading to differences in the species composition of bats among the

areas.

2) On the local scale, Kuhl’s pipistrelle activity levels are higher near artificial

water sources than near ephemeral natural water sources, and newly established

water source will attract Kuhl’s pipistrelle and therefore their activity level will

increase.

3) Where bodies of open water are available, a larger proportion of the total bat

activity is that of Kuhl’s pipistrelle, than of desert-dwelling species. At sites far

from bodies of open water, a higher proportion of the total bat activity is of desert-

dwelling species then of Kuhl’s pipistrelle.

2.2 Methods

2.2.1 Study area

Field sampling of bat community compositions took place in areas that are

estimated to be the front edges of expansion for the Mediterranean bat species, i.e.

the Ramon region. The 200 m high Makhtesh Ramon south-facing cliff, extends

over 40 km from the North-East (3069’N, 3495’E) to the South-West (3050’N,

3464’E). The cliff is a natural boundary between two climatic zones, arid at the top

(700 m -1000m above sea level) and hyper-arid in the lower part of the

14

Makhtesh(450 m - 550 m above sea level), and is also the southern edge of

agricultural and water developments in the Negev highlands (Mazor 2001).

Incidental observations and preliminary recordings indicated that the Mediterranean

bat species are present north of the cliff but not south of it (Korine unpublished data

2008, BioGis 2012). In this desert region rain occurs during winter with large

annual differences in total precipitation and in its temporal and spatial distribution.

Ta are highest during summer and are lowest in winter, with a daily mean of 25 °C

in August, and 10-11 °C in January (Hillel and Tadmor, 1962).

Bodies of open water in the Ramon region are limited, and most of them are

associated with human activity. The region can be divided into three areas (Fig 2.2),

based on climatic and ecological differences between them, following Ward et al.

(2000):

A. Central Negev Highlands

This area consists of the highest mountaintops in the Negev region (Ramon

Mountain - 1071 m), with occasional snow during some winters. These mountains

are the origin of the big wadis that drain the entire area (Zin, Nitzana, Ramon). The

area has been found to receive more annual rainfall (100 mm) than its surroundings

(Ward and Olsvig-Whittaker, 1993). The area is unpopulated and uncultivated, and

the only sources of anthropogenic disturbances to wildlife are limited to hiking,

military camps, and military activity.

B. Mitzpe Ramon and its surroundings

The town of Mitzpe Ramon was established on the northern cliff of Makhtesh

Ramon and is part of the Central Negev highlands, with a slightly lower altitude

(880-900m) the their western side. This is an area with higher density of human

15

development than the Makhtesh and the rest of the Central Negev Highlands. The

town of Mitzpe Ramon and the surrounding farms and water developments create

an artificial oasis of modified habitat, in extreme contrast to the arid landscape

surrounding it.

C. Makhtesh Ramon

This area is much lower (from 800 m and down to 420 m) and drier than the rest

of the Ramon region, with an average of 56 mm of precipitation per year in the

eastern, lower part, and up to 80 mm average annual precipitation at the western

edge (Ward and Olsvig-Whittaker 1993). The vegetation is limited to the wadis.

The area is unpopulated with no agriculture, and most of it is a nature reserve with

limited hiking trails and dirt roads (Mazor 2001).

2.2.2 Surveyed species

In the field, all bats species were recorded, but the following three species --

Kuhl’s pipistrelle, Bodenheimer’s pipistrelle, Botta's serotine that belong to the

“background-cluttered foraging" guild (Korine and Pinshow 2004) were the focus

of my research. The fourth member of this guild, Rueppell’s pipistrelle, was never

recorded In the Ramon region. Since Rueppell’s pipistrelle is not easy to

differentiate acoustically from Bodenheimer’s pipistrelle, I chose to neglect the

possibility that an occasional individual was present in my study sits and it’s passes

were counted as though it was a Bodenheimer’s pipistrelle.

A. Kuhl’s pipistrelle (Pipistrellus kuhlii) – family Vespertilionidae, subfamily

Vespertilioninae. It is a small (body mass ~ 6.5-7.5 g) and common insectivorous

bat, and it occurs in a wide variety of natural and anthropogenic habitats from

16

southwest Asia to southern and eastern Africa. In Israel, it is the most common bat

and is found from the north of the rift valley and Mediterranean down to Eilat and

in some desert habitats (Yom-Tov and Kadmon 1998). It is an aerial forager,

tending to use artificial lights and bodies of water for foraging (Korine and

Pinshow 2004). Most of its activity occurs in the early hours of the evening.

Echolocation calls used by this bat while foraging are most intense at around 35-

40 kHz (Whitaker et al. 1994, Berger-Tal et al. 2008).

B. Bodenheimer’s pipistrelle (Hypsugo bodenheimeri) – family

Vespertilionidae, subfamily Vespertilioninae. It is the smallest bat species in Israel

(body mass 2.5-2.9 g). It occurs in the desert regions of Arabia and Egypt (Benda

et al. 2008), and in Israel, it is found in the rift valley from Ein Geddi to Eilat and

in the southern Negev (Yom-Tov and Mendelssohn 1988). It is an aerial forager

that tends to fly low over the ground, around trees and, mainly around bodies of

water. Most activity occurs in the early hours of the evening. The echolocation

calls of this bat are most intense at around 44-46 kHz (Whitaker et al. 1994,

Riskin 2009).

C. Botta's serotine (Eptesicus bottae)- family Vespertilionidae, subfamily

Vespertilioninae. It is a small/medium-sized (body mass 7-9 g) insectivorous bat.

It occurs in dry habitats, from India and western China to Egypt and the Eastern

Mediterranean (Benda et al. 2008), and is found in Israel from the north of the rift

valley down to Eilat and the Negev Desert (Yom-Tov and Mendelssohn 1988). It

is an aerial forager, commonly foraging above vegetation (Korine and Pinshow

2004). The echolocation calls of this bat are most intense around 32-33 kHz

(Benda et al. 2008, Holderied et al. 2005).

17

2.2.3 Field sampling methods

I measured the activity levels of the different bat species by sampling their

echolocation calls, using Bat detectors (AnaBat II, Titley Electronics, Australia). I

defined bat activity as the number of bat passes per hour of recording at each

pond, whereby a pass is a sequence of bat calls (Fenton 1970). The echolocation

call frequencies of the bat species found in the Negev do not overlap (Benda et al.

2008). Therefore, I was able to distinguish individual calls at the species level.

Only echolocation calls emitted during foraging were used to distinguish between

the species and to determine activity levels for each species.

I studied the bat community compositions in each of the three areas, Makhtesh

Ramon, the Central Negev Highlands and the Mitzpe Ramon area. In each area, I

sampled three locations, which represent three different habitats (Table 2.1): natural

sources of water such as water holes, artificial bodies of water (see description

below), and, as a control, wadis without any known source of water but with

vegetation cover. Due to the extremely dry winter of 2008-2009, with only 50% of

the annual mean precipitation and the last rain event occurring in March, all the

natural water sources surveyed did not contain ponds throughout the entire survey.

Natural water sources were expected to attract high bat activity levels even without

water, due to the fact that they were surrounded by dense vegetation. Sites

surrounded by dense vegetation, studied by Korine and Pinshow (2004) have higher

bat activity levels compared with the sites categorized as dry sites.

Sampling locations were chosen based on the home ranges of similar bat

species, Common pipistrelle and the Soprano pipistrelle, that were found to forage

up to 2-3 km away from their roosts, and never more than 3.8 km away (Nicholls

18

and Racey 2006). These distances were recorded for non-desert species in Great

Britain, a fact that may affect their relevance to my study, yet they allow a rough

estimate of commuting distances. To lower the probability that bats from the same

populations would be recorded in different areas, I made sure that the distances

among the areas exceed 8 km, namely, more than twice the maximum estimated

commuting range of Kuhl’s pipistrelle. The distances between the sites within areas

were chosen to exceed 3 km to minimize the probability of counting the same bat at

two sites in one night. All sampling sites were located in wadis with dense

vegetation compared to the surrounding slopes.

One new artificial pool was built in the Mitzpe Ramon area, within the vicinity

of the Alpaca farm (30°36'36"N, 34°46'39"E). The other two artificial pools that

already existed in Makhtesh Ramon and the Negev highlands north of Mitzpe

Ramon (NPA troughs) were reconstructed to the same dimensions, thus controlling

for pond characteristics (shape and size) that may affect bat activity (Razgour et al.

2010). They were all located at ground level, rectangular in shape, and 3 2.5 m

minimal size. To construct the pools, I dug 30-50 cm deep ditches and covered

them with 0.8 mm thick, black P.V.C sheets (Sera GmbH,Heinsberg, Germany).

The water level was kept 20-30 cm deep using a valve with a float, allowing inflow

to compensate for evaporation.

Each site was sampled at least three nights in each of three seasons during 2009:

spring (March-May), summer (June-August), and fall (September-November), from

dusk (17:45-19:00, GMT+2) to dawn (04:30-05:45). In 2010 and 2011, partial

sampling was performed at the artificial site in the Alpaca farm. All sampling nights

took place during and close to the new moon, to prevent possible bias caused by the

19

increased lunar illumination that reduces bat activity, as was reported for some

species (reviewed by Lang et al. 2006). Each sampling night consisted of recording

bat calls using an AnaBat detector, positioned at a distance of approximately 5 m

from the pond at a vertical angle of 45°, and measuring proximate Ta (±0.5 °C)

every hour using iButton® data loggers (Dallas Semiconductor, Maxim Integrated

Circuits, Dallas, USA), which were sensitive. iButtons were placed approximately

1m above the ground, sheltered within vegetation from the wind.

Table 2.1- Field sampling sites in the Ramon region, Israel. Sites were designed to cover

the three different areas by surveying three types of sites in each area.

Makhtesh Ramon High Negev Mountain Mitzpe Ramon area

Natural

water

sources

1. Saharonim – A natural

spring. 30°36'13.22"N

34°56'14.11"E

4. Lotz – One of the

Borot Lotz cisterns.

30°30'22.56"N

34°36'45.64"E

7. Hemet – A cistern

8 km west of the

town. 30°35'36.85"N

34°42'35.73"E

Dry

Wadis

2. Ramon – On the banks

of the eastern Ramon

Wadi. 30°36'58.87"N

34°55'18.85"E

5. Dorban – A tributary

of the Nitzana Wadi.

30°33'8.97"N

34°38'59.91"E

8. Zin – On a tributary

of the Zin Wadi, 6 km

west of the town.

30°35'59.78"N

34°45'12.61"E

Artificial

bodies of

water

3. Afor - An NPA trough

in the Afor Wadi.

30°36'44.98"N

34°54'39.87"E

6. Nitzana – An NPA

trough in the Nitzana

Wadi.

30°32'32.30"N

34°38'43.40"E

9. Alpaca farm – A

pool set up for the

experiment.

30°36'36.11"N

34°46'39.07"E

20

Figure 2.2 - Field sampling sites in the Ramon region: Makhtesh Ramon area (1-3), Central Negev

Highlands (4-6), Mitzpe Ramon area (7-9). The different site types are represented by colors:

natural water sources (green), dry wadis (red), and artificial bodies of water (blue). The broken line

represents the northern cliff of Makhtesh Ramon.

2.2.4 Data analysis of field sampling

I analyzed the calls using the software AnalookW 3.3q. Each AnaBat recording

file contains at least one bat pass. I counted the number of passes for each species

for each night and standardized the data by dividing the total passes for each night

by the number of hours from dusk to dawn. Since the activity levels for all species

were not normally distributed and their among-group variance lacked homogeneity,

and since many ‘0’ values were recorded, some of the statistical tests required

10 Km

Makhtesh Ramon Mitzpe Ramon area

Central Negev Highlands

21

log(x+1) transformation and the use of non-parametric statistics. I used the multi-

variant statistics “one way ANOSIM” which is a tool that uses non-parametric

multi-dimensional scaling (nMDS) to present the degree of similarity compared to a

null hypothesis of identity. 0 < R < 1 is defined as the degree of similarity, and

ranks from R = 0 (different) to R = 1 (identical). I used ANOSIM to test the

differences between the species compositions in different areas with PAST software

(Hammer et al. 2001). All other analyses were done with STATISTICA7 software.

Results were considered significant at p < 0.05.

2.3 Results

Bat recordings were obtained from nine sites, during 10 months (February-

November) in 2009. Data from ten nights were lost due to vandalizing of equipment

or technical problems. Therefore, data are available from 80 full night recordings

(Table 2.2). Distribution of sampling nights between seasons was even, except for

the Zin site where data for the spring was lost. Partial recordings continued in 2010

and 2011 at the Alpaca farm for studying the long term effects of the new water

source.

Table 2.2 – The number of night recordings for which data were available from

each site during survey of 2009 in the Ramon region, Israel:

Central Negev

Highlands

Mitzpe Ramon

area

Makhtesh

Ramon

Total nights

per site type

Natural Lotz

7

Hemet

9

Saharonim

9

25

Artificial Nitzana

11

Alpaca farm

9

Afor

10

30

Wadi without

water

Dorban

10

Zin

6

Ramon

9

25

Total per Area 28 24 28 80

22

2.3.1 Total bat activity and the activity level of Kuhl’s pipistrelle in the

Ramon region

Activity varied considerably among nights and sites (min = 0, max = 59.88).

Variations among the different sites were considerable (max in Afor = 14.59, and

min in Zin = 0.03, Mann–Whitney U-test, p = 0.0048, Table 2.3). Total activity was

positively correlated with Ta (Spearman rank correlation: r2 = 0.4, p = 0.0002).

Table 2.3 – Activity levels of Pipistrellus kuhlii, Eptesicus bottae, Hypsugo

bodenheimeri at each of the 2009 survey sites in the Ramon region, Israel.

Activity is presented in average pass per hour ±SD.

Area Site name P. kuhlii E.bottae H. bodenheimeri

A. Makhtesh

Ramon

1.Saharonim 0 2.68±3.21 5.43±4.56

2. Ramon 0 0.16±0.29 1.45±2.11

3. Afor 0.06±0.08 3.67±3.94 10.88±17.51

B. Central

Negev

HIghlands

4. Lotz 0.14±0.16 0 0

5. Dorban 2.13±4.01 0 0.01±0.03

6. Nitzana 8.64±15.02 0.18±0.50 1.20±3.91

C. Mitzpe

Ramon area

7. Hemet 1.82±2.63 0.14±0.26 0.03±0.06

8. Zin 0.015±0.037 0.02±0.05 0

9. Alpaca 0.013±0.04 0.05±0.09 0.01±0.03

Total 1.68±6.29 0.83±2.17 2.305±7.27

Kuhl’s pipistrelle was highly active in the areas of the Central Negev Highlands

(mean activity level= 4.19 ± 10.14 pass/hour), less so at the Mitzpe Ramon sites

(0.67 ± 1.79), and nearly absent from the Makhtesh sites (0.02 ± 0.05). In fact, only

two Kuhl’s pipistrelle passes were recorded with certainty south of the Makhtesh

cliff during the entire year of the survey and in additional recordings during the

following year (Table 2.3 and Figure 2.3). The differences between the Central

Negev Highlands and the Makhtesh areas were significant (two-tailed Kruskal-

23

Wallis ANOVA on transformed data: H1,55 =16.90, p = 0.0003). The differences

between the Central Negev Highlands and Mitzpe Ramon were also significant

(H1,51 = 6.42, p = 0.019), but the differences were not significant between Mitzpe

Ramon and the Makhtesh (H1,51= 2.07, p = 0.30).

Figure 2.3 – Average activity levels ±SD of Pipistrellus kuhlii throughout

the survey in the three different areas of the Ramon region, Israel. Different

letters above bars indicate significant differences (Kruskal-Wallis ANOVA

p< 0.02).

Species composition was significantly different between the High Negev

Mountains and the Makhtesh Ramon (one-way ANOSIM: p < 0.002, R = 0.73).

Kuhl’s pipistrelle dominated the Central Negev Highlands community and was

absent from the Makhtesh, while Bodenheimer’s pipistrelle dominated the

Makhtesh community and was less common in the Central Negev Highlands and

Mitzpe Ramon areas. The community of the Mitzpe Ramon area was intermediate

A

B B

0

1

2

3

4

5

6

7

8

Area

Pas

ses

/ ho

ur

Central Negev Highlands Mitzpe Ramon area Makhtesh Ramon

Pas

ses

/ h

ou

r

24

in the proportion of each species between the Central Negev Highlands and

Makhtesh communities, but was closer in its composition to the community of the

Central Negev Highlands (one-way ANOSIM: p < 0.002, R = 0.08) than to the

Makhtesh community (one-way ANOSIM: p < 0.002, R = 0.47) (Figure 2.4).

Figure 2.4 – The species composition of bats from the "background cluttered

space" guild, is illustrated by the proportion of each species from the total

activity in each area of the Ramon region, Israel. The three assemblages differ

significantly from each other (ANOSIM, p < 0.002).

25

2.3.2 Activity levels in the different site types

Total activity levels varied between the different site types, with the highest

total bat activity at the artificial water sources (8.57 ± 14.87 pass/hour), lower at the

natural water sources (3.67 ± 5.25), and lowest at the dry wadi sites (1.45 ± 2.89)

(Figure 2.5). Differences were significant only between the natural water sources

and the dry sites (two-tailed Kruskal-Wallis ANOVA H1,49 = 4.25, p = 0.043).

Figure 2.5 – Total activity levels for all three species (Pipistrellus kuhlii,

Eptesicus bottae and Hypsugo bodenheimeri) at different types of sites in

the three areas of the Ramon region, Israel. Letters above bars indicate

significant differences.

Kuhl’s pipistrelle’s activity levels, excluding the Makhtesh area from which it

was absent, were highest in artificial bodies of open water (4.76 ± 11.75), but were

not significantly different from the activity levels at natural water sources (1.09 ±

A

B

26

2.11) or dry wadi sites (1.34 ± 3.28). Two-tailed Kruskal-Wallis ANOVA on

transformed data: H1,51= 0.34, p = 0.89.

2.4 Discussion

2.4.1 Species composition of the "background-cluttered space" guild in the

Ramon region

The factors determining species composition at a given location can be divided

into large scale processes and local factors. The large scale processes, such as

geographical and historical phenomena, set the upper limit on species diversity,

while local factors, such as environmental characteristics and biological

interactions, can determine the site-specific species composition (Gaston 2000,

Ricklefs 2004 for all species; Ford 2005 for bats). The activity levels of all bat

species surveyed in the Ramon region followed the well-established correlation

with Ta (for instance, Williams 1961, Hayes 1997). This correlation is associated

with the seasonal pattern of activity levels that increases with the advancement of

the spring and summer, and then drops in the autumn (Korine and Pinshow 2004).

The climatic requirements of the Kuhl’s pipistrelle enable its presence in all of the

Negev (Yom-Tov and Kadmon 1998) on a large scale, but it seems to be absent

from some habitats, at least from Makhtesh Ramon as shown in Table 2.3. The

other two desert-dwelling species, Botta’s serotine and Bodenheimer’s pipistrelle,

were found in all surveyed areas. This could be due to water availability, which is

much lower in the Makhtesh, where there are no artificial water sources, and where

the hotter and dryer conditions may induce greater drinking needs than in the

northern areas of the region. In addition to the limited number of bodies of open

27

water available for the Kuhl’s pipistrelle to drink from in this region, its absence

may also be associated with interspecific competition over the use of the few bodies

of water that are available (Razgour et al. 2011). It seems that in areas that have

sufficient water available within the home range of Kuhl’s pipistrelle’s populations

their activity is not limited to the water sites alone, but they are active in dry sites as

well (Figure 2.5). This is an important observation since it may indicate that water

sources that can support Kuhl’s pipistrelle populations in dry habitats may become

stepping stones for range expansion of the Kuhl’s pipistrelle to surrounding dry

sites as well.

The choice of sites for this survey was far from optimal. The scarcity of bodies

of open water in the study area forced me to compare sites, without an effective

method to control for important factors, such as the isolation of each site, its

distance from roosts, or the flight paths that may be of more importance than actual

aerial distance. For example, the Hemet site is closer to the two Mitzpe Ramon sites

than to any Central Negev Highlands site, but it is drained by Wadi Nitzana that

shares its drainage basin with the two Mitzpe Ramon sites. Since it is the site with

the highest activity in the Mitzpe Ramon area, changing its classification would

affect the analysis. Unexpected disturbances to ponds (for example: several dry outs

of the troughs) and to equipment (for example: a bat recorder was vandalized at the

Zin site causing data loss and the changing of the site location), resulted in a

reduced sample size and some questionable reliability of some data points. The

results are also limited by the small number of ponds - just one of a type for each

area.

28

2.4.2 Range expansion of Kuhl’s pipistrelle in the Ramon region

With caution, I assume that, at least at the time of the study, Kuhl’s pipistrelle

had not yet established populations in Makhtesh Ramon. This may change in the

future. The phenomenon of time lag between the introduction of an invasive species

and the time of population outburst has been documented for many organisms,

including several vertebrates, and was explained by either changes in the

environment or genetic adaptations to the new environment (reviewed by

Simberloff 2010). The presence of a permanent body of water (NPA trough) and

development plans for visitor facilities in Makhtesh Ramon, which include an

artificial lake (NPA project manager, personal communication), may be enough to

allow Kuhl’s pipistrelle to further expand its range into areas from which it is yet

absent in the Negev. This threat is supported by the limited data collected from the

trough I set in the Alpaca Farm. The trough was monitored irregularly for bat

activity during the three years of its existence, and Kuhl’s pipistrelle activity was

apparently higher each succeeding year.

Another type of disturbance that is associated with anthropogenic development

and that gives a competitive advantage to Kuhl’s pipistrelle over desert-dwelling

bats is artificial light. Kuhl’s pipistrelles forage near street lamps while the Botta’s

serotine avoids the light (Polak et al. 2011). The Negev has been undergoing

intensive development over the past decades and more development is planned for

the coming years (Avigdor 2004). Anthropogenic development of the Negev, as

manifested by water developments and artificial illumination, provides at least two

known benefits to Kuhl’s pipistrelles. This process may support the establishment

of Kuhl’s pipistrelle populations around settlements and even in rural areas, and

29

may be causing increased competition among local desert-dwelling bat

communities (Razgour et al 2011), sharing similar ecological requirements

(Feldman et al. 2000, Korine and Pinshow 2004) .The abundance of Kuhl’s

pipistrelle in all of Israel’s climatic areas enables this species to expand its range to

all habitats, opposed to the rest of its guild that are more limited in their distribution

(Yom-Tov and Kadmon 1998). I assume that not much can be done on the regional

scale to prevent the competition of invading Kuhl’s pipistrelle with desert-dwelling

species, and the knowledge acquired on the importance of pond characteristics in

shaping the local species composition (Razgour et al. 2010), encouraged me expand

my research to identify possible management schemes that will limit the

distribution of Kuhl’s pipistrelle. Such an approach is described in Chapter Three.

30

3 Chapter Three – Managing bat communities' species composition by placing

obstructions to hamper Kuhl’s pipistrelle’s ability to drink from desert ponds

3.1 Introduction

3.1.1 Management of invasive bat species

Bat conservation is traditionally concerned with the protection of endangered

bat species by protecting either their roosting sites by placing gates at the entrance

of caves to prevent cave activities during the breeding or hibernation periods

(Martin et al. 2000), preventing habitat loss (Russo and Jones 2003), or by

conserving major foraging habitats (Sparks et al. 2005). Several bat conservation

plans have considered both roosting and foraging habitat protection (Schmidt 2003,

Ellison et al. 2003, Wermundsen 2010). Alien invasive species are considered a

major threat to biodiversity (Simberloff 2010), and bat species are threatened by

invasive terrestrial mammals such as rats (McClelland 2002). Almost all

conservation research has focused on the threats to bat species and their protection

needs. However, as diverse and abundant as they are worldwide, no insectivorous

bat has ever been declared an invasive species. This is mainly because, as opposed

to other vertebrates that humans tend to deliberately relocate (pets etc.), bats lack

such a relationship with humans in most cultures and are not deliberately

transported (Clout and Russell 2008). This might explain the absence of literature

regarding the management of invasive bat populations.

There are a few exceptions of studied cases of bats endangering other bat

species (Baagøe 1986, Arlettaz et al. 2000). In these cases, the population growth or

range expansion of the threatening populations were due to their use of artificial

31

lights for foraging, thus obtaining a competitive advantage over bats that did not

benefit from this disturbance. The same phenomenon was observed by Polak et al.

(2011) studying Kuhl’s pipistrelle in the Negev Desert. Both Arlettaz et al. (2000)

and Polak et al. (2011) suggested the control of light pollution and habitat

preservation as management tools. However, these studies did not deal with the

competitive effect of the threatening species beyond the disturbed areas. Kuhl’s

pipistrelle may pose a threat to local species beyond the surroundings of bodies of

water (Section 2.4) and therefore local bat communities may benefit from

management of Kuhl’s pipistrelle as an invasive species. Due to the lack of studies

on invasive species of bats, I explored the literature on invasive vertebrates and

birds, in particular.

The most effective approach when confronting the dangers of invasive species

is prevention -- either of the introduction itself or of further range expansion

(Wittenberg and Cock 2001, Simberloff 2010). Once an invasive species has been

established, and especially one with high maneuverability such as flying vertebrates

that may be impossible to keep away, constant monitoring and effective

management may still be useful in reducing the invader’s impact on local

ecosystems (Wittenberg and Cock 2001, Simberloff 2010). Any form of

management must take into account the species’ effect on other species in the

habitat and the interspecific interactions with other endangered species (Gumm et

al. 2011). The hierarchy of management tools suggested both by Wittenberg and

Cock (2001) and Simberloff (2010) starts with the eradication of the invader as the

most effective tool, when possible. The eradication of large scattered populations of

flying vertebrates is not easily carried out. In Europe for example, international

32

efforts for over 20 years to eradicate the American Ruddy duck (Oxyura

jamaicensis) that poses a threat to local duck species in several European countries,

are unsuccessful (Genovesi 2005). I found only one documented case of successful

bird eradication, on private farm land in Nevada Starlings (Sturnus vulgaris) had

competitively excluded local birds, and with their removal, the local bird

community has restored itself naturally (Weitzel 1988). Containment of the

invasion in a well-defined territory or exclusion from vulnerable habitats have been

discussed mostly in regard to slow growing or slow moving organisms (Brown and

Carter 1998) and are not feasible tools for flying vertebrates (Tidemann 2001).

When none of the above methods of controlling the population size and range of the

invader are feasible, mitigation has been used. Mitigation focuses on the native

species and involves taking measures to reduce the effect of the invader on native

endangered species, rather than confronting the invading species population directly

(Wittenberg and Cock 2001). In some cases, as with the common myna

(Acridotheres tristis) in Australia (Tidemann 2001), and with the protection of sea

birds (Wilcox and Donlan 2007), means of population control were used as a

successful type of mitigation. This type of management often involves a

development of artificial roosting or feeding sites for local species when threatened

by feeding competition or habitat degradation. Mitigation by the construction of

artificial roosting sites is a common practice in bat conservation, but in the context

of habitat loss and not interspecific competition (Appleton 2003, Briggs 2004).

Management practices designed to keep only specific animals away are found in the

form of fences (Andrew et al. 1997, Larsen et al. 2011) but are irrelevant for bats.

33

The activity of Kuhl’s pipistrelle is repeatedly found to be at higher levels than

any other bat species in the Negev (Korine and Pinshow 2004, Razgour et al 2010,

current survey figure 2.4), suggesting that it may be the most abundant

insectivorous bat in the region. Moreover, Kuhl’s pipistrelle is the only bat species

in the Negev that is assumed to have increased its population size over the past

decade (2000-2010) (Shalmon 2010, Korine unpublished data 2011). If the species

richness of Negev bats is to be conserved, the increasing competitive load presented

by populations of Kuhl’s pipistrelle on the rest of the desert-dwelling bat

community (Polak et al. 2011, Razgour et al 2011) requires management. Since

direct reduction of the competitive load of Kuhl’s pipistrelle by eradication or

population control methods are hardly possible, mitigation can be used to locally

reduce Kuhl’s pipistrelle activity and provide local ìcompetition freeî sites for

desert-dwelling bats.

3.1.2 Interspecific differences in the use of bodies of water as the key for the

management of desert bat communities

All four species of the “background-cluttered spaceî guild -- Kuhl's pipistrelle,

Ruppell's pipistrelle, Bodenheimer's pipistrelle, and Botta's serotine-- share the

same families of insects in their diet (Feldman et al. 2000) and foraging behavior

(Korine and Pinshow 2004). Considering that Kuhl’s pipistrelle in the Negev is an

invasive species, it seems as if any attempt at eradication or population control will

necessarily harm the endangered desert-dwelling bats. One possible mitigation

scheme may therefore lie within their different drinking behaviors, which may

influence local species assemblages.

34

Bats worldwide have been found to leave day roosts after sunset and commute

to productive foraging sites, peaking their activity levels within the first hours after

sunset and then again near dusk (Hayes 1997, Nicholls and Racey 2006). In arid

habitats, open bodies of water are preferred by bats as foraging and drinking sites

(O'Farrell and Riddle 2006, Razgour et al. 2010). For example, Myotis species in

arid habitats, were found fly shortly after sunset to nearby water holes (Adams and

Thibault 2006). The spatial distribution of foraging usually consists of commuting

from the roost to a patchy foraging site where longer amounts of time are spent and

finally returning to the same roost (Nicholls 2006). Bat species with similar diets

and morphology (body mass 4-8g) to my study species were studied in the British

Islands: the Common pipistrelle (P. pipistrellus) and the Soprano pipistrelle (P.

pygmaeus). Both species were found to commute between 0.5 and 3 km (3.8 km

max) from roosting sites to foraging areas (Nicholls and Racey 2006). The time

allocation between foraging sites or patches is determined by the resource value of

the patch, compared to the costs of obtaining the resource (Arlettaz 1996). Both

values depend on bat population density and can be influenced by competition,

promoting spatial partitioning (Kunz 1973, Arlettaz 1999, Razgour et al. 2011) and

temporal partitioning (Kunz 1973, Adams and Thibault 2006, Razgour et al. 2011)

between bat species with similar diets and foraging habitats.

Bats in the Zin Valley in the Negev Desert differ in the way they use ponds

(Razgour et al. 2010, 2011), with the Mediterranean species (the European free-

tailed bat and Kuhl’s pipistrelle) drinking seven and four (respectively) times more

frequently than the desert-dwelling bat species. Competition over the use of bodies

of water is found between the two sets of bats (Razgour et al. 2011): 1) The

35

Mediterranean species, the European free-tailed bat and Kuhl’s pipistrelle, compete

for drinking space over bodies of water, and seem to be obliged to drink. 2) Within

the “background-cluttered space” guild, Kuhl’s pipistrelle, Bodenheimer’s

pipistrelle and Rueppell’s pipistrelle compete for foraging space over bodies of

water. This competition results in spatial partitioning between the Kuhl’s pipistrelle,

which is associated with small temporary ponds, and the European free-tailed bat,

which is associated with large permanent ponds and tends not to visit medium and

small temporary ponds. These two species are temporally partitioned, with their

activity level speaking at different times. Similar partitioning was found among the

“background cluttered space” species that compete for foraging space. Rueppell’s

pipistrelle was found to be associated with medium ponds and its activity peaks in

the fifth and sixth hours of the night. Bodenheimer’s pipistrelle is associated with

large permanent ponds and its activity peaks during the first and last hours of the

nightand Kuhl’s pipistrelle prefers small temporary ponds and its activity peaks

between the second and fourth hours of the night. Razgour et al. (2011) suggested

that these partitioning mechanisms may prevent the expected competition due to the

high niche overlap between the two sets of competing bat species.

The Mediterranean species drink more frequently than the desert-dwelling

species, and therefore, the Kuhl’s pipistrelle's competitive load may be higher in

foraging sites that allow drinking. Several studies have reported reduced drinking

activity when the surface area of the water body was reduced in size experimentally

(Tuttle et al. 2006, Razgour et al. 2010), or when the swoop zone was limited

(Tuttle et al. 2006). Nonetheless, in both studies, the manipulations caused an

increase in the total activity at the sites, possibly due to repeated unsuccessful

36

drinking attempts. The different drinking frequencies of Mediterranean and desert-

dwelling bats, reported by Razgour et al. (2010), may suggest that different species

are affected more than others by such manipulations of drinking area. Tuttle et al.

(2006) do not differentiate between the responses of different species to the

manipulation, but Razgour et al. (2010) do. They found that the manipulation,

leaving 8 m of pond free for drinking, reduced the activity of all four species of the

ìbackground-cluttered spaceî guild. These results may have been due to the short

and intensive periods of disturbance caused by the manipulation. The results do not

shed light on the question of whether a stronger manipulation, preventing Kuhl’s

pipistrelle from drinking, would have a stronger effect on Mediterranean species.

My experimental approach was an attempt to show whether such a disturbance can

prevent Kuhl’s pipistrelle from drinking, thus reducing competitive loads from the

desert-dwelling species, which will benefit from a higher value foraging site.

I hypothesized that placing obstructions above open bodies of water would

hamper the drinking ability of Kuhl’s pipistrelle and would have a lesser effect on

desert-dwelling bat species since the former drink more frequently than the latter.

The resulting reduction in Kuhl’s pipistrelle activity would create "competition

free" foraging sites for desert-dwelling bats.

I did three different field experiments and tested the following predictions:

1) Increasing the density of obstructions over a body of water, used only by

Kuhl’s pipistrelle and only for drinking, reduces the drinking frequency of the bats

and, eventually, prevents it.

37

2) The activity level of Kuhl’s pipistrelle (measured by their echolocation

calls) is higher in un-manipulated ponds (explained hereinafter) than in the same

ponds, once manipulated. The activity level of Ruppell’s pipistrelle,

Bodenheimer’s pipistrelle and Botta’s serotine is higher in manipulated ponds

(explained hereinafter) than in the same un-manipulated ponds due to reduced

competition from Kuhl’s pipistrelle. Field experiment II (manipulation of natural

ponds in the Zin valley (Section 3.2.2)), and the following year field experiment

III (manipulation of isolated natural ponds (Section 3.2.3)) were designed to test

this prediction.

The manipulation was formed by obstructions constructed over open bodies of

water to hamper bat drinking, as is explained in detail (Section 3.2.1).

3.2 Methods

3.2.1 Field experiment I: Manipulation of the swimming pool

Study area and site

This experiment took place in Midreshet Ben Gurion in the Negev Desert, in the

summer of 2010. In this desert region rain occurs during the winter and averages 97

mm per year. Ta are highest during summer and are lowest in winter, with a daily

mean of 25.3 ºC in August, and 9.7 ºC in January in Midreshet Ben-Gurion

(Meteorology unit BIDR 2010). The study site used was the local swimming pool,

25 m by 25 m and ‘L’ shaped (Figure 3.1), located at 30°51'13.57"N, 34°47'08.53"E

and at 478 m elevation. This pool was chosen since Kuhl’s pipistrelle had

previously been observed drinking in it and it is located ~ 100 m from a known

Kuhl’s pipistrelle roosting site (Berger-Tal et al. 2008). Kuhl’s pipistrelle is

38

apparently the only species that uses this site (Korine, personal communication, and

verified by vocalizations recordings during the experiment).

Experimental design

The swimming pool was covered with a grid of black Polypropylene string with

an average diameter of 2 mm, limiting the water surface to sections of 1 1 m. This

was done to prevent the Kuhl’s pipistrelle from drinking from the pool. The

experimental plot was created by leaving the non-experimental area covered by the

string grid and uncovering a 9 9 m experimental area (Figure 3.1). Four different

treatments were done on the 9 9 m section of the pool. For the control treatment,

the experimental area was left uncovered, and for the other treatments, this area was

sectioned into varying sized squares: 4.5 4.5 m, 3 3 m and 1 1 m (Figure 3.1).

The string was elevated 10 cm above the surface of the water.

12.5m

25m

1 m

3 m

4.5 m

Figure 3.1 A diagram of the

swimming pool, at which the

experiment was carried out, is

shown. It measured 25 m in length

and 12.5 m in width. The black line

represents the 1 1 m fixed grid

covering the non-experimental area.

The area without a black line

represents the 9 9 m experimental

area (also used as the 9 9 m control

treatments). The other colored lines,

green (1 1 m), red (3 3 m) and

blue (4.5 4.5 m) represent the three

different treatments.

39

The four treatments were scheduled to be done at different times over five

evenings, to control for the fact that bat activity varied at different times throughout

the night. This also controlled for possible affects on drinking due to a preceding

treatment, as each treatment was preceded by a different treatment on each of the

nights. The experiment began each night 30 min after sunset (~19:30), and the

periods used for the different treatments each night were: 20:10 - 20:30, 21:30 -

21:50, 22:10 - 22:30 and 22:40 - 23:00.

I counted the number of passes and drinking events over the experimental area

in a set period of time. This was done by direct un-aided observations. In all of the

trials, there was one artificial light source, set 10 m away from the pool, and

observers used head lamps. I used an AnaBat detector to monitor the experimental

area acoustically. This allowed an auditory prompt for bat arrival, and the data

recorded were analyzed using AnaLookW (3.3) software to ascertain that all

observations were of Kuhl’s pipistrelle.

Data analysis

Since the data on the number of drinking events were not normally distributed,

their variance among treatments lacked homogeneity and since many ‘0’ values

were recorded, some of the statistical tests required transforming the activity data

by log(x+1). A repeated-measures ANOVA was used to determine whether there

was a difference between nights. When the different nights were found to have no

effect on bat activity, a one-way ANOVA was used to determine whether there was

a difference between treatments. All tests done for all three field experiments were

two-tailed, and the results were considered to be significant at p < 0.05. All tests in

the three field experiments were done using STATISTICA 7.

40

3.2.2 Field experiment II: Manipulation of natural ponds in the Zin Valley

Study area and sites

This field experiment examined drinking ability of bats from manipulated ponds

in areas where Mediterranean bat species are well established – the Zin Valley in

the proximity of Midreshet Ben-Gurion (30º52' N, 34º47' E) in the Negev Desert

Highlands (Climatic description in section 3.2.1). The field experiment took place

in temporary and permanent open ponds located in the Zin and Akkev Wadis. In

both wadis, springs sustain permanent ponds all year round, while temporary ponds

dry out during the spring or the beginning of the summer (Table 3.1).

Table 3.1 – Sites chosen for experiment II. All sites are located within the Zin

Valley in the Negev, Israel. The measures marked by * changed during the

three week period of the experiment. The values in the table are the measured

maxima at the time of the experiment.

Site name Description Compass

Coordinates

Pond

length

(m)*

Distances

from nearest

pond(m)*

A.White Cliff East Temporary pond

inWadi Zin

30°50'37.69"N

34°47'24.29"E

8 50

B. Akkev East Temporary pond in the

Akkev Wadi

30°49'35.38"N

34°48'38.82"E

7 10

C. White Plains A Temporary pond in the

Akkev Wadi

30°49'54.71"N

34°48'46.51"E

19 5

D. Bor Nabati Temporary pond in the

Zin Wadi

30°50'31.50"N

34°48'42.32"E 10

20

E. White Plains B Temporary pond in the

Akkev Wadi

30°49'50.54"N

34°48'46.61"E 15

0

F. White Cliff West Temporary pond in the

Zin Wadi

30°50'36.53"N

34°47'22.37"E 5

50

G. Australia Pond Temporary pond in the

Zin Wadi

30°50'30.24"N

34°46'37.11"E 6

150

H. Conglomerate Pond Temporary pond in the

Zin Wadi

30°50'14.24"N

34°46'30.17"E 8

50

J. Heart Pond Temporary pond in the

Akkev Wadi

30°49'3.59"N

34°48'45.04"E 3

30

K. Salon Pond Temporary pond in the

Zin Wadi

30°50'38.82"N

34°46'59.49"E 3

5

L. Akkev Horseshoe Temporary pond in the

Akkev Wadi

30°49'33.13"N

34°48'35.42"E 5

50

41

Figure 3.2 - Field experiment temporary pond sites in the Zin Valley.

Experimental design

From among many temporary ponds in the Zin Valley that were filled with

water by the winter floods of January 2010 (Israel Meteorology Service 2010), I

chose 11 that held water for several months. This allowed for the experiment to

begin in April when bat activity levels are sufficiently high. I chose ponds that were

large (mean pond length: 7.67±4.96 m) and deep enough (>15 cm), so I could

expect them not to dry out during the three weeks of the experiment, and that were

at least 3 m long to ensure that they would facilitate drinking activity and not only

foraging. The ponds were not isolated (mean distance from nearest other pond:

47.5±51.67 m, min=0 m, max=150 m), but were chosen to make distances between

ponds as far as possible, to prevent calls from a nearby pond from being mistakenly

accounted for in the analysis. The experiment was done in three ponds

simultaneously, but I manipulated only one pond at a time (Table 3.1). This was

42

done to control for temporal factors