Modelling the Impact of Roads on Regional Population of Moor Frogs

) Rana arvalis Rana Maj-Britt Pontoppidan Department of Biology Modelling the impact of roads on Modelling impact the on of roads of Moor population frogs regional ( university copenhagen of

Maj-Britt Pontoppidan Modelling the impact of roads on regional populations of Moor frogs (Rana arvalis) ) Rana arvalis Rana Maj-Britt Pontoppidan populations Modelling the impact on regional of roads ( of Moor frogs PhD Thesis Thesis PhD 2013/1 ISBNXXX-XX-XXXXX-XX-X FACULTY OF SCIENCE UNIVERSITY OF COPENHAGEN

PhD thesis

Maj-Britt Pontoppidan Section of Ecology & Evolution Department of Biology

Modelling the impact of roads on regional populations of Moor frogs (Rana arvalis)

A thesis submitted to the University of Copenhagen in accordance with the requirements for the degree of the PhD at the Graduate School of Science, Faculty of Science, University of Copenhagen, Denmark to be defended publicly before a panel of examiners

Academic advisor: Gösta Nachman

Submitted: January 2013

Preface

In this thesis I present my work carried out during a 3-year PhD fellowship funded by the Danish Road Directorate. The objective of the project has been to develop a management tool for assessment of the impact of roads on Moor frog populations. During my PhD-work, I have been based at the Department of Biology, Section for Ecology and Evolution and I have been supervised by Dr. Gösta Nachman.

The end product of the project is an individual based model called SAIA (Spatial Am- phibian Impact Assessment). The model has evolved in close and continuous dialogue with the project group, which contains members from the Danish Nature Agency, the Road Direc- torate as well as specialists on environmental impact assessments (EIA) and amphibians. Not being a herpetologist nor road ecologist myself, there is always the danger that the dazzling model you have come up with is just a tiny bit far reached. During the design process, it has been extremely important for me continuously to have the opportunity to give my model and its components a "reality check". Hence, discussions in the project group with amphibian experts and end users on the validity and usefulness of the model behaviour and output has been an integrated part of the model development.

The thesis consists of a synopsis and three chapters. In the synopsis, I give an overview of the background theory and the model development. The chapters each contain a manuscript submitted to a scientific journal. The manuscript in chapter one has been peer reviewed and the enclosed version is now under revision. The two remaining manuscripts are in the process of peer review. The appendix contains examples of SAIA’s output files.

Maj-Britt Pontoppidan

Copenhagen, January 2013

4 Index

Index

ACKNOWLEDGEMENTS ...... 7

ENGLISH SUMMARY ...... 9

DANSK RESUME ...... 11

SYNOPSIS ...... 13

BACKGROUND ...... 15

Fragmentation ...... 15

Connectivity ...... 16

Objective ...... 18

DESIGNING SAIA ...... 19

Conceptual model ...... 19

The habitat patch ...... 21

Dispersal behaviour ...... 22

SAIA v1.0 ...... 23

CONCLUSION ...... 25

REFERENCES ...... 27

CHAPTER ONE ...... 33

CHAPTER TWO ...... 59

CHAPTER THREE ...... 95

APPENDIX ...... 143

6 Acknowledgements

Acknowledgements

This project would not have been possible without the goodwill and assistance of many peo- ple. I would like to express my heartfelt thanks to all of them: to the Danish Road Directorate for funding the project and giving me this wonderful opportunity. to the members of the project group: Marianne Ujvari, Martin Schneekloth, Agnete Jør- gensen and Martin Hesselsøe. It’s been a joy working with you. to AmphiConsult for sharing your expertise with me. to Volker Grimm and Uta Berger for introducing me to the intriguing world of NetLogo and Individual Based Modelling as well as the beautiful region of Swiss Saxony. Special thanks to Uta for encouraging and inspiring talks and for keeping me on the IBM-track. to Bjørn Hermansen for patiently helping me with GIS. to Henning Bang Madsen and Ruth Bruus Jakobsen for your readiness to help out and your generous limousine service. to Marianne Philipp for providing refuge in stressful times and for sharing your anemones with me. to all my colleagues at the section of Ecology & Evolution for good company and for so gen- erously letting me pick your brains and books.

And, last but not least, to my supervisor Gösta Nachman for embarking on this journey with me and for always having an open door. I’ve enjoyed our time together and I’ll miss all your anecdotes.

8 English summary

English summary

Over the last decade a growing amount of literature has documented the severe impacts of transport infrastructure on biodiversity, population persistence and gene flow, and there is an increasing awareness of the importance of finding agreement between nature conservation and land use. To ensure ecologically sustainable road planning conservation measures must be taken into consideration already in the earliest phases of road development. This requires adequate tools for assessment, prevention and mitigation of the impacts of infrastructure. For this reason the Danish Road Directorate decided to finance a PhD project with the objective of developing a management tool that could be used to substantiate that the conservation status of annex IV species would be unaffected by the a given road project. The purpose of the project was to provide a standardized and scientifically well founded basis for decisions concerning road lay-out and mitigation measures. As model species was chosen the Moor frog (Rana arvalis). Populations of Moor frogs are assumed to follow a pattern of metapopulation dynamics, with colonisation, extinction and recolonisation of suitable habitat patches. Thus, road constructions must be expected to have implication on both local and regional persistence; the former due to habitat destruction, the latter because of disrupted dispersal between subpopulations due to barrier effects.

The result of the project was the development of the model presented in this thesis. The model, called SAIA (Spatial Amphibian Impact Assessment), considers a landscape mosaic of breeding habitat, summer habitat and uninhabitable land. As input I use a GIS-map of the landscape with information on land cover. In addition, data on observed frog populations in the survey area are needed. The dispersal of juvenile frogs is simulated by means of individual-based modelling, while a population-based model is used for simulating long-term population dynamics. In combination, the two types of models generate output on landscape connectivity and population viability. To assess road impacts two scenarios have to be constructed and analysed. The first scenario should be a map of the area as it is before the planned road construction (scenario 0). This analysis measures the ecological performance of the original landscape and is a reference against which other scenarios are to be compared. The second map (scenario 1) should show the landscape as it is expected to be after the road constructions. In combination, the analyses of scenario 0 and scenario 1 make it possible to assess the effect of road construction on connectivity and population persistence. The analyses

9 English summary also constitute the basis for planning of mitigation measures. Analyses and comparisons of several alternative road projects can identify the least harmful solution. The effect of mitigation measures, such as new breeding ponds and tunnels, can be evaluated by incorporating them in the maps, thereby enhancing the utility of the model as a management tool in Envi- ronmental Impact Assessments.

The thesis consists of a synopsis and three manuscripts for scientific journals. An appendix contains examples of the result files SAIA produces. In the synopsis, I give an overview of the background theory and the conceptual model development.

In the first manuscript I introduce an alternative patch concept, the complementary habitat patch, and use a simple model to explore how intra-patch heterogeneity affects immigration and emigration probabilities. I find that the realised connectivity depends on internal structure of both the target and the source patch as well as on how habitat quality is affected by patch structure. Although fragmentation is generally thought to have negative effects on connectivity, the results suggest that, depending on patch structure and habitat quality, positive effects on connectivity may occur.

The second manuscript uses a light-version of SAIA and explores how changes in road mortality and road avoidance behaviour affect local and regional connectivity in a population of Moor frogs. The results indicate that road mortality has a strong negative effect on regional connectivity, but only a small effect on local connectivity. Regional connectivity is positively affected by road avoidance and the effect becomes more pronounced as road mortality decreases. Road avoidance also has a positive effect on local connectivity. When road avoidance is total and the road functions as a 100% barrier regional connectivity is close to zero, while local connectivity exhibit very elevated values. The results suggest that roads may affect not only regional or metapopulation dynamics but also have a direct effect on local population dynamics.

The third manuscript describes the full SAIA model. By means of a case study I demonstrate how SAIA can be used for assessment of road impact and evaluation of which management measures would be best to mitigate the effect of landscape fragmentation caused by road constructions.

10 Dansk resumé

Dansk resume

En stadigt tættere infrastruktur præger vores landskaber og er blevet en kraftig trussel mod biodiversiteten. Der er en stigende bevidsthed om nødvendigheden af at anlæggelse af veje må være bæredygtig. For at opnå dette er det nødvendigt allerede i de tidligste faser af vejpro- jekter at inddrage overvejelser omkring bevarelsesforanstaltninger. Dette kræver, at der er passende værktøjer til rådighed, hvormed det er muligt at vurderer effekten på naturen af så- vel den kommende vej som mulige afværgeforanstaltninger. På denne baggrund besluttede Vejdirektoratet at finansierer et PhD projekt, hvis formål var at udvikle et modelværktøj, der kunne underbygge et ensartet og fagligt baseret beslutningsgrundlag ved valg af linjeføring. Endvidere skulle værkstøjet kunne understøtte beslutninger vedrørende afværgeforanstaltnin- ger, deres antal og placering. Spidssnudet frø (Rana arvalis) blev valgt som model-art. En population af Spidssnudet frø antages at bestå af et netværk af delpopulationer, samt at følge en metapopulationsdynamik med kontinuert kolonisering, udryddelse og rekolonisering af egnede habitatområder. Nye vejanlæg kan forventes at påvirke en populations levedygtighed, både lokalt ved at ødelægge habitatområder og regionalt ved at virke som en barriere for spredning af individer mellem delpopulationerne.

Resultatet af projektet blev modellen som præsenteres i denne afhandling. Modellen, kaldet SAIA (Spatial Amphibian Impact Assessment), tager udgangspunkt i et landskab be- stående af en mosaik af ynglehabitat, sommer habitat og ubeboeligt habitat. Som input til modellen bruger jeg GIS-kort, der indeholder informationer om arealanvendelse. Derudover skal der bruges data på populationen af Spidssnudet frø i undersøgelsesområdet. Jeg bruger individ-baseret modellering til at simulerer spredningen af nyforvandlede frøer samt en populati- onsbaseret model til at simulerer populationsdynamikken i de enkelte populationer. Tilsam- men genererer de to modeller output om landskabets konnektivitet og om populationens levedygtighed.

For at kunne evaluerer konsekvenserne af et kommende vejanlæg kræves to scenarier. Det første scenarie fungerer som reference og skal være et kort over landskabet, som det ser ud før det planlagte vejanlæg. Det andet scenarie er et kort over landskabet, som det forventes at se ud, når vejprojektet er udført. Ved at sammenligne resultaterne fra de to analyser er det muligt at vurdere, hvordan vejanlægget vil påvirker landskabets konnektivitet og frø-

11 Dansk resumé populationens levedygtighed. Analyserne kan samtidigt danne basis for planlægning af af- værgeforanstaltninger. Analyser og evaluering af scenarie med alternative lineføringer eller forskellige afværgeforanstaltninger giver mulighed for identificerer de bedste løsninger og resultaterne kan indgå i f.eks. VVM-undersøgelser.

Afhandling består af en synopsis, tre videnskabelige artikler samt et appendiks det indeholder eksempler på de resultat-filer SAIA genererer. I synopsen giver jeg et overblik over den konceptuelle udvikling af SAIA-modellen samt den bagvedliggende teori.

I den første artikel beskriver jeg, hvordan et habitatområde kan betragtes som sammen- sat af forskellige habitat typer og introducerer et alternativt habitat begreb, det komplementæ- re habitatområde. Jeg bruger en simpel model til at udforske, hvordan sammensætningen af et komplementært habitat påvirker sandsynligheden for immigration til og emigration fra habita- tet. Resultaterne viser, at strukturen såvel som kvaliteten i et habitatområde har stor betydning for landskabets konnektivitet. Desuden finder jeg, at fragmentering under visse forhold kan have en positiv effekt på konnektiviteten.

I den anden artikel bruger jeg en forenklet version af SAIA til at afsøge, hvordan æn- dringer i vejdødelighed og dyrs evne til at undvige veje påvirke konnektiviteten mellem bestande, både lokalt og regional. Resultaterne viser, at vejdødeligheden har en kraftig negativ effekt på regional konnektivitet, men kun lille effekt lokalt. Afværgeadfærd har en positiv effekt på både regional og lokal konnektivitet - effekten er dog mest udtalt, når vejdødelighe- den er lav. Hvis afværgeadfærden er så kraftig, at vejen reelt fungerer som en 100 % barrierer, er den regional konnektivitet dog tæt på nul, mens det lokale konnektivitet opnår meget høje værdier. Resultaterne peger på, at veje kan påvirke populationsdynamikken både lokalt og regionalt.

Den tredje artikel beskriver den fulde SAIA model. Ved hjælp af et case-studie demon- strerer jeg ,hvordan modellen kan anvendes til vurdere effekten af et planlagt vejanlæg på en bestand af spidssnudet frø, samt hvilke afværgeforanstaltninger der kan modvirke effekten

SYNOPSIS

14 Synopsis

Background

Roads are everywhere. An extensive and expanding infrastructural network connects human activities; it enables us to reach the furthest parts of the world and gives us access to the resources we need. They bring us to our friends and family, our working places and recreational activities. We use them to go shopping and to enjoy a walk in the forest. But while infrastructure binds the human society together, roads also act as barriers – cutting through home ranges of animals and crossing their migration or dispersal routes. Roads restrict animals’ access to resources and affect their behaviour and movement patterns [1, 2].

Today’s huge network of roads is a major threat to many species. Animals trying to cross a road experience a very high mortality risk, and many populations of (mostly large) mammals, birds and amphibians are negatively affected by road killings [3-7]. Especially species with large ranges or with high mobility seem to be most vulnerable to road mortality [8, 9]. Increased mortality caused by roads may not only reduce population sizes, but – at least in amphibians – also shift age distributions toward younger age classes resulting in reduced reproduction [10]. Even though road killings reduce road crossing some movement across the road may occur; the barrier effect will not be total unless road mortality is 100%.

Many species are able to detect roads and thereby actively avoid them [11]. While this behaviour reduces the risk of road-killing, it also limits access to resources and further isolates populations on one side of the road [12]. Thus, animals’ behavioural responses to roads may enhance the barrier effect of the road. Consequently, road effects on population persistence may depend on the interaction between road mortality and road avoidance [13, 14].

Fragmentation The barrier effect of roads fragments the natural habitat of species; consistently dividing continuous areas of habitat into smaller and more isolated fragments. Today, fragmentation and habitat loss are considered the greatest threat to biodiversity and population persistence [15]. All else being equal, habitat loss will reduce the amount of resources and consequently also population sizes [16]. Furthermore, fragmentation divides a population into several smaller subpopulations. Small populations are more vulnerable to environmental and demographic stochasticity and thus have a higher risk of extinction. [17-20] Conversion of a continuous habitat area into several smaller also results in more edge area and a higher perimeter:area

15 Synopsis ratio. This affects the quality of a habitat patch and the area effectively available to a population may be reduced [21]. The fragmentation process changes the landscape composition by substituting habitat with non-habitat, and this may affect the movements of animals. Individu- als may not move into non-habitat at all, or if they do the changes in the spatial arrangement of habitat fragments may increase the time it takes to find resources, sometimes causing significantly higher transit mortalities [22-26]. Hence, fragmentation impedes movement and isolates habitat fragments from each other. However, movement is an important part of many organisms’ ecology. Individuals need to move to find necessary resources such as food, pro- tection, mates, breeding sites or space, and the success in finding these resources will determine the density and distribution of a population [27, 28]. Thus, the viability of a population will depend on how well resource patches are linked together, and the term “connectivity” is frequently used to describe the strength of those linkages.

Connectivity Connectivity depends on the patchiness and the spatial structure of the landscape and it is therefore a central concept in “spatial” disciplines like Metapopulation ecology and Land- scape ecology [29, 30]. Even though both disciplines are concerned with the effect of connectivity on population persistence, their focus is not quite the same [31].

Metapopulation ecology considers populations consisting of a network of subpopulations. Some or all of the subpopulations repeatedly experience decreasing population sizes or even extinction and may be rescued or recolonized by immigrants from neighbouring subpopulations. The persistence of the whole population relies on the dispersal of individuals among subpopulations. Within the metapopulation framework subpopulations are considered to inhabit patches of homogenous habitat embedded in a homogenous matrix of non-habitat. Connectivity is regarded as the probability of a local population receiving an immigrant from another population. In its basic form connectivity depends on the distance to and the size of the donor population (often measured as the area of the habitat patch) [29, 32-35], and thus connectivity is defined as a property of the subpopulation (or habitat patch).

In Landscape ecology the focus is on the composition of the landscape. Habitat patches do not exist in a homogenous background matrix but is part of a landscape mosaic of different habitats and structures [27]. The movement paths of individuals depend on the spatial arrangement of habitats. Some habitat types are avoided, others are preferred; structures may

16 Synopsis obstruct accessibility or incur high mortality. Thus, accessibility of resource patches does not only depend on distance but also on the properties and configuration of the matrix in between patches [22, 36]. Landscape ecology defines connectivity as the degree to which the landscape facilitates or impedes movement among resource patches [30]. Connectivity is regarded as a landscape property describing the permeability of the mosaic, and opposite to metapopulation ecology does not contain any demographic indicators [37].

Maintaining connectivity is generally regarded as an essential goal of environmental conservation [38], and methods for quantifying connectivity are, thus, important. Improved computer power, advancing use of remote sensing and GIS have allowed for increasing use of spatially explicit methods, and there is now a range of tools available for assessing fragmentation or connectivity [39-41]. Graph and circuit theoretical approaches have been used to construct networks with habitat patches represented as nodes and links between nodes (edges) representing inter-node connectivity [42-44]. Network characteristics can then be used as metrics of landscape connectivity. Other methods are based on metapopulation theory and use incidence functions to model connectivity [45, 46]. In all of the above mentioned approaches connectivity between habitat patches can be based on dispersal distance alone [47], but other species specific parameters can be included. This can be indices on population size or habitat resistance to movement [48]. Furthermore, least cost analysis can be used to substitute Euclidean distances with optimal movement paths between patches [49-51].

Recently there has been a growing interest in individual (or agent) based methods (IBM) in ecological modelling [52-55]. Recognising that movement patterns and, thus, also connectivity depend on individual behaviour, landscape configuration and their interaction [22, 40] individual based modelling seems a promising method for assessment of connectivity. IBM has been used to assess the effect of roads on dispersal success of for example Eurasian lynx [56] and Elk [57]. Graf, et al. [58] assessed the connectivity between patchy populations of Capercaillie embedded in a mountainous landscape. A generic individual based model to simulate dispersal (J-Walk) has been developed by Gardner and Gustafson [59] and used to estimate connectivity of a wide range of species, e.g. Delmarva fox squirrel [60], Black bears [61] and American marten [62]. FunCon [63] is another individual based connectivity tool which can be applied on bird species.

17 Synopsis

Objective Over the last decade a growing amount of literature has documented the severe impacts of transport infrastructure on biodiversity, population persistence and gene flow [1, 2, 12, 64- 66], and there is an increasing awareness of the importance of finding agreement between nature conservation and land use. In Europe, the EU Habitats directive enjoins member states to safeguard the ecological performance of breeding sites and resting places of species protected by the Habitat directive annex IV [67]. Furthermore, according to EU legislation, all major projects, including infrastructure projects, are subject to Environmental Impact As- sessment (EIA) [68]. Infra Eco Network Europe (IENE), a network of specialists, governmen- tal agencies, scientists and NGOs, enables cross-boundary and interdisciplinary cooperation on issues regarding ecologically sustainable transportation systems. A prominent result from IENE has been the European review of Habitat Fragmentation due to Linear Transportation Infrastructure [69] as well as the European handbook on sustainable road planning [70]; both published by the European Communities.

To ensure ecologically sustainable road planning, conservation measures must be taken into consideration already in the earliest phases of road development. This requires adequate tools for assessment of both the impacts of infrastructure and the effect of mitigation measures [71-73]. For this reason the Danish Road Directorate decided to finance a PhD project with the objective of developing a management tool which could be used to substantiate that the conservation status of annex IV species would remain unaffected by a given road project [67]. The Moor frog (Rana arvalis) was chosen as the model species. This pond breeding amphibian is listed in annex IV of the Habitats directive, but is relatively common, at least in the eastern part of Denmark. Therefore, there will often be a need to assess the impact of new road constructions on local populations of Moor frogs.

The purpose of the project was to provide a standardized and scientifically well founded basis for decisions concerning road lay-out and mitigation measures. The management tool should support decision making by enabling caseworkers

 to find the optimal location and road lay-out for a specific species  to assess the need for mitigation measures, such as tunnels, fences and compensation habitat for a specific species  to identity the best location for tunnels, fences and compensation habitat

18 Synopsis

 to evaluate the effect of mitigation measure on ecological performance

The project resulted in the development of a spatially explicit and individual based model called SAIA (Spatial Amphibian Impact Assessment).

Designing SAIA

When assessing ecological performance the most obvious metrics are population size and persistence [74]. In general, roads can affect amphibian populations in three ways – by destruction and fragmentation of habitat, by road kills, or by disruption of movement patterns [5, 7, 75]. Thus, the first step in the model development has been to construct a conceptual model of the possible effects of road construction on a population of Moor frogs.

Conceptual model The Moor frog is a pond breeding amphibian that needs aquatic as well as terrestrial habitat to complete its life cycle. The first phase of the life cycle, as egg and larva, takes place in shal- low often ephemeral ponds. The remaining part takes place in terrestrial habitat while ponds are only visited during breeding [76, 77]. The life cycle of the Moor frog is characterized by two types of movement: migration, the seasonal intrapopulational movement of adult individuals between summer habitat and breeding ponds and dispersal, the interpopulational movement of the newly metamorphosed frogs away from their natal pond [77-81]. Many amphibian populations are considered to be organised as a metapopulation [82-85]. This has not been studied explicitly for Moor frogs, but it is generally assumed by experts (pers. comm.) that Moor frogs form regional networks of subpopulations. Thus, regional population persistence depends on successful dispersal between subpopulations.

Given this background, it seems reasonable to assume that roads can affect the persistence of Moor frog populations in several ways (Fig 1):

 destruction of aquatic habitat  destruction of terrestrial habitat  fragmentation of terrestrial habitat  impaired migration between aquatic and terrestrial habitat  impaired dispersal between subpopulations

19 Synopsis

The first effect prevents the subpopulation from reproducing and is considered equal to destruction of the subpopulation. The next three effects reduce the amount and quality of accessible terrestrial habitat, and hence, the size of the population that can be sustained by the habitat. The last effect reduces the probability of (re)colonization of habitat patches, resulting in isolation of subpopulations.

Figure 1 Conceptual model of road effects on a regional Moor frog population. Dotted lines delimit subpopulations, blue dots represent breeding ponds and green areas are summer habitat fragments. Processes affected by roads are outlined in red

To incorporate local as well regional population dynamics into the model, I combine individual based modelling with a population dynamics model. The local population dynamics in each pond are simulated by use of an age-based Leslie matrix and are affected by the size and quality of the breeding pond and summer habitat. Regional effects are assessed by simulating dispersing frogs’ behavioural responses to land cover and structure while moving through the landscape. This provides estimates of immigration probabilities between subpopulations. These estimates reflect the connectivity of the landscape to be entered into the local dynamics of each pond as immigration rates (Fig. 2).

Figure 2

Elements of SAIA

20 Synopsis

The connectivity measure of SAIA does not adhere strictly to either of the definitions used in metapopulation or landscape ecology. Rather it is an attempt to find a “third” way [31]. SAIA’s connectivity measure considers dispersal between subpopulations and, thus, adopts metapopulation ecology’s patch based focus. However, as in landscape ecology, population size (or any other demographic indicator) does not enter into the connectivity measure. In SAIA, connectivity is solely a function of landscape configuration and animal behaviour and is measured as the probability of an individual finding its way from habitat patch A to habitat patch B. Moreover, SAIA’s connectivity measure is an index of the potential connectivity between all habitat patches, whether they are populated or not. The population based model links the potential connectivity with local population dynamics, and estimated abundances and persistence probabilities can be regarded as a result of the realised connectivity.

The habitat patch An important characteristic of SAIA is how the habitat patch of a subpopulation of Moor frogs is represented. In most studies measuring or modelling connectivity in regional populations of amphibians, the breeding pond is used as the spatial unit of a subpopulation [85]. As Moor frogs mostly breed in the same pond every year, SAIA also considers the pond as the potential site of a subpopulation [86]. However, the attributes of the breeding pond alone will not be an adequate descriptor of a subpopulation’s habitat patch. Outside the breeding season, the frogs reside in adequate terrestrial habitat (summer habitat) usually within a distance of 400 m from the pond (defined as migration distance) [77, 78]. Whereas the size and quality of the breeding pond may affect the reproductive output [87], it is reasonable to assume that adult abundance will depend on the amount and quality of the terrestrial habitat in which the frogs live during summer. Consequently, in SAIA the habitat patch of a subpopulation of Moor frogs is defined as complementary habitat patch consisting of a breeding pond and all accessible summer habitat fragments within migration distance. Accessibility is important in this context. Roads (or other impregnable structures) can function as barriers preventing access to resources on the opposite side [88]. Thus, the summer habitat available for the frogs is restricted by linear infrastructure. Conversely, construction of underpasses can re-establish the connection with isolated habitat fragments (Fig. 3).

21 Synopsis

A Figure 3 Illustration of how accessible summer habitat is identified. Blue circle is a pond; dotted circle represents maximum migration distance. Green areas are accessible summer habitat while shaded areas are inaccessible summer habitat. B A) All summer habitat within migration distance is regarded as accessible B) Road traversing the habitat prevents access to summer habitat on the opposite side of the road C) Structures breaking the road such as underpasses again C permits access to summer habitat on the opposite side of the road.

In the model, the carrying capacity of a habitat patch is determined by the area of accessible summer habitat, and adult survival is modelled as depending on the frog density of the summer habitat. The effective area of the summer habitat depends on the degree of fragmentation and, thus, the area of summer habitat is weighted by the amount of edges [89]. In real landscapes summer habitat is often shared by several breeding ponds, and in the model the frog density of a single summer habitat cell, therefore, depends on the population sizes of all the breeding ponds sharing the summer habitat.

Dispersal behaviour Individual based modelling is a little like story telling. Based on the available research on behaviour, patterns of abundance, distribution etc, you try to think like your species. And the model becomes your story of how you believe individuals of the modelled species experience and react to the set of conditions and circumstances constituting their environment. In more scientific terms, models can be regarded as hypotheses [55] and, thus, SAIA is my hypothesis on how newly metamorphosed frogs leave their natal pond and move through the landscape until they settle in a new habitat patch.

Young frogs have no prior knowledge of the landscape they disperse into. While dispersing frogs may have an innate urge to move away from their natal pond, their movements

22 Synopsis are also assumed to be affected by their immediate concern of staying alive [77, 80, 90]. Thus, movement decisions are usually centred on the environmental cues, which guide the animals into habitat where survival probabilities are high. Very little is known about the dispersal of juveniles. Dispersal distances are recorded to be between a few hundred meters up to 1-2 kilometres [77, 91, 92], but what triggers the decision to stop and settle down?

Adult frogs show a high degree of site fidelity in regard to breeding pond and summer habitat, and juvenile frogs appear to inhabit the same summer habitat as the adults [77, 86, 93, 94]. During breeding migrations adults often exhibit quite goal-oriented movements, and some juveniles seem to follow adult frogs towards the breeding ponds although they do not enter the ponds themselves [77, 90]. Juvenile frogs have to stay alive for at least two years before they start breeding. The above observations suggest that juvenile frogs spend the first couple of years in the summer habitat learning to know and navigate in the habitat patch. Thus the primary goal of dispersing juveniles must be to find summer habitat where they can survive until maturity. The next step will then be to find a suitable breeding pond. Modelling the dispersal behaviour, I therefore assume that it is the presence of summer habitat rather than breeding ponds that triggers the settling behaviour. When a dispersing frog encounters summer habitat it may decide to stop moving and settle in the new habitat, without knowing whether there is a breeding pond nearby or not.

SAIA v1.0 The resulting model, SAIA v1.0, is meant to be a strategic management tool supporting decision-making. Furthermore, it is meant to be used by non-specialists. Therefore, it should be intuitively understandable, flexible, easy to use and with an output that can be interpreted without much effort.

The workflow is simple: a GIS map of the relevant area must be constructed and converted into a text file; then imported into SAIA and the analysis can be started. After the simulations, SAIA generates several types of output in the form of text files and shape files to be used in GIS (Fig. 4).

23 Synopsis

Figure 4 SAIA’s workflow

At least two scenarios have to be constructed to carry out a meaningful analysis. The first scenario should be a map of the area as it is before the planned road construction (scenario 0). This analysis measures the ecological performance of the original landscape and is a reference against which other scenarios are to be compared. The second map (scenario 1) should show the landscape as it is expected to be after the road constructions. This will typically involve drawing the new road or, in case of a road expansion, changing the properties of the original road. Breeding ponds destroyed by the construction work are removed from the map. More indirect effects such as reduced habitat quality close to the road or expected changes in traffic intensity (and thus road mortality) of adjacent roads can also be incorpo- rated in the map. In combination, the analyses of scenario 0 and scenario 1 make it possible to assess the effect of road construction on connectivity and population persistence which constitute the basis for planning of mitigation measures. Hereafter, additional scenarios with alternative suggestions for mitigation measures can be constructed, analysed and compared.

As input data SAIA needs two text file; one file to construct the land cover map and a one file containing data about the potential breeding ponds in the area. The data in the input files have to be structured in a specific way, but there are no special requirements on which software to be used when constructing the files. In this project, the land cover maps are based on several GIS layers describing roads, buildings, nature reserves, fallows, fields and so on, while data on breeding ponds originate from field surveys. The construction of land cover maps has not been entirely trivial as data had to be obtained from many different digital

24 Synopsis sources. The consultancy firm AmphiConsult has been responsible for the job and has developed a standard protocol for construction of land cover maps to be used with SAIA [95]. Even though, map construction has not been part of my PhD project I have been involved to ensure the compatibility with SAIA.

SAIA produces output regarding connectivity, population dynamics as well as the results of a cluster analysis based on the connectivity matrix (examples of the output files can be found in the appendix):

 A text file with descriptive statistics on regional connectivity, abundance and population persistence probability as well as descriptive statistics on abundance and persistence probability of individual pond populations.  A text file containing information on clusters and their pond members as well as connectivity within and between clusters

In addition, SAIA produces several GIS data files for graphic display and further analysis in GIS software:

 A point-data set with information on mean estimated abundance and population persistence probability of the ponds.  Vector-data set with information about immigration probability between ponds (connectivity network)  Vector-data set with information about cluster configuration

Conclusion

The following three chapters contain manuscripts representing different stages of the modelling process. In the first manuscript, I use a simple model to explore the concept of the complementary habitat patch and how intra-patch heterogeneity affects immigration and emigration probabilities. This manuscript has been submitted to the Open Access journal “Web Ecology” and been peer-reviewed. It is now under revision to be resubmitted soon. The second manuscript uses a light-version of SAIA and explores how changes in levels of road avoidance and road mortality affect connectivity locally as well as regionally. The third manuscript describes the full SAIA model. By means of a case study, I demonstrate how SAIA can be used for assessing which management measures would be best to mitigate the

25 Synopsis effect of landscape fragmentation caused by road constructions. These last two manuscripts are submitted to the Open Access journal “Nature Conservation”.

Modelling is a never-ending story and the name “SAIA v1.0” implies the possibility of a version 2.0. In the coming months, SAIA will be implemented as a planning tool in the Dan- ish Road Directorate and this will be the real test of SAIA. As a variety of landscapes are being analysed, the validity and usefulness of the output will be tested and through dialogs with the users, the model design may be adjusted. The functionality of SAIA may also be improved by incorporating other protected amphibian species with ecology similar to the Moor frog, like for instance Crested newt (Triturus cristatus).

SAIA is not only a planning tool. The model can also be used to explore other aspects of impact assessments or hypotheses concerning road ecology. By applying a “virtual ecologist” approach [96] different types of input data can be tested and compared. Of interest could be how substituting counts of egg masses with presence/absence data or using aerial photos to assess pond quality will affect model output. Additionally, virtual experiments with varying sizes of the survey area could give insights about effective sampling schemes.

26 Synopsis

References

1. Coffin, A.W. (2007). From roadkill to road ecology: A review of the ecological effects of roads. Journal of Transport Geography 15, 396-406. 2. Fahrig, L., and Rytwinski, T. (2009). Effects of Roads on Animal Abundance: an Em- pirical Review and Synthesis. Ecology and Society 14. 3. Bruinderink, G., and Hazebroek, E. (1996). Ungulate traffic collisions in Europe. Con- servation Biology 10, 1059-1067. 4. Garriga, N., Santos, X., Montori, A., Richter-Boix, A., Franch, M., and Llorente, G.A. (2012). Are protected areas truly protected? The impact of road traffic on vertebrate fauna. Biodiversity and Conservation 21, 2761-2774. 5. Elzanowski, A., Ciesiolkiewicz, J., Kaczor, M., Radwanska, J., and Urban, R. (2009). Amphibian road mortality in Europe: a meta-analysis with new data from Poland. Euro- pean Journal of Wildlife Research 55, 33-43. 6. Hels, T., and Buchwald, E. (2001). The effect of road kills on amphibian populations. Biological Conservation 99, 331-340. 7. Pukey, M. (2006). Amphibian road kills: a global perspective. 8. Rytwinski, T., and Fahrig, L. (2012). Do species life history traits explain population responses to roads? A meta-analysis. Biological Conservation 147, 87-98. 9. Carr, L.W., and Fahrig, L. (2001). Effect of road traffic on two amphibian species of differing vagility. Conservation Biology 15, 1071-1078. 10. Karraker, N.E., and Gibbs, J.P. (2011). Contrasting road effect signals in reproduction of long-versus short-lived amphibians. Hydrobiologia 664, 213-218. 11. Benítez-López, A., Alkemade, R., and Verweij, P.A. (2010). The impacts of roads and other infrastructure on mammal and bird populations: A meta-analysis. Biological Con- servation 143, 1307-1316. 12. Forman, R.T.T., and Alexander, L.E. (1998). Roads and their major ecological effects. Annual Review of Ecology and Systematics 29. 13. Jaeger, J.a.G., Bowman, J., Brennan, J., Fahrig, L., Bert, D., Bouchard, J., Charbonneau, N., Frank, K., Gruber, B., and von Toschanowitz, K.T. (2005). Predicting when animal populations are at risk from roads: an interactive model of road avoidance behavior. Ecological Modelling 185, 329-348. 14. Jaeger, J.a.G., and Fahrig, L. (2004). Effects of Road Fencing on Population Persis- tence. Conservation Biology 18, 1651-1657. 15. Sala, O.E., Chapin, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber- Sanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., et al. (2000). Biodiversity - Global biodiversity scenarios for the year 2100. Science 287, 1770-1774. 16. Fahrig, L. (2001). How much habitat is enough? Biological Conservation 100, 65-74. 17. Didham, R.K. (2010). Ecological Consequences of Habitat Fragmentation. In eLS. (John Wiley & Sons, Ltd).

27 Synopsis

18. Fahrig, L. (2003). Effects of habitat fragmentation on biodiversity. Annual Review of Ecology Evolution and Systematics 34, 487-515. 19. Wiegand, T., Revilla, E., and Moloney, K.A. (2005). Effects of habitat loss and fragmentation on population dynamics. Conservation Biology 19, 108-121. 20. Fraterrigo, J.M., Pearson, S.M., and Turner, M.G. (2009). Joint effects of habitat configuration and temporal stochasticity on population dynamics. Landscape Ecology 24, 863-877. 21. Ries, L., Fletcher, R.J., Battin, J., and Sisk, T.D. (2004). Ecological responses to habitat edges: Mechanisms, models, and variability explained. Annual Review of Ecology Evo- lution and Systematics 35, 491-522. 22. Fahrig, L. (2007). Non-optimal animal movement in human-altered landscapes. Func- tional Ecology 21, 1003-1015. 23. Fahrig, L., and Nuttle, W.K. (2005). Population ecology in spatially heterogeneous en- vironments. Ecosystem Function in Heterogeneous Landscapes, 95-118. 24. King, A.W., and With, K.A. (2002). Dispersal success on spatially structured landscapes: when do spatial pattern and dispersal behavior really matter? Ecological Model- ling 147, 23-39. 25. Revilla, E., and Wiegand, T. (2008). Individual movement behavior, matrix heterogeneity, and the dynamics of spatially structured populations. Proceedings of the National Academy of Sciences of the United States of America 105, 19120-19125. 26. Ricketts, T.H. (2001). The matrix matters: Effective isolation in fragmented landscapes. American Naturalist 158, 87-99. 27. Wiens, J.A., Stenseth, N.C., Vanhorne, B., and Ims, R.A. (1993). Ecological mechanisms and landscape ecology. Oikos 66, 369-380. 28. Dunning, J.B., Danielson, B.J., and Pulliam, H.R. (1992). Ecological processes that affect populations in complex landscapes. Oikos 65, 169-175. 29. Hanski, I. (1999). Habitat connectivity, habitat continuity, and metapopulations in dy- namic landscapes. Oikos 87, 209-219. 30. Taylor, P.D., Fahrig, L., Henein, K., and Merriam, G. (1993). Connectivity is a vital element of landscape structure. Oikos 68, 571-573. 31. Wiens, J.A. (1997). Metapopulation Dynamics and Landscape Ecology. In Metapopula- tion Biology: ecology, genetics, and evolution, I. Hanski and M.E. Gilpin, eds. (Aca- demic press, Inc.). 32. Hanski, I. (1999). Metapopulation Ecology, (Oxford University Press). 33. Hanski, I., and Simberloff, D. (1997). The Metapopulation Approach, Its history, Con- ceptual domain, and Application to Conservation. In Metapopulation Biology: ecology, genetics, and evolution, I. Hanski and M.E. Gilpin, eds. (Academic press, Inc.). 34. Hanski, I. (1994). A practical model of metapopulation dynamics. Journal of Animal Ecology 63, 151-162. 35. Hanski, I., and Ovaskainen, O. (2000). The metapopulation capacity of a fragmented landscape. Nature 404, 755-758.

28 Synopsis

36. Fahrig, L., Baudry, J., Brotons, L., Burel, F.G., Crist, T.O., Fuller, R.J., Sirami, C., Siriwardena, G.M., and Martin, J.L. (2011). Functional landscape heterogeneity and animal biodiversity in agricultural landscapes. Ecology letters 14, 101-112. 37. Tischendorf, L., and Fahrig, L. (2000). On the usage and measurement of landscape connectivity. Oikos 90, 7-19. 38. Crooks, K.R., and Sanjayan, M. eds. (2006). Connectivity Conservation (Cambridge University Press). 39. Calabrese, J.M., and Fagan, W.F. (2004). A comparison-shopper's guide to connectivity metrics. Frontiers in Ecology and the Environment 2, 529-536. 40. Kindlmann, P., and Burel, F. (2008). Connectivity measures: a review. Landscape Ecol- ogy 23, 879-890. 41. Kadoya, T. (2008). Assessing functional connectivity using empirical data. Population Ecology 51, 5-15. 42. Galpern, P., Manseau, M., and Fall, A. (2011). Patch-based graphs of landscape connectivity: A guide to construction, analysis and application for conservation. Biological Conservation 144, 44-55. 43. Laita, A., Kotiaho, J.S., and Monkkonen, M. (2011). Graph-theoretic connectivity measures: what do they tell us about connectivity? Landscape Ecology 26, 951-967. 44. McRae, B.H., Dickson, B.G., Keitt, T.H., and Shah, V.B. (2008). Using circuit theory to model connectivity in ecology, evolution, and conservation. Ecology 89, 2712-2724. 45. Moilanen, A. (2004). SPOMSIM: software for stochastic patch occupancy models of metapopulation dynamics. Ecological Modelling 179, 533-550. 46. Quintana, S.M., Ramos, B.M., Martinez, M.A.C., and Pastor, I.O. (2010). A model for assessing habitat fragmentation caused by new infrastructures in extensive territories - Evaluation of the impact of the Spanish strategic infrastructure and transport plan. Jour- nal of environmental management 91, 1087-1096. 47. Minor, E.S., and Urban, D.L. (2008). A graph-theory frarmework for evaluating landscape connectivity and conservation planning. Conservation Biology 22, 297-307. 48. Vasas, V., Magura, T., Jordan, F., and Tothmeresz, B. (2009). Graph theory in action: evaluating planned highway tracks based on connectivity measures. Landscape Ecology 24, 581-586. 49. Adriaensen, F., Chardon, J.P., De Blust, G., Swinnen, E., Villalba, S., Gulinck, H., and Matthysen, E. (2003). The application of 'least-cost' modelling as a functional landscape model. Landscape and urban planning 64, 233-247. 50. Pinto, N., Keitt, T.H., and Wainright, M. (2012). LORACS: JAVA software for modeling landscape connectivity and matrix permeability. Ecography 35, 388-392. 51. Janin, A., Léna, J.-P., Ray, N., Delacourt, C., Allemand, P., and Joly, P. (2009). Assess- ing landscape connectivity with calibrated cost-distance modelling: predicting common toad distribution in a context of spreading agriculture. Journal of Applied Ecology 46, 833-841.

29 Synopsis

52. DeAngelis, D.L., and Mooij, W.M. (2005). Individual-based modeling of ecological and evolutionary processes. In Annual Review of Ecology Evolution and Systematics, Vol- ume 36. pp. 147-168. 53. Guzy, M.R., Smith, C.L., Bolte, J.P., Hulse, D.W., and Gregory, S.V. (2008). Policy Research Using Agent-Based Modeling to Assess Future Impacts of Urban Expansion into Farmlands and Forests. Ecology and Society 13. 54. McLane, A.J., Semeniuk, C., McDermid, G.J., and Marceau, D.J. (2011). The role of agent-based models in wildlife ecology and management. Ecological Modelling 222, 1544-1556. 55. Grimm, V., and Railsback, S.F. (2005). Individual-based Modeling and Ecology, (Princeton University Press ). 56. Kramer-Schadt, S., Revilla, E., Wiegand, T., and Breitenmoser, U. (2004). Fragmented landscapes, road mortality and patch connectivity: modelling influences on the dispersal of Eurasian lynx. Journal of Applied Ecology 41, 711-723. 57. Frair, J.L., Merrill, E.H., Beyer, H.L., and Morales, J.M. (2008). Thresholds in landscape connectivity and mortality risks in response to growing road networks. Journal of Applied Ecology 45, 1504-1513. 58. Graf, R.F., Kramer-Schadt, S., Fernandez, N., and Grimm, V. (2007). What you see is where you go? Modeling dispersal in mountainous landscapes. Landscape Ecology 22, 853-866. 59. Gardner, R.H., and Gustafson, E.J. (2004). Simulating dispersal of reintroduced species within heterogeneous landscapes. Ecological Modelling 171, 339-358. 60. Lookingbill, T.R., Gardner, R.H., Ferrari, J.R., and Keller, C.E. (2010). Combining a dispersal model with network theory to assess habitat connectivity. Ecological Applica- tions 20, 427-441. 61. Morzillo, A.T., Ferrari, J.R., and Liu, J. (2011). An integration of habitat evaluation, individual based modeling, and graph theory for a potential black bear population re- covery in southeastern Texas, USA. Landscape Ecology 26, 69-81. 62. Wasserman, T.N., Cushman, S.A., Shirk, A.S., Landguth, E.L., and Littell, J.S. (2012). Simulating the effects of climate change on population connectivity of American marten (Martes americana) in the northern Rocky Mountains, USA. Landscape Ecology 27, 211-225. 63. Pe'er, G., Henle, K., Dislich, C., and Frank, K. (2011). Breaking Functional Connec- tivity into Components: A Novel Approach Using an Individual-Based Model, and First Outcomes. PLoS ONE 6. 64. Spellerberg, I.F. (1998). Ecological effects of roads and traffic: a literature review. Global Ecology and Biogeography 7, 317-333. 65. Trombulak, S.C., and Frissell, C.A. (2000). Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14, 18-30. 66. Holderegger, R., and Di Giulio, M. (2010). The genetic effects of roads: A review of empirical evidence. Basic and Applied Ecology 11, 522-531.

30 Synopsis

67. European-Council (1992). Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora. (European Communities). 68. European-Council (1997). Council Directive 97/11/EC of 3 March 1997 amending Di- rective 85/337/EEC on the assessment of the effects of certain public and private projects on the environment. (European Communities). 69. Trocmé, M., Cahill, S., de Vries, J.G., Farrall, H., Folkeson, L.G., Hichks, C., and Pey- men, J. eds. (2003). COST 341 – Habitat Fragmentation due to Transportation Infra- structure (Office for Official Publications of the European Communities, Luxembourg). 70. Iuell, B., Bekker, G.J., Cuperus, R., Dufek, J., Fry, G., Hicks, C., Hlaváˇc, V., Keller, V., B., R., C., Sangwine, T., et al. eds. (2003). Wildlife and Traffic: A European Hand- book for Identifying Conflicts and Designing Solutions (Office for official publications of the European Communities, Luxembourg). 71. Beckmann, J.P. (2010). Safe passages, highways, wildlife, and habitat connectivity, (Washington). 72. Gontier, M., Mörtberg, U., and Balfors, B. (2010). Comparing GIS-based habitat models for applications in EIA and SEA. Environmental Impact Assessment Review 30, 8- 18. 73. Forman, R.T.T., Sperling, D., Bissonette, J.A., Clevenger, A.P., Cutshall, C.D., Dale, V.H., Fahrig, L., France, R., Goldman, C.R., Heanue, K., et al. (2003). Road ecology: science and solutions. Road ecology: science and solutions., Chp. 6. 74. Boyce, M.S. (1992). Population Viability Analysis. Annual Review of Ecology and Systematics 23, 481-506. 75. Andrews, K.M., Gibbons, J.W., and Jochimsen, D.M. (2008). Ecological effects of roads on amphibians and reptiles: a literature review. 76. Glandt, D. (2008). Der Moorfrosch (Rana arvalis): Erscheinungsvielfalt, Verbreitung, Lebensräume, Verhalten sowie Perspectiven für den Artenschutz. In The Moor Frog, Volume Zeitschrift für Feldherpetologie, Supplement 13, D. Glandt and R. Jehle, eds. (Bielefeld: Laurenti-Verlag). 77. Hartung, H. (1991). Untersuchung zur terrestrischen Biologie von Populationen des Moorfrosches (Rana arvalis NILSSON 1842) unter besonderer Berücksichtigung der Jahresmobilität. In Fachbereiches Biologie, Volume PhD. (Hamburg: Universität Ham- burg). 78. Kovar, R., Brabec, M., Vita, R., and Bocek, R. (2009). Spring migration distances of some Central European amphibian species. Amphibia-reptilia 30, 367-378. 79. Semlitsch, R.D. (2008). Differentiating migration and dispersal processes for pond- breeding amphibians. Journal of Wildlife Management 72, 260-267. 80. Sinsch, U. (1990). Migration and orientation in anuran amphibians. Ethology Ecology & Evolution 2, 65-79. 81. Hartel, T., Sas, I., Pernetta, A.P., and Geltsch, I.C. (2007). The reproductive dynamics of temperate amphibians: a review. North-Western Journal of Zoology 3, 127-145. 82. Hels, T. (2002). Population dynamics in a Danish metapopulation of spadefoot toads Pelobates fuscus. Ecography 25, 303-313.

31 Synopsis

83. Marsh, D. (2008). Metapopulation viability analysis for amphibians. Animal Conserva- tion 11, 463-465. 84. Marsh, D.M., and Trenham, P.C. (2001). Metapopulation dynamics and amphibian conservation. Conservation Biology 15, 40-49. 85. Smith, M.A., and Green, D.M. (2005). Dispersal and the metapopulation paradigm in amphibian ecology and conservation: are all amphibian populations metapopulations? Ecography 28, 110-128. 86. Loman, J. (1994). Site tenacity, within and between summers, of Rana arvalis and Rana temporaria. Alytes 12, 15-29. 87. Loman, J., and Lardner, B. (2006). Does pond quality limit frogs Rana arvalis and Rana temporaria in agricultural landscapes? A field experiment. Journal of Applied Ecology 43, 690-700. 88. Eigenbrod, F., Hecnar, S.J., and Fahrig, L. (2008). Accessible habitat: an improved measure of the effects of habitat loss and roads on wildlife populations. Landscape Ecology 23, 159-168. 89. Watts, K., and Handley, P. (2010). Developing a functional connectivity indicator to detect change in fragmented landscapes. Ecological Indicators 10, 552-557. 90. Sjögren-Gulve, P. (1998). Spatial movement patterns in frogs: Differences between three Rana species. Ecoscience 5, 148-155. 91. Baker, J.M.R., and Halliday, T.R. (1999). Amphibian colonization of new ponds in an agricultural landscape. Herpetological Journal 9, 55-63. 92. Vos, C.C., and Chardon, J.P. (1998). Effects of habitat fragmentation and road density on the distribution pattern of the moor frog Rana arvalis. Journal of Applied Ecology 35, 44-56. 93. Loman, J. (1978). Macrohabitat and microhabitat distribution in Rana-arvalis and Rana- temporaria (amphibia, anura, ranidae) during summer. Journal of Herpetology 12, 29- 33. 94. Loman, J. (1984). Density and survival of Rana arvalis and Rana temporaria. Alytes 3, 125-134. 95. Hassingboe, J., Neergaard, R.S., and Hesselsøe, M. (2012). Manual til produktion af GIS raster kort til:”EDB-værktøj til at vurdere skader på bestande af padder /økologisk funktionalitet”. (Amphi Consult). 96. Zurell, D., Berger, U., Cabral, J.S., Jeltsch, F., Meynard, C.N., Munkemuller, T., Nehr- bass, N., Pagel, J., Reineking, B., Schroder, B., et al. (2010). The virtual ecologist approach: simulating data and observers. Oikos 119, 622-635.

CHAPTER ONE

EFFECTS OF WITHIN-PATCH

HETEROGENEITY ON CONNECTIVITY

IN POND-BREEDING AMPHIBIANS

STUDIED BY MEANS OF AN

INDIVIDUAL-BASED MODEL

Submitted to Web Ecology

October 2012

Under revision

34 Chapter One

Effects of within-patch heterogeneity on connectivity in pond-breeding amphibians studied by means of an individual-based model

M.-B. Pontoppidan and G. Nachman

Section for Ecology and Evolution, Department of Biology, University of Copenhagen

Universitetsparken 15, DK-2100 Copenhagen

Correspondence: M.-B. Pontoppidan ([email protected])

Abstract

The metapopulation framework presumes the habitat of a local population to be continuous and homogenous, and patch area is often used as a proxy for population size. Many populations of pond-breeding amphibians are assumed to follow metapopulations dynamics, and connectivity is mostly measured between breeding ponds. However, the habitat of pond- breeding amphibians is not only defined by the pond but, typically, consists of a breeding pond surrounded by clusters of disjoint summer habitat patches interspersed with an agricultural/semi-urban matrix. We hypothesize that the internal structure of a habitat patch may change connectivity in two ways: i) by affecting animal movements and thereby emigration and immigration probabilities; ii) by affecting habitat quality and population size. To test our hypotheses, we apply a spatially explicit individual-based model of Moor frog dispersal. We find that the realised connectivity depends on internal structure of both the target and the source patch as well as on how habitat quality is affected by patch structure. Although fragmentation is generally thought to have negative effects on connectivity, our results suggest that, depending on patch structure and habitat quality, positive effects on connectivity may occur.

35 Chapter One

Introduction

Within the framework of metapopulations, inter-patch connectivity is modelled as an incidence function measuring the dispersal success between two habitat patches (Moilanen and Nieminen 2002). The essential components of the incidence function models are emigration and immigration rates. The number of emigrating individuals is assumed to depend on the population size of the donor patch and the probability of an individual actually leaving the patch. Likewise, the number of immigrants depends on the probability that dispersing individuals will find the target patch (Hanski and Simberloff 1997; Moilanen and Hanski 2006; Wiens 1997). A patch is assumed to constitute a continuous and homogenous habitat area with all the necessary resources needed for the persistence of a local population. The incidence function usually models the emigration and immigration rates as linear functions of donor and target patch area, respectively. The survival probability during the transit between two patches is modelled as a function of distance (Hanski and Simberloff 1997; Kindlmann and Burel 2008; Moilanen and Hanski 2001; Moilanen and Hanski 2006; Moilanen and Nieminen 2002). However, it is questionable to what extent the above assumptions apply to real populations of many species.

Regional populations of pond-breeding amphibians are frequently considered to be structured as metapopulations (Hels 2002; Marsh 2008; Marsh and Trenham 2001; Smith and Green 2005). Pond-breeding amphibians need ponds for breeding and development of tad- poles, but otherwise live most of their life in terrestrial habitat (also called summer habitat). Proximity between the required habitat types (landscape complementation) is important for population size and persistence (Dunning et al. 1992; Haynes et al. 2007; Johnson et al. 2007; Pope et al. 2000). However, as a consequence of increased landscape fragmentation, the summer habitat of many subpopulations does not form one continuous patch. Typically, a subpopulation of pond-breeding amphibians occupies a landscape consisting of breeding ponds surrounded by fragments of summer habitat interspersed with an agricultural/semi- urban matrix (Hamer and McDonnell 2008; Hartung 1991; Pillsbury and Miller 2008; Pope et al. 2000; Sjögren-Gulve 1998; Tramontano 1998). Thus, the metapopulation premise of a continuous and homogenous habitat patch is compromised, which might have consequences for patch connectivity and the way it is measured (Rothermel 2004).

36 Chapter One

Numerous studies, empirical as well as modelling, have shown that structure and composition of the habitat matrix can have strong effects on animal movement and dispersal success (Bender and Fahrig 2005; Chin and Taylor 2009; Gustafson and Gardner 1996; Haynes and Cronin 2006; Prevedello and Vieira 2010; Ricketts 2001; Vandermeer and Carvajal 2001; Watling et al. 2011). Similar effects may be found within heterogeneous habitat patches, such as those of pond-breeding amphibians. At the core of the habitat patch is the breeding pond surrounded by satellites of summer habitat fragments separated by matrix habitat. The summer habitat fragments within the habitat patch work as a collective, functioning as a filter catching dispersers which will then eventually find their way to the breeding pond. Emigra- tion and immigration probabilities may thus be influenced by the spatial distribution of the summer habitat fragments within the habitat patch.

Metapopulation theory usually assumes that the size of a subpopulation is proportional to the area of the patch it inhabits. However, in some cases, the quality of the occupied habitat may be a better predictor of patch carrying capacity (Jaquiéry et al. 2008; Moilanen and Han- ski 1998). One of the factors that may affect habitat quality is the degree of habitat fragmentation. Thus, a fragmented habitat may not be able to sustain as large a population as a non- fragmented habitat of equal area due to a combination of negative edge effects and reduced landscape complementation (Dunning et al. 1992; Haynes et al. 2007; Johnson et al. 2007; Lehtinen et al. 2003; Pope et al. 2000; Ries et al. 2004).

The internal structure of a habitat patch may therefore change inter-patch connectivity in two ways: i) by affecting animal movements and thereby emigration and immigration probabilities; ii) by affecting habitat quality and population size. To test how intra-patch structur- ing may influence dispersal success and connectivity we apply a spatially explicit individual- based model of how the Moor frog (Rana arvalis Nilsson) moves in a heterogeneous landscape. The model is part of a larger study aiming at modelling the effect of roads on regional persistence of Moor frog metapopulations (Pontoppidan and Nachman in prep.). With this model, we test the following

 Does the distance between the breeding pond and the summer habitat within a habitat patch affect inter-patch connectivity?  Does the degree of summer habitat fragmentation within a habitat patch affect inter-patch connectivity?

37 Chapter One

 Do effects of pond distance and summer habitat fragmentation on inter-patch connectivity interact?  Does the quality of the habitat patch affect inter-patch connectivity?

Methods

Model species Long distance dispersal in Moor frogs takes place predominantly during the juvenile life- stage. Shortly after metamorphosis, the young frogs leave the natal pond and disperse into the surrounding landscape seeking out suitable summer habitat. Dispersal distances are between a few hundred meters up to 1-2 kilometres (Baker and Halliday 1999; Hartung 1991; Sinsch 2006; Vos and Chardon 1998). The juveniles stay in terrestrial habitat 2-3 years until they reach maturity. During early spring, the adults move to the breeding ponds. Soon after breeding, the frogs return to the summer habitat, which lies mostly within a 400 m radius from the breeding pond Adult frogs show a high degree of site fidelity and often use the same breeding pond and summer habitat patch from year to year (Hartung 1991; Loman 1984, 1994; Sem- litsch 2008; Tramontano 1998).

Model overview The model considers nine subpopulations of Moor frogs within a spatially explicit landscape matrix. Each landscape cell represents an area of 10 x 10 meters, which can be either summer habitat or matrix habitat. Each habitat type is associated with a daily survival probability (s) and an index of attractiveness (a) (table 1). The area inhabited by a subpopulation is defined by the habitat patch, which comprises a breeding pond, all the summer habitat fragments located within migration distance from the pond as well as the intermediate matrix habitat (Fig. 1).

The model simulations mimic the dispersal of juvenile Moor frogs. Successful dispersal requires two events: 1) movement of a juvenile frog to summer habitat outside its natal habitat patch and 2) subsequent movement from the new summer habitat to a nearby breeding pond. In real life dispersal starts just after metamorphosis in early summer and lasts until hibernation in the autumn. The second part of the dispersal event takes place in the spring 2.5 years later. For simplicity, we simulate the two events, as if they take place in the same year.

38 Chapter One

At the start of a simulation, 500 frogs are created at each breeding pond. Each individual has an inherent random direction, which characterizes its preferred direction of movement. This direction does not change unless summer habitat is found. At each time step, frogs move to one of its n neighbouring cells according to the following movement rules: A frog has a sensing area of 225 degrees placed symmetrically around its preferred direction and within this area it moves to cell i with a probability that depends on the cell’s attractiveness (ai). The

probability of moving into cell i is calculated as where ∑ is the total attrac- ∑ tiveness of the n cells located within the sensing area. A uniform pseudorandom number is selected to choose the cell to move to. Once a cell is chosen, the frog moves to a random position within the cell. The movement rules generate a biased random walk away from the natal pond and in the preferred direction.

The time step of the model is one day and the simulated period is 120 days. Dispersal continues until the frog either settles, dies or the time runs out. Frogs encountering a cell with summer habitat randomly choose whether to settle in the cell or not. Settled frogs stay in the summer habitat until day 75. Hereafter, they start moving towards the breeding pond associated with the habitat patch in which they have spent the summer. For each day, the survival probability of every frog is based on the daily survival rate associated with the habitat type it occupies. Netlogo (Wilensky 1999) is used as modelling environment (freely downloadable at http://ccl.northwestern.edu/netlogo). Full model documentation following the ODD-template suggested by Grimm et al. (2006; 2010) is found in supplementary material, Appendix A.

Scenarios A landscape consisting of 300x300 cells is constructed as a torus with nine evenly spaced habitat patches. We define the habitat patch of a subpopulation as a complementary habitat patch containing not only the breeding pond but also all summer habitat fragments within 400 m from the pond. Five habitat patches are configured with one coherent summer habitat fragment randomly placed in a distance of 100 meters from the pond. These habitat patches serve as control patches as they have non-fragmented summer habitat and high landscape complementation. The remaining four habitat patches are test patches, in which the number of summer habitat fragments and their distance to the breeding pond are determined by the chosen parameter values for the number of fragments (P) and distance to the pond (d) (fig. 1, table 1). In each of the nine habitat patches the total area of the summer habitat sums to 0.81 ha (81

39 Chapter One cells) irrespective of fragmentation. The same set of parameter values is applied to all test patches in a given landscape scenario. We run 100 simulations for every combination of the parameter values for P and d given in table 1. For each of the 100 simulations, the position of the test and control patches is randomly shuffled. At the end of each simulation we record the number of dispersers that has settled at each breeding pond, and the origin of the dispersers (i.e. their natal pond).

Connectivity In metapopulation ecology, most measures of connectivity between patch i and j are based on the formula:

, where j is the source patch and i the target patch. Ai and Aj denote the area of patch i and j, respectively, and b and c are parameters scaling patch area to emigration and immigration rates, respectively. D(dij, α) is a species-specific function describing the dispersal ability of the species (Kindlmann and Burel 2008; Moilanen and Hanski 2001; Moilanen and Nieminen

2002). In this study we substitute the dispersal parameter D(dij, α) with the actual dispersal success between patches obtained by the individual-based simulations and compute connectivity as:

(1) where pij is the probability of a disperser getting from habitat patch j to habitat patch i. Aj and

Ai are the total area of summer habitat fragments within the two habitat patches.

Fragmentation of summer habitat within the habitat patch is expected to have a negative effect on the quality of habitat patch so that a habitat patch with highly fragmented summer habitat supports fewer individuals even though the total area is the same (Fahrig 2003). To compensate for the effect of summer habitat fragmentation on the quality of the habitat patch we introduce a quality-weighted connectivity measure (Moilanen and Hanski 1998):

′ 2 where Q denotes the quality of the habitat patch. In principle, Q corresponds to A in eq. 1, but only if patch quality is independent of fragmentation of summer habitat within the habitat

40 Chapter One patch. If this is not the case, Q is computed as the total area of summer habitat fragments within the habitat patch weighted by the degree of fragmentation (Jaeger 2000):

/ ∑ , z 0 3

th where P is the number of summer habitat fragments in habitat patch i; Ak is the area of the k fragment of summer habitat, and z is a scaling factor indicating the effect of fragmentation on quality. If z = 1 then Qi = Ai; if k > 1 and z > 1 then Qi < Ai and if k > 1 and 0 < z < 1 then Qi >

Ai.

In order to evaluate the effects of fragmentation and pond distance on connectedness at the landscape level we compute mean connectivity of the test patches as (Goodwin and Fahrig 2002; Kindlmann and Burel 2008):

∑∑ ̅ i j 4 where nt is the number of test patches (4). In all simulations, the structure within the control patches is the same and the contribution of the control patches to the connectivity measure remains constant. This allows us to distinguish between the effect of structure within source and target patches on connectivity:

Mean connectivity from control to test patches. Evaluates the effect of pond distance and fragmentation within the target habitat patch on connectivity

∑∑ ̅ where nc is the number of control patches.

Mean connectivity from test to control patches. Evaluates the effect of pond distance and fragmentation within the source patch on connectivity. It is computed as

∑∑ ̅

We compute ̅ ,̅ and ̅ for each scenario and average them across replicates. stcendfrmentiwiintdonocstcnvity. tisoutMean connectivity values adjusted for quality are denoted ′, ′ and ′, respectively, and are computed by replacing A with Q in the above equations. We combine habitat patch fragmentation (P) with pond distance (d) in a fully factorial design, using the following values of P = 4 and 9 and of d = 100, 200 and 300. For each of the six combinations of P and d, we construct a

41 Chapter One

series of ′ and ′ with z-values ranging from 0.7 to 2.0. We transform the ′ values into a relative connectivity value (R) by dividing with the corresponding mean connectivity found when P = 1 and d = 100, i.e. R = ′(P,d) / ′ (P=1, d=100). For any given combination of pond distance and habitat fragmentation, this value expresses the relative effect habitat quality has on mean connectivity.

We use a two-way ANOVA to test for the effect of the number of summer habitat fragments within the habitat patch and their distance to the breeding pond on mean connectivity values. We also use an ANOVA to test for effects of patch structure on the proportion of juveniles settling in their natal patch.

Results

Mean connectivity between test patches is clearly affected by the structure of the habitat patches, showing strong interactions between pond distance and fragmentation (F8,891 = 37.5, p < 0.0001) (fig. 2a, table 2a). Pond distance has a positive effect on the connectivity between highly fragmented habitat patches, while the effect is the opposite between non-fragmented patches. Moreover, fragmentation of habitat patches has a positive effect on connectivity, especially between habitat patches with long pond distance. Intra-patch structure also affects the number of juveniles being intercepted by summer habitat within the natal habitat patch and, thus, prevented from dispersing. The probability of staying in the home patch increases with fragmentation and decreases with distance (F8,891 = 19196, p < 0.0001) (fig. 2b, table 2d).

Distinguishing between outward and inward movements between test and control patches enables us to tease apart the effects target and source patch structure, respectively, have on connectivity. Mean connectivity from test to control patches shows that decreasing fragmentation and increasing pond distance within source patches promotes connectivity

(F8,891 = 58.1, p < 0.0001). The highest connectivity is found in non-fragmented source patches with long pond distance. Lowest connectivity is found in fragmented source patches with short pond distance (fig. 3a, table 2b). Mean connectivity from control to test patches reveals target patch structure to have an opposite effect on connectivity. Fragmentation of summer habitat in the target patches promotes connectivity while distance between breeding pond and summer habitat has a negative effect (F8,891 = 154.5, p < 0.0001) (fig. 3b, table 2c).

42 Chapter One

The relative effect of habitat quality on mean connectivity (R) will decrease with increasing z-values. When R > 1, the quality-weighted connectivity for a given scenario is greater than the mean connectivity in a non-fragmented habitat patch with high landscape complementation (i.e. P = 1, d = 100 m). At R-values below 1 the effect of habitat patch structure reduces connectivity compared to the control habitat patches. The threshold z-values at which R shifts below 1 depends on patch structure. In less fragmented target patches threshold-values range from 1 – 1.6 as pond distance decreases. Target patches with nine fragments exhibit a much narrower range of thresholds with z-values between 1.2 and 1.4 (fig. 4a). Con- nectivity is negatively affected by fragmentation in source patches, which again is reflected in the z-thresholds (fig. 4b). Here most thresholds are less than 1, indicating that the negative effect of patch structure on connectivity only can be counterbalanced if fragmentation has a positive effect on habitat quality.

Discussion

Inter-patch distance is widely recognised as a key factor for dispersal success (Goodwin and Fahrig 2002; Gustafson and Gardner 1996; Hanski 1998; Prevedello and Vieira 2010). All else being equal, increasing inter-patch distances means more time spent in inhospitable matrix habitat, with consequently higher mortality rates. In our model, the setup ensures equal inter-pond distances, thus, if no other factors interfered we would expect dispersal success to be the same between all pond pairs. Our results show that this is not the case; dispersal success varies depending on the structure of the habitat patches. The distribution of summer habitat fragments within source as well as target patches is important for emigration and immigration probabilities.

Emigration probability depends on the chances of not being retained by summer habitat within the home (habitat) patch. This probability increases the further away from the breeding pond the summer habitat is found and decreases the more fragmented the summer habitat is. Conversely, the proportion of juveniles that are retained and thus return to their natal pond increases with fragmentation of summer habitat and decreases with pond distance. The opposite pattern is found when looking at immigration, i.e. the probability of a dispersing juvenile finding summer habitat in a new patch. Immigration probability increases with fragmentation of summer habitat but decreases with pond distance. Thus, the combination of effects creates

43 Chapter One a complex pattern of dispersal success, depending on the structure of source and target patches.

Bowman et al. (2002) suggest that for non-searching dispersers, immigration probability will be proportional to the linear dimensions of the target patch. In a simulation study, Pfen- ning et al. (2004) found immigration rate to increase with perimeter-to-area ratio; dispersers using correlated (straight) walk having the strongest effect. The dispersal patterns found in this study can be explained by similar statistical reasoning. As fragmentation of summer habitat increases, the perimeter:area ratio of summer habitat fragments also increases. In this study, the linear dimension of summer habitat fragments increases from ca 100 meters to ca 270 meters as the summer habitat gets more fragmented. Increasing pond distance causes the gabs between the summer habitat fragments to become wider. Consequently, the probability of dispersers encountering summer habitat becomes relatively smaller as pond distance increases. The effects of the p:a ratio and gab size will interact. At any given pond distance, the probability of finding summer habitat patches will be proportional to the ratio between the linear dimension of summer habitat patches and gabs. This is the same whether the movement is outbound or inbound. Successful dispersal will depend on the probability of a disperser to avoid summer habitat within the home patch and the probability of finding summer habitat in the target patch.

Exchange of individuals between habitat patches is important for the persistence of a regional population of pond-breeding amphibians (Marsh and Trenham 2001). Our results suggest that fragmentation of summer habitat within target habitat patches can have positive effects on dispersal success. However, intra-patch structure may also affect the persistence of the local population; habitat fragmentation is in general thought to have a negative effect on habitat quality (Jaeger 2000; Pillsbury and Miller 2008; Vos and Chardon 1998). The same spatial distribution that promotes regional persistence, thus, seems to impair local persistence. The results suggest that adjusting connectivity for the effect of fragmentation on target quality may offset the positive effects of fragmentation on dispersal success. This, however, will depend on how strongly fragmentation is assumed to affect habitat quality (i.e. the z-value) and the structure of the target patch. We find that the threshold values in target patches very much depend on patch structure. For some landscapes a downscaling of effective area to 0.48 ha is needed before positive effects are turned into negative. In source patches fragmentation reduces connectivity and this pattern is not changed when adjusting for habitat quality, unless

44 Chapter One we assume positive quality effects of fragmentation and up-scale effective area to 2 ha. Our simulations also reveal an increase in dispersers settling in their natal patch as the habitat gets more fragmented. This will benefit the local population by increasing the population size. Thus, the negative effects fragmentation can have on habitat quality may be reduced by an increased recruitment of juvenile frogs.

Like fragmentation, pond distance may affect habitat quality. In the breeding season, mature frogs move between the summer habitat and the breeding pond. Thus, longer distances through the matrix may induce higher mortality. Furthermore, breeding ponds with high quality summer habitat in the immediate surroundings tend to have higher juvenile survival and thus more dispersers (Hamer and McDonnell 2008; Puglis and Boone 2012; Todd and Rothermel 2006). For the sake of simplicity, we have chosen not to incorporate these effects into the quality-weighted connectivity measure. We expect, though, that the negative effect a long pond distance will have on local population size, at least partly, will negate the positive effect on dispersal success.

Conclusion

To our knowledge, this is the first study looking at the effect of intra-patch structure on connectivity. We find that the realised connectivity depends on internal structure of both the target and the source patch as well as on how habitat quality is affected by patch structure. Al- though fragmentation is generally thought to have negative effects on connectivity, our results suggest that, depending on patch structure and habitat quality, positive effects on connectivity may occur. Connectivity is frequently used in conservation planning and studies on pond breeding amphibian often use distance between breeding ponds as a measure of dispersal ability. This study emphasises that complex interactions between individuals and landscape elements in both source and target patches determine the connectivity between habitat patches.

Acknowledgments

This study is part of a PhD-project funded by the Danish Road Directorate. We thank Uta Berger and Volker Grimm for valuable comments on the model ODD.

45 Chapter One

References

Baker JMR, Halliday TR (1999) Amphibian colonization of new ponds in an agricultural landscape. Herpetological Journal 9(2):55-63 Bender DJ, Fahrig L (2005) Matrix structure obscures the relationship between interpatch movement and patch size and isolation. Ecology 86(4):1023-1033 Bowman J, Cappuccino N, Fahrig L (2002) Patch size and population density: the effect of immigration behavior. Conservation Ecology 6(1) Chin KS, Taylor PD (2009) Interactive effects of distance and matrix on the movements of a peatland dragonfly. Ecography 32(5):715-722 Dunning JB, Danielson BJ, Pulliam HR (1992) Ecological processes that affect populations in complex landscapes. Oikos 65(1):169-175 Fahrig L (2003) Effects of habitat fragmentation on biodiversity. Annual Review of Ecology Evolution and Systematics 34:487-515 Goodwin BJ, Fahrig L (2002) How does landscape structure influence landscape connectivity? Oikos 99(3):552-570 Grimm V, Berger U, Bastiansen F, Eliassen S, Ginot V, Giske J, Goss-Custard J, Grand T, Heinz SK, Huse G, Huth A, Jepsen JU, Jorgensen C, Mooij WM, Muller B, Pe'er G, Piou C, Railsback SF, Robbins AM, Robbins MM, Rossmanith E, Ruger N, Strand E, Souissi S, Stillman RA, Vabo R, Visser U, DeAngelis DL (2006) A standard protocol for describing individual-based and agent-based models. Ecological Modelling 198: 115-126 Grimm V, Berger U, DeAngelis DL, Polhill JG, Giske J, Railsback SF (2010) The ODD protocol: A review and first update. Ecological Modelling 221:2760-2768 Gustafson EJ, Gardner RH (1996) The effect of landscape heterogeneity on the probability of patch colonization. Ecology 77(1):94-107 Hamer AJ, McDonnell MJ (2008) Amphibian ecology and conservation in the urbanising world: A review. Biological Conservation 141(10):2432-2449 Hanski I (1998) Metapopulation dynamics. Nature 396(6706):41-49 Hanski I, Simberloff D (1997) The Metapopulation Approach, Its history, Conceptual domain, and Application to Conservation. In: Hanski I. and Gilpin M. E. (eds), Metapopulation Biology: ecology, genetics, and evolution. Academic press, Inc., Hartung H (1991) Untersuchung zur terrestrischen Biologie von Populationen des Moorfro- sches (Rana arvalis NILSSON 1842) unter besonderer Berücksichtigung der Jahresmobilität. Universität Hamburg Haynes KJ, Cronin JT (2006) Interpatch movement and edge effects: the role of behavioral responses to the landscape matrix. Oikos 113(1):43-54 Haynes KJ, Diekotter T, Crist TO (2007) Resource complementation and the response of an insect herbivore to habitat area and fragmentation. Oecologia 153(3):511-20 Hels T (2002) Population dynamics in a Danish metapopulation of spadefoot toads Pelobates fuscus. Ecography 25(3):303-313

46 Chapter One

Jaeger JAG (2000) Landscape division, splitting index, and effective mesh size: new measures of landscape fragmentation. Landscape ecology 15(2):115-130 Jaquiéry J, Guélat J, Broquet T et al (2008) Habitat-quality effects on metapopulation dynamics in greater white-toothed shrews, Crocidura russula. Ecology 89:2777-85 Johnson JR, Knouft JH, Semlitsch RD (2007) Sex and seasonal differences in the spatial terrestrial distribution of gray treefrog (Hyla versicolor) populations. Biological Conservation 140(3-4):250-258 Kindlmann P, Burel F (2008) Connectivity measures: a review. Landscape Ecology 23(8):879-890 Lehtinen RM, Ramanamanjato JB, Raveloarison JG (2003) Edge effects and extinction proneness in a herpetofauna from Madagascar. Biodiversity and Conservation 12(7):1357- 1370 Loman J (1984) Density and survival of Rana arvalis and Rana temporaria. Alytes 3:125-134 Loman J (1994) Site tenacity, within and between summers, of Rana arvalis and Rana temporaria. Alytes 12(1):15-29 Marsh D (2008) Metapopulation viability analysis for amphibians. Animal Conservation 11(6):463-465 Marsh DM, Trenham PC (2001) Metapopulation dynamics and amphibian conservation. Con- servation Biology 15(1):40-49 Moilanen A, Hanski I (1998) Metapopulation dynamics: Effects of habitat quality and landscape structure. Ecology 79(7):2503-2515 Moilanen A, Hanski I (2001) On the use of connectivity measures in spatial ecology. Oikos 95(1):147-151 Moilanen A, Hanski I (2006) Connectivity and metapopulation dynamics in highly fragmented landscapes. In: Crooks K. R. and Sanjayan M. (eds), Connectivity Conservation. Cambridge University Press, Moilanen A, Nieminen M (2002) Simple connectivity measures in spatial ecology. Ecology 83(4):1131-1145 Pfenning B, Hovestadt T, Poethke HJ (2004) The effect of patch constellation on the exchange of individuals between habitat-islands. Ecological Modelling 180(4):515-522 Pillsbury FC, Miller JR (2008) Habitat and landscape characteristics underlying anuran com- munity structure along an urban-rural gradient. Ecological Applications 18(5):1107-1118 Pontoppidan M-B, Nachman G (in prep.) SAIA – a management tool for assessment of road effects on regional populations of Moor frogs (Rana arvalis) Pope SE, Fahrig L, Merriam NG (2000) Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81(9):2498-2508 Prevedello JA, Vieira MV (2010) Does the type of matrix matter? A quantitative review of the evidence. Biodiversity and Conservation 19(5):1205-1223 Puglis HJ, Boone MD (2012) Effects of terrestrial buffer zones on amphibians on golf courses. PLoS ONE 7(6):e39590

47 Chapter One

Ricketts TH (2001) The matrix matters: Effective isolation in fragmented landscapes. Ameri- can Naturalist 158(1):87-99 Ries L, Fletcher RJ, Battin J, Sisk TD (2004) Ecological responses to habitat edges: Mecha- nisms, models, and variability explained. Annual Review of Ecology Evolution and Systemat- ics 35:491-522 Rothermel BB (2004) Migratory success of juveniles: A potential constraint on connectivity for pond-breeding amphibians. Ecological Applications 14(5):1535-1546 Semlitsch RD (2008) Differentiating migration and dispersal processes for pond-breeding amphibians. Journal of Wildlife Management 72(1):260-267 Sinsch U (2006) Orientation and navigation in Amphibia. Marine and Freshwater Behaviour and Physiology 39(1):65-71 Sjögren-Gulve P (1998) Spatial movement patterns in frogs: Target-oriented dispersal in the pool frog, Rana lessonae. Ecoscience 5(1):31-38 Smith MA, Green DM (2005) Dispersal and the metapopulation paradigm in amphibian ecology and conservation: are all amphibian populations metapopulations? Ecography 28(1):110- 128 Todd BD, Rothermel BB (2006) Assessing quality of clearcut habitats for amphibians: Effects on abundances versus vital rates in the southern toad (Bufo terrestris). Biological Conserva- tion 133(2):178-185 Tramontano R (1998) The post-breeding migration of the European common frog, Rana temporaria: effects of landscape structure and seasonal weather. Lund University Vandermeer J, Carvajal R (2001) Metapopulation dynamics and the quality of the matrix. American Naturalist 158(3):211-220 Vos CC, Chardon JP (1998) Effects of habitat fragmentation and road density on the distribution pattern of the moor frog Rana arvalis. Journal of Applied Ecology 35(1):44-56 Watling JI, Nowakowski AJ, Donnelly MA, Orrock JL (2011) Meta-analysis reveals the importance of matrix composition for animals in fragmented habitat. Global Ecology and Bio- geography 20(2):209-217 Wiens JA (1997) Metapopulation Dynamics and Landscape Ecology. In: Hanski I. and Gilpin M. E. (eds), Metapopulation Biology: ecology, genetics, and evolution. Academic press, Inc., Wilensky U (1999) NetLogo. Center for Connected Learning and Computer-Based Modeling, Northwestern University, Evanston, IL. , http://ccl.northwestern.edu/netlogo,

48 Chapter One

Tables

Table 1 Default settings of parameters and the range of parameter values used in the simulations

Parameter Description Default Test values

Effect of summer habitat frag- z 1 0.7 – 2.0 mentation on habitat quality

A Area of summer habitat 0.81 ha (81 cells )

Distance between pond and d 100m (10 cells) 100, 200, 300 m summer habitat

Number of summer habitat P 1 1, 4, 9 fragments

Weight of habitat type on direc- Matrix = 1 a tion Summer habitat = 2

Matrix = 0.985 s Daily survival Summer habitat = 0.995

49 Chapter One

Table 2 Anova test results. Effect of pond distance (d) and number of summer habitat frag-

ments (P) on mean connectivity ̅ , ̅ ,̅ and the proportion of non-dispersing juveniles (1)

Source df Model a Model b Model c Model d

F p F p F p F p d 2 36.9 < 0.0001 5.1 0.0004 3.6 0.006 49871.7 < 0.0001

P 2 6.7 0.0013 107.2 < 0.0001 160.9 < 0.0001 20113.4 < 0.0001

P*d 4 69.4 < 0.0001 115.1 < 0.0001 449.7 < 0.0001 3399.4 < 0.0001

(1) Model a evaluates the overall effect of patch structure while model b and c analyses the effect of source patch structure and target patch structure, respectively. Model d tests the effect of source patch structure on the proportion of dispersing juveniles settling in their natal habitat patch.

50 Chapter One

Figure legends

Figure 1:

Model landscape with nine habitat patches. Each habitat patch is defined by a central breeding pond (black dot), fragments of summer habitat (dark grey shapes) and matrix (light gray) within the habitat patch boundary. The model landscape contains five control (habitat) patches and four test (habitat) patches (see text). Distance between breeding pond and summer habitat is denoted d.

Figure 2:

Effect of habitat patch structure on a) mean connectivity between test patches, b) proportion of dispersing juveniles settling in their natal habitat patch.

Figure 3:

Effect of (a) target patch structure and (b) source patch structure on mean connectivity between habitat patches serving as test and control patches, respectively.

Figure 4:

Relative effect of habitat quality on mean connectivity (R) in a) target patches and b) source patches for different landscape scenarios. At a z-value equal to 1 (dotted, vertical line), the effective area equals the real area. R-values along this line represent the relative effect a particular patch structure has on mean connectivity. When R = 1 (dashed, horizontal line), the quality weighted connectivity for a given landscape corresponds to the mean connectivity between habitat patches serving as control patches (i.e. scenario with P = 1, d = 100). The z- value of the intercept between the dashed line and a given curve can be interpreted as the threshold at which the habitat quality is low enough to shift a positive effect on connectivity into a negative effect.

51 Chapter One

Figures

Figure 1

52 Chapter One

Figure 2

(a) (b)

Figure 3

(a) Target patch (b) Source patch

53 Chapter One

Figure 4

(a) Target patch (b) Source patch

54 Chapter One

Supplementary material - Appendix A

Model ODD

1. Purpose The purpose of the model is to analyse the effect of intra-patch structure on inter-patch dispersal success and connectivity.

2. Entities, state variables, and scales Breeding ponds are treated as stationary agents. Each pond is characterized by a unique id- number, the habitat patch type (test or control patch) and the quality-weighted area of associated summer habitat. Frog-agents are characterized by the pond in which they are hatched and the breeding pond they immigrate to. The extent of the model landscape is 300 x 300 grid cells, and each grid cell represents 10 x 10 m. Grid cells belong either to Matrix habitat or Summer habitat. Each habitat type is associated with a habitat-attraction (indicating how will- ing frogs will be to enter the habitat) and the frogs’ daily survival probability in the habitat. The simulation runs for 120 time steps, each representing one day.

3. Process overview and scheduling 500 Frog-agents are located at each of the Pond-agents. During each time-step the following procedures are executed: Settle (evaluates if a Frog-agent stops dispersing), Move (movement of dispersing Frog-agents), Homing (movement of settled Frog-agents towards breeding pond) and Survival (evaluates if a Frog-agent survives). The simulation stops at time step 120 and the procedure Connectivity is run, computing dispersal rates and connectivity.

4. Design concepts

Emergence Immigration rates will emerge as a response to the landscape configuration.

Adaptation & Objectives To avoid desiccation and thereby increase survival Frog-agents are assumed to move in response the moistness of its surroundings. The moister, in general, a habitat is the more attractive the habitat is for the frog as indicated by the habitat-attraction parameter a. Dispersing juvenile Moor frogs have an innate tendency to move away from their natal pond. In the simu-

55 Chapter One lation the movement of the Frog-agents is thus oriented in a random direction away from the pond, and they are not allowed to backtrack.

Sensing Frog-agents are assumed to be aware of their own state-variables. Frog-agents are also aware of the habitat-attraction of the grid cells, as well as the identity of the ponds.

Interaction There is no interaction between frog-agents. Movement decisions of the Frog-agents depend on the habitat type of the neighbouring cells. Survival of the Frog-agents depends on the daily survival rates of the traversed habitat.

Stochasticity Which cell to move to is chosen randomly from neighbouring cells with the probability of being chosen weighted by the habitat-attraction and the number of neighbouring cells with the same value. If Frog-agents occupy a cell with summer habitat they will stop dispersing with a certain probability. The probability increases with time.

Observation At the end of each run the number of dispersers settled at each breeding pond and their natalpond is registered and the following are calculated (see also Method-section in main text):

 Mean connectivity between test patches  Mean connectivity from control to test patches  Mean connectivity from test to control patches  Proportion of dispersers settled in home habitat patch

5. Initialization Nine habitat patches are constructed according to the test scheme (see Method section in main text). Map-scan procedure is run calculating the weighted area of summer habitat in each habitat patch. 500 Frog-agents are located on each pond-agent, their direction set randomly.

6. Input data No input data are used

56 Chapter One

7. Submodels

Map-scan The procedure delimits individual summer habitat fragments within the habitat patch and computes the effective (quality-weighted) area of summer habitat. Interconnected summer habitat cells define an individual summer habitat fragment. The area (A) of each summer habitat fragment is calculated as the sum of cells within the patch. The effective area of summer / th habitat is, subsequently, calculated as ∑ where Ak is the area of the k summer habitat fragment, P the number of summer habitat fragments within the habitat patch, and z a constant weighting the effect of fragmentation on habitat quality (z > 0).

Settle Dispersing frogs occupying a summer habitat cell has a certain probability of settling and cease moving. The initial settling probability is 0, increasing every time step with 0.04 and ending at 0.96.

Move Assuming the frog to head in the direction it was assigned when it left the natal pond, a frog can move to one of its neighbouring cells located within ±1100 from the preferred direction. Based on the habitat-attraction of the neighbouring cells frogs decide which cell-type they want to move to. The probability of moving into cell i is a function of habitat attraction (a) of

the accessible neighbouring cells (n): . A uniform pseudorandom number is se- ∑ lected to choose the cell to move to. Once a cell is chosen the frog moves to a random position within the cell. The direction of the frogs does not change.

Homing After time step 75 settled frogs have their direction set towards their breeding pond and they start moving again following the Move-procedure. If they get within a distance of 2 cells from the breeding pond, the frogs move directly to the breeding pond and stay there.

Survival For each frog a pseudo-random number is drawn between 0 and 1. If the number exceeds the daily survival probability of the frog’s current cell, the frog dies.

57 Chapter One

Connectivity At the end of each simulation, dispersal probabilities are computed for all pair wise combinations of habitat patches and connectivity measures are computed as described in the method section.

CHAPTER TWO

CHANGES IN BEHAVIOURAL RESPONSES

TO INFRASTRUCTURE AFFECTS LOCAL

AND REGIONAL CONNECTIVITY —

A SIMULATION STUDY ON POND BREEDING

AMPHIBIANS

Submitted to Nature Conservation

December 2012

60 Chapter Two

Changes in behavioural responses to infrastructure affects local and regional connectivity — a simulation study on pond breeding amphibians

Maj-Britt Pontoppidan, Gösta Nachman

Section for Ecology and Evolution Department of Biology University of Copenhagen Universitetsparken 15 DK-2100 Copenhagen

Corresponding author: M-B. Pontoppidan email: [email protected] phone: +45 51518791

61 Chapter Two

Abstract

An extensive and expanding infrastructural network destroys and fragments natural habitat and has detrimental effect on abundance and population viability of many amphibian species. Roads functions as barriers in the landscape. They separate local populations from each other or prevent access to necessary resources. Therefore, road density and traffic intensity in a region may have severe impact on regional as well as local connectivity. Amphibians may be able to detect and avoid unsuitable habitat. Individuals’ ability to avoid roads can reduce road mortality but at the same time road avoidance behaviour, can increase the barrier effect of the road and reduce connectivity. We use an individual based model to explore how changes in road mortality and road avoidance behaviour affect local and regional connectivity in a population of Moor frogs (Rana arvalis). The results indicate that road mortality has a strong negative effect on regional connectivity, but only a small effect on local connectivity. Regional connectivity is positively affected by road avoidance and the effect becomes more pronounced as road mortality decreases. Road avoidance also has a positive effect on local connectivity. When road avoidance is total and the road functions as a 100% barrier regional connectivity is close to zero, while local connectivity exhibit very elevated values. The results suggest that roads may affect not only regional or metapopulation dynamics but also have a direct effect on local population dynamics.

62 Chapter Two

Introduction

All over the world, amphibian populations are declining and many amphibian species are listed in the IUCN as threatened or vulnerable (IUCN 2012). The causes for the decline are hypothesized to be (combinations of) factors such as climate change, diseases, predation and UV-radiation. But the main factor, especially in the western world, is thought to be the increasing urbanisation (Alford and Richards 1999; Beebee and Griffiths 2005; Collins and Storfer 2003; Gardner et al. 2007). The negative effect of urbanisation is not only due to changes in land use and destruction of habitat. A huge infrastructural network functions as barriers to movement and causes the death of a huge number of amphibians every year (An- drews et al. 2008; Hamer and McDonnell 2008). Road density in an area as well as traffic density on individual roads have been shown to have a negative effect on amphibian populations (Eigenbrod et al. 2009; Fahrig and Rytwinski 2009; Hels and Buchwald 2001; Vos and Chardon 1998). Veysey et al. (2011) even found road density to have a stronger effect on population size than habitat availability, while Carr and Fahrig (2001) found more vagile species to be more vulnerable to road mortality.

Very little literature exists on amphibians’ reactions to road. Amphibians are able to recognise and avoid unsuitable habitat. Although there are species specific variations, individuals tend to prefer more shady and moist habitat types (Mazerolle 2005; Mazerolle and Desrochers 2005; Popescu and Hunter 2011; Vos et al. 2007). In more open and dry habitats like fields and clear-cuts, water loss is bigger and survival lower resulting in avoidance of such habitats (Rothermel and Semlitsch 2002; Todd and Rothermel 2006). Individuals also tend to move more quickly in inhospitable habitats (Hartung 1991; Tramontano 1997). These observations suggest that amphibians should be able to avoid roads, however, the number of road kills suggests the opposite (Elzanowski et al. 2009) and the only study on this topic did not find any indication of road avoidance in Rana pipiens (Bouchard et al. 2009).

Pond breeding amphibians require both terrestrial and aquatic habitat to complete their life cycle. Proximity between the required habitat types is important for the survival of the population. Loss of, or diminished access to, one or both habitats will affect population size and persistence probability (Dunning et al. 1992; Haynes et al. 2007; Johnson et al. 2007; Pope et al. 2000). Moreover, populations of pond-breeding amphibians are frequently considered to be structured as a regional network or a metapopulation, making dispersal between

63 Chapter Two subpopulations essential to regional population persistence (Hels 2002; Marsh 2008; Marsh and Trenham 2001; Smith and Green 2005). Thus the barrier effect caused by roads may have severe consequences for populations of pond breeding amphibians.

We have developed an individual based model to assess the effects of infrastructure on landscape connectivity. The model is part of a larger study concerning road effects on regional populations of Moor frogs (Rana arvalis). In this paper we present our model and explore how behavioural responses to infrastructure may affect local and regional connectivity. The ability to avoid roads may diminish the amount of road kills. This behaviour will prevent dispersal across the road but at the same time it may affect connectivity locally. Lower levels of road avoidance can reduce the road’s barrier effect but this will probably depend on the level of road mortality. We hypothesize that

 Regional connectivity will be inhibited by high levels of road avoidance and high road mortality and will depend on interactions between the degree of road avoidance and road mortality  Local connectivity will be promoted by high levels of road avoidance but not be affected by road mortality

We use a real Danish landscape with a population of Moor frogs traversed by a large road to test how regional and local connectivity are affected by changes in road mortality and road avoidance.

Methods

We use an individual based model to simulate the movements of juvenile Moor frogs and estimate immigration probabilities between habitat patches. The purpose of the model is to measure the connectivity of the landscape. In the following we use the terms dispersal and migration as defined by Semlitsch (2008), i.e. dispersal is “interpopulational, unidirectional movements from natal sites to other breeding sites” and migration is “intrapopulational, round-trip movements toward and away from aquatic breeding sites”. The habitat of pond breeding amphibians as the Moor frog includes terrestrial as well as aquatic habitat. Therefore we define the habitat patch of a subpopulation as a complementary habitat patch containing not only the breeding pond but also all accessible summer habitat within migration distance from the pond (Dunning et al. 1992; Pontoppidan and Nachman In review; Pope et al. 2000).

64 Chapter Two

Model species Moor frogs spend most of their life in terrestrial habitat; aquatic habitat is only used during the breeding season, which takes place in the early spring (Elmberg 2008; Glandt 2008; Har- tung 1991). Soon after breeding, the frogs return to the summer habitat, which lies mostly within a 400 m radius from the breeding pond (Elmberg 2008; Hartung 1991; Kovar et al. 2009). Adult frogs show a high degree of site fidelity and often use the same breeding pond and summer habitat from year to year (Loman 1994). Long distance dispersal in Moor frogs takes place predominantly during the juvenile life-stage (Semlitsch 2008; Sinsch 1990; 2006). Shortly after metamorphosis, the young frogs leave the natal pond and disperse into the surrounding landscape seeking out suitable summer habitat. Dispersal distances are between a few hundred meters up to 1-2 kilometres (Baker and Halliday 1999; Hartung 1991; Sinsch 2006; Vos and Chardon 1998). The juveniles stay in terrestrial habitat 2-3 years until they reach maturity, although some observations indicate that juvenile frogs follow the adults during the spring migration, without entering the breeding ponds (Hartung 1991; Sjögren-Gulve 1998).

Model overview The model considers a regional population of Moor frogs within a spatially explicit landscape matrix. The landscape is constructed from a 600 x 800 cell GIS raster map, each cell representing an area of 10 x 10 meters. A raster cell is characterised by a set of variables defining the habitat type and its value in regard to the different aspects of the life cycle and behaviour of the Moor frog (Table 1). Potential sites for subpopulations of Moor frogs are represented by a GIS point-data set of ponds surveyed during field work. Each pond is defined by an ID- number, a quality index and the summer habitat fragments located within migration distance from the pond (Table 1).

Immigration requires two events: 1) the successful dispersal of a juvenile frog to summer habitat outside its natal habitat patch and 2) subsequent successful migration from the new summer habitat to a nearby breeding pond. In real life dispersal starts just after metamorphosis in early summer and lasts until hibernation in the autumn. The second part of the immigration event, migration, takes place in the spring 2.5 years later. For simplicity, we simulate the dispersal and breeding migration, as if they take place in the same year.

65 Chapter Two

The time step of the model is one day and the simulated period for dispersal as well as migration is 120 days each. At the start of a simulation, 500 frog agents are created at each pond. Each agent is assigned a random direction, which determines its preferred direction of movement. This direction does not change unless summer habitat is found. At each time step, a random daily travelling distance is chosen for each agent; the distance depending on the attractiveness of the current habitat. The distance is travelled one cell at a time, moving to one of the neighbouring cells, depending on the relative attractiveness of the cells, although back- wards movement is not allowed. The movement rules generate a biased random walk away from the natal pond and in the preferred direction. During dispersal, frog agents encountering a cell with summer habitat will have a certain probability of settling in the habitat and stop dispersing. This probability will increase with time. At the end of the dispersal period all frog agents that have not settled in summer habitat are removed. Starting the migration phase, the remaining frog agents move toward the breeding pond associated with their summer habitat; in case several breeding ponds are available one is chosen randomly weighted by pond quality. After each time step, the survival probability of every frog agent is assessed, based on the daily survival rates associated with the habitat type traversed during the day. Full model documentation is found in Appendix 1 in the supplementary material, following the ODD- template suggested by Grimm et al. (2006; 2010). Netlogo v.4.1.3 (Wilensky 1999) is used as modelling environment (freely downloadable at http://ccl.northwestern.edu/netlogo).

Input data We use GIS data sets from a road project in Denmark, supplied by the Danish Road Director- ate and Amphi Consult. The project concerns an area in north-western part of Zealand, 10 km east of the city of Kalundborg (55° 40.14’ N 11° 17.85’ E) (Fig. 1). The area is characterised as semi-urban and agricultural landscapes, traversed by creeks and wetlands. A project data set contains a land cover map of the area and a point-data set of potential breeding ponds found during field surveys. The land cover maps are constructed following a protocol de- signed by amphibian experts (Hassingboe et al. 2012), in which a range of different habitat types are identified. Each habitat type has been assessed and ranked on a scale from 1- 5, for the following three variables: the habitat’s relative suitability as summer habitat (Hq), its relative attraction to frogs during movement (Ha) and the relative survival probability (Hs) in the habitat. In the model the survival index (Hs) is converted into a daily survival probability (Ds) (see appendix 1 for details). Infrastructural elements like roads and railways are processed as

66 Chapter Two any other habitat type and assigned values of habitat attraction and daily survival. However, in the literature the terms “road avoidance” and “road mortality” are more commonly used. To avoid confusion when discussing these effects, we therefore convert (Ha) and (Ds) to road avoidance (Ra) and road mortality (Rd), respectively, and invert the ranking, i.e. Ra = 6 - Ha and Rd = 1 - Ds.

The point-data set contains information on the location of the potential breeding pond, its ID-number as well as a quality index (Q). Pond qualities ranges from 0.1 – 1 and relates to the suitability of the pond and the immediate surroundings in regard to egg and larval survival and are estimated by experts during field work. In this paper we have excluded low-quality ponds (Q < 0.6), since they per definition have a low probability of maintaining a population on their own. The extent of the map is 6x8 km and it contains 40 ponds.

Scenarios

We create scenarios with increasing values of road avoidance, Ra = [1; 2; 3; 4; 5], and road mortality, Rd = [0.1; 0.3; 0.5; 0.7; 0.9], of the two major roads cutting through the map (Fig.1, roads shown in red). We run 25 simulations for every combination of the parameter values of

Ra and Rd. As Ra increases the willingness of the frog agents to enter the road will decrease, while the probability of surviving will increases with decreasing values of Rd.

Output At the end of each simulation, the natal pond and the breeding pond of all frog agents are registered and immigration probability (pij) between all pair-wise ponds is calculated. Landscape connectivity (S) is then found as

∑∑ , (eq.1)

Local populations are identified by grouping ponds into clusters depending on their mutual connectivity, using the method of unweighted, arithmetic, average clustering as described by Legendre and Legendre (1998). Since, immigration probabilities between any two ponds are not necessarily symmetric, i.e. pij ≠ pji, we use summed immigrations probabilities as similarity measure (m): mij = pij + pji. The threshold at which a given pond or cluster no longer can be added to another cluster is set to mij ≤ 0.01. We define local connectivity as the connectivity within a cluster and regional connectivity is defined as the connectivity between all

67 Chapter Two pair-wise combinations of clusters. Based on the clustering result we compute within-cluster connectivity (Sc) for each cluster as

∑∑ , (eq. 2) where nc is the number of ponds belonging to cluster c. Connectivity between clusters (Sb) is then found as Sb = S – Sc. However, to be able to detect changes in local connectivity, the ponds constituting a cluster must be the same in all scenarios. Therefore, we use the cluster configuration found when Ra is set to 5 to define clusters, and use this in all calculations of within-cluster connectivity.

We use a multiple regression model, with the general form y = β0 + β1Rd + β2Ra +

β3RdRa + ε., to test for the effect of road avoidance (Ra), road mortality (Rd ) and their interaction on landscape connectivity (S), within-cluster connectivity (Sc) and between-cluster connectivity (Sb). Sequential Holm-Bonferroni correction is used to adjust p-values. When Ra is set to 5, frog agents to do not enter the road, therefore the level of road mortality is inconsequential. Moreover, preliminary tests showed extreme connectivity values when the road is 100% blocked. Both of these factors risk masking the statistical effect of road mortality and road avoidance on connectivity at other levels of Ra. Consequently, the results from the scenarios with Ra = 5 are excluded from the statistical testing.

Results

Analyses of the scenarios with Ra = 5 identifies seven clusters (Fig. 2A Table 2). Cluster c1 contains four ponds and is located rather remotely in the top of the map. Clusters c2 and c3 are found in areas close to where the two test roads cross and contains four, respectively, five ponds. Cluster c4 and cluster c5 contains seven and nine ponds, respectively. These are more widespread clusters situated on either side of the road in the middle of the map. The last two clusters c6 and c7 are placed near the bottom of the map and contain two and six ponds. As described in the method section we use this cluster configuration as a reference for all scenarios when calculating within-cluster connectivity. Nonetheless, the analyses show that cluster configurations do not change with the different scenarios except when road mortality is set to 0.1. In this case dispersal success is sufficiently high between cluster c4 and c5 and they fuse into one cluster (Fig. 2B).

68 Chapter Two

When Ra ≤ 4, road mortality has strong negative effect on the connectivity between clusters (Sb) while road avoidance has a positive effect. Furthermore, there is an interaction effect; the effect of road avoidance becomes more pronounced as road mortality decreases

(Table 3). These effects are all statistically significant (F3,496 = 1814, p <0.001). When Ra = 5, between-cluster connectivity yields its lowest values (Fig. 3A). In these scenarios all dispersal across the road is impossible. Hence, the measured connectivity must represent the connectivity between clusters on the same side of the road.

Overall landscape connectivity (S) exhibits very elevated values when road avoidance is total. Otherwise landscape connectivity decreases with road mortality and increases with road avoidance, the effect of road avoidance being strongest at lower values of road mortality (Fig.

3B). All effects are statistically significant (F3,496 = 1297, p <0.001) (Table 3).

In general, the same trends are found in within-cluster connectivity (Sc) (Table 3). There are, however, some differences. Within-cluster connectivity of cluster c1 and c7 is affected neither by changes in road mortality nor road avoidance. These two clusters are also furthest away from the test roads (Table 2). Clusters c2 trough c6 all exhibit very elevated connectivity values when Ra = 5. Connectivity within these clusters also tend to increase with road avoidance and decrease with road mortality when Ra ≤ 4 (Fig. 4). However, the effect of road mortality is only statistically significant in clusters c4 and c5 (and after Bonferroni correction only c5). Both are large clusters with several pond members very close to the road (Table 2, Table 3). Clusters c2 - c5 are all significantly affected by road avoidance; in contrast to c6 in which the majority of member ponds are further away from the road (Table 2, Table 3).

Discussion

As hypothesized road mortality has a negative effect on between-cluster connectivity. How- ever, contrary to our expectation we find that road avoidance can promote connectivity across roads. An explanation could be that roads actually functions as traps if road avoidance is low. With this model roads are not just lines to cross; they are considered the same way as other kinds of habitat. Hence, when habitat attraction of the roads is higher (and thus road avoidance low) than the surrounding habitat, the road may actually be the preferred habitat. More- over, the survival probability on roads is always lower than in any other type of habitat. Thus, at high levels of road avoidance, frog agents only rarely enter the roads, but if they do they

69 Chapter Two quickly leave it again and only suffer the high mortality for at short time. When road avoidance is low, frog agents enter the roads more willingly and will tend to stay there, suffering from the higher mortality for a longer time. The severity of the “trap” effect will depend on road mortality as, all else being equal, successful dispersal across the road depends on the survival probability. The results do in fact show a strong interaction; the positive effect of road avoidance on between-cluster connectivity getting more pronounced as road mortality increases.

In accordance with our second hypothesis, we find that road avoidance has a positive effect on local connectivity. In particular, when road avoidance is set to five, connectivity shows considerable elevated values. The strong effect on within-cluster connectivity of a 100 % barrier may seem surprising, but can be explained as a “deflection” effect. When the road is inaccessible road mortality is no longer an issue and a larger proportion of frog agents will survive. Moreover, the blockage forces the agents to move along the road instead of crossing. Taken together, this has the effect that a larger proportion of frog agents stay within the local area for a longer time, which increases the probability of an agent settling within the cluster, enhancing within-cluster connectivity. We did not expect road mortality to have an effect on within-cluster connectivity, but we do find a negative, although week, response. This is probably because there will be a small proportion of frog agents entering the road and return- ing to the same side. The survival probability of these returnees will depend on road mortality.

The seven clusters identified in this study do not all respond in the same way on changes in road avoidance and road mortality. Two of the clusters, c1 and c7, are not affected at all; these are also the clusters furthest away from the road. Road mortality only significantly affects larger clusters with several pond members very close to the road; maybe because only these clusters have sufficient number of returnees for the effect to be detectable. Road avoidance, on the other hand, affects also clusters further away; only clusters with a minimum distance to road above 300 m are unaffected by road avoidance. The result suggests that if the road is within the summer habitat of some of the member ponds, then road avoidance will affect within-cluster connectivity.

In this study, scenarios with road avoidance set to five, corresponds to real life situations where fencing along roads prevents access to the road. Our results suggest that fencing

70 Chapter Two can result in highly increased local connectivity, even between ponds not in immediate proximity to the road. Thus, fencing may not just mitigate road induced mortality but may actually enhance local population persistence. Depending on number, quality and connectivity between subpopulation on the same side of the road a strengthening of subpopulations adjacent to road fences can potentially improve regional population persistence (Hels and Nachman 2002). However, fencing also separates a population into several smaller and more isolated groups of subpopulations, each of which may have a higher risk of extinction and a lower probability of recolonisation. In a simulation experiment with a local population of virtual animals Jaeger and Fahrig (2004) found that fencing could prolong persistence time but had little effect on persistence probability, and in most cases the population only survived on one side of the road.

The series of scenarios are hypothetical and all may not correspond to real life situations but road mortality can indeed range between very low and very high values, depending on traffic intensity. Extreme low values of road avoidance, to the point where the road becomes more attractive than the surrounding landscape, may seem very unrealistic. However, behavioural responses to traffic like immobilisation (Mazerolle et al. 2005) can have similar effects; and after rain fall wet, dark roads may appear deceptively attractive to frogs (Andrews et al. 2008).

Our study concerns a specific landscape and a specific species, but still it is possible to draw some general conclusion. First of all our results emphasize that connectivity is context dependent. The behaviour of the focal species, the structure of the landscape and their interaction are essential to how connectivity is realized. Furthermore, our simulations indicate that roads not only affect dispersal across roads. Even between ponds located on the same side of the road, dispersal success can be highly susceptible to road avoidance and road mortality, depending on the distance to the road. This suggests that roads may affect not only regional or metapopulation dynamics but also have a direct effect on local population dynamics.

Acknowledgements

The work was funded by the Danish Road Directorate. The Amphi Consult group has pro- vided amphibian expertise as well as map data. Marianne Ujvári, Martin Hesselsøe, Agnete Jørgensen and Martin Schneekloth have given valuable feed-back during the model develop-

71 Chapter Two ment. Uta Berger has given precious help during the work and kept the main author on the IBM-track. We are very thankful to Carolyn Bauer for linguistic help.

References

Alford RA, Richards SJ (1999) Global amphibian declines: A problem in applied ecology. Annual Review of Ecology and Systematics 30: 133-165. doi:10.1146/annurev.ecolsys.30.1.133 Andrews KM, Gibbons JW, Jochimsen DM (2008) Ecological effects of roads on amphibians and reptiles: a literature review. 121-143 pp. Baker JMR, Halliday TR (1999) Amphibian colonization of new ponds in an agricultural landscape. Herpetological Journal 9: 55-63 Beebee TJC, Griffiths RA (2005) The amphibian decline crisis: A watershed for conservation biology? Biological Conservation 125: 271-285. doi:10.1016/j.biocon.2005.04.009 Bouchard J, Ford AT, Eigenbrod FE, Fahrig L (2009) Behavioral Responses of Northern Leopard Frogs (Rana pipiens) to Roads and Traffic: Implications for Population Persistence. Ecology and Society 14: Carr LW, Fahrig L (2001) Effect of road traffic on two amphibian species of differing vagility. Conservation Biology 15: 1071-1078 Collins JP, Storfer A (2003) Global amphibian declines: sorting the hypotheses. Diversity and Distributions 9: 89-98. doi:10.1046/j.1472-4642.2003.00012.x Dunning JB, Danielson BJ, Pulliam HR (1992) Ecological processes that affect populations in complex landscapes. Oikos 65: 169-175. doi:10.2307/3544901 Eigenbrod F, Hecnar SJ, Fahrig L (2009) Quantifying the Road-Effect Zone: Threshold Ef- fects of a Motorway on Anuran Populations in Ontario, Canada. Ecology and Society 14: Elmberg J (2008) Ecology and natural history of the moor frog (Rana arvalis) in boreal Swe- den. In: Glandt D, Jehle R (Eds) The Moor Frog Laurenti-Verlag, Bielefeld, 179-194 Elzanowski A, Ciesiolkiewicz J, Kaczor M, Radwanska J, Urban R (2009) Amphibian road mortality in Europe: a meta-analysis with new data from Poland. European Journal of Wild- life Research 55: 33-43. doi:10.1007/s10344-008-0211-x Fahrig L, Rytwinski T (2009) Effects of Roads on Animal Abundance: an Empirical Review and Synthesis. Ecology and Society 14: 21 Gardner TA, Barlow J, Peres CA (2007) Paradox, presumption and pitfalls in conservation biology: The importance of habitat change for amphibians and reptiles. Biological Conserva- tion 138: 166-179. doi:10.1016/j.biocon.2007.04.017 Glandt D (2008) Der Moorfrosch (Rana arvalis): Erscheinungsvielfalt, Verbreitung, Lebens- räume, Verhalten sowie Perspectiven für den Artenschutz. In: Glandt D, Jehle R (Eds) The Moor Frog. Laurenti-Verlag, Bielefeld, Grimm V, Berger U, Bastiansen F, Eliassen S, Ginot V, Giske J, Goss-Custard J, Grand T, Heinz SK, Huse G, Huth A, Jepsen JU, Jorgensen C, Mooij WM, Muller B, Pe'er G, Piou C,

72 Chapter Two

Railsback SF, Robbins AM, Robbins MM, Rossmanith E, Ruger N, Strand E, Souissi S, Stillman RA, Vabo R, Visser U, DeAngelis DL (2006) A standard protocol for describing individual-based and agent-based models. Ecological Modelling 198: 115-126. doi:10.1016/j.ecolmodel.2006.04.023 Grimm V, Berger U, DeAngelis DL, Polhill JG, Giske J, Railsback SF (2010) The ODD protocol: A review and first update. Ecological Modelling 221: 2760-2768. doi:10.1016/j.ecolmodel.2010.08.019 Hamer AJ, McDonnell MJ (2008) Amphibian ecology and conservation in the urbanising world: A review. Biological Conservation 141: 2432-2449. doi:10.1016/j.biocon.2008.07.020 Hartung H (1991) Untersuchung zur terrestrischen Biologie von Populationen des Moorfro- sches (Rana arvalis NILSSON 1842) unter besonderer Berücksichtigung der Jahresmobilität. Hamburg: Universität Hamburg. Hassingboe J, Neergaard RS, Hesselsøe M (2012) Manual til produktion af GIS raster kort til:”EDB-værktøj til at vurdere skader på bestande af padder /økologisk funktionalitet”. Am- phi Consult. Haynes KJ, Diekotter T, Crist TO (2007) Resource complementation and the response of an insect herbivore to habitat area and fragmentation. Oecologia 153: 511-520. doi:10.1007/s00442-007-0749-4 Hels T (2002) Population dynamics in a Danish metapopulation of spadefoot toads Pelobates fuscus. Ecography 25: 303-313. doi:10.1034/j.1600-0587.2002.250307.x Hels T, Buchwald E (2001) The effect of road kills on amphibian populations. Biological Conservation 99: 331-340. doi:10.1016/S0006-3207(00)00215-9 Hels T, Nachman G (2002) Simulating viability of a spadefoot toad Pelobates fuscus metapopulation in a landscape fragmented by a road. Ecography 25: 730-744. doi:10.1034/j.1600- 0587.2002.250609.x IUCN (2012) IUCN Red List of Threatened Species. Version 2012.2. Jaeger JaG, Fahrig L (2004) Effects of Road Fencing on Population Persistence. Conservation Biology 18: 1651-1657. doi:10.1111/j.1523-1739.2004.00304.x Johnson JR, Knouft JH, Semlitsch RD (2007) Sex and seasonal differences in the spatial terrestrial distribution of gray treefrog (Hyla versicolor) populations. Biological Conservation 140: 250-258. doi:10.1016/j.biocon.2007.08.010 Kovar R, Brabec M, Vita R, Bocek R (2009) Spring migration distances of some Central European amphibian species. Amphibia-reptilia 30: 367-378 Legendre P, Legendre L (1998) Numerical ecology. Elsevier Loman J (1994) Site tenacity, within and between summers, of Rana arvalis and Rana temporaria. Alytes 12: 15-29 Marsh D (2008) Metapopulation viability analysis for amphibians. Animal Conservation 11: 463-465. doi:10.1111/j.1469-1795.2008.00223.x Marsh DM, Trenham PC (2001) Metapopulation dynamics and amphibian conservation. Con- servation Biology 15: 40-49

73 Chapter Two

Mazerolle MJ (2005) Drainage Ditches Facilitate Frog Movements in a Hostile Landscape. Landscape Ecology 20: 579-590. doi:10.1007/s10980-004-3977-6 Mazerolle MJ, Desrochers A (2005) Landscape resistance to frog movements. Canadian Jour- nal of Zoology-Revue Canadienne De Zoologie 83: 455-464. doi:10.1139/z05-032 Mazerolle MJ, Huot M, Gravel M (2005) Behavior of amphibians on the road in response to car traffic. Herpetologica 61: 380-388. doi:10.1655/04-79.1 Pontoppidan M-B, Nachman G (In review) Effects of within-patch heterogeneity on connectivity in pond-breeding amphibians studied by means of an individual-based model. Web- ecology Pope SE, Fahrig L, Merriam NG (2000) Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81: 2498-2508 Popescu VD, Hunter ML, Jr. (2011) Clear-cutting affects habitat connectivity for a forest amphibian by decreasing permeability to juvenile movements. Ecological Applications 21: 1283- 1295. doi:10.1890/10-0658.1 Rothermel BB, Semlitsch RD (2002) An experimental investigation of landscape resistance of forest versus old-field habitats to emigrating juvenile amphibians. Conservation Biology 16: 1324-1332 Semlitsch RD (2008) Differentiating migration and dispersal processes for pond-breeding amphibians. Journal of Wildlife Management 72: 260-267. doi:10.2193/2007-082 Sinsch U (1990) Migration and orientation in anuran amphibians. Ethology Ecology & Evolu- tion 2: 65-79 Sinsch U (2006) Orientation and navigation in Amphibia. Marine and Freshwater Behaviour and Physiology 39: 65-71. doi:10.1080/10236240600562794 Sjögren-Gulve P (1998) Spatial movement patterns in frogs: Differences between three Rana species. Ecoscience 5: 148-155 Smith MA, Green DM (2005) Dispersal and the metapopulation paradigm in amphibian ecology and conservation: are all amphibian populations metapopulations? Ecography 28: 110- 128 Todd BD, Rothermel BB (2006) Assessing quality of clearcut habitats for amphibians: Effects on abundances versus vital rates in the southern toad (Bufo terrestris). Biological Conserva- tion 133: 178-185. doi:10.1016/j.biocon.2006.06.003 Tramontano R (1997) Continuous radio tracking of the common frog, Rana temporaria. Her- petologia Bonnensis: 359-365 Veysey JS, Mattfeldt SD, Babbitt KJ (2011) Comparative influence of isolation, landscape, and wetland characteristics on egg-mass abundance of two pool-breeding amphibian species. Landscape Ecology 26: 661-672. doi:10.1007/s10980-011-9590-6 Vos CC, Chardon JP (1998) Effects of habitat fragmentation and road density on the distribution pattern of the moor frog Rana arvalis. Journal of Applied Ecology 35: 44-56 Vos CC, Goedhart PW, Lammertsma DR, Spitzen-Van der Sluijs AM (2007) Matrix permeability of agricultural landscapes: an analysis of movements of the common frog (Rana temporaria). Herpetological Journal 17: 174-182

74 Chapter Two

Wilensky U (1999) NetLogo. Center for Connected Learning and Computer-Based Modeling, Northwestern University, Evanston, IL. , http://ccl.northwestern.edu/netlogo

Figure legends

Figure 1

Study area. A) Location of two study areas in Denmark. KaB is an area near Kalundborg on Zealand and HoB is near Holstebro in Jutland. Only KaB is used in the present analysis, but both areas are used for the parameterisation of the model. B) KaB map used in the analysis. Black dots are breeding ponds, test roads are marked with red.

Figure 2

Results from cluster analyses with two different parameter settings. A) Ra = 5, Rd = 0.1 and B)

Ra = 4, Rd = 0.9

Figure 3

Effect of road avoidance (Ra) and road mortality (Rd) on a) between-cluster connectivity (Sb) and b) landscape connectivity (S)

Figure 4

Effect of road avoidance (Ra) and road mortality (Rd) on within-cluster connectivity (Sc) in cluster c2 – c6

75 Chapter Two

Figures

Figure 1

76 Chapter Two

Figure 2

Figure 3

77 Chapter Two

Figure 4

78 Chapter Two

Tables

Table 1 List of variables characterizing the agents in the model

Value Agent Variable Notation Description range type

DailySurvival Ds Cell Daily survival probability

The cell’s relative attraction to frogs during HabitatAttraction Ha 1-5 Cell movement

HabitatCode Hc Cell Cell code for habitat type

HabitatSurvival Hs 1-5 Cell The cell’s relative survival index

The cell’s relative suitability as summer SummerQuality Hq 1-5 Cell habitat

BreedingPond Frog Breeding pond of frog agent

NatalPond Frog Natal pond of frog agent

PondID Pond ID number

PondQuality Q 0.1-1 Pond Quality index of the pond

Summer habitat cells associated with the SummerHabitat A Pond pond

79 Chapter Two

Table 2 Descriptive statistics of the identified clusters.

The cluster’s ID, number of ponds in the cluster, mean distance from the ponds to a road, the distance to the pond closest to then road and the number of pond members no more than 200 m from the road. Furthermore it is shown whether the cluster exhibits extreme connectivity values when Ra=5, its response to road avoidance and its response to road mortality

Cluster characteristics Responds patterns

Cluster Cluster Mean dis- Min distance Ponds Ra= Avoid- Mortal- Id size tance (m) ( m) 200m 5 ance ity

c1 4 1322 1082 0 N N N

c2 4 184 110 3 Y Y N

c3 5 345 76 1 Y Y N

c4 7 431 61 2 Y Y Y

c5 9 223 71 6 Y Y Y

c6 6 323 98 1 Y N N

c7 2 385 318 0 N N N

80 Chapter Two

Table 3 Statistical results of multiple regression models

Statistical significance of variables and interactions in multiple regressions on landscape con-

nectivity (S), within-cluster connectivity (Sc) of cluster c1 – c7 and between-cluster connec-

tivity (Sb). Sequential Holm-Bonferroni correction is used to adjust p-values. Statistically significant values are shown in bold.

Full model Road mortality Rd Road avoidance, Ra Interaction Ra * Rd

Dependent Para- Para- Para- df F p R2 p p p factor meter meter meter

Sc1 496 0.30 0.82 0.002 -0.003 0.873 0.001 0.738 0.001 0.884

Sc2 496 7.20 <0.001 0.042 0.018 0.281 0.012 <0.001 -0.014 0.018

Sc3 496 39.31 <0.001 0.192 -0.024 0.240 0.021 <0.001 -0.020 0.007

Sc4 496 255.1 <0.001 0.607 -0.053 0.014 0.064 <0.001 -0.012 0.121

Sc5 496 487.9 <0.001 0.747 -0.094 <0.001 0.092 <0.001 -0.006 0.522

Sc6 496 0.85 0.45 0.005 -0.014 0.496 0.002 0.739 0.002 0.775

Sc7 496 0.81 0.49 0.005 -0.005 0.584 0.001 0.676 0.0001 0.976

Sb 496 1814 <0.001 0.917 -0.831 <0.001 0.114 <0.001 -0.096 <0.001

S 496 1297 <0.001 0.887 -1.005 <0.001 0.307 <0.001 -0.145 <0.001

81 Chapter Two

Supplementary material - Appendix 1

Model ODD

Purpose The purpose of the model is to measure the connectivity of the landscape. In this study connectivity between any two subpopulations is measured as the probability of successful immigration. The habitat of pond breeding amphibians as the Moor frog includes terrestrial as well as aquatic habitat. Therefore, the habitat patch of a subpopulation is modelled as a complementary habitat patch containing not only the breeding pond but also all accessible summer habitat within migration distance from the pond (Dunning et al. 1992; Pontoppidan and Nachman In review; Pope et al. 2000). Immigration, thus, requires two events: 1) the successful dispersal of a juvenile frog to summer habitat outside its natal habitat patch and 2) subsequent successful migration to a breeding pond associated with the new summer habitat. In real life these two events is 2 year apart, but, for simplicity, we only simulate the dispersal and migration events not the intervening years.

Entities, state variables, and scales Breeding ponds are treated as stationary agents. Each pond agent is characterized by a unique id-number, pond-quality and the associated summer habitat. Frog agents are characterized by the pond in which they are hatched and the breeding pond they immigrate to. The extent of the model landscape is 600 x 800 grid cells, and each grid cell represents 10 x 10 m. Grid cells are defined by their relative attraction to dispersing frogs, a habitat survival index and a daily survival probability, the habitat type and the relative value as summer habitat (see Table 1 in main text). The first part of the simulation mimics the dispersal of newly metamorphosed frogs, starting in mid-summer until hibernation in autumn. The second part considers the spring movement of juveniles from the summer habitat towards their future breeding pond and back to their summer habitat. Each part runs for 120 time steps, one step representing one day.

Process overview and scheduling At the start of a simulation, 500 frog agents are located at each pond agent and the frog variable NatalPond is updated with the ID-number of the pond. In the first part of the simulation

82 Chapter Two

(dispersal) the following procedures are executed each time-step: Move (movement of frog agents), Settle (evaluates if a frog agent stops dispersing and assigns frogs to breeding ponds) and Survival (evaluates if a frog agent survives). At day 120 the dispersal stops; frog agents that have not settled are removed and the migration simulation starts. The two procedures Move and Survival are run every time step and settled frogs start moving again, this time towards their breeding pond. When a frog reaches its assigned breeding pond, its direction is set towards one of the summer habitat fragments associated with the breeding pond. The simulation stops at day 240 and at each pond the model counts the number of immigrants from each of the other pond agents, computing immigration probabilities between all pairs of ponds. The simulation is repeated 25 times.

Design concepts

Emergence Immigration rates will emerge as a response to the landscape configuration.

Adaptation & Objectives To avoid desiccation and thereby increase survival, frog agents are assumed to move in response to the moistness of its surroundings. In general, the moister a habitat is the more attractive the habitat is for the frog as indicated by the habitat-attraction parameter Ha. Dispers- ing juvenile Moor frogs have an innate tendency to move away from their natal pond. Each frog agent is assigned a random direction to move, but during dispersal the frog adjusts its path to the encountered habitat. Adjustments are centred about the preferred direction in a way that prevents backtracking.

Sensing Frog agents are assumed to be aware of their own state-variables. Frog agents are also aware of the habitat attraction of the grid cells, as well as the identity of the pond agents.

Interaction There is no interaction between frog agents. Movement decisions of the frog agents depend on the habitat attraction of the neighbouring cells. Survival of the frog agents depends on the daily survival rates of the traversed habitat.

83 Chapter Two

Stochasticity Which cell to move to, is chosen randomly among the neighbouring cells with the probability of being chosen weighted by the habitat-attraction of the neighbouring cells. If frog agents occupy a cell suitable as summer habitat they will stop dispersing with a certain probability; the probability increases with time. The breeding pond of a settled frog agent is chosen randomly among accessible pond agents weighted by pond-quality.

Observation At the end of each simulation, the natal pond and breeding pond of all frog agents are registered and immigration probability (pij) between all pair-wise pond agents is calculated.

Initialization A landscape is constructed based on a GIS-raster data set. Each cell contains information about habitat type, habitat attraction, habitat survival and summer quality. A data set with information on location, ID-number and pond quality of surveyed ponds is used to create pond agents. Once the landscape is created, the Map-scan procedure is run to identify all accessible summer habitat cells associated with each pond agents and the pond variable A is updated. Habitat survival is converted into daily survival probabilities and the cell variable Ds is updated. 500 frog agents are located on each pond agent, their direction set randomly.

Submodels

Map-scan

The cell variable DailySurvival (Ds) is set as the probability of a frog agent surviving one time step in the cell. This depends on the habitat code (Hc) and the habitat survival index (Hs) of the cell. Cells belonging to roads, Hc = [2, 3, 4 5], are assign Ds values specific for their habitat code (see Appendix 1, Parameterisation). All other cells are modelled as

1∨ 6 , where σ0 and σ1 are species specific constants.

Local populations of pond-breeding amphibians inhabit a kind of composite habitat patch. At its core the breeding pond is located, surrounded by satellites of summer habitat fragments separated by matrix habitat (Pontoppidan and Nachman In review). The Map-scan procedure delimits the extent of the habitat patch as all accessible cells within a 40 cells (400 m) radius from the pond. Accessible cells are defined as cells with habitat attraction (Ha) greater than 1

84 Chapter Two

and daily survival (Ds) greater than 0.3. This excludes structures such as buildings and large roads. Furthermore, inaccessible cells functions as a barrier blocking access to the habitat be- hind (Eigenbrod et al. 2008) (Fig.A1).

Move Each day frog agents move a randomly chosen distance depending on the habitat attraction

(Ha) of its current cell. The distance is drawn from a normal distribution with a mean of c and standard deviation of s. Assuming the frog to head in the direction it was assigned when it left the natal pond, it moves to one of its neighbouring cells located within ±900 from the preferred direction. Cells with Ha = 1 is considered inaccessible. Based on habitat attraction of the neighbouring, accessible cells (n), frog agents first decide which kind of habitat they want to move to. The probability of moving into one of the cells with habitat attraction Ha is found

as 1 , where n is the number of neighbouring cells with habitat attrac- a ∑ tion Ha and Hai is the habitat attraction of cell i. A uniform pseudorandom number is selected to choose the type of habitat. If there is more than one neighbouring cell with the chosen habitat attraction, one of them is chosen randomly. The frog agent then moves to a random position within the cell, without changing its direction. This routine is repeated until the chosen distance for the day is traversed.

During the second part of the simulation, as frog agents get within a distance of two cells from their destination (breeding pond agent or summer habitat cell), the frogs move directly to it. At the breeding pond frog agents are randomly assigned a summer habitat cell associated with the habitat patch. When the frogs reach their summer habitat cell they stop moving. Frog agents reaching the boundary of the landscape are removed. If frogs agents are cornered with no accessible habitat to move to, their direction is permanently changed either 35 degrees to the left or to the right.

Settle Dispersing frog agents encountering a summer habitat cell has a certain probability of settling. The probability depends on the day number of the year (t) and is found as:

, where ν0 and ν1 are species specific constants

85 Chapter Two

When the summer habitat cell in which the frog agent has settled is part of a habitat patch, the frog is assigned the associated pond agent. If the summer habitat cell is shared by several pond agents, the probability of being assigned a pond agent i is a function of the pond quality

(Q) of the available pond agents (n): . Once a frog agent is settled it stops moving ∑ until the second part of the simulation.

Survival For each frog agent a pseudo-random number is drawn between 0 and 1. If the number exceeds the geometric average of daily survival (Ds) of the cells traversed during the day, the agent dies.

Parameterisation

All parameter values are listed in Table A1.

Habitat attraction (Ha) Terrestrial amphibians are assumed to prefer habitat in which the water content is high thereby minimizing the risk of desiccation. In an experiment with northern green and northern leopard frogs in peatlands Mazerolle and Desrochers (2005) found that 18 out of 25 frogs (72%) avoided barren surface. Hartung (1991) found that Moor frogs avoided areas with sparse or low vegetation, and recorded the ratio between densities in grass areas and densities in moor lands, hedges, ditches and forests to be 1:3.5.

In the model, the probability of a frog agent choosing one type of cell above another during movement therefore depends on the attractiveness of the cells habitat type. The habitat attraction (Ha) of the different habitat types in the GIS maps was guesstimated by amphibian specialists (Table A2). We tested three different expressions of Ha to enter into the Move- 2 procedure: a) Ha, b) (Ha) and c) exp(Ha) and compared the results with the above mentioned empirical findings. In addition, we also ran a simulation where movement was independent on habitat attraction.

As test landscapes we used GIS data sets from two different road projects in Denmark, supplied by the Danish Road Directorate and Amphi Consult. The first project (KaB) concerns an area in north-western part of Zealand, 10 km east of the city of Kalundborg (55° 40.14’ N 11° 17.85’ E); the second project (HoB) is from central Jutland, ca. 5 km east of

86 Chapter Two

Holstebro (56° 19.66’ N 8° 44.65’ E) (Fig. 1). Both areas are characterised as semi-urban and agricultural landscapes, traversed by creeks and wetlands. In the HoB map 36% of all cells were classified as attractive habitat (Ha > 3) and the KaB map contained 51% attractive habitat. The model was run for 40 time steps without the Settle-procedure, and the ratio between frog agents in attractive (Ha = 4 or 5) and unattractive (Ha = 2 or 3) habitat was computed as well as the percentage of frog agents in attractive habitat. Although there were differences between the maps, we chose the expression exp(Ha) as being the best to reproduce the empirical patterns (Table A3).

Distance travelled pr. day In a radio tracking experiment with common frog (Rana temporaria) Tramontano (1997) found adult frogs moving through a rye field to cover 148 m in one week, corresponding to ca 20 m pr day. In a study on dispersing juvenile moor frogs Hartung (1991) reported daily travelling distances of 12.5–18.8 m (mean 15.5 m) in attractive habitat as moors and 39.9–40.9 m in unattractive areas as pine forests. The daily travelling distance is, thus, assumed to depend on habitat attraction of the current cell and is modelled as following a normal distribution with a mean of c/Ha and standard deviation s.

Two homogenous landscapes were constructed with a habitat attraction of 2 or 4, respectively. In each landscape frog agents were allowed to move according to the Move- procedure for 40 time steps. When a simulation ended the straight distance between start and end point for all frog agents was measured and the mean daily travelling distance computed. The simulations were conducted for varying values of c and s, each combination repeated 50 times. A parameter set was sought where daily travelling distance

 in attractive habitat exhibited values in the range between 12 m and 19 m and with a mean around 15 m  in unattractive habitat exhibited values in a range between 20 m and 40 m

The parameter set (c=7, s=0.5) was chosen as the best to fulfil the conditions.

Settle Little data has been found on dispersal distances. Hartung (1991) found dispersal distances of newly metamorphosed moor frogs up to 1200 m, but mean dispersal distance is guesstimated by experts to be 200-300 m. As the new frogs are assumed to have an innate urge to move

87 Chapter Two away from their natal pond, settling probability is expected to be low in the beginning. The dispersal drive is expected to subside with time, with settling probability increasing accord- ingly, creating an s-shaped probability function (Fig.A2). The realized dispersal distances will depend on the landscape configuration. With the chosen parameter set (ν0, ν1) mean dispersal distance of frog agents in the HoB map is 363 m and max. dispersal distance is 2068 m. 68% of the frog agents settle within their own habitat patch. Less than 2% of the frog agents disperse more than 1000 m. In the KaB map mean and maximum dispersal distances are 291 m and 2151 m, respectively. 83% settle within the home patch and less than 0.5% disperse more than 1000 m (Fig.A3).

Daily survival The survival probability of dispersing frogs is assumed to depend on the habitat. No empirical data on daily survival probabilities were found but annual survival rates for young adult frogs has been estimated to be between 55% (Fog and Hesselsøe 2009) and 63% (Loman 1984). These rates are not habitat specific but are realized during annual movements in a heterogeneous landscape.

All land cover categories in GIS maps were ranked according to amphibian survivability by specialist and assigned a value of relative survival index (Hs) (Table A2) which is subsequently converted into daily survival probabilities (Ds). The parameter set (σ0, σ1) gives the functional relationship between Hs and Ds. The parameters were found by iteration. The model was run with varying combinations of parameter values on both test maps until a set was found with which the annual survival rates were within 55% and 63%. With the chosen parameter set the realized annual survival rates was between 56 % and 57 % in both maps. Table A4 shows the resulting habitat specific daily survival probability.

Road survival Hels and Buchwald (2001, fig. 5) found the probability of a Moor frog getting killed when crossing a road ranged from ca 35% to ca 90% depending on traffic intensity. We assume traffic intensity to be correlated with road width and let daily survival probability (Ds) depend on road category as shown in Table A5.

88 Chapter Two

References

Dunning JB, Danielson BJ, Pulliam HR (1992) Ecological processes that affect populations in complex landscapes. Oikos 65: 169-175. doi:10.2307/3544901

Eigenbrod F, Hecnar SJ, Fahrig L (2008) Accessible habitat: an improved measure of the effects of habitat loss and roads on wildlife populations. Landscape Ecology 23: 159-168. doi:10.1007/s10980-007-9174-7

Fog K, Hesselsøe M (2009) Udvikling af prototypemodel til brug for forvaltning af spidssnudet frø i forbindelse med vejanlæg. Amphi Consult, pp.

Hartung H (1991) Untersuchung zur terrestrischen Biologie von Populationen des Moorfro- sches (Rana arvalis NILSSON 1842) unter besonderer Berücksichtigung der Jahresmobilität. Hamburg: Universität Hamburg.

Hels T, Buchwald E (2001) The effect of road kills on amphibian populations. Biological Conservation 99: 331-340. doi:10.1016/S0006-3207(00)00215-9

Loman J (1984) Density and survival of Rana arvalis and Rana temporaria. Alytes 3: 125- 134

Mazerolle MJ, Desrochers A (2005) Landscape resistance to frog movements. Canadian Jour- nal of Zoology-Revue Canadienne De Zoologie 83: 455-464. doi:10.1139/z05-032

Pontoppidan M-B, Nachman G (In review) Effects of within-patch heterogeneity on connectivity in pond-breeding amphibians studied by means of an individual-based model. Web- ecology:

Pope SE, Fahrig L, Merriam NG (2000) Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81: 2498-2508

Tramontano R (1997) Continuous radio tracking of the common frog, Rana temporaria. Her- petologia Bonnensis: 359-365

89 Chapter Two

Tables & Figures

Table A1. List of parameters, their default values and the procedure in which they appear.

Parameter Value Procedure

ν0 1.50E-06 Settle

ν1 0.06 Settle

σ0 2.2 Map-scan

σ1 4 Map-scan c 7 Move s 0.5 Move

90 Chapter Two

Table A2. 3 Land cover categories and the associated values of Habitat survival, Habitat attraction and Summer quality

Habitat Habitat Summer Habitat Code Description Attraction Survival Quality (Hc) (Ha) (Hs) (Hq) 2 4-lane motorway 2 N/A 1 3 2-lane motorway 2 N/A 1 4 Road, width > 6m 3 N/A 1 5 Road, width 3-6 m 3 N/A 1 6 Other roads 3 2 2 8 Pathway 4 4 3 11 Multiple surface 3 3 3 11 Railway 4 2 3 12 Building 1 N/A N/A 15 Other made surface 2 3 2 18 Wetlands 5 5 5 20 Running water 4 4 3 22 Meadows 5 5 5 24 Grassland 4 4 4 25 Lakes 1 N/A N/A 28 Hedgerow 4 4 4 29 Heath land 5 5 4 32 Woodland 4 4 4 34 Stand of trees 4 4 3 36 Bare surface 2 2 1 40 Fallow land 4 4 4 42 Field crops 2 2 2

91 Chapter Two

Table A3 Observed patterns and the patterns emerging using three different expressions with

Ha entering into the Move procedure as well as random movement. Results are shown for two different maps, Kalundborg (KaB) and Holstebro (HoB)

2 Random Ha Ha exp(Ha) Pattern Observed KaB HoB KaB HoB KaB HoB KaB HoB

Ratio between frog densities in good 0.29 0.39 0.48 0.46 0.64 0.43 0.53 0.38 0.46 and bad habitat (Hartung 1991)

Percentage of individual choosing good habitat (Maze- 72% 72% 67% 68% 61% 70% 66% 73% 68% rolle and Desrochers 2005)

Table A4 Habitat survival index (Hs), the corresponding daily survival probability (Ds) and Ds converted into annual survival probability.

Annual survival

Hs Ds probability

1 0.9820 0.01

2 0.9960 0.38

3 0.9984 0.68

4 0.9991 0.81

5 0.9995 0.88

92 Chapter Two

Table A5 Road categories with the corresponding habitat code (Hc) and daily survival probability (Ds)

Habitat code (Hc) Road category Ds

2 4-lane motorway 0.1

3 2-lane motorway 0.2

Road width > 6 4 0.5 m

Road width 3-6 5 0.8 m

Figure A1 Illustration of how accessible summer habitat is identified.

Blue circle is a pond; dotted circle represents maximum migration distance. Green areas are accessible summer habitat while shaded areas are inaccessible summer habitat. a) All summer habitat within migration distance is regarded as accessible b) Road traversing the habitat prevents access to summer habitat on the opposite side of the road c) Structures breaking the road such as underpasses again permit access to summer habitat on the opposite side of the road. a b c

93 Chapter Two

Figure A2 Settling probability.

If a frog agent encounters a summer habitat cell, the probability of settling in the cell will depend on the time, measured as day number

1.0

0.8

0.6

0.4 Settle probability

0.2

0.0

170 190 210 230 250 270 290 310 Daynumber

Figure A3 Dispersal distances

The frequency distribution of dispersal distances with chosen settle-parameters for a) HoB map b) KaB map. Black line shows accumulated frequencies. a b

25 25

20 20

15 15 % %

10 10

5 5

0 0 0.10.20.30.40.50.60.70.80.91.01.11.21.31.41.5Mor 0.10.20.30.40.50.60.70.80.91.01.11.21.31.41.5Mor Dispersal distance (km) Dispersal distance (km)

CHAPTER THREE

SAIA – A MANAGEMENT TOOL FOR

ASSESSMENT OF ROAD EFFECTS ON

REGIONAL POPULATIONS OF

MOOR FROGS (RANA ARVALIS)

Submitted to Nature Conservation

December 2012

96 Chapter Three

SAIA – a management tool for assessment of road effects on regional populations of Moor frogs (Rana arvalis)

Maj-Britt Pontoppidan, Gösta Nachman

Both: Section for Ecology and Evolution Department of Biology University of Copenhagen Universitetsparken 15 DK-2100 Copenhagen

Corresponding author: M-B. Pontoppidan email: [email protected] phone: +45 51518791

97 Chapter Three

Abstract

An expanding network of roads and railways fragments natural habitat affecting the amount and quality of habitat and reducing connectivity between habitat patches with severe consequences for biodiversity and population persistence. To ensure an ecologically sustainable transportation system it is essential to find agreement between nature conservation and land use. GIS are frequently used to support management decisions in landscape planning. They are used to produce land-use maps and provide various tools for measuring connectivity etc., mostly based on graph theory or least cost analysis considering distances and landscape permeability. However, GIS map do not provide information about the particular dispersal, survival and establishment of the animals, which depend not only on the quality of the habitat but also on the behaviour of the animals, their response to habitat conditions and landscape elements. Individual based modelling has been proven to be suitable for describing such processes. The study presented, combines GIS with IBM in order to merge the strengths of both approaches, since a combination of land use analysis with animal behaviour is essential for an effective planning of landscapes, providing for the urban use by humans as well as the survival of endangered species. The model, called SAIA (Spatial Amphibian Impact Assess- ment), provides information on connectivity as well as estimates of population persistence. By means of a case study dedicated to pond breeding amphibians (Rana arvalis) we demonstrate how SAIA can be used for assessing which management measures would be best to mitigate the effect of landscape fragmentation caused by road constructions.

98 Chapter Three

Introduction

Over the last decade a growing amount of literature has documented the severe impacts of transport infrastructure on biodiversity, population persistence and gene flow. An expanding network of roads and railways divides natural habitat into smaller and smaller fragments, affecting the amount and quality of habitat and reducing connectivity between habitat patches (Coffin 2007; Fahrig and Rytwinski 2009; Forman and Alexander 1998; Holderegger and Di Giulio 2010; Spellerberg 1998; Trombulak and Frissell 2000). To ensure an ecologically sustainable transportation system it is essential to find agreement between nature conservation and land use. In Europe and the US, programs and policies are being developed addressing this need in strategic and environmental impact assessments (Brown 2006; Iuell et al. 2003; Trocmé et al. 2003), However, sustainable road planning requires adequate tools for assessment, prevention and mitigation of the impacts of infrastructure (Beckmann 2010; Forman et al. 2003; Gontier et al. 2010).

The persistence of a population depends on the amount and accessibility of its required resources and, within a metapopulation framework, also on sufficient dispersal between subpopulations (Dunning et al. 1992; Wiens 1997). Therefore, measures of connectivity have been used as indicators of a landscape’s capability to sustain a population. Likewise, GIS has proved to be an important tool when assessing the impact of roads on landscape fragmentation and/or connectivity (Beckmann 2010; Brown 2006; Calabrese and Fagan 2004). Different species have both different habitat requirements and behaviours. Therefore, connectivity must in essence be species specific. Methods using least cost modelling (Adriaensen et al. 2003; Epps et al. 2007) or graph theoretical approaches (Bunn et al. 2000; Minor and Urban 2008; Zetterberg et al. 2010) usually combine GIS data with some species specific data such as dispersal distances or habitat suitability. However, neither of these methods considers the particular dispersal, survival and establishment of the animals, which depend not only on the quality of the habitat but also on the behaviour of the animals, their responses to habitat conditions, landscape elements, interactions with other animals and many other factors. Individ- ual based models (IBMs) have proved to be suitable for describing such processes (Grimm 1999; McLane et al. 2011) and recently there has been an increase in IBM case studies dem- onstrating the potential for analysing population dynamics emerging from the interactions

99 Chapter Three between landscape settings and animal behaviour (e.g. Graf et al. 2007; Kramer-Schadt et al. 2004; Pe'er et al. 2011).

We have developed a strategic management tool to be used in assessment and mitigation of road effects on a regional population of pond-breeding amphibians. The model, called SAIA (Spatial Amphibian Impact Assessment), combines the use of GIS land cover maps with IBM and provides information on connectivity as well as estimates of population persistence. SAIA is to be used by the Danish road authorities when assessing how new road construction may affect Moor frogs (Rana arvalis). In this paper we demonstrate how SAIA can be used for assessing which management measures would be best to mitigate the effect of landscape fragmentation caused by the construction of a road ca 90 km west of Copenhagen, Denmark. To achieve this goal the following specific research questions were addressed:

 What is the structure of the regional habitat network before road construction?  How is the habitat network affected by the new road?  Which mitigation strategies are best suited to preserve the overall persistence of the regional population of Moor frogs?

Methods

SAIA combines an individual based model with a population based model. The IBM is basi- cally identical to the model described in Pontoppidan and Nachman (In prep.) It simulates the movement of newly metamorphosed frogs from their natal ponds to new habitat patches. Here, we use the terms dispersal and migration as defined by Semlitsch (2008), i.e. dispersal is interpopulational, unidirectional movements from natal sites to other breeding sites and migration is intrapopulational, round-trip movements toward and away from aquatic breeding sites. The habitat of pond breeding amphibians, such as the Moor frog, includes terrestrial as well as aquatic habitat. Therefore, we define an adequate habitat patch of a subpopulation as containing not only the breeding pond but also all accessible summer habitat within migration distance from the pond (Dunning et al. 1992; Pontoppidan and Nachman In review; Pope et al. 2000).

100 Chapter Three

Model species Moor frogs spend most of their life in terrestrial habitat; aquatic habitat is only used during the breeding season in early spring (Elmberg 2008; Glandt 2008; Hartung 1991). Soon after breeding, the frogs return to the summer habitat, which lies mostly within a 400 m radius from the breeding pond (Elmberg 2008; Hartung 1991; Kovar et al. 2009). Adult frogs show strong site fidelity and often use the same breeding pond and summer habitat from year to year (Loman 1994). Long distance dispersal takes place predominantly during the juvenile life-stage (Semlitsch 2008; Sinsch 1990; 2006). Shortly after metamorphosis, the young frogs leave the natal pond and disperse into the surrounding landscape seeking suitable summer habitat. Dispersal distances are between a few hundred meters up to 1-2 kilometres (Baker and Halliday 1999; Hartung 1991; Sinsch 2006; Vos and Chardon 1998). The juveniles stay in terrestrial habitat 2-3 years until they reach maturity, although some observations indicate that juvenile frogs follow the adults during the spring migration, without entering the breeding ponds (Hartung 1991; Sjögren-Gulve 1998) .

Model overview As in its predecessor (Pontoppidan and Nachman In prep.), SAIA is based on a GIS raster map. Each raster cell contains information about the cell’s land cover or habitat type (Hc) and relative suitability as summer habitat (Hq), its relative attraction to frogs during movement

(Ha) and the relative survival index (Hs) associated with the cell. A point-data set containing information on potential breeding ponds was obtained from extensive field surveys in the area. Each pond is characterized by an ID-number, the perimeter of the pond (O), the number of egg masses found during the field surveys (N0) and a quality index (Q).

The GIS map is imported into the model and the model landscape is constructed. The point-data set is used to create stationary pond agents. After import, the map is processed and additional variables are added to the raster cells and pond agents. Daily survival probabilities

(Ds) associated with each cell are computed based on the cell values for HabitatCode and HabitatSurvival. Cells with high values of SummerQuality are classified as summer habitat. Summer habitat cells can be completely surrounded by other summer habitat cells (core cells) or have one or more neighbouring cells which are not summer habitat (edge cells). To account for edge effects, core cells are given the area value (W) of 1 while W is 0.5 for edge cells (Watts and Handley 2010). Pond agents are updated with the summer habitat cells within mi-

101 Chapter Three gration distance (A) as well as with the effective area of the summer habitat (A´). The carrying capacity of the summer habitat (K) is estimated, based on the amount of available summer habitat and number of egg masses found during field work. Table 1 shows a full list of model variables.

At the start of a simulation 250 frog agents are created in each pond agent. The frogs disperse through the landscape in random directions from the ponds; the movement of the frogs depends on the attractiveness of neighbouring cells and the cells’ suitabilities as summer habitat. Survival probabilities depend on the traversed habitat types. Unlike the former model, movement behaviour in SAIA also depends on weather conditions. When daily precipitation exceeds a given threshold (α), the variables HabitatAttraction and DailySurvival of all accessible cells are given the highest value. An exception is paved roads where only HabitatAttrac- tion, but not DailySurvival, is changed. After the simulation, immigration probabilities between all pairs of subpopulations are calculated and an immigration matrix is constructed. The immigration matrix is used to compute connectivity. Apart from being a landscape attribute in itself, the matrix also enters into to the population-based procedure PopulationDynamics (PD).

The PD-procedure estimates population sizes in each pond through 40 iterations of a life cycle model. The elements of the life cycle model are 1) Reproduction, 2) Survival and 3) Immigration. For simplicity, we only look at the female part of the population. We assume a sex ratio of 0.5 and that females always become mated. Pond subpopulations are grouped by age from 0 through 6 years, and survival and reproductive rates come from life-table data constructed by amphibian experts (Table 2). The number of egg masses found in the surveyed ponds (N0) is expected to equal the number of breeding females in the subpopulation. This number is set as the initial population size of the pond. After each iteration, the pond variables Froglets and AgeClassList are updated with the reproductive output and the number of surviving individuals in each age class, respectively. The number of immigrants is added to age class 0 in the AgeClassList.

Individuals can reproduce starting in age 3. As in Hels and Nachman (2002), the expected egg production of a female is assumed to follow a negative binomial distribution with mean and clumping parameter k. is the mean number of eggs produced by a female of a given age. The number of newly metamorphosed frogs, ready to disperse is considered as the

102 Chapter Three reproductive output. This involves the survival of egg and larvae, as well as the survival of the young frogs the first two weeks after metamorphosis. The overall probability that an egg develops into a frog that survives until dispersal time is assumed to be affected by two factors: Density of eggs in the pond and the quality of the pond. The conditional probability that a frog survives from age a to age a+1 is assumed to depend on age. Furthermore, survival is assumed to depend on the frog density in the summer habitat. For simplicity, this is modelled as a “culling” process when frog density exceeds the carrying capacity of summer habitat. Immigration probabilities between all pairs of subpopulations are obtained from the immigration matrix. The actual number of immigrants a subpopulation receives depends on the variable Froglets of each of the other ponds and the corresponding immigration probability. Emi- gration rates are not modelled explicitly. For a full model description according to the protocol suggested by Grimm et al. (2006); (2010) see Appendix 1 in supplementary material and Pontoppidan and Nachman (In prep.). Netlogo v.4.1.3 (Wilensky 1999) was used as modelling environment (freely downloadable at http://ccl.northwestern.edu/netlogo).

Output At the end of a simulation, the following output was recorded: the number of surviving frogs, the natal and breeding ponds of all frogs, and the immigration probabilities (pij) between all pair-wise ponds. Landscape connectivity (S) is found as

∑∑ , . (Eq. 1)

By the end of the PD-procedure, the population size of each pond is estimated as the resulting numbers of frogs of ages 2 through 6. The model was run 50 times and mean connectivity with 95% confidence interval (CI) were computed. For each pond, as well as for the whole landscape, we computed mean population size with 95% confidence intervals (CI) and the proportion of replicates where the predicted population size was positive. Mean number of populated ponds with 95% CI was also calculated.

Ponds were grouped into clusters depending on their mutual connectivity, using unweighted, arithmetic, average clustering as described by Legendre and Legendre (1998).

Since immigration probabilities between ponds are not necessarily symmetric, i.e. pij ≠ pji, we used summed immigrations probabilities as similarity measures (m), i.e. mij = pij + pji. The threshold at which a given pond or cluster no longer can be added to another cluster was set to mij ≤ 0.01. Connectivity between any pairs of clusters (Sk,l) is found as

103 Chapter Three

, ∑∑ , (Eq. 2)

where nk and nl are the number of ponds in clusters k and l, respectively.

Scenarios We constructed five different scenarios. The analyses of scenario 0 and scenario 1 were used to plan a series of suggestions for mitigation measures which are put into effect in scenarios 2, 3a and 3b.

Scenario 0: Before the road project This is an analysis of the landscape before the planned road construction and it works as a reference against which the other analyses are compared. As input data we use a GIS data set from a road project in Denmark, supplied by the Danish Road Directorate and Amphi Con- sult. The project concerns an area in the north-western part of Zealand, 10 km east of the city of Kalundborg (55° 40.14’ N 11° 17.85’ E) (fig. 1). All cell values of Ha, Hs and Hq are ranked on a scale from 1-5, following the protocol of Hassingboe et al. (2012) (Table 3). The extent of the map is 600 x 800 cells, and each cell is 10 x 10 m. The point data set contains information about potential breeding ponds found during the field survey. Pond qualities (Q) range from 0.1 – 1 and relate to the suitability of the pond and the immediate surroundings in regard to egg and larval survival. These values were based on field work conducted by amphibian experts. The data set contains 121 ponds, of which 23 ponds are of high quality (Q > 0.6). In total, 106 egg masses were found distributed among 6 ponds (Fig. 2A).

Scenario 1: After the road project The landscape in scenario 0 is modified according to the planned road project. The changes include a broadening of an existing 2-lane motorway into a 4-lane motorway as well as an extension of the motorway. This changes the survival parameter Ds of the road from 0.20 to 0.10. The construction involves removal of five ponds along the road (Fig 3A).

Scenario 2: Construction of underpasses and drift fences Three underpasses are added to scenario 1. Drift fences are established along the road for 100 m on each side of the underpass, except for underpass 2 which has a 300 m drift fence to the south (Fig 4A).

104 Chapter Three

Scenario 3: Construction of artificial breeding ponds Eight new breeding ponds are added to scenario 1. Each breeding pond is assigned a pond quality of 0.7. Scenarios 3a and 3b represent two alternative locations of the eight ponds (Fig 4B and C).

Results

Initial analyses

Scenario 0 The average proportion of ponds populated during a simulation was 32%, although only 22% of the ponds have a more permanent status (pond persistence probability > 0.75). The abundance of adult female frogs was estimated to be 157; annual survival probability is 57%, and landscape connectivity was 55 (Fig 5).

The cluster analysis identified 13 clusters, cluster sizes ranging from 2-20 ponds (Fig 2B, Table 4). The six populated ponds found during field surveys were distributed on four different clusters. One pond with only one adult female was found in cluster 5 (c5). Another pond belongs to c4 and two other ponds are found in c8. In each of these ponds, the initial population size was set to 5 adult females. Cluster c11 contains the remaining two populated ponds with an initial total population of 90 adult females. Apart from c5, all of these initially populated clusters exhibited high viability. Clusters c4, c8 and c11 have mean pond persistence probabilities between 77% - 93% and estimated cluster abundances from 23-51 adult females. Cluster c9 also shows high values of abundance and persistence. Although initially unpopulated, c9 contains several high quality ponds and is connected with c8 and c11 which may promote colonisation and establishment. In c6 and c7, the mean pond persistence probability is considerably lower (29-35%) as is the estimated cluster abundance. While the two clusters, especially c6, are connected with other populated clusters, they lack high -quality ponds and the clusters may function as sinks. In the remaining clusters the estimated abundance is less than one individual.

105 Chapter Three

Scenario 1 After construction of the road, the proportion of populated ponds was reduced to 26% and permanent ponds are now down to 16%. Annual survival rate is 56% and estimated abundance is 136 adult females. Landscape connectivity decreased to 51 (Fig 5).

The number of clusters was unchanged but connectivity between clusters was reduced (Table 5). Connectivity from c7 and c9 to their primary source (c8) decreased more than 80%. Moreover, three ponds were lost in c7 and c9 due to the road construction. Estimated abundance and mean pond persistence probability decreased in c7 and c9 and these clusters were no longer able to uphold viable populations (fig 3B). However the initially populated clusters c4, c8 and c11 were not affected by the road construction.

Mitigation planning The first analyses revealed that the landscape contains three viable populations (c4, c8 & c11) centred on the initially populated ponds. These populations appear not to be affected by the road construction and in the simulations the clusters seemed to function as sources enabling colonisation and establishment of populations in c9 and c7. Cluster c4 has a large and viable population, but even though it is well connected with the neighbouring clusters their qualities were not high enough to enable establishment of new populations. Since c4 is not connected with c7 and c9, its potential as source cluster is low. Cluster c8 seems to be the primary source cluster to c9 and c7; however, expansion of the road heavily reduces its value as a source. Furthermore, the removal of three ponds between c9 and c7 may diminish the connectivity between these clusters. Cluster c11 has a viable population and although situated somewhat remotely there is still some connectivity to c7 and c9.

The results indicate that, in order to compensate or mitigate the effect of the road project, the best strategies will be either to re-establish connectivity across the road between c8 and c7/c9 and between c7 and c9 or to take advantage of the viability of c11 and its source potential. Based on this, we created and analysed the following scenarios:

Scenario 2: Connectivity across the road is re-established by constructing three underpasses and drift fences along the middle section of the motorway (Fig. 4A). The expectation is that connectivity between c8 and c9/c7 will improve and enable establishment of populations in c9.

106 Chapter Three

Scenario 3a: The quality of c9 and c11 is improved by establishing three, and then five, new high -quality ponds within the range of the clusters (Fig. 4B). The three new ponds in c9 are expected to improve the probability of successful establishment of immigrants as well as reconnect c9 with c7.We expected an increase in abundance in c11, and hence increased immigration to and colonisation of c9.

Scenario 3b: This is a modification of scenario 3a. The quality of c9 and c11 is still improved but with only one and two ponds, respectively. The remaining five ponds are used to create a dispersal corridor between c11 and c9 (Fig. 4C). This strategy is expected to enhance the abundance in c11 and to improve connectivity to c9, thereby increasing the probability of colonisation.

Mitigation analyses

Scenario 2 Quite unexpectedly, the creation of drift fences and underpasses did not improve the condition of the landscape. The mean proportion of populated ponds is 26% and permanent ponds is 16% as in scenario 1. However, the estimated abundance of female adults decreased to 115, landscape connectivity is 48 and annual survival rate 53% (Fig. 5).

Two of the underpasses (including drift fences) were placed between c6 and c7; the third between c8 and c9. As expected, connectivity between c8 and 9 was greatly improved. Cluster c9 now spans the road and it annexed one of the ponds in the periphery of c8 (Fig. 4A). Abundance and mean pond persistence probability of c9 increased; this, however, was due to the inclusion of a pond from c8. Persistence and abundance did not improve on the original configuration of c9 (Table 6). Apart from c9, connectivity between initially populated clusters and other clusters did not improve. Connectivity to c4 and c11 were unchanged, while connectivity to c8 actually decreased. Finally, the abundance of frogs in c4 and c11 decreased to 20% even though connectivity both within the cluster and to other clusters was unchanged.

Scenario 3a Establishment of eight new ponds had a positive effect on landscape condition. The estimated number of adult females increased to 143, proportion of populated ponds is 31%, of which 19% are permanently populated. Landscape connectivity is 56 and annual survival rate 56% (Fig. 5).

107 Chapter Three

The five new ponds in c11 performed well and contained permanent populations. How- ever, the performance of the cluster did not improve, apart from a slightly higher persistence probability (Table 6). Cluster c9 seemed to benefit from the additional ponds, although none of the new ponds contained permanent populations. Cluster abundance and connectivity were nearly restored to their original conditions although mean pond persistence probability was still below 50%. Connectivity from c7 to other ponds improved somewhat, but not enough to restore the cluster to its former performance (Fig. 4B).

Scenarios 3b With this strategy we succeeded in restoring the landscape to its original ecological performance. The number of adult female frogs is 159. Mean proportion of populated ponds is 32% and 22% are populated permanently. Annual survival rate is 56% and landscape connectivity is 55 (Fig. 5).

Three of the new ponds are now part of c9 while the remaining five new ponds belong to c11. Connectivity between c9 and c11 is strong and six of the new ponds contain permanent populations (Fig. 4C). The abundance and mean persistence probability of c11 increased and are now better than before the road construction. Conditions in c9 also improved, compared to scenario 1, but its original performance in not quite restored. The performance of c7 did not change and is still at the same level as found in scenario 1 (Table 6).

Discussion

This study demonstrates how initial analyses of the landscape before and after the planned road constructions can help identify which areas will be most affected by construction. The analysis enables the user to recognise the colonisation potential of the clusters and to identify source or sink clusters and to use this knowledge for planning mitigation measures. In the present case study, the simulations indicated that the population recorded during the field survey will be largely unaffected. Nevertheless, the road construction will severely impair the colonisation potential of cluster c8, thereby reducing the ecological performance of the landscape. Of the three mitigation strategies tested, the analysis shows that scenario 3b is the best solution. This strategy of connecting clusters c9 and c11 restores the landscape to its former ecological performance. Even though not all individual ponds or clusters would be in the same condition as before, the strategy promotes viable populations on both sides of the road.

108 Chapter Three

The strategy is not strictly aimed at mitigating the impaired connectivity across the road, but rather tries to compensate for the effects of construction by improving other areas. Still, the populations on either side of the road are not totally isolated from each other; some dispersal does take place making genetic exchange possible.

Comparing the results from the analyses of scenarios 3a and 3b suggests that the location of compensating new ponds is not trivial. In both scenarios c11 gets five new ponds, all of high quality. Nevertheless, the results differ quite a lot. Scenario 3a places the new ponds within the cluster sharing the summer habitat of other ponds. Even though the new ponds are colonized and support viable populations, the abundance of frogs within the cluster does not improve. In scenario 3b, where the abundance of frogs within cluster c11 increases, the new ponds were placed between c9 and c11 and only partly share summer habitat with other ponds. This result emphasizes that for the Moor frog the carrying capacity of an area is not improved by adding new ponds, only new or better summer habitat can achieve this. Hence, we may improve cluster performance by creating new ponds in unutilized summer habitat within dispersal distance.

In scenarios 3a and 3b, c9 is also enlarged with three new ponds. In these cases there was no difference in frog abundance in the cluster whether the new ponds were placed in un- used summer habitat or not. In both scenarios, though, mean pond persistence probability greatly improved compared to scenario 1. So, while adding ponds to a cluster did not improve carrying capacity, it ensured a more viable cluster population.

The analysis of scenario 2 shows that drift fences and underpasses had negative effects on the ecological performance of this landscape. This result is highly surprising as well as controversial since fences and underpasses are standard mitigation measures used in many road projects (Iuell et al. 2003). Even though fences and underpasses should prevent road mortality and promote connectivity, the overall annual survival rate, as well as connectivity, decreased. These effects are probably mostly due to the fences. Underpasses per se do not change movement patterns, but fences do. Moreover, we did see increased connectivity locally across the road between c8 and c9.

Fences may force individuals to move along the road exposing them to low quality habitat for a longer time. Furthermore, the mitigation measures may be counterproductive if the combination of fences and underpasses lead individuals into low quality habitat or areas

109 Chapter Three without ponds to colonize. The population dynamics in the ponds is an emergent property, dependent on local conditions as well as regional dynamics. The change in connectivity and movement patterns caused by the migration measures, therefore, seems to be able to affect abundances even in clusters farther away.

Very little is known about the effects of mitigation measures in general. Once mitigation measures are implemented, efforts are seldom put into discovering how well they work. Recordings of animals using wild life passages reveal nothing about effects on local and regional persistence (Lesbarreres and Fahrig 2012). In a simulation study, Jaeger and Fahrig (2004) found that fencing, while preventing road mortality, did not necessarily improve population persistence and they recommended fencing only when road mortality is 100 %. In a study on moose (Alces alces), Olsson and Widen (2008) found that fences resulted in decreased use of wildlife passages. Our simulation results underscore the need for a better un- derstanding of how mitigation measures affect animal behaviour and population dynamics.

Conclusion

When planning road constructions, it is important to integrate mitigation measures right from the start. Often there are economic constraints on which measures are possible, certain structures as viaducts or bridges may already be in place or land available for compensation measures is restricted. SAIA offers a tool to evaluate different scenarios to find the best combination of mitigation measures for a given set of conditions. The model is meant to be used by non-specialists – all that is needed are GIS maps of the different scenarios. We attempted to find a balance between detailed and yet intuitive and easy interpretable output. Even though SAIA was developed for the Danish Road Directorate, its use is not restricted to road constructions but can be applied to other structures affecting the landscape and their potential impacts on wildlife.

Acknowledgments

The work was funded by the Danish Road Directorate. We thank Amphi Consult for providing us with amphibian expertise and field data. We are grateful for continuous and enthusias- tic feed-back from M. Ujvári, M. Hesselsøe, A. Jørgensen and M. Schneekloth during model

110 Chapter Three development. Special thanks are due to Uta Berger for encouraging and inspiring discussions. We thank Michal J. Reed for linguistic assistance.

References

Adriaensen F, Chardon JP, De Blust G, Swinnen E, Villalba S, Gulinck H, Matthysen E (2003) The application of 'least-cost' modelling as a functional landscape model. Landscape and urban planning 64: 233-247. doi:10.1016/s0169-2046(02)00242-6 Baker JMR, Halliday TR (1999) Amphibian colonization of new ponds in an agricultural landscape. Herpetological Journal 9: 55-63 Beckmann JP (2010) Safe passages, highways, wildlife, and habitat connectivity. Washing- ton, pp. Brown JW (2006) Eco-logical: An Ecosystem Approach to Developing Infrastructure Pro- jects. Federal Highway Administration, Washington, DC 20590, FHWA-HEP-06-011, pp. Bunn AG, Urban DL, Keitt TH (2000) Landscape connectivity: A conservation application of graph theory. Journal of environmental management 59: 265-278. doi:10.1006/jema.2000.0373 Calabrese JM, Fagan WF (2004) A comparison-shopper's guide to connectivity metrics. Fron- tiers in Ecology and the Environment 2: 529-536. doi:10.1890/1540- 9295(2004)002[0529:acgtcm]2.0.co;2 Coffin AW (2007) From roadkill to road ecology: A review of the ecological effects of roads. Journal of Transport Geography 15: 396-406. doi:10.1016/j.jtrangeo.2006.11.006 Dunning JB, Danielson BJ, Pulliam HR (1992) Ecological processes that affect populations in complex landscapes. Oikos 65: 169-175. doi:10.2307/3544901 Elmberg J (2008) Ecology and natural history of the moor frog (Rana arvalis) in boreal Swe- den. In: Glandt D, Jehle R (Eds) The Moor Frog Laurenti-Verlag, Bielefeld, 179-194 Epps CW, Wehausen JD, Bleich VC, Torres SG, Brashares JS (2007) Optimizing dispersal and corridor models using landscape genetics. Journal of Applied Ecology 44: 714-724. doi:10.1111/j.1365-2664.2007.01325.x Fahrig L, Rytwinski T (2009) Effects of Roads on Animal Abundance: an Empirical Review and Synthesis. Ecology and Society 14: 21 Forman RTT, Alexander LE (1998) Roads and their major ecological effects. Annual Review of Ecology and Systematics 29: 207-+. doi:10.1146/annurev.ecolsys.29.1.207 Forman RTT, Sperling D, Bissonette JA, Clevenger AP, Cutshall CD, Dale VH, Fahrig L, France R, Goldman CR, Heanue K, Jones JA, Swanson FJ, Turrentine T, Winter TC (2003) Road ecology: science and solutions. Road ecology: science and solutions: Chp. 6 Glandt D (2008) Der Moorfrosch (Rana arvalis): Erscheinungsvielfalt, Verbreitung, Lebens- räume, Verhalten sowie Perspectiven für den Artenschutz. In: Glandt D, Jehle R (Eds) The Moor Frog. Laurenti-Verlag, Bielefeld,

111 Chapter Three

Gontier M, Mörtberg U, Balfors B (2010) Comparing GIS-based habitat models for applications in EIA and SEA. Environmental Impact Assessment Review 30: 8-18. doi:10.1016/j.eiar.2009.05.003 Graf RF, Kramer-Schadt S, Fernandez N, Grimm V (2007) What you see is where you go? Modeling dispersal in mountainous landscapes. Landscape Ecology 22: 853-866. doi:10.1007/s10980-006-9073-3 Grimm V (1999) Ten years of individual-based modelling in ecology: what have we learned and what could we learn in the future? Ecological Modelling 115: 129-148. doi:10.1016/S0304-3800(98)00188-4 Grimm V, Berger U, Bastiansen F, Eliassen S, Ginot V, Giske J, Goss-Custard J, Grand T, Heinz SK, Huse G, Huth A, Jepsen JU, Jorgensen C, Mooij WM, Muller B, Pe'er G, Piou C, Railsback SF, Robbins AM, Robbins MM, Rossmanith E, Ruger N, Strand E, Souissi S, Stillman RA, Vabo R, Visser U, DeAngelis DL (2006) A standard protocol for describing individual-based and agent-based models. Ecological Modelling 198: 115-126. doi:10.1016/j.ecolmodel.2006.04.023 Grimm V, Berger U, DeAngelis DL, Polhill JG, Giske J, Railsback SF (2010) The ODD protocol: A review and first update. Ecological Modelling 221: 2760-2768. doi:10.1016/j.ecolmodel.2010.08.019 Hartung H (1991) Untersuchung zur terrestrischen Biologie von Populationen des Moorfro- sches (Rana arvalis NILSSON 1842) unter besonderer Berücksichtigung der Jahresmobilität. Hamburg: Universität Hamburg. Hassingboe J, Neergaard RS, Hesselsøe M (2012) Manual til produktion af GIS raster kort til:”EDB-værktøj til at vurdere skader på bestande af padder /økologisk funktionalitet”. Am- phi Consult, pp. Hels T, Nachman G (2002) Simulating viability of a spadefoot toad Pelobates fuscus metapopulation in a landscape fragmented by a road. Ecography 25: 730-744. doi:10.1034/j.1600- 0587.2002.250609.x Holderegger R, Di Giulio M (2010) The genetic effects of roads: A review of empirical evidence. Basic and Applied Ecology 11: 522-531. doi:10.1016/j.baae.2010.06.006 Iuell B, Bekker GJ, Cuperus R, Dufek J, Fry G, Hicks C, Hlaváˇc V, Keller V, B. R, C., Sangwine T, Tørsløv N, Wandall BlM (Eds) (2003) Wildlife and Traffic: A European Hand- book for Identifying Conflicts and Designing Solutions. Office for official publications of the European Communities, Luxembourg, pp. Jaeger JaG, Fahrig L (2004) Effects of Road Fencing on Population Persistence. Conservation Biology 18: 1651-1657. doi:10.1111/j.1523-1739.2004.00304.x Kovar R, Brabec M, Vita R, Bocek R (2009) Spring migration distances of some Central European amphibian species. Amphibia-reptilia 30: 367-378 Kramer-Schadt S, Revilla E, Wiegand T, Breitenmoser U (2004) Fragmented landscapes, road mortality and patch connectivity: modelling influences on the dispersal of Eurasian lynx. Journal of Applied Ecology 41: 711-723. doi:10.1111/j.0021-8901.2004.00933.x Legendre P, Legendre L (1998) Numerical ecology. Elsevier, pp.

112 Chapter Three

Lesbarreres D, Fahrig L (2012) Measures to reduce population fragmentation by roads: what has worked and how do we know? Trends in Ecology & Evolution 27: 374-380. doi:10.1016/j.tree.2012.01.015 Loman J (1994) Site tenacity, within and between summers, of Rana arvalis and Rana temporaria. Alytes 12: 15-29 McLane AJ, Semeniuk C, McDermid GJ, Marceau DJ (2011) The role of agent-based models in wildlife ecology and management. Ecological Modelling 222: 1544-1556. doi:10.1016/j.ecolmodel.2011.01.020 Minor ES, Urban DL (2008) A graph-theory frarmework for evaluating landscape connectivity and conservation planning. Conservation Biology 22: 297-307. doi:10.1111/j.1523- 1739.2007.00871.x Olsson MPO, Widen P (2008) Effects of highway fencing and wildlife crossings on moose Alces alces movements and space use in southwestern Sweden. Wildlife Biology 14: 111-117. doi:10.2981/0909-6396(2008)14[111:eohfaw]2.0.co;2 Pe'er G, Henle K, Dislich C, Frank K (2011) Breaking Functional Connectivity into Compo- nents: A Novel Approach Using an Individual-Based Model, and First Outcomes. PLoS ONE 6. doi:10.1371/journal.pone.0022355 Pontoppidan M-B, Nachman G (In prep.) Changes in behavioural responses to infrastructure affects local and regional connectivity – a simulation study on pond-breeding amphibians. Pontoppidan M-B, Nachman G (In review) Effects of within-patch heterogeneity on connectivity in pond-breeding amphibians studied by means of an individual-based model. Web- ecology: Pope SE, Fahrig L, Merriam NG (2000) Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81: 2498-2508 Semlitsch RD (2008) Differentiating migration and dispersal processes for pond-breeding amphibians. Journal of Wildlife Management 72: 260-267. doi:10.2193/2007-082 Sinsch U (1990) Migration and orientation in anuran amphibians. Ethology Ecology & Evolu- tion 2: 65-79 Sinsch U (2006) Orientation and navigation in Amphibia. Marine and Freshwater Behaviour and Physiology 39: 65-71. doi:10.1080/10236240600562794 Sjögren-Gulve P (1998) Spatial movement patterns in frogs: Differences between three Rana species. Ecoscience 5: 148-155 Spellerberg IF (1998) Ecological effects of roads and traffic: a literature review. Global Ecol- ogy and Biogeography 7: 317-333. doi:10.1046/j.1466-822x.1998.00308.x Trocmé M, Cahill S, de Vries JG, Farrall H, Folkeson LG, Hichks C, Peymen J (Eds) (2003) COST 341 – Habitat Fragmentation due to Transportation Infrastructure. Office for Official Publications of the European Communities, Luxembourg, pp. Trombulak SC, Frissell CA (2000) Review of ecological effects of roads on terrestrial and aquatic communities. Conservation Biology 14: 18-30. doi:10.1046/j.1523- 1739.2000.99084.x

113 Chapter Three

Vos CC, Chardon JP (1998) Effects of habitat fragmentation and road density on the distribution pattern of the moor frog Rana arvalis. Journal of Applied Ecology 35: 44-56 Watts K, Handley P (2010) Developing a functional connectivity indicator to detect change in fragmented landscapes. Ecological Indicators 10: 552-557. doi:10.1016/j.ecolind.2009.07.009 Wiens JA (1997) Metapopulation Dynamics and Landscape Ecology. In: Hanski I, Gilpin ME (Eds) Metapopulation Biology: ecology, genetics, and evolution. Academic press, Inc., Wilensky U (1999) NetLogo. Center for Connected Learning and Computer-Based Modeling, Northwestern University, Evanston, IL. , http://ccl.northwestern.edu/netlogo, pp. Zetterberg A, Mortberg UM, Balfors B (2010) Making graph theory operational for landscape ecological assessments, planning, and design. Landscape and urban planning 95: 181-191. doi:10.1016/j.landurbplan.2010.01.002

114 Chapter Three

Tables

Table 1 List of variables characterizing the agents in SAIA Value Agent Variable Notation Description range type AreaValue W 0.5; 1 Cell Effective area of the cell

DailySurvival Ds Cell Daily survival probability FrogDensity D Cell Mean number of frogs in the cell The cell’s relative attraction to frogs HabitatAttraction H 1-5 Cell a during movement

HabitatCode Hc Cell The land cover category of the cell

HabitatSurvival Hs 1-5 Cell The cell’s relative survival index The cell’s relative suitability as summer SummerQuality H 1-5 Cell q habitat BreedingPond Frog Breeding pond of frog agents NatalPond Frog Natal pond of frog agents PondID Pond ID number PondPerimeter O Pond Perimeter of the pond PondQuality Q 0.1-1 Pond Quality index of the pond Number of egg masses found in the PopulationSize N Pond 0 pond during survey Summer habitat cells associated with SummerHabitat A Pond the pond Effective area of associated summer SummerHabitatArea A' Pond habitat Froglets Pond Number of newly metamorphosed frogs Number of surviving individuals in age AgeClassList Pond class 0-6

115 Chapter Three

Table 2 Life table for Rana arvalis and estimated age distribution Percentage of popula- Stage/Age Survival probability Fecundity ( eggs pr. female) tion Egg/larvae 0.005 - - 0 0.55 0 46.7 % 1 0.55 70 25.7 % 2 0.55 945 14.1% 3 0.55 1190 7.8% 4 0.50 1250 3.9% 5 0.40 1300 1.6 % 6 0.20 1300 0.3%

116 Chapter Three

Table 3 Land cover categories and the associated values of HabitatSurvival, HabitatAttraction and SummerQuality HabitatCode HabitatAttraction HabitatSurvival SummerQuality Description (Hc) (Ha) (Hs) (Hq) 2 4-lane motorway 2 N/A 1 3 2-lane motorway 2 N/A 1 4 Road, width > 6m 3 N/A 1 5 Road, width 3-6 m 3 N/A 1 6 Other roads 3 2 2 8 Pathway 4 4 3 11 Multiple surface 3 3 3 11 Railway 4 2 3 12 Building 1 N/A N/A 15 Other made surface 2 3 2 18 Wetlands 5 5 5 20 Running water 4 4 3 22 Meadows 5 5 5 24 Grassland 4 4 4 25 Lakes 1 N/A N/A 28 Hedgerow 4 4 4 29 Heath land 5 5 4 32 Woodland 4 4 4 34 Stand of trees 4 4 3 36 Bare surface 2 2 1 40 Fallow land 4 4 4 42 Field crops 2 2 2 50 Drift fence 1 N/A N/A 60 Underpass 4 4 1

117 Chapter Three

Table 4 Results from analysis of scenario 0 (before road construction)

Cluster ID c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 Number of 9 5 2 8 20 13 19 7 7 11 12 5 3 ponds in cluster Number of high quality ponds 2 0 0 4 2 0 1 7 3 0 3 1 0 (Q> 0.6) Connectivity to 0.45 0.54 0.30 0.57 2.19 1.13 2.94 0.36 1.20 0.07 0.10 0.05 0.06 other clusters Estimated clus- 0 0 0 49 0 3 11 23 14 0 51 0 0 ter abundance Mean pond persistence 0 0 0 0.82 0.03 0.29 0.35 0.93 0.73 0.01 0.77 0 0.13 probability Connectivity to 0 0.09 0.04 - 0.01 0.44 0 0 0 0 0 0 0 c4 Connectivity to 0 0 0 0 0 0.23 0.04 - 0.10 0 0 0 0 c8 Connectivity to 0 0 0 0 0 0.00 0 0 0.02 0.02 - 0 0.06 c11

Table 5 Results from analysis of scenario 1 (after road construction)

Cluster ID c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 Number of ponds 9 5 2 8 19 13 17 7 6 9 12 5 3 in cluster Number of high quality ponds 2 0 0 4 2 0 1 7 2 0 3 1 0 (Q> 0.6) Connectivity to 0.02 0.10 0.27 0.52 1.69 0.67 1.92 0.23 0.55 0.06 0.09 0.05 0.06 other clusters Estimated cluster 0 0 0 50 2 2 1 24 5 0 49 0 0 abundance Mean pond persistence probabil- 0 0 0.03 0.84 0.05 0.24 0.08 0.93 0.26 0 0.78 0 0.15 ity Connectivity to 0 0.08 0 - 0 0.44 0 0 0 0 0 0 0 c4 Connectivity to 0 0 0 0 0 0.20 0.01 - 0.02 0 0 0 0 c8 Connectivity to 0 0 0 0 0 0 0 0 0.03 0 - 0 0.06 c11

118 Chapter Three

Table 6 Results from analysis of mitigation measures. Scenario 2: Drift fences and underpasses; Scenario 3a: Implementing new ponds in clusters c9 and c11; Scenario 3b: Implementing new ponds as corridor between c9 and c11. Cluster Estimated cluster Mean pond persistence Connectivity to c8 Connectivity to c11 ID abundance probability S2* S3a S3b S2* S3a S3b S2* S3a S3b S2 S3a S3b c4 40 48 49 0.82 0.82 0.82 0 0 0 0 0 0 0.002 c7 1 3 4 0.08 0.11 0.11 0 0 0 0 0 (0.004) 17 0.90 c8 23 23 0.90 0.92 - - - 0 0 0 (21) (0.95) 6 0.38 0.19 c9 11 10 0.44 0.58 0.02 0.02 0.025 0.03 0.22 (2) (0.27) (0.29) c11 40 50 61 0.74 0.80 0.85 0 0 0 - - - *) In scenario 2 one pond originally belonging to c8 is annexed by c9. Entries in parentheses are values based on the original cluster configurations.

119 Chapter Three

Figure legends Figure 1: Location of two study areas in Denmark. KaB is an area near Kalundborg on Zealand and HoB is near Holstebro in Jutland. Only KaB is used in the present analysis, but both areas are used for the parameterisation of the model Figure 2: Scenario 0. A) Map of the landscape before road constructions. Black dots represent potential breeding ponds. Small dots are ponds with pond quality (Q) ≤ 6; large dots are ponds with Q ≥ 7. Populated ponds are indicated with a star shape. B) Result of cluster analyses showing clusters c1–c13. Ponds linked with black lines belong to the same cluster. Pond size and colour indicate the result of the PopulationDynamics procedure. Yellow circles represent ponds with an estimated population size ≥ 1. Ponds with larger yellow circles have a persistence probability > 0.75. Figure 3: Scenario 1. A) Map of the landscape after road constructions (red road). Black dots represent potential breeding ponds. Small dots are ponds with pond quality (Q) ≤ 6; large dots are ponds with Q ≥ 7. Populated ponds are indicated with a star shape. Pink ponds are ponds removed in connection with the constructions. B) Result of cluster analyses showing clusters c1–c13. Ponds linked with black lines belong to the same cluster. Pond size and colour indicate the result of the PopulationDynamics procedure. Yellow circles represent ponds with an estimated population size ≥ 1. Ponds with larger yellow circles have a persistence probability > 0.75. Figure 4: Analyses of mitigation measures. Result of cluster analyses showing clusters c3–c11. Ponds linked with black lines belong to the same cluster. Pond size and colour indicate the result of the PopulationDynamics procedure. Yellow circles represent populated ponds. Ponds with larger yellow circles have a persistence probability > 0.75. A) Scenario 2 - Location of underpasses is shown with red arrows. B) Scenario 3a – Three new ponds in cluster c9 and five new ponds in c11 are shown with red dots C) Scenario 3b – Eight new ponds connecting c9 and c11 are shown with red dots Figure 5: Key results from analysis of the five scenarios. Upper and lower 95% confidence limits are indicated with black triangles.

120 Chapter Three

Figures Figure 1

121 Chapter Three

Figure 2

Figure 3

122 Chapter Three

Figure 4

Figure 5

123 Chapter Three

Supplementary material - Appendix 1

Model ODD

Purpose Road construction and implementation of mitigation measures may change landscape features and thereby affect amphibians like the Moor frog. The purpose of SAIA is to assess how such changes will influence landscape connectivity and the persistence of a regional population of Moor frogs. The habitat of pond breeding amphibians as the Moor frog includes terrestrial as well as aquatic habitat. Therefore, the habitat patch of a subpopulation is modelled as a complementary habitat patch containing not only the breeding pond but also all accessible summer habitat within migration distance from the pond (Dunning et al. 1992; Pontoppidan and Nachman In review; Pope et al. 2000). Immigration, thus, requires two events: 1) the successful dispersal of a juvenile frog to summer habitat outside its natal habitat patch and 2) subsequent successful migration to a breeding pond associated with the new summer habitat. In real life these two events is 2 year apart, but, for simplicity, we only simulate the dispersal and migration events not the intervening years.

Entities, state variables, and scales Breeding ponds are treated as stationary agents. Each pond is characterized by a unique ID- number, population size, pond quality, pond perimeter, the associated summer habitat and the quality-weighted area of the summer habitat. Frog agents are characterized by the pond in which they are hatched and the breeding pond they immigrate to. The extent of the model landscape is 600 x 800 grid cells, and each grid cell is 10 x 10 m. Grid cells are defined by their relative attraction to dispersing frogs, habitat survival index and a daily survival probability, the habitat type and the cell’s relative value as summer habitat (see Table 1 in main text). The first part of the simulation mimics the dispersal of newly metamorphosed frogs, starting in mid-summer until hibernation in autumn. The second part considers the spring movement of juveniles from the summer habitat towards their future breeding pond and back to their summer habitat. Each part runs for 120 time steps, one step representing one day.

124 Chapter Three

Process overview and scheduling At the start of a simulation, 250 frog agents are located at each pond agent and the frog variable NatalPond is updated with the ID-number of the pond. In the first part of the simulation (dispersal) the following procedures are executed each time-step: Move (movement of frog agents), Settle (evaluates if a frog agent stops dispersing and assigns frogs to breeding ponds) and Survival (evaluates if a frog agent survives). At day 120 the dispersal stops; frog agents that have not settled are removed and the migration simulation starts. The two procedures Move and Survival are run every time step and settled frogs start moving again, this time towards their breeding pond. When a frog reaches its assigned breeding pond, its direction is set towards one of the summer habitat fragments associated with the breeding pond. The simulation stops at day 240 and at each pond the model counts the number of immigrants from each of the other pond agents, computing immigration probabilities between all pairs of ponds. These immigration probabilities then enter the population-based PopulationDynamics (PD) procedure, which is executed. The simulation is repeated 50 times.

Design concepts

Emergence Immigration rates emerge as a response to the landscape configuration.

Sensing Frog agents are assumed to be aware of their own state variables. Frog agents are also aware of the habitat attraction of the grid cells as well as the identities of the ponds.

125 Chapter Three

Interaction There are no interactions between frog agents. Movement decisions of the frog agents depend on the habitat attraction of the neighbouring cells. Survival of the frog agents depends on the daily survival rates associated with the traversed habitat. The cell-variables HabitatAttraction and DailySurvival depend on daily precipitation.

Stochasticity Which cell to move to is chosen randomly among the neighbouring cells with the probability of being chosen weighted by the habitat attraction of the neighbouring cells. If frog agents occupy a cell suitable as summer habitat they will stop dispersing with a certain probability; the probability increases with time. The breeding pond of a settled frog is chosen randomly among accessible breeding ponds weighted by pond-quality. Demographic, but not environmental, stochasticity is included in the PopulationDynamics procedure.

Observation At the end of each simulation, the natal pond and breeding pond of all frog agents are registered and immigration probabilities (pij) between all pair-wise pond agents are calculated. The PD-procedure is run and the estimated population size of each pond agent is recorded.

Initialization A landscape is constructed based on a GIS-raster data set. Each cell contains information about habitat type, habitat attraction, habitat survival and summer quality. A data set with information on location, ID-number, pond quality, population size, and pond perimeter of the surveyed ponds is used to create pond agents. Once the landscape is created, the Map-scan procedure is run to identify all accessible summer habitat associated with each breeding pond and the pond variables A and A’ are updated. Habitat survival is converted to daily survival probabilities and the cell variable Ds is updated. A random year is chosen from a climate database and a data set containing daily precipitation is constructed. 250 frog agents are located on each pond agent and their dispersal direction is set randomly. The precipitation threshold α is set.

Input data A database containing data on daily precipitation measured in Copenhagen for 21 consecutive years (1985-2005) is used to reflect natural weather patterns. At the start of a simulation, a

126 Chapter Three random year is chosen from this database and at each time step, information on precipitation is drawn for the simulated day of the year. Data were supplied by the Danish Meteorological Institute (Cappelen 2009).

Submodels

Map-scan

The cell variable DailySurvival (Ds) is set as the probability of a frog agent surviving one time step in the cell. This depends on the habitat code (Hc) and the habitat survival index (Hs) of the cell. Cells belonging to roads, Hc = [2, 3, 4 5] are assigned Ds values specific to their habitat code (see Appendix 1, Parameterisation). All other cells are modelled as

1∨ 6 , where σ0 and σ1 are species -specific constants.

Local populations of pond breeding amphibians inhabit a composite habitat patch. The breeding pond is located at its core, surrounded by satellites of summer habitat fragments separated by matrix habitat (Pontoppidan and Nachman In review). The Map-scan procedure delimits the extent of the habitat patch as all accessible cells within a 40-cell (400 m) radius of the pond. Accessible cells are defined as cells with habitat attraction (Ha) greater than 1 and daily survival (Ds) greater than 0.3. This excludes structures such as buildings and large roads. Fur- thermore, inaccessible cells function as barriers blocking access to the habitat beyond (Eigen- brod et al. 2008) (Fig. A1). Next, the area of summer habitat within the habitat patch is found.

Summer habitat cells are defined as cells with a summer quality (Hq) higher than 3. A summer habitat cell can be completely surrounded by other summer habitat cells (core cells) or have one or more neighbouring cells which are not summer habitat (edge cells). To account for edge effects, core cells are given the area value (W) of 1 while W is 0.5 for edge cells (Watts and Handley 2010). The effective area of the summer habitat (A´) within the habitat patch is then computed as:

∑ , where n is the number of summer habitat cells belonging to the habitat patch and is the area value of cell i.

Move Each day frog agents move a randomly chosen distance depending on the habitat attraction

(Ha) of its current cell. The travelling distance is drawn from a normal distribution with a mean of c and a standard deviation of s. Assuming the frog to head in the direction it was as-

127 Chapter Three signed when it left the natal pond, it moves to one of its neighbouring cells located within 0 ±90 from the preferred direction. Cells with Ha = 1 are considered inaccessible. Based on the habitat attraction of the neighbouring and accessible cells (n), frogs first decide which kind of habitat they want to move to. The probability of moving into one of the cells with habitat at-

traction H is found as 1 , where na is the number of neighbouring cells a ∑ with habitat attraction Ha and Hai is the habitat attraction of cell i. A uniform pseudorandom number is selected to choose the type of habitat. If there is more than one neighbouring cell with the chosen habitat attraction, one of them is chosen randomly with equal probability. The frog agent then moves to a random position within the cell, without changing its direction. This routine is repeated until the chosen travelling distance for the day is traversed. If daily precipitation exceeds the threshold value (α), habitat attraction is inconsequential and the frog moves randomly to one of its accessible neighbours.

During the second part of the simulation, as frog agents get within two cells from their destination (breeding pond or summer habitat), the frogs move directly to it. At the breeding pond, frog agents are randomly assigned a summer habitat cell within the habitat patch. When the frogs reach their summer habitat they stop moving. Frog agents reaching the boundary of the landscape are removed. If a frog agent does not have an accessible cell within moving range, its direction is permanently changed either 35 degrees to the left or to the right.

Settle Dispersing frogs encountering a summer habitat cell have a certain probability of settling. The probability depends on the Julian day (t) and is found as:

, where ν0 and ν1 are species specific constants

When the summer habitat cell in which the frog agent has settled is part of a habitat patch, the frog is assigned the associated breeding pond. If the summer habitat cell is shared by several ponds, the probability of being assigned a breeding pond i is a function of the pond quality

(Q) of the available breeding ponds (n): . Once a frog is settled, it stops moving ∑ until the second part of the simulation.

128 Chapter Three

Survival For each frog agent a pseudo-random number is drawn between 0 and 1. If the number exceeds the geometric average of DailySurvival (Ds) of the cells traversed during the day, the agent dies. When daily precipitation exceeds the threshold value (α), Ds is temporarily set to the highest value (Table A1), except for cells belonging to the road categories, Hc = [2, 3, 4 5].

PopulationDynamics (PD) The PD-procedure estimates frog population size in each pond running through 40 iterations of a life cycle model. The elements of the life cycle model are 1) Reproduction, 2) Survival and 3) Immigration. For simplicity we only model the female part of the population. We assume a sex ratio of 0.5 and that females always are mated. Pond populations are grouped by age 0 through 6 years, and preliminary iterations of the life cycle produced an estimate of the age distribution (see Table 2 in main text). The initial population sizes of the ponds are the number of egg masses found in the surveyed ponds (N0). This is expected to equal the number of adult females of ages 2 - 6. Based on the age distribution, the initial number of frogs of ages 0-6 was estimated and AgeClassList is updated. After each iteration, the pond variables Froglets and AgeClassList are updated with the reproductive output and the number of surviving individuals in each age class, respectively. The number of immigrants is added to age class 0 in AgeClassList.

Reproduction

The number of newly metamorphosed frogs, ready to disperse, is considered to be the reproductive output. This is the product of egg production, egg and larval survival, as well as the survival of the frogs the first 2 weeks after metamorphosis. The reproductive output is, thus, modelled as follows. Individuals can reproduce starting at age 3. A mated female produces R eggs (R = 0, 1, 2 ...) with probability P(R) which is assumed to follow a negative binomial distribution with mean and clumping parameter k, i.e.

k k  __  (k  R)  k   R  P(R)    (R1) R!k  __   R  k   R k   

varies with age so the expected egg production of a female of age ai is modelled as

129 Chapter Three

(for a ≥3 , otherwise Rai  0 ) (R2) where ρ0 and ρ1 are parameters expressing the effect of age on reproduction. Rm is the reproduction if ρ0 = ρ1 = 0.

The overall probability that an egg develops into a young frog that survives until dispersal is assumed to be affected by two factors: intraspecific competition and pond quality. In- traspecific competition is modelled as

F (R3) where ψ0 is a species specific constant and d is egg density in the pond. Note that density is assumed to be proportional with the perimeter of the pond (O) rather than with the area. The effect of pond quality (Q) is modelled as

F R4

In combination, survival probability until the first two weeks after metamorphosis is found as

Survival

The conditional probability that a frog survives from age a to age a+1 is assumed to depend on age, which is modelled with a logistic model as

2 e 0 1a2a P a 1a  s    a a2 1 e 0 1 2 S1 where β0, β1 and β2 are species-specific constants.

Furthermore, survival is assumed to depend on the frog density in the summer habitat. For simplicity, this is modelled as a “culling” process when frog density (D) exceeds the carrying capacity (Kj) of summer habitat cell j. Kj is determined as WK, where W is the cell’s area value and K is the estimated maximum carrying capacity per cell.

Frog density is calculated as ∑ , where h is the ponds sharing summer habitat cell j, and zi is the number of frogs in age 0 to 6 in pond i divided by the number of summer habitat cells belonging to pond i. The number of frogs in pond i after “culling” is calculated as:

130 Chapter Three

∑ , where n is summer habitat cells belonging to pond i, m j  zi 1 D j / K j , when Dj > Kj and mj = zi, when Dj ≤ Kj. Culling is done proportionally on all ages.

Migration

Immigration probabilities between all pairs of ponds are obtained from the individual based simulation. The number of immigrants (I) arriving at pond i is modelled as:

∑ , where sij is the immigration probability of an individual moving from pond j to pond i and nj is the reproductive output of pond j given by the pond variable Fro- glets. Emigration rates are not modelled explicitly.

Parameterisation

All parameter values are listed in table A2.

Habitat attraction (Ha) Terrestrial amphibians are assumed to prefer habitat in which the water content is high, thereby minimizing the risk of desiccation. In an experiment with Northern green frogs and Northern leopard frogs in peatlands, Mazerolle and Desrochers (2005) found that 18 out of 25 frogs (72%) avoided barren surfaces. Hartung (1991) found that Moor frogs avoided areas with sparse or low vegetation, and recorded the ratio between densities in grass areas and densities in moor lands, hedges, ditches and forests to be 1:3.5.

In the model, the probability of a frog agent choosing one type of cell above another during movement therefore depends on the attractiveness of the cell’s habitat type. The habitat attraction (Ha) of the different habitat types in the GIS maps was approximated by amphibian specialists (see Table 3 in main text). We tested three different expressions of Ha to enter 2 into the Move-procedure: a) Ha, b) (Ha) and c) exp(Ha) and compared the results with the above-mentioned empirical findings. In addition, we ran a simulation where movement was independent of habitat attraction.

As test landscapes we used GIS data sets from two different road projects in Denmark, supplied by the Danish Road Directorate and Amphi Consult. The first project (KaB) concerns the area described in the main text while the second project (HoB) is from central Jut- land, ca. 5 km east of Holstebro (56° 19.66’ N 8° 44.65’ E N) (Fig. 1). Both areas are charac-

131 Chapter Three terised as semi-urban and agricultural landscapes, traversed by creeks and wetlands. In the HoB map 36% of all cells were classified as attractive habitat (Ha > 3) and the KaB map contained 51% attractive habitat cells.

The model was run for 40 time steps without the Settle-procedure, and the ratio between frog agents in attractive (Ha = 4 or 5) and unattractive (Ha = 2 or 3) habitat was computed as the percentage of frog agents in attractive habitat. Although there were differences between the maps, we chose the expression exp(Ha) as being the best to reproduce the empirical patterns (Table A3).

Distance travelled per day In a radio tracking experiment with Common frogs (Rana temporaria), Tramontano (1997) found that adult frogs moving through a rye field covered 148 m in one week, corresponding to ca. 20 m per day. In a study on dispersing juvenile Moor frogs, Hartung (1991) reported daily travelling distances of 12.5 – 18.8 m (mean 15.5) m in attractive habitat (moors) and 39.9 – 40.9 m in unattractive areas (pine forests). The daily travelling distance was thus assumed to depend on habitat attraction of the current cell and is modelled as normal distribution with a mean of and standard deviation s.

Two homogenous landscapes were constructed with a habitat attraction of 2 or 4, respectively. In each landscape frog agents were allowed to move according to the Move- procedure for 40 time steps. When a simulation ended, the straight distances between start and end point for all frog agents were measured and the mean daily travelling distance computed. The simulations were conducted for varying values of c and s, each combination repeated 50 times. A parameter set was sought where the daily travelling distance

 in attractive habitat takes values between 12 m and 19 m and with a mean around 15 m  in unattractive habitat takes values in the range between 20 m and 40 m

The parameter set (c=7, s=0.5) was chosen as the one that best fulfilled the conditions.

Settle Few data have been reported on dispersal distances. Hartung (1991) found dispersal distances of newly metamorphosed Moor frogs up to 1200 m, but mean dispersal distance is estimated by experts to be 200-300 m. As the young frogs are assumed to have an innate urge to move away from their natal pond, settling probability is expected to be low in the beginning. The

132 Chapter Three dispersal drive is expected to subside with time, with settling probability increasing accord- ingly, creating an s-shaped probability function (Fig A2). The realized dispersal distances will depend on the landscape configuration. With the chosen parameter set (ν0, ν1), the mean dispersal distance of frog agents in the HoB map is 363 m and the maximum dispersal distance is 2068 m. From this, 68% of the frog agents settle within their own habitat patch. Less than 2% of the frog agents disperse more than 1000 m. In the KaB map, the mean and maximum dispersal distances are 291 m and 2151 m, respectively. From this, 83% settle within the home patch and less than 0.5% disperse more than 1000 m (Fig. A3).

Daily survival The survival probability of frogs is assumed to depend on the habitat. We were unable to find any empirical data on daily survival probabilities, but annual survival rates of young adult frogs have been estimated to be between 55% (Fog and Hesselsøe 2009) and 63% (Loman 1984). These rates are not habitat specific but are realised during annual movements in a heterogeneous landscape.

All land cover categories in GIS maps were ranked according to survivability by amphibian specialists and assigned a value of relative survival index (Hs) (see Table 3 in main text), which was subsequently converted into daily survival probabilities (Ds). The parameter set (σ0, σ1) gives the functional relationship between Hs and Ds. The parameters were found by iteration. The model was run with varying combinations of parameter values on both test maps until a set was found for which the annual survival rates were between 55% and 63%. With the chosen parameter set, the realised annual survival rates were between 56 % and 57 % in both maps. Table A1 shows the resulting habitat specific daily survival probabilities.

Road survival Hels and Buchwald (2001, fig. 5) found the probability of a Moor frog being killed when crossing a road ranged from ca. 35% to ca. 90%, depending on traffic intensity. We assumed traffic intensity to be correlated with road width and let daily survival probability (Ds) depend on road category as shown in table A4.

PopulationDynamics Parameterisation of reproduction and survival are primarily based on life-table data constructed by amphibian experts (see Table 2 in main text) as well as experts’ best guesses on quality-dependent survival (Table A5).

133 Chapter Three

Reproduction and egg survival

The reproduction parameters (Rm, ρ0, ρ1) are estimated by fitting equation R2 to life-table data on fecundity (Fig. A4). The estimation of clumping factor k is based on data from Lyapkov (2008) (Table A6).

The quality-dependent survival parameters ψ1 and ψ2 are estimated by fitting equation R4 to data in table A5. These survival rates are expected to be realized only when the effect of egg density is negligible. Based on survey data from other Danish road projects (unpub- lished), egg density can be expected to range from 1 to 900 eggs per meter of pond perimeter with a mean density of ca. 80 eggs/m. The density-dependent parameter ψ0 is estimated by iteration, finding the value at which survival probability in high-quality ponds and with an egg density of 1000 eggs/m is ca. 3% (Fig A5).

Frog survival

The survival parameters (β0, β1, β2) are estimated by fitting equation S1 to life-table data on survival (Fig. A6). Maximum carrying capacity (K) of a summer habitat cell depends on the actual landscape map being analysed and is determined at the start of the PD-procedure as the mean initial frog density of summer habitat cells associated with the populated ponds.

134 Chapter Three

References

Cappelen J (Ed) (2009) DMI Daily Climate Data Collection 1873-2008, Denmark, The Faroe Islands and Greenland - including Air Pressure Observations 1874-2008 (WASA Data Sets). Danish Meteorological Institute, Copenhagen, pp. Dunning JB, Danielson BJ, Pulliam HR (1992) Ecological processes that affect populations in complex landscapes. Oikos 65: 169-175. doi:10.2307/3544901 Eigenbrod F, Hecnar SJ, Fahrig L (2008) Accessible habitat: an improved measure of the effects of habitat loss and roads on wildlife populations. Landscape Ecology 23: 159-168. doi:10.1007/s10980-007-9174-7 Fog K, Hesselsøe M (2009) Udvikling af prototypemodel til brug for forvaltning af spidssnudet frø i forbindelse med vejanlæg. Amphi Consult, pp. Hartung H (1991) Untersuchung zur terrestrischen Biologie von Populationen des Moorfro- sches (Rana arvalis NILSSON 1842) unter besonderer Berücksichtigung der Jahresmobilität. Hamburg: Universität Hamburg. Hels T, Buchwald E (2001) The effect of road kills on amphibian populations. Biological Conservation 99: 331-340. doi:10.1016/S0006-3207(00)00215-9 Loman J (1984) Density and survival of Rana arvalis and Rana temporaria. Alytes 3: 125- 134 Lyapkov SM (2008) A long-term study on the population ecology of the moor frog (Rana arvalis) in Moscow province, Russia. In: Glandt D, Jehle R (Eds) The Moor Frog Laurenti- Verlag, Bielefeld, 211-230 Mazerolle MJ, Desrochers A (2005) Landscape resistance to frog movements. Canadian Jour- nal of Zoology-Revue Canadienne De Zoologie 83: 455-464. doi:10.1139/z05-032 Pontoppidan M-B, Nachman G (In review) Effects of within-patch heterogeneity on connectivity in pond-breeding amphibians studied by means of an individual-based model. Web- ecology: Pope SE, Fahrig L, Merriam NG (2000) Landscape complementation and metapopulation effects on leopard frog populations. Ecology 81: 2498-2508 Tramontano R (1997) Continuous radio tracking of the common frog, Rana temporaria. Her- petologia Bonnensis: 359-365 Watts K, Handley P (2010) Developing a functional connectivity indicator to detect change in fragmented landscapes. Ecological Indicators 10: 552-557. doi:10.1016/j.ecolind.2009.07.009

135 Chapter Three

Tables

Table A1 Habitat survival index (Hs), the corresponding daily survival probability (Ds) and Ds converted into annual survival probability.

Hs Ds Annual survival probability 1 0.9820 0.01 2 0.9960 0.38 3 0.9984 0.68 4 0.9991 0.81 5 0.9995 0.88

Table A2 List of parameters, their default values and the procedure in which they appear Parameter Value Procedure

ν0 1.50E-06 Settle

ν1 0.06 Settle

σ0 2.2 Map-scan

σ1 4 Map-scan

τ0 0.76 Map-scan

τ1 -5.89 Map-scan c 7 Move s 0.5 Move α 5 mm Move k 20.5 PD (reproduction)

Rm 2.61 PD (reproduction)

ρ0 0.168 PD (reproduction)

ρ1 -0.014 PD (reproduction)

ψ0 0.03 PD (egg survival))

ψ1 2.20E-05 PD (egg survival)

ψ2 8.0E-4 PD (egg survival)

β0 0.1 PD (survival)

β1 0.28 PD (survival)

β2 -0.08 PD (survival) K 0.04 PD (survival)

136 Chapter Three

Table A3 Observed and emerging patterns using three different expressions for entering Ha into the Move procedure as well as random movement independent of Ha. Results are shown for two different maps, Kalundborg (KaB) and Holstebro (HoB) 2 Random Ha Ha exp(Ha) Pattern Observed KaB HoB KaB HoB KaB HoB KaB HoB Ratio between frog densities in good and bad 0.29 0.39 0.48 0.46 0.64 0.43 0.53 0.38 0.46 habitat (Hartung 1991) Percentage of individuals choosing good 72% 72% 67% 68% 61% 70% 66% 73% 68% habitat (Maze- rolle and Des- rochers 2005)

Table A4 Road categories with the corresponding habitat code (Hc) and daily survival probability (Ds)

HabitatCode (Hc) Road category Ds 2 4-lane motorway 0.1 3 2-lane motorway 0.2 Road width > 6 4 0.5 m Road width 3-6 5 0.8 m

137 Chapter Three

Table A5 Estimated survival probability from egg until 2 weeks after metamorphosis depending on pond quality (Q) Pond quality (Q) Survival probability 0.1 0.00001 0.2 0.00015 0.3 0.00032 0.4 0.00067 0.5 0.0014 0.6 0.003 0.7 0.0064 0.8 0.0136 0.9 0.029 1.0 0.061

Table A6 Age specific mean value of fecundity and coefficient of variance of Rana arvalis from a study in Moscow, Russia (Lyapkov 2008) Average number of Coefficient of Variation Age eggs (%) 3 1057 22,3 4 1193 22,3 5 1267 19,9 6 1332 25,5

138 Chapter Three

Figures Figure A1 Illustration of how accessible summer habitat is identified. Blue circle is a pond; dotted circle represents maximum migration distance. Green areas are accessible summer habitat while shaded areas are inaccessible summer habitat. A) All summer habitat within migration distance is regarded as accessible B) Road traversing the habitat prevents access to summer habitat on the opposite side of the road C) Structures breaking the road such as underpasses again permits access to summer habitat on the opposite side of the road. A B C

139 Chapter Three

Figure A2 Settling probability. If a frog agent encounters a summer habitat cell, the probability of settling in the cell will depend on day number (Julian day).

1.0

0.8

0.6

0.4 Settle probability

0.2

0.0

170 190 210 230 250 270 290 310 Daynumber

Figure A3 Dispersal distances The frequency distribution of dispersal distances with the chosen settling parameters for a) HoB map b) KaB map. Black line shows accumulated frequencies.

A B

25 25

20 20

15 15 % %

10 10

5 5

0 0 0.10.20.30.40.50.60.70.80.91.01.11.21.31.41.5Mor 0.10.20.30.40.50.60.70.80.91.01.11.21.31.41.5Mor Dispersal distance Dispersal distance (km)

140 Chapter Three

Figure A4 Age dependent fecundity Number of produced eggs per female in each age class. Black dots are life table data. Black line shows the modelled function

1300

1200

1100 Number of eggs of Number

1000

900 23456 Age

Figure A5 Egg and larval survival Survival probabilities as functions of pond quality for four different egg densities.

D1 0.06 D100 D500 D1000

0.04

Survival probability 0.02

0.00

0.0 0.2 0.4 0.6 0.8 1.0 Pond quality

141 Chapter Three

Figure A6 Adult survival Age dependent adult survival probability. Black dots are life table data, black line is the modelled function.

0.6

0.5

0.4 Survival probability 0.3

0.2

0123456 Age class

142

APPENDIX

SAIA OUTPUT FILES

144 Appendix

SAIA output files

Text file with descriptive statistics on regional connectivity, abundance and population persistence probability as well as descriptive statistics on abundance and persistence probability of individual pond populations.

145 Appendix

Text file containing information on clusters and their pond members as well as connectivity within and between clusters

146 Appendix

GIS point-data set with information on mean estimated abundance and population persistence probability of the ponds. How the information is displayed is up to the user and the facilities in the chosen GIS software. Here the size of the dots represents persistence probability and the colour represents estimated population size.

147 Appendix

GIS vector-data set with information about immigration probability between ponds (connectivity network). Connections are represented as lines, and the intensity of the colour indicates the strength of the link.

148 Appendix

GIS vector-data set with information about cluster configuration. Black lines connect ponds belonging to the same cluster and cluster-ID is shown beside the clusters.

149