U N I V E R S I T Y O F C O P E N H A G E N F A C U L T Y O F S C I E N C E

PhD thesis Lars Lønsmann Iversen

Spatial distribution of Aquatic Insects

Academic advisor: Professor Kaj Sand-Jensen

Submitted: November 2016

Spatial distribution of Aquatic Insects

PhD thesis

Lars Lønsmann Iversen

Freshwater Biological Laboratory

Institute of Biology

University of Copenhagen

2

Author Lars Lønsmann Iversen

Freshwater Biological Laboratory

University of Copenhagen, Denmark

lliversen@bio.ku.dk

Supervisor Kaj Sand‐Jensen, Professor

Freshwater Biological Laboratory

University of Copenhagen, Denmark

Committee Jens Borum, Associate Professor (Chair)

Freshwater Biological Laboratory

University of Copenhagen, Denmark

Signe Normand, Associate Professor

Institute of Bioscience

University of Aarhus, Denmark

Christoffer Hassall, Associate Professor

School of Biology

University of Leeds, UK

Financial support Amphi Consult International

Danish Ministry of Higher Education and Science

(Project 0604-02230B)

3

4

Table of Contents

Summary 5

Resume 7

Acknowledgement 9

List of publications 10

Synopsis 13

Chapter 1 41

Are latitudinal richness gradients in European freshwater species

only structured according to dispersal and time?

Chapter 2 57 Variable history of land use reduces the relationship to specific

habitat requirements of a threatened aquatic insect

Chapter 3 79 Time restricted flight ability influences dispersal and colonization

rates in a group of freshwater beetles

Chapter 4 105

Community composition of diving beetles across a human altered

microhabitat gradient in a raised bog system

5

Summary

Species associated with freshwater ecosystems are currently undergoing

severe global declines and freshwater ecosystems are regarded as some of

the most endangered ecosystems in the world. These declines are a

consequence of decades of human overexploitation, pollution and climate

change. If adequate conservation actions are to be implemented, there is an

urgent need for improving our understanding of the specific habitat

requirements and the ecological niches of species in freshwater ecosystems.

This thesis explores current challenges in our understanding of how spatial

processes structure and shape the habitat requirements and distribution of one

of the most affected groups of freshwater species: aquatic insects. It

comprises four chapters each addressing different spatial factors in relation to

the occurrence of aquatic insects in Europe.

Chapter I examine two spatial ecological processes (time since glacial

disturbance and habitat stability) and question the generality of these

processes for the understanding of species richness gradients in European

rivers. Using regional distributions of European mayflies, stoneflies, and

caddisflies this chapter demonstrates that differences in latitudinal richness

patterns between headwater streams and rivers are as great as the differences

between river and lake habitats. Thus, the paper questions former spatio-

temporal hypotheses stating that the dichotomic diversity patterns between

rivers and lakes in Europe are solely created by post-glacial dispersal and

differences in species dispersal capacity between the two habitat types. The

results instead suggest that integrating effects caused by environmental

filtering or other factors would expand our understanding of the mechanisms

causing the current diversity patterns within these habitats.

Chapters II exemplifies that national changes in land use can influence the

perception of a species’ realized niche across different landscapes. By

observing the environmental niche of a threatened European dragonfly in

landscapes with different land use history it is shown that if a species’ realized

6

niche is derived from local distribution patterns, without incorporating

landscape history it can lead to an erroneous niche definition.

Chapter III provides some of the first evidence for differences in dispersal

phenology related to flight potential in aquatic insects. The chapter highlights

seasonality and resource availability as factors leading to differences in flight

potential within a group of aquatic beetles. Among the study species the

documented differences in dispersal scale to the dependency of spatial

connectivity of freshwater habitats when colonizing newly created ponds

Chapter IV documents the species composition of diving beetles across a

spatial gradient from the core to the edge of one of the most threatened

European aquatic habitats: Bog pools. The chapter identifies the key

environmental factors structuring communities along the core-edge gradient

and highlights that when undergoing environmental change, species

communities associated with open bog pool habitats are affected by a

multitude of factors. From this finding it is concluded that bog pool habitats are

affected by the same core-edge gradients and face the same conservation

challenges as other isolated unique habitats. The chapter highlights rewetting

schemes followed by tree removal as potential restoration measures for

pristine bog pool communities.

7

Resumé

Arter knyttet til ferske vande forsvinder globalt. Dette skal ses som en

konsekvens af årtier med menneskelig overudnyttelse, forurening og

nuværende klimaforandringer. Hvis tilbagegangen af biodiversiteten i ferske

vande skal stoppes effektivt er det nødvendigt at etablere en forståelse for de

primære processer og mekanismer der bestemmer og forklarer arters

tilstedeværelse i akvatiske økosystemer.

Hovedfokus i denne phd-afhandling har været at undersøge hvordan rumlige

processer påvirker vores forståelse af habitatkrav og udbredelse hos en af de

dyregrupper der påvirkes kraftigst af nuværende globale ændringer:

ferskvandsinsekter. Afhandlingen indeholder fire kapitler der hver især

beskriver hvordan rumlige faktorer påvirker tilstedeværelsen af

ferskvandsinsekter i Europa.

I kapitel I undersøges det hvordan to rumlige og tidslige processer (tid siden

sidste istid og habitat stabilitet) kan forklare regional artsrigdom i Europæiske

vandløb. Ved at bruge udbredelse og habitatpræferencer for døgnfluer,

slørvinger og vårfluer i Europa vises det at artsdiversitetsgradienten fra syd

mod nord varierer mellem små opstrøms vandløb og større floder. Disse fund

indikerer at andre faktorer, såsom habitat tilstedeværelse, sammen med post-

glacial indvandring kan have skabt de nuværende diversitets mønstre inden for

Europæiske vandløb.

Kapitel II viser hvordan observerede habitat præferencer hos en truet

guldsmed kan være et produkt af landskabets historie. Ved at kvantificere den

realiserede miljø-niche i restaurerede og tilgroede områder vises det at hvad

der ligner en generalistart faktisk er en konsekvens af en forsinket uddøen i et

hurtigt ændrende landskab og den egentlige habitat præference er langt mere

specifik.

I kapitel III præsenteres et af de første eksempler på hvordan forskelle i

årsvarierende sprednings aktivitet er relateret til et spredningspotentiale hos

ferskvandsinsekter. Ved at undersøge flugt hos en gruppe af vandkalve med

forventet variation i spredningspotentiale vises det at lav mobile arter flyver

8

inden for en ganske kort periode, lige efter forvandlingen til voksne biller.

Modsat, flyver de spredningsstærke arter på alle tider af året. Disse forskelle

kan skaleres til at beskrive arternes evne til at kolonisere nye vandhuller og

præsentere en mulig forklaring på hvorfor nogle vandkalve forsvinder fra det

europæiske landskab.

Kapitel IV forsøger at kortlægge vigtige miljøfaktorer der bestemmer

udbredelsen af vandkalve knyttet til damme i højmoser og hvilken effekt

påvirkningen fra det omkring liggende landskab har på disse. Det vises det at

arter knyttet til højmose damme forsvinder og erstattes af almindelige arter når

påvirkningen fra de omgivende arealer stiger. Kapitlet identificerer ikke kun

naturlig hydrologi men også åbne træfrie områder som betydende for sikringen

af artssamfund knyttet til højmose damme.

9

Acknowledgement

First and foremost I would like to thank Kaj Sand-Jensen for his unconditional

support, supervision, and guidance throughout my PhD. His commitment to my

projects and always encouragement to pursuit my own ideas have been

essential for the outcome of this PhD. I’m ever grateful for his support and help

when facing challenges reaching beyond the everyday problems in a PhD

student’s life.

I want to thank Lars Briggs for believing in the value of integrating research

into applied conservation consultancy, a priority fundamental to the existence

of this PhD project. I want to emphasize that my gratitude to him goes beyond

the support of this project, introducing me to the life of a field biologist 15 years

ago has most certainty shaped the ecologist I am today.

The list of people who have helped during the last three years is too long to list

and I would like to thank everybody who has contributed to the outcome of this

project. Notably I would like to thank my friends and colleagues at the section

of Freshwater Biology and Amphi Consul, for inspiring discussions and a lot of

fun over the years. Additionally, I would like to thank Riinu Rannap for

academic and practical support of my research in Estonia, and Jean-Philippe

Lessard for hosting me in Montreal and taking the time to make my stay

profitable.

Last but not least, I thank friends and family for their support and coping with me

in general. I’m deeply indebted to the support and love I receive on an everyday

basis from Mette and Bille. Without you, life would be different and I value your

presence every single day.

10

List of publications

Publications and manuscripts included in the thesis: • Iversen L.L., Jacobsen, D. & Sand-Jensen K. (2016): Are latitudinal

richness gradients in European freshwater species only structured

according to dispersal and time? Ecography DOI: 10.1111/ecog.02183

• Iversen L.L., Rannap R., Briggs L. & Sand-Jensen K. (2016): Variable

history of land use reduces the relationship to specific habitat

requirements of a threatened aquatic insect. Population Ecology DOI:

10.1007/s10144-015-0516-z

• Iversen L.L., Rannap R., Briggs L. & Sand-Jensen K.: Time restricted

flight ability influences dispersal and colonization rates in a group of

freshwater beetles. In review in Ecology and Evolution.

• Iversen L.L., Kristensen E., Briggs L., Edvardsson J., Jarašius L.,

Sendžikaitė J., & Sand-Jensen K.: Community composition of diving

beetles across a human altered microhabitat gradient in a raised bog

system. In review in Hydrobiologia.

Additional publications published but not included in the thesis:

• Iversen L.L., Kielgast J. & Sand-Jensen K. (2015): Monitoring of

animal abundance by environmental DNA - an increasingly obscure

perspective: A reply to Klymus et al., 2015. Biological Conservation

DOI: 10.1016/j.biocon.2015.09.024

• Rannap R., Kaart T., Iversen L.L., Briggs L. & Vries W. (2015):

Geographically Varying Habitat Characteristics of a Wide-ranging

Amphibian, the Common Spadefoot Toad (Pelobates fuscus), in

Northern Europe. Herpetological Conservation and Biology 10(3):904–

916

• Baastrup-Spohr L., Iversen L. L., Borum J. & Sand-Jensen K. (2015):

Niche specialization and functional traits regulate the rarity of

11

charophytes in the Nordic countries. Aquatic Conservation: Marine and

Freshwater Ecosystems. DOI: 10.1002/aqc.2544

• Iversen L. L., Rannap R., Kielgast J., Thomsen P.F. & Sand-Jensen K.

(2013): How do low dispersal species establish large range sizes? – the

case of the water beetle Graphoderus bilineatus. Ecography 36: 770–

777

• Baastrup-Spohr L., Iversen, L.L., Dahl-Nielsen J. & Sand-Jensen K.

(2013): Seventy years of changes in the abundance of Danish

charophytes. Freshwater Biology 58: 1682–1693

12

13

Synopsis

14

Introduction

The duality of Niches Inferring species environmental requirements from their spatial distribution is

one of the most important attempts to understand the factors influencing the

presence of species (Gaston 2003, Peterson 2011). The conceptual idea, that

species distribution reflects the environment in which they can exist, originates

from the theory of ecological niches (Soberón and Nakamura 2009). Building

on the work of Joseph Grinnell and Charles Elton (Wake et al. 2009), G.

Evelyn Hutchinson defined the ecological niche as the N-dimensional

environmental hyperspace in which a species can survive and maintain a

positive rate of population growth (Hutchinson 1957, Hutchinson 1978). The

concept reflects a connection between multidimensional niche space and

biotope space of species (Colwell and Rangel 2009). The connection enables

inference of species niches based on their spatial distribution and vice versa,

creating the foundation of species distribution models and contemporary

biogeography research (Guisan and Thuiller 2005, Hirzel and Le Lay 2008).

Hutchinson also introduced the distinction between the fundamental niche,

reflecting the abiotic requirements of a species, and the realized niche, i.e. the

set of observed conditions occupied by the species in the presence of other

species acting as grazers, predators, competitors and facilitators (Hutchinson

1978). Given that the spatial distribution of a species is the product of biotic

and abiotic interactions (Wisz et al. 2013), potential problem arises when

inferring abiotic requirements from species distribution only (Araujo and

Guisan 2006). Unoccupied space of a species might not be a consequence of

unsuitable local (abiotic) conditions but simply reflect that interactions with

other species prevent the focal species from being present (Soberón and

Nakamura 2009). Even though biotic interactions had no effect on a species

occurrence the duality between niche and biotic space might still be obscured

by spatial disequilibrium processes (Jackson and Sax 2010, Svenning et al.

2015). Following environmental change, dispersal limitation among isolated

habitats can restrict the geographical range of a species, leaving areas that

15

correspond to parts of the species’ fundamental niche unoccupied (Devictor et

al. 2012, Svenning and Sandel 2013). The potential effect of disequilibrium

states on present-day patterns of biodiversity has been acknowledge for more

than a decade (Pearson and Dawson 2003, Wiens and Donoghue 2004).

However understanding the effect of historical legacies and taking them into

account when defining species’ ecological niches from their spatial

distributions still remain one of the main challenges in modern ecological

research (Sutherland et al. 2013, Svenning et al. 2015, Perring et al. 2016).

The state of aquatic insects Species associated with freshwater ecosystems are currently undergoing

severe global declines and freshwater ecosystems are regarded as some of

most endangered ecosystems in the world (Jenkins 2003, Dudgeon et al.

2006). The declines of freshwater species is forecasted to continue at an even

faster pace than their terrestrial and marine counterparts (Ricciardi and

Rasmussen 1999, Sala et al. 2000). The reasons for the declines have been

linked to global warming, multiple anthropogenic stressors and the interaction

between the two (Heino et al. 2009, Ormerod et al. 2010). It has been

highlighted that freshwater species should be especially susceptible to global

warming because: 1) the fragmented nature of freshwaters creates limited

possibilities to disperse as the environment changes; 2) availability of aquatic

habitats and water temperature itself are climate-dependent; and 3), many

aquatic ecosystems are already exposed to many anthropogenic stressors

enhancing the effect of global warming (Sala et al. 2000, Woodward et al.

2010).

Several studies on aquatic insects have shown the negative effects of

contemporary and historic global changes. Habitat destruction and degradation

are often seen as the major cause of decline or change in species composition

of aquatic insects (Richter et al. 1997, Strayer 2006). Especially in the

temperate regions changes of land use (i.e. nature to cultivated areas and

extensive farming to intensive farming) have led to loss of natural freshwaters

habitats and associated species (Naiman and Decamps 1997, Brinson and

Malvárez 2002). Also, species richness decreases with changes in habitat

16

conditions following pollution (e.g. by increased nutrient loadings from

surrounding farmland) (Harding et al. 1998, Azevedo et al. 2013). Climate

change are expected to enhance the negative effect of intensive land use and

to structure macroinvertebrate communities by changing local conditions of

temperature, oxygen and water flow (Daufresne et al. 2004, Durance and

Ormerod 2007, Chessman 2009). In Europe climate change has led to a

reduction in suitable habitats for cold-tolerant species in the northern or high

altitude regions and warm-tolerant species in southern regions (Bonada et al.

2009, Bálint et al. 2011).

This ongoing decline of aquatic insects and invertebrates in general has led to

a pledge for an increase in conservation actions targeting these organisms

(Strayer 2006, Cardoso 2012). Specific actions are needed for aquatic insects

since the threats of and response to global change are highly group specific

(Heino et al. 2009, Schuldt and Assmann 2010) and invertebrates are usually

not supported by popular conservation actions aimed at birds and mammals

(Westgate et al. 2014, Guareschi et al. 2015).

As a consequence several species of dragonflies and two diving beetles have

been included in the Habitats Directive and associated NATURA 2000

protection areas (Council of the European Union 1992). Together with the Bird

Directive, the Habitats Directive represents the legislation regarding Europe’s

nature conservation policy. However, the funding allocated to invertebrate

conservation within this programme is lacking behind the vertebrate priorities

(Cardoso et al. 2011). The lack of ecological knowledge and habitat

preferences of the different species included in the Habitats Directive has been

stressed as one of the main causes of the slow implementation of insects

species in the Directive (Cardoso et al. 2011, Cardoso 2012). Thus, there is an

urgent need for improving our understanding of specific habitat requirements

and the ecological niches of aquatic insects in order to implement conservation

actions (Cardoso et al. 2011) and attain favorable conservation status of the

aquatic insects in the Habitats Directive (Mehtälä and Vuorisalo 2007).

17

Study objectives The overall objective of my PhD thesis is to contribute to a better

understanding of how spatial processes influence the distribution and

presence of species in aquatic ecosystems. The four chapters of the thesis

aim at exploring current patterns of species distributions and species richness

in relation to habitat characteristics and spatial factors. In particular, my goal

was to:

• explore the potential impacts of disequilibrium states on species richness

gradients and ecological niches (chapters I and II);

• explore the importance of flight plasticity in species occupying isolated and

fragmented habitats (chapter III)

• identify the processes of community change along periphery-to-core

gradients in aquatic habitats (chapter IV)

Disequilibrium states and biodiversity patterns The distribution of species is a product of local conditions in the past and today

(Svenning et al. 2015). Given that species’ responses to environmental

changes may not be instantaneous or complete, local community composition

may not accurately reflect present day abiotic conditions (Svenning and

Sandel 2013). This environmental disequilibrium, the presence of communities

of species with niches less close to the local observed conditions, influences

our perception of the filtering role of the environment in present biodiversity

patterns and how species respond to global change (Blonder et al. 2015,

Svenning et al. 2015). Disequilibrium states are driven by the processes of

delayed extinction and colonization credit (Jackson and Sax 2010). Delayed

extinction is when one or more extinctions caused by a specific event do not

coincide precisely in time with that event, but instead follows it after a distinct

delay (Kuussaari et al. 2009). An example is, when single standing trees or

small tree parcels persist even though the environmental conditions are not

suited for the species anymore (Vellend et al. 2006) or when plants maintain a

population for a period even though the local climate has become unsuitable

due to climate change (Dullinger et al. 2012). Colonization credit is the time lag

18

from when suitable environmental conditions arises at a site and until the

species fitted to this environment colonize the given site (Jackson and Sax

2010). This lag period vary in space and time given the spatial extent of former

negative or present positive events, biotic interactions and species dispersal

abilities (Bakker and Berendse 1999, Svenning et al. 2015).

Figure 1 demonstrates how disequilibrium states might influence our

perception on species community responses to present day global change. In

the example there is a constant rate of global change (e.g. change in climatic

conditions or land use) causing a linear change in the expected community

given the local abiotic conditions at a specific time. However, the species in the

community has a delayed response to the changing environment and sustain

populations even though the local conditions might not be suitable anymore.

Ad a certain threshold the initial species community collapses and changes

towards a new set of species reflecting the present environmental conditions. If

this collapse happens within a study period of interest the observed community

change will only partly be a consequence of the global change within this

period. A proportion of the observed community change will be a product of

historical events (the green bar in Fig 1), partly independent of the present day

changes. These disequilibrium states challenge our understanding of the

mechanisms casing observed changes in species communities (Hylander and

Ehrlén 2013) and our predictions of future ecosystem patterns (Perring et al.

2016). It should be noticed that the global change response and the

disequilibrium response in Figure 1 aren’t necessarily independent processes

as there might be an additive effects or even modification of the two (Harpole

et al. 2016).

19

Figure 1: Hypothetical example of a community showing a delayed response to a constant

global change. In the period of interest the observed community change is caused partly by

the global change and partly by the disequilibrium state of the community (the delayed

response to historical global change).

The interplay between dispersal, ecological niche and local distribution

The spatial occurrence and temporal persistence of species are closely linked

to dispersal ability and dispersal strategy (Bowler and Benton 2005, Kubisch et

al. 2014). Interacting with large scale disturbances (such as the Quatennary

glaciations) dispersal properties can shape the ranges of species and

determine continental diversity gradients (Svenning and Skov 2004, Geber

2011, Baselga et al. 2012). Abiotic dispersal barriers are often considered to

be the main evolutionary drivers of different dispersal strategies (Kubisch et al.

2014). Usually, environmental disturbance increases dispersal provided that

suitable habitats are evenly distributed (Gadgil 1971, Levin et al. 1984,

Poethke et al. 2003). This pattern can, however, be counterweighted in low

20

dispersive species by an optimized and adaptive utilization of their natal

conditions (Mathias et al. 2001, Kubisch et al. 2014). As a consequence

species with regional distinct distributions will have a strong link between

dispersal capacity and their fundamental niches.

Figure 2 demonstrates the relationship between dispersal, local distribution

and the niche in habitats undergoing global changes. In the example four

spatially separated habitats are undergoing a period of succession from open

to closed,dense vegetation cover, but the number of occupied sites remains

constant across the entire period. Species with full dispersal capacity can track

the environment they are adapted to, by dispersing to natal sites as they

become suitable. As such the niche remains stable and for both the

fundamental and realized niche conservatism is observed (Wiens and Graham

2005). It should be noted that niche conservatism is not a prerequisite for high

dispersal abilities in species, but if species show adaptive or broad niches

together with full dispersal capacity they would not show local restricted

distributions as depicted in Figure 2 (Kotze and O'hara 2003, Henle et al.

2004). Species with limited or no dispersal capacity cannot track the suitable

habitat by dispersing to neighboring patches (Figure 2). As such the realized

niche has to adapt to the changes in local conditions (in this case vegetation

cover) in order to prevent local extinction (Ackerly 2003). These changes in

niches can be a consequence of niche adaptation where the fundamental

niche does change in a changing environment (Holt 2009).

Niche adaptation has been suggested to differentiate populations of species in

their niche attributes both at continental scales (Rehfeldt et al. 1999) and

within local communities (Logan et al. 2014, Logan et al. 2016). An alternative

to niche adaptions would be the presence of niche vacancy. Some regions of

the fundamental niche may not correspond to any contemporary biotope and

may as such not be visible in the realized niche at a given time slice (Colwell

and Rangel 2009). Thus, the fundamental niche of the species are constant

but only a fraction is utilized in the realized part of the niche allowing the

species to sustain a population as local habitat conditions change (Williams et

al. 2007, Sax et al. 2013)

21

Figure 2: The relationship between dispersal, local distribution and the niche. In panel A and

B there is a period of succession from open (light green) to dense (dark green) vegetation in

four spatially separated patches. The hatched patches are occupied by a species with strong

dispersal capacity (A) and limited dispersal capacity (B) at a given time slices. The occupied

space among the four patches corresponds to a niche space of suitable temperatures for the

given species at a given time point (the realized niches are marked in dark red and the

fundamental niche in light red). With full dispersal capacity the species can track the change in

habitat conditions by colonizing new patches and as such sustain a constant niche (lower left

box). If tracking habitat changes are prohibited by insufficient dispersal capacity a species can

only sustain a population by adapting its thermal niche to the increase in vegetation cover

(lower central box) or by having a broad fundamental niche potentially suitable for a range of

different vegetation covers (lower right box)

Edge effect in isolated habitats In isolated habitats there is a change in environmental characteristics and

associated community composition from the core to the border of the habitat

(Ries et al. 2004, Ewers and Didham 2008). This gradient is caused by

22

“environmental leaking” in which the environmental conditions of the

surrounding landscape (often phrased as the matrix) penetrate into the habitat

(Ewers and Didham 2006). This edge effect depends on the contrast between

the habitat and the surrounding landscape and varies between different habitat

types (Murcia 1995). However independent of habitat type an increasing edge

effect following fragmentation or habitat degradations leads to an increase in

species turnover (Haddad et al. 2015).

From the core to the edge of a habitat the change in species communities

often leads to a peak in species richness at the edge of a habitat (Ewers and

Didham 2006). This peak is seen as a consequence of a mixing between

fragment and matrix species, giving rise to a zone of overlap of species giving

rise to an overall higher species richness (Magura 2002, Ewers and Didham

2008). However, it has been highlighted that different underlying mechanism

might structure the core to edge richness gradient (Ries et al. 2004). It has

been demonstrated that fragment specialist species decrease with an increase

in edge effect (Davies et al. 2001, Barbosa and Marquet 2002) or that matrix

species only partly penetrate into the habitat edge (Ewers and Didham 2007)

creating either decreasing or stable core to edge richness gradients. It has

been argued that when the contrast between a habitat and the surrounding

landscape becomes greater it has a proportional larger impact on specialist

versus generalist species (Ewers and Didham 2006). As such there is an

interaction between habitat-matrix contrast and the edge effect given that

edges with a high contrast between the habitat and matrix are more likely to

generate stronger edge effects than low-contrast edges (Franklin 1993,

Laurance et al. 2011).

23

Chapter synopsis

Are latitudinal richness gradients in European freshwater

species only structured according to dispersal and time? (Chapter I) The latitudinal gradient in species richness of the European freshwater fauna

is usually depicted as a contrasting pattern between lentic (standing waters)

and lotic (running waters) habitats. Species richness in lotic habitats is

suggested to decrease linearly with higher latitude, while richness in lentic

habitats has an intermediate peak in central Europe (Hof et al. 2008). It has

been proposed that the interaction between colonization ability and time

elapsed since the last major glaciation has been responsible for the

differences in latitudinal richness gradients in lentic and lotic habitats in Europe

(Ribera et al. 2003, Hof et al. 2006, Marten et al. 2006).

We argue that this hypothess to a certain extend ignores the sorting effect of

environmental conditions which are so obviously structuring species

communities at local scales (Vannote et al. 1980). Using the distribution and

habitat preference of the European fauna of the three insect orders:

Ephemeroptera, Plecoptera, and Trichoptera (EPT) across 22 biogeographic

regions in Europe we demonstrate that:

(1). Regional species richness can be explained solely from the hypervolume

defined by the variety of lotic habitats preferences in a region (Figure 3). This

finding suggests that higher species richness is related to regions with a broad

spectrum of habitats and not to the spatial location of a given region.

(2). EPT-species richness within lotic zonation classes showed a change in the

relationship to latitude from headwaters to lower reaches. In headwater

classes, richness decreased linearly with an increase in latitude. But in the

lower reaches the river zonation classes showed a unimodal richness pattern

with latitude (resembling lentic habitats)

In summary this chapter questions the strict dichotomic split in latitudinal

richness patterns between lentic and lotic habitats in Europe. We do this by

demonstrating that the differences in richness patterns between headwater

24

streams and rivers are as great as the differences between lentic and lotic

habitats.

Figure 3: Parameter estimates and 95 % confidence intervals of the selected geographical

and habitat niche variables. The estimates are derived from an average model of four possible

models explaining the species richness of European EPT fauna. Redrawn from Iversen et al.

2016a.

Variable history of land use reduces the relationship to

specific habitat requirements of a threatened aquatic insect (Chapter II) A common practice in conservation science is use of the local realized niche to

define important local habitat characteristics for a species of conservation

interest (Margules and Pressey 2000). However, this can be problematic if

local populations are in disequilibrium with the local environment (e.g. if local

occurrence is solely a consequence of former land use practices (Harding et

al. 1998, Cramer et al. 2008). Furthermore, the environmental niche of the

species might vary between regions (Oliver et al. 2009), challenging the

potential use of the observed niches of a species across regions.

25

In this chapter we explored whether national changes in land use influence the

realized niche of a threatened dragonfly in a significant manner across

different landscapes in a given region. Using contrasting landscape history (25

years of national afforestation and areas in which former semi-open farmland

had been recreated), we recorded the presence/absence of the white-faced

darter Leucorrhinia pectoralis and measured seven habitat variables for 140

lakes and ponds located in one restored and three un-restored landscapes in

Estonia. The results suggested that lake size and proportion of short riparian

vegetation were significantly positive parameters determining the presence of

L. pectoralis across landscape types. However, the species was much more

habitat specific in the restored landscape, with larger influence of other habitat

parameters (Figure 4). Our data suggests that the realized niche of the species

in the un-restored landscapes was constrained by the present-day habitats.

The chapter demonstrates that if a species’ realized niche is derived from local

species distributions without incorporating landscape history it can lead to an

erroneous niche definition. We show that landscape restorations can provide

knowledge on the species’ habitat dependencies before habitat degradations

occurred, provided that restoration mitigation reflects the former landscape

characteristics

26

Figure 4: Habitat characteristics associated with the presence of the dragonfly Leucorrhinia

pectoralis. The bars correspond to the independent contributions (percentage of the total

independent explained variance) of habitat predictor variables of the restored landscape (white

bars) and un-restored landscapes (black bars). Redrawn from Iversen et al. 2016b

Time restricted flight ability influences dispersal and

colonization rates in a group of freshwater beetles (chapter II) The ecology and seasonal variability of flight activity, and subsequently

dispersal of individuals, arekey-elements for understanding both current and

future patterns of species distributions (Clobert et al. 2004, Bowler and Benton

2005)

In this chapter we examine how non-morphological flight polymorphism might

vary across seasons and between two closely related genera of freshwater

beetles with similar geographical ranges, life histories and dispersal related

morphology.

By combining flight experiments of more than 1,100 specimens with

colonization rates in a meta-community of 54 ponds in Southern Estonia, we

analyzed the relationship between flight potential and spatio-environmental

distribution of the study genera.

27

We document profound differences in flight potential between the two genera

across seasons. High flight potential for Acilius (97%) and low for Graphoderus

(14%) corresponded to the different colonization rates of newly created ponds.

While Acilius dispersed throughout the season, flight activity in Graphoderus

was restricted to stressed situations immediately after the emergence of

adults. Regional colonization ability of Acilius was independent of spatial

connectivity and mass effect from propagule sources. In contrast,

Graphoderus species were closely related to high connectivity between ponds

in the landscape (Figure 5).

The results from this chapter suggest that different dispersal potential can

account for different local occurrences of Acilius and Graphoderus. In general

our findings provide some of the first insights into the understanding of

seasonal restrictions in flight patterns of aquatic beetles and their

consequences for species distributions.

Figure 5: Probability of presence in 54 newly created Estonian ponds of Graphoderus (blue)

and Acilius (green) in relation to landscape proximity index. The central tendency is shown by

the solid line and 95 % confidence limits are shaded.

28

Community composition of diving beetles across a human-

altered microhabitat gradient in a raised bog system (chapter IV) Ombrotrophic bogs are ecosystems with a hydrology regime decoupled from

the surrounding landscape creating unique environmental conditions

compared to the surroundings (Van Breemen 1995, Renou-Wilson et al.

2011). Given this landscape contrast it is expected that ombrotrophic bogs will

experience a strong edge effect along the perimeter of the bog habitats (Rydin

and Jeglum 2013). In Europe this effect are enhanced by the isolated nature of

ombrotrophic bogs and human induced habitat degradation (Ewers and

Didham 2006, Tscharntke et al. 2012). Within microhabitats in bog systems a

reported response to increasing edge effect is a decrease in bog specialist

species and an increase in total species abundance and richness (Verberk et

al. 2006, Flynn et al. 2016). However, the underlying processes of changes in

community compositions along these core-edge gradients in ombrotrophic

bogs still remain greatly understudied. In this chapter we report on a field

campaign documenting the relative role of general and bog specific

environmental factors in structuring diving beetle communities in 50

ombrotrophic bog pools. This study was performed by quantifying changes in

abundance within two groups of diving beetles representing bog pool and

forest/matrix species, respectively.

We show that the community turnover reflects the expected core-edge

patterns. Bog pool species decreased and edge-species increased in

abundance towards the edge of the bog system. However, the environmental

variables causing these changes in abundance differed between the two study

groups. Edge-species were primarily structured by environmental variables

associated to general habitat changes, not specific to the study system. In

contrasty, core-species were structured by environmental variables associated

to both general and ombrotrophic bog specific habitat changes.

29

Even though these conclusions are drawn from a single bog system we show

that when undergoing environmental change species communities associated

to ombrotrophic bog pools are affected by a multitude of factors leading to a

graduate decrease in abundance and loss of species.

Conclusions and Outlook

Collectively the work presented in the thesis highlight the impact off spatial

factors when understanding the distribution and ecological niches of aquatic

insects. The work supports the notion that the fragmented nature of aquatic

habitats enhances environmental disequilibrium states in the species

communities associated to these habitats. The thesis documents that

environmental history affects current species occurrences both at local and

continental scales. The strong link between dispersal ecology and spatial

connectivity shown in chapter III underlines the challenges aquatic insects’

faces when tracking environmental and climatic changes.

The results obtained from the thesis can be followed up by further work

addressing newly generated questions or outstanding conservation issues. For

the individual chapters, the following next steps are proposed.

Chapter I: Integrating detailed environmental factors and not just dispersal

ability when attempting to predict freshwater species ranges might expand our

understanding of the mechanisms structuring the current diversity patterns

within these habitats. Furthermore, the structuring effect of the evolutionary

constraints between dispersal and environmental preference when colonizing

post-glaciated areas still remains largely unknown. Such constraints could be

addressed by studying the dispersion of relevant functional traits of EPT

species across the latitudinal gradient in Europe.

Chapter II: The importance of incorporating landscape history when extracting

relevant habitat components from local distributions has significant

conservation implications. Restoration recommendations and management

strategies might be significantly flawed if build on the census of habitat

30

dependencies from single populations not integrating the potential effect of

land use legacies. Disentangling the actual environmental preferences from

random associations to local habitat attributes is essential if adequate and

effective restorations are to be conducted. Given the huge amount of funds

allocated into conservation actions of the species included in the EU habitat-

directive (such as the dragonfly studied in chapter II) robust estimates of

habitat components essential for these species are of currently needed.

Chapter III: Further expanding the landscape component presented in this

chapter might elucidate drivers creating the observed differences in flight

potential. Landscape scale studies on the distribution of the studied species

might expose dispersal related dependencies of habitat stability and resource

utilization. From site specific abundance data and data on potential flight

distances two regulating mechanisms could be studied for species with

seasonal dispersal constraints differences

Chapter IV: The impact of rewetting schemes and deforestation as possible

mitigation measures for climate change and land degradation effects in bog

pools should be explicitly tested. In a counterfactual study design the response

of bog pool communities to combined rewetting and tree removal actions

should be quantified. It should also be ensured that the restoration actions

have a dual effect on both species communities and the environmental

conditions. Thus, besides changes in species assemblages the ecological

response in bog pools following the proposed conservation actions should be

monitored.

31

References

• Ackerly, D. D. 2003. Community assembly, niche conservatism, and

adaptive evolution in changing environments. International Journal of Plant

Sciences 164:S165-S184.

• Araujo, M. B., and A. Guisan. 2006. Five (or so) challenges for species

distribution modelling. Journal of Biogeography 33:1677-1688.

• Azevedo, L. B., R. van Zelm, P. M. F. Elshout, A. J. Hendriks, R. S. E. W.

Leuven, J. Struijs, D. de Zwart, and M. A. J. Huijbregts. 2013. Species

richness-phosphorus relationships for lakes and streams worldwide.

Global Ecology and Biogeography 22:1304-1314.

• Bakker, J. P., and F. Berendse. 1999. Constraints in the restoration of

ecological diversity in grassland and heathland communities. Trends in

Ecology & Evolution 14:63-68.

• Bálint, M., S. Domisch, C. Engelhardt, P. Haase, S. Lehrian, J. Sauer, K.

Theissinger, S. Pauls, and C. Nowak. 2011. Cryptic biodiversity loss linked

to global climate change. Nature Climate Change 1:313-318.

• Barbosa, O., and P. A. Marquet. 2002. Effects of forest fragmentation on

the beetle assemblage at the relict forest of Fray Jorge, Chile. Oecologia

132:296-306.

• Baselga, A., J. M. Lobo, J.-C. Svenning, P. Aragón, and M. B. Araújo.

2012. Dispersal ability modulates the strength of the latitudinal richness

gradient in European beetles. Global Ecology and Biogeography 21:1106-

1113.

• Blonder, B., D. Nogués-Bravo, M. K. Borregaard, I. Donoghue, C. John, P.

M. Jørgensen, N. J. Kraft, J.-P. Lessard, N. Morueta-Holme, and B.

Sandel. 2015. Linking environmental filtering and disequilibrium to

biogeography with a community climate framework. Ecology 96:972-985.

• Bonada, N., C. Múrria, C. Zamora‐Muñoz, M. El Alami, J. M. Poquet, T.

Puntí, J. L. Moreno, N. Bennas, J. Alba‐Tercedor, and C. Ribera. 2009.

Using community and population approaches to understand how

32

contemporary and historical factors have shaped species distribution in

river ecosystems. Global Ecology and Biogeography 18:202-213.

• Bowler, D. E., and T. G. Benton. 2005. Causes and consequences of

animal dispersal strategies: relating individual behaviour to spatial

dynamics. Biological Reviews 80:205-225.

• Brinson, M. M., and A. I. Malvárez. 2002. Temperate freshwater wetlands:

types, status, and threats. Environmental conservation 29:115-133.

• Cardoso, P. 2012. Habitats Directive species lists: urgent need of revision.

Insect Conservation and Diversity 5:169-174.

• Cardoso, P., T. L. Erwin, P. A. Borges, and T. R. New. 2011. The seven

impediments in invertebrate conservation and how to overcome them.

Biological Conservation 144:2647-2655.

• Chessman, B. C. 2009. Climatic changes and 13‐year trends in stream

macroinvertebrate assemblages in New South Wales, Australia. Global

Change Biology 15:2791-2802.

• Clobert, J., R. Ims, and F. Rousset. 2004. Causes, mechanisms and

consequences of dispersal. Ecology, genetics, and evolution of

metapopulations:307-335.

• Colwell, R. K., and T. F. Rangel. 2009. Hutchinson's duality: the once and

future niche. Proc Natl Acad Sci U S A 106 Suppl 2:19651-19658.

• Cramer, V. A., R. J. Hobbs, and R. J. Standish. 2008. What's new about

old fields? Land abandonment and ecosystem assembly. Trends in

Ecology & Evolution 23:104-112.

• Daufresne, M., M. Roger, H. Capra, and N. Lamouroux. 2004. Long‐term

changes within the invertebrate and fish communities of the Upper Rhône

River: effects of climatic factors. Global Change Biology 10:124-140.

• Davies, K. F., B. A. Melbourne, and C. R. Margules. 2001. Effects of

within‐and between‐patch processes on community dynamics in a

fragmentation experiment. Ecology 82:1830-1846.

• Devictor, V., C. Van Swaay, T. Brereton, D. Chamberlain, J. Heliölä, S.

Herrando, R. Julliard, M. Kuussaari, Å. Lindström, and D. B. Roy. 2012.

Differences in the climatic debts of birds and butterflies at a continental

scale. Nature Climate Change 2:121-124.

33

• Dudgeon, D., A. H. Arthington, M. O. Gessner, Z. Kawabata, D. J.

Knowler, C. Leveque, R. J. Naiman, A. H. Prieur-Richard, D. Soto, M. L.

Stiassny, and C. A. Sullivan. 2006. Freshwater biodiversity: importance,

threats, status and conservation challenges. Biol Rev Camb Philos Soc

81:163-182.

• Dullinger, S., A. Gattringer, W. Thuiller, D. Moser, N. E. Zimmermann, A.

Guisan, W. Willner, C. Plutzar, M. Leitner, and T. Mang. 2012. Extinction

debt of high-mountain plants under twenty-first-century climate change.

Nature Climate Change 2:619-622.

• Durance, I., and S. J. Ormerod. 2007. Climate change effects on upland

stream macroinvertebrates over a 25-year period. Global Change Biology

13:942-957.

• Ewers, R. M., and R. K. Didham. 2006. Confounding factors in the

detection of species responses to habitat fragmentation. Biol Rev Camb

Philos Soc 81:117-142.

• Ewers, R. M., and R. K. Didham. 2007. The effect of fragment shape and

species' sensitivity to habitat edges on animal population size.

Conservation Biology 21:926-936.

• Ewers, R. M., and R. K. Didham. 2008. Pervasive impact of large-scale

edge effects on a beetle community. Proceedings of the National Academy

of Sciences 105:5426-5429.

• Flynn, C., C. T. Griffin, J. Coll, and C. D. Williams. 2016. The diversity and

composition of moth assemblages of protected and degraded raised bogs

in Ireland. Insect Conservation and Diversity.

• Franklin, J. F. 1993. Preserving biodiversity: species, ecosystems, or

landscapes? Ecological Applications 3:202-205.

• Gadgil, M. 1971. Dispersal: population consequences and evolution.

Ecology 52:253-261.

• Gaston, K. J. 2003. The structure and dynamics of geographic ranges.

Oxford University Press on Demand.

• Geber, M. A. 2011. Ecological and evolutionary limits to species

geographic ranges. Am Nat 178 Suppl 1:S1-5.

34

• Guareschi, S., D. Bilton, J. Velasco, A. Millán, and P. Abellán. 2015. How

well do protected area networks support taxonomic and functional diversity

in non-target taxa? The case of Iberian freshwaters. Biological

Conservation 187:134-144.

• Guisan, A., and W. Thuiller. 2005. Predicting species distribution: offering

more than simple habitat models. Ecology Letters 8:993-1009.

• Haddad, N. M., L. A. Brudvig, J. Clobert, K. F. Davies, A. Gonzalez, R. D.

Holt, T. E. Lovejoy, J. O. Sexton, M. P. Austin, and C. D. Collins. 2015.

Habitat fragmentation and its lasting impact on Earth’s ecosystems.

Science Advances 1:e1500052.

• Harding, J., E. Benfield, P. Bolstad, G. Helfman, and E. Jones. 1998.

Stream biodiversity: the ghost of land use past. Proceedings of the

National Academy of Sciences 95:14843-14847.

• Harpole, W. S., L. L. Sullivan, E. M. Lind, J. Firn, P. B. Adler, E. T. Borer,

J. Chase, P. A. Fay, Y. Hautier, and H. Hillebrand. 2016. Addition of

multiple limiting resources reduces grassland diversity. Nature 537:93-96.

• Heino, J., R. Virkkala, and H. Toivonen. 2009. Climate change and

freshwater biodiversity: detected patterns, future trends and adaptations in

northern regions. Biological Reviews 84:39-54.

• Henle, K., K. F. Davies, M. Kleyer, C. Margules, and J. Settele. 2004.

Predictors of species sensitivity to fragmentation. Biodiversity &

Conservation 13:207-251.

• Hirzel, A. H., and G. Le Lay. 2008. Habitat suitability modelling and niche

theory. Journal of Applied Ecology 45:1372-1381.

• Hof, C., M. Brändle, and R. Brandl. 2006. Lentic odonates have larger and

more northern ranges than lotic species. Journal of Biogeography 33:63-

70.

• Hof, C., M. Brändle, and R. Brandl. 2008. Latitudinal variation of diversity

in European freshwater animals is not concordant across habitat types.

Global Ecology and Biogeography 17:539-546.

• Holt, R. D. 2009. Bringing the Hutchinsonian niche into the 21st century:

ecological and evolutionary perspectives. Proceedings of the National

Academy of Sciences 106:19659-19665.

35

• Hutchinson, G. E. 1957. Cold spring harbor symposium on quantitative

biology. Concluding remarks 22:415-427.

• Hutchinson, G. E. 1978. An introduction to population ecology.

• Hylander, K., and J. Ehrlén. 2013. The mechanisms causing extinction

debts. Trends in Ecology & Evolution 28:341-346.

• Iversen, L. L., D. Jacobsen, and K. Sand-Jensen. 2016a. Are latitudinal

richness gradients in European freshwater species only structured

according to dispersal and time? Ecography 39:001-003.

• Iversen, L. L., R. Rannap, L. Briggs, and K. Sand-Jensen. 2016b. Variable

history of land use reduces the relationship to specific habitat

requirements of a threatened aquatic insect. Population Ecology 58:155-

164.

• Jackson, S. T., and D. F. Sax. 2010. Balancing biodiversity in a changing

environment: extinction debt, immigration credit and species turnover.

Trends in Ecology & Evolution 25:153-160.

• Jenkins, M. 2003. Prospects for biodiversity. Science 302:1175-1177.

• Kotze, D. J., and R. B. O'hara. 2003. Species decline—but why?

Explanations of carabid beetle (Coleoptera, Carabidae) declines in

Europe. Oecologia 135:138-148.

• Kubisch, A., R. D. Holt, H.-J. Poethke, and E. A. Fronhofer. 2014. Where

am I and why? Synthesizing range biology and the eco-evolutionary

dynamics of dispersal. Oikos 123:5-22.

• Kuussaari, M., R. Bommarco, R. K. Heikkinen, A. Helm, J. Krauss, R.

Lindborg, E. Öckinger, M. Pärtel, J. Pino, and F. Roda. 2009. Extinction

debt: a challenge for biodiversity conservation. Trends in Ecology &

Evolution 24:564-571.

• Laurance, W. F., J. L. Camargo, R. C. Luizão, S. G. Laurance, S. L. Pimm,

E. M. Bruna, P. C. Stouffer, G. B. Williamson, J. Benítez-Malvido, and H.

L. Vasconcelos. 2011. The fate of Amazonian forest fragments: a 32-year

investigation. Biological Conservation 144:56-67.

• Levin, S. A., D. Cohen, and A. Hastings. 1984. Dispersal strategies in

patchy environments. Theoretical population biology 26:165-191.

36

• Logan, M. L., R. M. Cox, and R. Calsbeek. 2014. Natural selection on

thermal performance in a novel thermal environment. Proceedings of the

National Academy of Sciences 111:14165-14169.

• Logan, M. L., M. Duryea, O. R. Molnar, B. J. Kessler, and R. Calsbeek.

2016. Spatial variation in climate mediates gene flow across an island

archipelago. Evolution.

• Magura, T. 2002. Carabids and forest edge: spatial pattern and edge

effect. Forest ecology and management 157:23-37.

• Margules, C. R., and R. L. Pressey. 2000. Systematic conservation

planning. Nature 405:243-253.

• Marten, A., M. Brandle, and R. Brandl. 2006. Habitat type predicts genetic

population differentiation in freshwater invertebrates. Mol Ecol 15:2643-

2651.

• Mathias, A., È. Kisdi, and I. Olivieri. 2001. Divergent evolution of dispersal

in a heterogeneous landscape. Evolution 55:246-259.

• Mehtälä, J., and T. Vuorisalo. 2007. Conservation policy and the EU

Habitats Directive: favourable conservation status as a measure of

conservation success. European Environment 17:363-375.

• Murcia, C. 1995. Edge effects in fragmented forests: implications for

conservation. Trends in Ecology & Evolution 10:58-62.

• Naiman, R. J., and H. Decamps. 1997. The ecology of interfaces: riparian

zones. Annual review of ecology and systematics:621-658.

• Oliver, T., J. K. Hill, C. D. Thomas, T. Brereton, and D. B. Roy. 2009.

Changes in habitat specificity of species at their climatic range boundaries.

Ecology Letters 12:1091-1102.

• Ormerod, S., M. Dobson, A. Hildrew, and C. Townsend. 2010. Multiple

stressors in freshwater ecosystems. Freshwater Biology 55:1-4.

• Pearson, R. G., and T. P. Dawson. 2003. Predicting the impacts of climate

change on the distribution of species: are bioclimate envelope models

useful? Global Ecology and Biogeography 12:361-371.

• Perring, M. P., P. De Frenne, L. Baeten, S. L. Maes, L. Depauw, H.

Blondeel, M. M. Carón, and K. Verheyen. 2016. Global environmental

37

change effects on ecosystems: the importance of land‐use legacies.

Global Change Biology 22:1361-1371.

• Peterson, A. T. 2011. Ecological niches and geographic distributions

(MPB-49). Princeton University Press.

• Poethke, H. J., T. Hovestadt, and O. Mitesser. 2003. Local extinction and

the evolution of dispersal rates: causes and correlations. The American

Naturalist 161:631-640.

• Rehfeldt, G. E., C. C. Ying, D. L. Spittlehouse, and D. A. Hamilton. 1999.

Genetic responses to climate in Pinus contorta: niche breadth, climate

change, and reforestation. Ecological Monographs 69:375-407.

• Renou-Wilson, F., T. Bolger, C. Bullock, F. Convery, J. Curry, S. Ward, D.

Wilson, and C. Müller. 2011. BOGLAND: Sustainable Management of

Peatlands in Ireland. STRIVE Report Series:181.

• Ribera, I., G. N. Foster, and A. P. Vogler. 2003. Does habitat use explain

large scale species richness patterns of aquatic beetles in Europe?

Ecography 26:145-152.

• Ricciardi, A., and J. B. Rasmussen. 1999. Extinction rates of North

American freshwater fauna. Conservation Biology 13:1220-1222.

• Richter, B. D., D. P. Braun, M. A. Mendelson, and L. L. Master. 1997.

Threats to imperiled freshwater fauna. Conservation Biology 11:1081-

1093.

• Ries, L., R. J. Fletcher Jr, J. Battin, and T. D. Sisk. 2004. Ecological

responses to habitat edges: mechanisms, models, and variability

explained. Annual Review of Ecology, Evolution, and Systematics:491-

522.

• Rydin, H., and J. K. Jeglum. 2013. The Biology of Peatlands. OUP Oxford.

• Sala, O. E., F. S. Chapin, J. J. Armesto, E. Berlow, J. Bloomfield, R. Dirzo,

E. Huber-Sanwald, L. F. Huenneke, R. B. Jackson, and A. Kinzig. 2000.

Global biodiversity scenarios for the year 2100. Science 287:1770-1774.

• Sax, D. F., R. Early, and J. Bellemare. 2013. Niche syndromes, species

extinction risks, and management under climate change. Trends Ecol Evol

28:517-523.

38

• Schuldt, A., and T. Assmann. 2010. Invertebrate diversity and national

responsibility for species conservation across Europe–a multi-taxon

approach. Biological Conservation 143:2747-2756.

• Soberón, J., and M. Nakamura. 2009. Niches and distributional areas:

concepts, methods, and assumptions. Proceedings of the National

Academy of Sciences 106:19644-19650.

• Strayer, D. L. 2006. Challenges for freshwater invertebrate conservation.

Journal of the North American Benthological Society 25:271-287.

• Sutherland, W. J., R. P. Freckleton, H. C. J. Godfray, S. R. Beissinger, T.

Benton, D. D. Cameron, Y. Carmel, D. A. Coomes, T. Coulson, and M. C.

Emmerson. 2013. Identification of 100 fundamental ecological questions.

Journal of Ecology 101:58-67.

• Svenning, J.-C., W. L. Eiserhardt, S. Normand, A. Ordonez, and B.

Sandel. 2015. The influence of paleoclimate on present-day patterns in

biodiversity and ecosystems. Annual Review of Ecology, Evolution, and

Systematics 46:551-572.

• Svenning, J.-C., and B. Sandel. 2013. Disequilibrium vegetation dynamics

under future climate change. American Journal of Botany 100:1266-1286.

• Svenning, J. C., and F. Skov. 2004. Limited filling of the potential range in

European tree species. Ecology Letters 7:565-573.

• Tscharntke, T., J. M. Tylianakis, T. A. Rand, R. K. Didham, L. Fahrig, P.

Batary, J. Bengtsson, Y. Clough, T. O. Crist, and C. F. Dormann. 2012.

Landscape moderation of biodiversity patterns and processes‐eight

hypotheses. Biological Reviews 87:661-685.

• Van Breemen, N. 1995. How Sphagnum bogs down other plants. Trends

in Ecology & Evolution 10:270-275.

• Vannote, R. L., G. W. Minshall, K. W. Cummins, J. R. Sedell, and C. E.

Cushing. 1980. The river continuum concept. Canadian Journal of

Fisheries and Aquatic Sciences 37:130-137.

• Vellend, M., K. Verheyen, H. Jacquemyn, A. Kolb, H. Van Calster, G.

Peterken, and M. Hermy. 2006. Extinction debt of forest plants persists for

more than a century following habitat fragmentation. Ecology 87:542-548.

39

• Verberk, W., G. Van Duinen, A. Brock, R. Leuven, H. Siepel, P.

Verdonschot, G. Van der Velde, and H. Esselink. 2006. Importance of

landscape heterogeneity for the conservation of aquatic macroinvertebrate

diversity in bog landscapes. Journal for Nature Conservation 14:78-90.

• Wake, D. B., E. A. Hadly, and D. D. Ackerly. 2009. Biogeography,

changing climates, and niche evolution. Proceedings of the National

Academy of Sciences 106:19631-19636.

• Westgate, M. J., P. S. Barton, P. W. Lane, and D. B. Lindenmayer. 2014.

Global meta-analysis reveals low consistency of biodiversity congruence

relationships. Nature communications 5.

• Wiens, J. J., and M. J. Donoghue. 2004. Historical biogeography, ecology

and species richness. Trends in Ecology & Evolution 19:639-644.

• Wiens, J. J., and C. H. Graham. 2005. Niche conservatism: integrating

evolution, ecology, and conservation biology. Annual Review of Ecology,

Evolution, and Systematics:519-539.

• Williams, J. W., S. T. Jackson, and J. E. Kutzbach. 2007. Projected

distributions of novel and disappearing climates by 2100 AD. Proceedings

of the National Academy of Sciences 104:5738-5742.

• Wisz, M. S., J. Pottier, W. D. Kissling, L. Pellissier, J. Lenoir, C. F.

Damgaard, C. F. Dormann, M. C. Forchhammer, J. A. Grytnes, A. Guisan,

R. K. Heikkinen, T. T. Hoye, I. Kuhn, M. Luoto, L. Maiorano, M. C. Nilsson,

S. Normand, E. Ockinger, N. M. Schmidt, M. Termansen, A. Timmermann,

D. A. Wardle, P. Aastrup, and J. C. Svenning. 2013. The role of biotic

interactions in shaping distributions and realised assemblages of species:

implications for species distribution modelling. Biol Rev Camb Philos Soc

88:15-30.

• Woodward, G., D. M. Perkins, and L. E. Brown. 2010. Climate change and

freshwater ecosystems: impacts across multiple levels of organization.

Philosophical Transactions of the Royal Society of London B: Biological

Sciences 365:2093-2106.

40

41

Chapter I

Are latitudinal richness gradients in

European freshwater species only

structured according to dispersal and time?

Early View (EV): 1-EV

richness patterns in European freshwater habitats, the lati-tudinal decrease in species richness should be independent of the sheer volume of diff erent species ’ habitat preferences for a given region (referred to as the habitat hypervolume). Moreover, the decreasing latitudinal richness patterns in lotic habitats from south to north should be constant across zona-tion classes from headwater streams to lower rivers reaches.

We tested these two assumptions across 22 biogeographic regions in Europe, using the distribution and habitat pref-erence of the European fauna of the three insect orders: Ephemeroptera, Plecoptera, and Trichoptera (EPT) (meth-ods in Supplementary materials Appendix 1). Th e diversity of the EPT fauna strongly correlates with total regional freshwater species richness ( β � 0.45, p-value � 0.001 and Supplementary materials Appendix 1, Fig. A1) and shows the same decreasing latitudinal richness pattern from south to north ( β � – 5.72, p-value � 0.001 and Supplementary materials Appendix 1, Fig. A2).

In an initial regression model including habitat hyper-volume, latitudinal midpoint, and two cofounding variables (altitude and log(region area)), only habitat hypervolume and altitude structured EPT species richness in an average model (Supplementary materials Appendix 1, Table A4). Th is result indicates that higher species richness is related to regions with a broad spectrum of habitats and not to the spatial location of a given region. Furthermore, EPT-species richness within lotic zonation classes showed a change in the relationship to latitude from headwaters to lower reaches (Fig. 1). Richness and latitude had a linear relationship for headwater classes, resembling the classic decreasing richness patterns to the latitudinal gradient (Hof et al. 2008). However, in larger streams and rivers between the Hyporhithral (grayling region) and Epipotamal (barbel region) zone this relationship shifted from being linear to becoming unimodal. When sub-setting the data by remov-ing habitat generalist species and using habitat specialist species only, the same shift in latitude – richness relationship was also evident (Supplementary materials Appendix 1, Fig. A3).

Ecography 39: 001–003, 2016 doi: 10.1111/ecog.02183

© 2016 Th e Authors. Ecography © 2016 Nordic Society Oikos Subject Editor: Jani Heino. Editor-in-Chief: Miguel Ara ú jo. Accepted 15 February 2016

Th e latitudinal gradient in species richness of the European freshwater fauna is usually depicted as a con-trasting pattern between lentic (standing waters) and lotic (running waters) habitats. Species richness decreases from south to north in lotic habitats, while lentic habi-tats show an intermediate richness peak in central Europe (Hof et al. 2008). In this Brevia we question the strict dichotomic split in latitudinal richness patterns between lentic and lotic habitats in Europe. We do this by demon-strating that the diff erences in richness patterns between headwater streams and rivers are as great as the diff er-ences between lentic and lotic habitats. We also show that species richness can be explained solely by the hypervol-ume defi ned by the variety of lotic habitats preferences in a region.

It has been proposed that the interaction between coloniza-tion ability and time elapsed since the last major glaciation has been responsible for the diff erences in latitudinal richness gradients of animal species between lentic and lotic habitats in Europe. Long-term stability of lotic habitats and a higher turnover rate of lentic habitats are thought to be the main evolutionary mechanism promoting high dispersive species in lentic habitats and low dispersive species in lotic habitats (Arribas et al. 2012). S á nchez-Fern á ndez et al. (2012) stated that the current ranges of lentic species are in equilibrium with their climatic niches, while lotic species are constrained to their pleistocene refugia due to dispersal limitations, creating the contrasting latitudinal species richness pattern in Europe. Th is spatio-temporal hypothesis of European freshwater species assumes that environmental habitat conditions do not infl u-ence the large-scale species richness patterns. However, it is well known that environmental conditions do in fact structure freshwater species communities in lotic habitats from head-waters to lower reaches (Vannote et al. 1980). Also, Azevedo et al. (2013) has shown that environmental factors structure freshwater species richness across large spatial scales.

If geological stability, together with colonization rates since the last glacial maximum, is the only driver of species

Are latitudinal richness gradients in European freshwater species only structured according to dispersal and time?

Lars L ø nsmann Iversen , Dean Jacobsen and Kaj Sand-Jensen

L. L. Iversen (lliversen@bio.ku.dk), D. Jacobsen and K. Sand-Jensen, Section for Freshwater Biology, Biological Inst., Univ. of Copenhagen, Universtitetsparken 4, DK-2100 K ø benhavn, Denmark. LLI also at: Amphi Consult ApS, International Sciencepark Odense, Forskerparken 10, DK-5230 Odense M, Denmark.

42

2-EV

Altogether, we fi nd very little support in a strict dispersal-driven split between lentic (Hof et al. 2008) and lotic habi-tats when studying species ranges and richness gradients. Species richness in the lower reaches of lotic habitats has the same relationship to latitude as lentic habitats. Moreover, lotic species richness patterns can solely be explained by the

regional spectrum of habitats, even when adjusting for the latitudinal placement of these regions.

Th ese results are in concordance with studies of fresh-water species richness in North America, documenting that environmental factors were always more important than spatial factors (latitude or longitude) in structuring species richness (B ê che and Statzner 2009). However, spe-cies ability to disperse undoubtedly plays a signifi cant role when recolonizing potential ranges from pleistocene refugia (Baselga et al. 2012). Our fi ndings suggest that processes aff ecting macroecological richness gradients of European freshwater species are more complex than just diff erences in geologically stability between habitat types. It has already been highlighted that interactions between dispersal and environmental fi lters organize metacommunities in aquatic systems (Henriques-Silva et al. 2013, Heino et al. 2015).

If our results are a consequence of such interactions, they suggest that the sorting eff ect of environmental condi-tions together with dispersal infl uence not only metacom-munities, but also continental species richness patterns in freshwaters (Heino 2011). Th e increasing availability of local community data from headwaters to lower reaches across the latitudinal range as well as fi ne-scale environ-mental data (elevation range, ruggedness, and freshwater habitat availability, see Domisch et al. (2015)) could bridge the current gap between processes sorting local and con-tinental richness gradients. Integrating other factors than dispersal ability when attempting to predict freshwater species ranges is likely to expand our understanding of the mechanisms structuring the current diversity patterns within these habitats.

Acknowledgement – We thank Christian Hof, Daniel Hering and Astrid Schmidt-Kloiber for facilitating and providing access to the data used. We also thank Charlotte Nekman for commenting on a previous version of the manuscript. Th e study was supported by the Danish Ministry of Higher Education and Science (project 0604-02230B).

References

Arribas, P. et al. 2012. Dispersal ability rather than ecological toler-ance drives diff erences in range size between lentic and lotic water beetles (Coleoptera: Hydrophilidae). – J. Biogeogr. 39: 984 – 994.

Azevedo, L. B. et al. 2013. Species richness – phosphorus relation-ships for lakes and streams worldwide. – Global Ecol. Biogeogr. 22: 1304 – 1314.

Baselga, A. et al. 2012. Dispersal ability modulates the strength of the latitudinal richness gradient in European beetles. – Global Ecol. Biogeogr. 21: 1106 – 1113.

B ê che, L. A. and Statzner, B. 2009. Richness gradients of stream invertebrates across the USA: taxonomy- and trait-based approaches. – Biodivers. Conserv. 18: 3909 – 3930.

Domisch, S. et al. 2015. Near-global freshwater-specifi c environ-mental variables for biodiversity analyses in 1 km resolution. – Sci. Data 2: 150073.

Heino, J. 2011. A macroecological perspective of diversity patterns in the freshwater realm. – Freshwater Biol. 56: 1703 – 1722.

Heino, J. et al. 2015. Metacommunity organisation, spatial extent and dispersal in aquatic systems: patterns, processes and prospects. – Freshwater Biol. 60: 845 – 869.

Figure 1. Relationship between latitude and species richness scores for Ephemeroptera, Plecoptera and Trichoptera species in European ecoregions. Each graph represents a specifi c river zonation class spanning from headwaters to lower reaches. Th e lower graph shows Δ AIC values from linear regression models within the eight river zonation categories. Th e values represent diff erences between a constant model and a unimodal (black circles) and a linear (grey triangles) relationship between EPT-species richness and latitude. High Δ AIC represents a better fi t for a given model.

43

3-EV

Henriques-Silva, R. et al. 2013. A community of metacommuni-ties: exploring patterns in species distributions across large geographical areas. – Ecology 94: 627 – 639.

Hof, C. et al. 2008. Latitudinal variation of diversity in European freshwater animals is not concordant across habitat types. – Global Ecol. Biogeogr. 17: 539 – 546.

S á nchez-Fern á ndez, D. et al. 2012. Habitat type mediates equilibrium with climatic conditions in the distribution of Iberian diving beetles. – Global Ecol. Biogeogr. 21: 988 – 997.

Vannote R. L. et al 1980. Th e river continuum concept. – Can. J. Fish Aquat. Sci. 37: 130 – 137.

Supplementary material (Appendix ECOG-02183 at < www.ecography.org/appendix/ecog-02183 > ). Appendix 1.

44

Appendix - Chapter I Material and Method Data compilation

Species distributions were defined as records within 25 biogeographic regions

presented in Limnofauna Europaea (Illies 1978). These regions have been

used in several studies on diversity of freshwater species in Europe (e.g. Hof

et al 2008, Dehling et al. 2010, Thieltges et al. 2011). Within the 1942

European EPT biogeographic regions distributions, species river zonation and

microhabitat preferences were derived from three volumes of Distribution and

Ecological Preferences of European Freshwater Organisms (Graf et al. 2008,

Graf et al. 2009, Buffagni et al. 2009), with recent updates by Schmidt-Kloiber

and Hering (2015). Altogether, river zonation preferences and microhabitat

preferences were derived for 71 % and 67 % of the European EPT species,

respectively. For each coded species a 10 point assignment system was used

to express habitat preference within 10 river zonation categories and 13

microhabitat categories (Table A1). These environmental preferences were

used in the following to define the position and diversity of habitat niches. The

diversity of EPT fauna do correlate with the total species diversity of entire

communities (Fig. A.1). Thus, we assume that habitat preferences gained from

EPT-fauna act as a valid proxy for general habitat preference in river

communities.

Patch size, geographic placement and maximum altitude for each

biogeographic region were obtained from Hof et al. (2008). From the original

set of 25 regions we excluded Iceland due to its isolated location and deviating

geological history. Additionally, the Caucasus and the Caspic Depression were

excluded due to an incomplete species lists (Conti et al. 2014). If not stated

differently, we omitted species restricted to brackish (hypopotamal) and lake

habitats (littoral and profundal), prior to any analyses. We also excluded

species only found in terrestrial microhabitats, leaving a total of 768 species.

The number of species from the pool of 768 species recorded in a region was

used as the measure of total EPT diversity in each of the 22 regions. In each

45

46

region we additionally estimated species richness within each of the river

zonation classes. Species richness estimate was defined as total sum of the

preference score across all species for each river zonation class within each

region. Species richness were calculated for all species, without generalist

species (excluding all preference scores below 5), and with specialist species

only (excluding all preference scores below 8).

Defining habitat niches and niche hypervolume

The placement of each EPT species on a river zonation and microhabitat

gradient was determined by means of a detrended correspondence analysis

(DCA; Hill and Gauch 1980). DCA analysis were preferred over other

commonly used ordination methods (e.g. PCA or NMDS) since we expected

unimodal rather than linear species response to the defined gradients and in

order to define primary gradients (axes) of species environmental spacing and

shifts in species composition among observations. Individual DCA analyses

were performed for river zonation and microhabitat preference identifying a

downstream gradient and changes in microhabitat structure (Table A.2). Both

DCAs had ordination axes correlated to river zonation and microhabitat

variables. Within river zonation, first axis scores were related to zonation

classes and higher scores represented a preference to downstream zonation

classes (Table A.2). The second axis of the microhabitat DCA demonstrated a

gradient in microhabitats preference from coarse organic substrates to fine-

grained inorganic substrates (Table A.2).

In order to explore possible links between species richness and the diversity of

environmental niche preference, we estimated a measure of niche diversity.

Species ordination scores in the two DCA analyses respectively were used to

express the diversity of niche positions within each biogeographic region. For

each region a hypervolume (Blonder et al. 2014) was calculated based on the

species found within each region. The hypervolume describe the diversity of

niche placements, i.e. the diversity of species with different habitat

preferences.

Hypervolume algorithms estimate the shape and volume of n-dimensional

objects via a defined kernel density estimate, i.e. they act as a measure of

niche breath. In its core structure the hypervolume is the niche deviance

46

47

between species pairs and is calculated independently of the species richness

within each region. In order to create euclidean distances and hypervolumes in

a functional space comparable within and across analyses all predictors were

transformed to z-scores. Hypervolumes were constructed by using total

probability densities, 1000 Monte Carlo samples per data point and a heuristic

bandwidth definition applied independently to each axis of the data. For

computational details of hypervolumes we refer to Blonder et al. (2014).

Data analysis:

Linear regression models were used in pairwise comparisons via an anova F-

test and p-values were evaluated at a 5% significance level. Normality and

equal variance within all initial models were checked by an inspection of the

residual vs. the fitted values and qq-plots of the standardized residuals. Based

on these inspections region area was log-transformed in order to homogenize

residual variance.

For EPT richness, model averaging was conducted in order to detect the

parameters affecting richness across regions. Initial models were constructed

containing richness as response variable and region size, maximum altitude,

latitudinal midpoint and the habitat hypervolume as explanatory variables. All

predictors were transformed to z-scores to standardize slope coefficients (β) to

compare the relative strength of the predictors. The individual models were

weighted based on AIC (Burnham and Anderson 2004, Tabel A.3) and we

performed model averaging by averaging the slope coefficients (β) across all

models with ΔAIC ≤2 from the model with the lowest AIC (as suggested by

Burnham & Anderson 2002). From the average model, explanatory variables

with unidirectional slope estimation (within the 95% confidence intervals) were

selected as variables affecting species richness. Estimates of parameters by

model averaging are robust in the sense that they reduce model selection bias

and account for model selection uncertainty (Johnson and Omland 2004).

Latitudinal richness patterns of the EPT fauna were calculated within each

section of zonation along the river continuum. This was done by calculating

AIC values from (linear) regressions modelling a linear, unimodal and constant

relationship between latitude and richness within each habitat class. The

strength of the relationships were expressed as ΔAIC(AICconstant model- AIClinear or

48

unimodal model

All analysis were conducted in R ver. 3.2.1 (www.r-project.org).

). Species richness estimates were analyzed separately based on

species pools including all species, no generalist species, and specialist

species only

49

Table A.1: Categories for stream zonation and microhabitat preferences

(following Graf et al. 2008, Graf et al. 2009, and Buffagni et al. 2009).

Stream zonation preference

Eucrenal (spring region)

Hypocrenal (spring-brook)

Epirhithral (upper trout region)

Metarhithral (lower trout region)

Hyporhithral (grayling region)

Epipotamal (barbel region)

Metapotamal (bream region)

Hypopotamal (brackish water)

Littoral (lake and stream shorelines, ponds, etc.)

Profundal (bottom of stratified lakes)

Microhabitat/substrate preference

Akal (fine to medium-sized gravel)

Algal

Argyllal (silt, loam, clay)

Hygropetric habitats

Macro/megalithal

Micro/mesolithal (coarsegravel to hand-sized cobbles)

Macrophytes (macrophytes, mosses, Characeae, living parts of terrestrial

plants)

Madicol (edge of water bodies, moist substrate)

Particulate organic matter (Coarse and fine particulate organic matter)

Pelal (mud)

Psammal (sand)

Woody debris (Woody debris, twigs, roots, logs)

Other habitats (E.g host of a parasite)

50

Table A.2: First and second axis scores of river zonation and microhabitat

DCA

Stream Zonation class DCA1 DCA2Eucrenal -2,56 0,04Hypocrenal -1,32 0,76Epirhithral -0,41 -1,25Metarhithral 0,53 -0,55Hyporhithral 1,10 0,83Epipotamal 1,93 1,55Metapotamal 3,25 -1,02

Mikrohabitat class DCA1 DCA2Macrophytes -0,21 -1,67Woody debris -0,13 -1,23Particulate organic matter -0,18 -0,56Macro/megalithal 0,09 -0,33Algae -1,30 0,12Hygropetric habitats 2,58 0,13Micro/mesolithal 0,08 0,43Pelal -0,25 1,23Akal -0,12 1,41Psammal -0,21 2,20Argyllal -0,25 2,94

51

Table A.3: AIC models from the initial set of four parameters.

β -estimate

Intercept Maximum altitude

Habitat hypervolume

Latitudinal midpoint

Log(region area) df logLik AIC ΔAIC

257.6 112.1 105.5

4 -112.8 233.6 0 257.6 89.6 90.7 -33.6

5 -111.8 233.6 0.02

257.6 108.5 108.3

-15.3 5 -112.4 234.9 1.33 257.6 90.1 93.4 -30.4 -7.1 6 -111.7 235.4 1.88 257.6

50.0 -92.0

4 -117.3 242.6 9.02

257.6

51.1 -90.7 -3.0 5 -117.3 244.6 11 257.6

-104.1

3 -119.5 245.0 11.5

257.6 34.9

-85.3

4 -118.8 245.5 12 257.6

-109.2 15.9 4 -119.3 246.6 13.1

257.6 38.2

-90.1 20.7 5 -118.4 246.8 13.3 257.6 81.0

3 -122.7 251.4 17.8

257.6

72.4

3 -123.5 253.0 19.4 257.6

82.5

-40.0 4 -122.6 253.1 19.6

257.6 82.3

4.4 4 -122.7 253.3 19.8 257.6

2 -126.3 256.5 23

257.6 -19.2 3 -126.1 258.2 24.6

Table A.4: Parameter estimates and 95 % confidence intervals of the selected

geographical and habitat niche variables. The estimates are derived from an

average model of four possible models explaining the species richness of

European EPT fauna.

Parameter Estimate

95% confidence

intervals

(Intercept) 257.6 [238.1,277.2]

Maximum altitude 100.8 [47.4,154.2]

Habitat

hypervolume 99.3 [51.2,147.4]

Latitudinal midpoint -15.6 [-88.6,23.2]

Log(region area) -3.7 [-57.0,33.4]

52

Fig A1: Relationship between EPT richness and Total species richness (minus

EPT richness) across the 22 biogeographic ecoregions used in the study. Total

species richness is compiled from Hof et al. 2008.

Fig A2: Relationship between latitudinal center and EPT species richness

across the 22 biogeographic ecoregions used in the study.

53

Fig A3: ΔAIC values from linear regression model within eight river zonation

categories based on generalist species excluded (top), and specialist species

only (bottom). The values represent differences from a constant model

compared to a unimodal (black circles) and a linear (grey triangles)

relationship between EPT-species richness and latitude. High ΔAIC represents

a better fit of a given model.

54

References • Blonder B. et al. 2014. The n-dimensional hypervolume. - Global Ecol.

Biogeogr.

• Buffagni, A. et al. 2009: Distribution and Ecological Preferences of

European Freshwater Organisms. Volume 3 - .phemeroptera. –

Pensoft.

23:595–609.

• Burnham K. P. and Anderson D. R. 2002. Model selection and

multimodel inference: a practical information-theoretic approach, 2nd

ed. – Springer.

• Burnham, K. P. and Anderson, D. R. 2004. Multimodel inference:

understanding AIC and BIC in model selection. - Sociol. Method. Res.

33: 261–304.

• Conti, L. et al. 2014. A trait-based approach to assess the vulnerability

of European aquatic insects to climate change. - Hydrobiologia 721:

297–315.

• Dehling, D.M. et al. 2010. Habitat availability cannot explain the species

richness patterns of European lentic and lotic freshwater animals. - J.

Biogeogr. 37:1919-1926.

• Graf, W. et al. 2008. Distribution and Ecological Preferences of

European Freshwater Organisms. Volume 1 - Trichoptera. – Pensoft.

• Graf, W. et al. 2009. Distribution and Ecological Preferences of

European Freshwater Organisms. Volume 2 - Plecoptera. - Pensoft.

• Hof, C. Brändle, M. and Brandl, R. 2008. Latitudinal variation of diversity

in European freshwater animals is not concordant across habitat types.

- Global Ecol. Biogeogr.

• Johnson, J. B. and Omland, K. S. 2004. Model selection in ecology and

evolution. – Trends Ecol. Evol. 19: 101 – 108.

17: 539–546.

• Schmidt-Kloiber, A. and Hering D. 2015. www.freshwaterecology.info -

an online tool that unifies, standardises and codifies more than 20,000

European freshwater organisms and their ecological preferences. –

Ecol. Indic 53: 271-282.

55

• Thieltges, D.W. et al., 2011. Host diversity and latitudinal gradients drive

trematode diversity patterns in the European freshwater fauna. - Global

Ecol. Biogeogr.

20: 675-682.

56

57

Chapter II

Variable history of land use reduces the relationship to specific habitat

requirements of a threatened aquatic insect

ORIGINAL ARTICLE

Variable history of land use reduces the relationship to specifichabitat requirements of a threatened aquatic insect

Lars Lønsmann Iversen1,2 • Riinu Rannap3 • Lars Briggs2 • Kaj Sand-Jensen1

Received: 15 October 2014 /Accepted: 18 September 2015

� The Society of Population Ecology and Springer Japan 2015

Abstract The hutchinsonian realized niche of a species is

the most common tool for selecting the actions needed

when restoring habitats and establishing conservation areas

of species. However, defining the realized niche of a spe-

cies is problematic due to variation across spatial and

temporal scales. In this study we tested the hypothesis that

habitat parameters defining the realized niche of a species

can be derived from a regional study and that national

changes in land use influence the perception of the realized

niche across different landscapes. We described the real-

ized habitat niche of the threatened dragonfly Leucorrhinia

pectoralis, in four Estonian landscapes which all have

undergone more than 20 years of habitat degradations. We

recorded the presence/absence of L. pectoralis and mea-

sured 7 habitat variables for 140 lakes and ponds located in

one restored and three un-restored landscapes. Lake size

and proportion of short riparian vegetation were signifi-

cantly positive parameters determining the presence of L.

pectoralis across landscape types. The species was much

more habitat specific in the restored landscape, with larger

influence of other habitat parameters. Our data suggest that

the realized niche of the species in the un-restored land-

scapes was constrained by the present-day habitats. The

study demonstrate that if a species realized niche is derived

from local distribution patterns without incorporating

landscape history it can lead to an erroneous niche defini-

tion. We show that landscape restoration can provide

knowledge on a species’ habitat dependencies before

habitat degradation has occurred, provided that restoration

mitigation reflects the former landscape characteristics.

Keywords Biodiversity � Extinction debt � Hutchinsonianrealized niche � Natura 2000 � Odonata � Speciesconservation

Introduction

Defining important local habitat characteristics in accor-

dance with the hutchinsonian realized niche concept of a

species (e.g., Jackson and Overpeck 2000; Colwell and

Rangel 2009) is the common tool used for selecting

appropriate actions for habitat restoration and establishment

of conservation areas of a species (Margules and Pressey

2000; Guisan and Thuiller 2005). This spatial template

referred to as ‘‘the habitat’’ contains the location where the

species undergoes its life cycle (Southwood 1977, 1988).

Thus, the habitat of a species is thereby defined by the

resources needed for a given individual to exist throughout

its different life stages (Dennis et al. 2003, 2014).

Electronic supplementary material The online version of thisarticle (doi:10.1007/s10144-015-0516-z) contains supplementarymaterial, which is available to authorized users.

& Lars Lønsmann Iversen

lliversen@bio.ku.dk

Riinu Rannap

riinu.rannap@ut.ee

Lars Briggs

lb@amphi.dk

Kaj Sand-Jensen

ksandjensen@bio.ku.dk

1 Freshwater Biological Laboratory, Biological Institute,

University of Copenhagen, Universitetsparken 4,

2100 Copenhagen Ø, Denmark

2 Amphi Consult ApS, International Sciencepark Odense,

Forskerparken 10, 5230 Odense M, Denmark

3 Institute of Ecology and Earth Sciences, University of Tartu,

Vanemuise 46, 51014 Tartu, Estonia

123

Popul Ecol

DOI 10.1007/s10144-015-0516-z

58

Defining resource-based habitat requirements of a cer-

tain species can be problematic for at least three reasons.

Firstly, when the environmental niche varies across the

species’ geographical range (Pearman et al. 2008; Oliver

et al. 2009). Secondly, when the presence and abundance of

the species at the local scale is strongly influenced by

patterns and processes operating at higher spatial scales

(Moilanen and Hanski 1998; Iversen et al. 2013). Thirdly,

when the presence and abundance of the species is the

result of long-term historical changes in land use which are

complex and often not quantitatively documented (Foster

et al. 2003; Cramer et al. 2008). Defining habitat require-

ments from empirical data on species presence may be

straightforward if the distribution is in equilibrium with

requirements (Peterson et al. 2011). However, more often

factors not directly linked to the fundamental niche of the

species do affect species distributions, including biotic

interactions (Gotelli et al. 2010), extinction debt (Schurr

et al. 2012) and time constraints on colonization (Svenning

and Skov 2007; Baselga et al. 2012).

Consequently, local conservation actions should be

linked to habitat dependencies within each geographic area

(Ferrier 2002; Rannap et al. 2012), including the influence

on the realized niche of the present-day structures of the

landscape (Peterson et al. 2011) and historical changes of

land use (Lindborg and Eriksson 2004; Kuussaari et al.

2009). This broad-scale approach is recommended to avoid

that habitat requirements deduced from single sites can

lead to erroneous conclusions because of current and his-

torical effects of land uses and land use changes (Davis

et al. 1998; Pulliam 2000).

In order to encompass these problems, habitat require-

ments should ideally be derived from studies of several

landscapes within a region to distinguish between species

relations to specific habitat characteristics and to broad-

scale landscape properties (Margules and Pressey 2000).

For threatened and declining species this ideal analysis

would often require studies on all national or all indepen-

dent populations of the species (e.g., Heikkinen et al. 2007;

Thomaes et al. 2008; Brito et al. 2009). So far studies

exploring whether the specific habitat requirements of a

species at larger spatial scales are appropriate for local

conservation management are seldomly reported, because

they require a priory knowledge of the species ecological

niche within the region (Pulliam 2000).

In this study we tested the hypothesis that habitat

parameters defining the realized niche of a threatened spe-

cies can be derived from a regional study. We also analyzed

whether national changes in land use influence the realized

niche in a significant manner across different landscapes of

a region. This was done by exploring the distribution and

breeding-habitat choices of the threatened dragonfly Leuc-

orrhinia pectoralis (Charpentier 1825) in four designated

conservation areas in the eastern part of Estonia (Fig. 1).

These areas have all experienced changes in land use during

the last 25 years, because of national changes in political,

Fig. 1 Study areas in the

restored landscape (in grey) and

three un-restored landscapes (in

black) in Estonia. Number of

lakes and ponds (n) is depicted

Popul Ecol

123

59

economical, and social structures (Peterson and Aunap

1998). To counteract negative effects on the ponds and

small lakes from these changes of land use (e.g., secondary

succession and afforestation of the cultivated areas), large

scale conservation actions have been conducted during the

last 10 years (Rannap et al. 2009).

The regional landscape history is used to test whether

the habitat dependencies of the species derived from the

un-restored landscapes reflect an ecological niche of L.

pectoralis in an equilibrium state. This is done by com-

paring habitat dependencies between the un-restored

landscapes and restored landscapes.

Methods

Study species and study sites

Leucorrhinia pectoralis (Charpentier 1825) (Odonata,

Libellulidae) is the largest of five European species in the

genus. Individuals range from 32 to 39 mm in body length.

The larvae are fully aquatic with a one- to two-year life

cycle (Brauner 2006; Corbet et al. 2006) with adult

emergence between spring and early summer (Dijkstra and

Lewington 2006). The species primarily inhabits sun-ex-

posed mesotrophic and slightly eutrophic waters with well-

developed riparian and aquatic vegetation (Schorr 1990;

Sternberg et al. 2000). However, the species seems to be

present in a wide range of different aquatic habitats. The

species’ ecological niche across its distribution range is not

well known (Foster 1996). It is often found in heteroge-

neous landscapes of mixed land use classes of forest and

extensive agriculture (Foster 1996). The species is dis-

tributed throughout Central and Northern Europe, though

the populations are often small and spatially isolated (Di-

jkstra and Lewington 2006). During the 20th century, L.

pectoralis has declined dramatically in its presence and

abundance, mainly because of the destruction of freshwater

habitats (Sahlen et al. 2004). Therefore, the species has

been included in the Annexes II and IV of the EU Habitats

Directive, given it a status as one of the most protected

insects within the European Union (Council of the Euro-

pean Union 1992).

The study was conducted in four nature reserves, in the

northern, eastern and southern parts of Estonia (Fig. 1).

Varangu Nature Protected Area (NPA) (59�030N, 26�100E)is the northernmost study site, located on the north-western

slope of the Pandivere Upland, the highest bedrock upland

in Estonia. Varangu NPA (total area 104.2 ha) is mainly

covered by pine Pinus sylvestris and spruce Picea abies

forests of varying soil humidity. An abandoned chalk

quarry, making up nearly 25 % of the reserve, provides

open environment and oligotrophic ponds and lakes.

Emajoe-Suursoo NPA (58�220N, 27�120E) is located at

the inlet of the Emajogi River, into Lake Peipus. Emajoe-

Suursoo is one of the oldest and largest delta swamps

(20,000 ha) in Estonia. Within the nature reserve there are

flooded marshes, lakes, fens and small ponds in the agri-

cultural areas at the outskirts and along settlements in the

central part of the nature reserve.

Karula National Park (NP) (57�420N, 26�290E) makes up

nearly a third (12,300 ha) of Karula Upland. The area has a

hummocky topography between the hills and is covered

with woodlands, fields and meadows. Several bogs, mires

and beaver floods are present. Also several lakes and

human-created ponds are found in the reserve.

Haanja Landscape Protected Areas (LPA) (57�430N,27�200E) is a large LPA (total area 16,900 ha) in southern

Estonia. The hummocky moraine landscape represents a

mosaic of forests (45 %), grasslands (21 %) and small

extensively used farmlands. Lakes, ponds, swamps and

small bogs are situated in depressions and valleys between

the hills (Rannap et al. 2009).

During the lasts 20 years the cultivation of these four

areas has been abandoned due to national changes in

political, economic, and social structure following the

independence of Estonia in 1991 (Peterson and Aunap

1998). These changes led to secondary succession and

afforestation of many former cultivated areas (Vares et al.

2001; Soo et al. 2009). Landscape changes have led to a

loss of farmland ponds and standing waters associated with

open land use practices in all four study areas (Rannap,

unpublished data). Therefore large scale conservation

actions were initiated to reactivate and create lost aquatic

habitats in Haanja LPA and reestablish the former mixed

use of the area (Rannap et al. 2009). From 2005 to 2008,

139 ponds (8.2 ponds/km2) were restored within the LPA.

In total 25 % of the ponds and the surrounding terrestrial

habitat within Haanja were restored. Partial or completely

overgrown ponds were cleaned from bushes and high dense

vegetation and mud sediment. In order to recreate the

shallow littoral zones the banks were levelled and very

small ponds were enlarged.

The large lakes in each study area act as population

sources to the surrounding landscape according to Wil-

dermuth (1993) and in agreement with the concept of

source-sink populations (Pulliam 1988). Determining

habitat dependence based on the contemporary distribution

of the species might be over- or underestimated due to the

presence of sink sub-populations not contributing to the

persistence of the species in the landscape (Beale and

Lennon 2012). Nonetheless, the smaller breeding habitats

do not act as true sink habitats because L. pectoralis is

known from several areas to sustain a viable population

within systems of exclusively small ponds (Toth 1983;

Dolny and Harabis 2012). This knowledge considered,

Popul Ecol

123

60

combined with the use of breeding sites of the species in

our study, we do expect that the ‘‘presence’’ sites contain

true breeding habitats, with a low amount of marginal

habitats, potentially confounding the niche determination

of the species (Pulliam 2000).

Data collection

A total of 140 localities were investigated during the first

half of June in 2010 (un-restored landscapes) and 2011

(restored landscape). In each study area, sites were selected

at random without any prior knowledge of the presence of

the study species. Sites were omitted if directly intercon-

nected by waterways to a site already examined. A stan-

dard dip-netting method (Iversen et al. 2013) was used to

record breeding sites. The larvae of L. pectoralis were

actively searched for during 45 min at each site by

sweeping a hand dipnet (40 9 40 cm frame) through

vegetation and surface sediment. Also, the riparian vege-

tation was searched for larvae exuviae of the species.

Thereby, the species was recorded as either present or

absent at each site. To rule out skill-related sampling bias,

all fieldwork was carried out by the same person (LLI).

Absence localities established by a 45 min active search

should represent a true absence of L. pectoralis. In a study

of 50 Swedish lakes, comparable to the Estonian lakes in

their physical appearance, all records of L. pectoralis were

obtained within the first 40 min of a 90 min long search

time [Fig. S1 in Electronic Supplementary Materials

(ESM)]. Thus, the likelihood of false negative results

during the 45 min-long search period of this study is

probably very small. Sampling across two generations of L.

pectoralis might be problematic if a certain study year

contained seasonal extremes creating abrupt changes in the

size of populations. Such changes in North European

dragonflies have been linked to average temperatures dur-

ing spring and summer months respectively (Dingemanse

and Kalkman 2008). Within the timespan of the two gen-

erations studied (2009–2011), we did not see any seasonal

differences in average temperatures in the spring or sum-

mer months (Section S1 in ESM). Therefore, we do not

expect great changes in the local populations studied, due

to climatic variations between two sampling years.

Habitat descriptions were conducted along with the

collection of species data. Instead of recording fluctuating

water-chemistry variables, we focused on more

stable physical and biological variables. Related to known

habitat dependencies of L. pectoralis (summarized in

Foster 1996; Sternberg et al. 2000) seven habitat characters

were assessed at each study site. They included habitat

characteristics related to sun exposure and water tempera-

ture [(1) shading from surrounding trees (%) and (2)

minimum edge slope]; larval habitat [(3) submerged

vegetation (%), (4) floating vegetation (%), and (5) riparian

vegetation (%)]; adult breeding habitat [(6) vegetation

higher than 1 m (%)] and site stability [(7) Lake surface

area] (Table S1 in ESM). These variables relate to the

ecological resources needed by L. pectoralis to complete its

lifecycle within the breeding site. Aquatic vegetation cover

is linked to shelter and potential foraging ground for the

larvae. Emergent vegetation enables the creation of

breeding territories between males and offers protection

from other more aggressive dragonfly species. Shading and

the slope of the bank edges are linked to water temperature

of the breeding sites because extensive zones with shallow

water and high sun exposure create areas with high water

temperatures suitable for larval growth. At each site,

shading was defined as the percentage of the shoreline

having standing threes of more than two meter in height,

present within three meters from the shoreline. The edge of

the bank slopes (�) was estimated for each of the four

cardinal banks. Larval and adult habitats were obtained by

visually estimating the percentage vegetation cover of the

total lake surface area. Surface area of ponds (\0.5 ha) was

measured in the field, while lakes ([0.5 ha) were measured

prior to the fieldwork, using high resolution orthophotos.

Statistical methods

For both landscape types initial logistic regression models

were established including all seven habitat variables as

explanatory variables and presence of L. pectoralis as the

dependent variable. Prior to this first selection step, lake

area was log-transformed to ensure that the probability of

L. pectoralis is presence would be zero when lake area is

zero. Model selection and parameter reduction then fol-

lowed two selection steps.

Firstly, all possible sub-models, created from the initial

models, were evaluated based on second-order Akaike

information criterion (AICc) values. This included an

evaluation of 127 different combinations of the seven

habitat descriptors plus the null model for the un-restored

and restored areas separately (Table S2 in ESM). We used

AICc and not AIC to evaluate the different sub-models due

to the low ratio between number of samples and number of

parameters (Symonds and Moussalli 2011). For the un-

restored and restored areas, the sub-models within two

units of the model with the lowest AICc value were iden-

tified (Burnham and Anderson 2002). Parameters in the

2DAICc selected models were chosen for further analysis,

while parameters not present in any of the 2DAICc models

were omitted.

Using these parameters we conducted, a second model

reduction based on the effects of the explanatory variables

on the presence of L. pectoralis. We tested for a linear

relationship on a log-odds scale in a multiple logistic

Popul Ecol

123

61

regression model containing the 2DAICc selected parame-

ters. The models followed a stepwise reduction until only

variables with a significant effect on the model result

remained. All P values corresponding to likelihood ratio

tests and the effect of the explanatory variables were

evaluated at a 5 % significance level. An initial data

exploration showed no relationship between the explana-

tory variables reducing the possibility of errors caused by

colinearity among explanatory variables (Whittingham

et al. 2006; Stephens et al. 2007).

It could be argued that the un-restored landscape model

should account for differences among the three study areas

(Fig. 1), for example in a mixed model environment.

However, we chose not to incorporate the study area into

the models since we wanted to describe habitat depen-

dencies of L. pectoralis across the study areas. Parameter

estimation becomes conditional when incorporating ran-

dom effects in a logistic regression model (Agresti 2010).

In our case this would restrict the parameter estimation to

be within and not between each of the study areas.

Moreover, we tested whether potential differences in

habitat relationships of L. pectoralis between un-restored

and restored landscapes were not just simple artifacts of

differences between the two landscape types. This was

done by creating null models for each area describing the

relationship between the significant parameters and a ran-

dom distribution of L. pectoralis. The null models used the

observed site occupation frequency of L. pectoralis in both

landscape types and randomly assigned these to the

observed environmental variables in each area. In a mul-

tiple logistic regression model the combined effect of the

selected variables and landscape type on the presence of L.

pectoralis was tested by a likelihood ratio test. A signifi-

cant effect modification in this model would indicate that

the differences in L. pectoralis’ response to a given

parameter, might just as likely be due to random patterns

caused by general differences among the two landscape

types. Average P values and corresponding 95 % CI were

generated based on 1000 null-model iterations.

The individual explanation power offered by the inde-

pendent variables in the two landscape types was quantified

by a hierarchical partitioning analysis. Hereby, the inde-

pendent contribution of a single variable and the joined

contribution with other variables can be quantified (Chevan

and Sutherland 1991; Mac Nally 2002). Variables selected

in the first model reduction step were chosen as having

potential independent explanation power and they were

included in the analyses of hierarchical partitioning (Mac

Nally and Walsh 2004).

We evaluated the predictive power of the models using

receiver operating characteristics (ROC curves, Robin et al.

2011). For each landscape type we produced ROC curves

based on a model containing only parameters with

significant individual effect on the presence of L. pectoralis

and a model containing the 2DAICc selected parameters.

Differences between the predictive abilities among models

within landscapes were evaluated based on their AICc

values using a method for paired ROC-curves (DeLong

et al. 1988).

All statistical analyses were performed in R v. 3.0.2

(http://www.r-project.org).

Results

Leucorrhinia pectoralis was found in 43 % (33 of 76) of

the visited sites in the un-restored landscapes and in 51 %

(33 of 64) of the sites in the restored landscape. The

occupation frequency confirmed that the species was not

significantly more abundant in one landscape type than in

the other (v2 test, P = 0.29).

By reducing parameters from the initial models using

the 2DAICc selected sub-models, the amount of shading

and tall vegetation ([1 m) in the un-restored landscape and

the amount of floating vegetation and tall vegetation in the

restored landscape were omitted from further analysis. The

single-step model reduction revealed that only lake size

and the amount of riparian vegetation (veg.\ 1 m tall) had

a significant effect on the presence of L. pectoralis in both

areas (Table 1). Also, there was no combined effect of lake

surface area and percentage of riparian vegetation in any of

the landscapes (Un-restored landscapes P = 0.84; Restored

landscape P = 0.10).

When comparing the individual explanation power from

the different parameters by hierarchical partitioning anal-

ysis, differences between the two landscape types appeared

(Fig. 2). In the un-restored landscape, lake surface area had

the main individual effect on the presence of L. pectoralis

(71 % in the 2DAICc selected model). Among the other

parameters, only short riparian vegetation (\1 m) con-

tributed with more than 10 % (i.e., 15 %) of individual

explanation power, while the other parameters had little

individual effect. In contrast, parameters in the restored

landscape had a more equal individual explanation power

in the 2DAICc selected model (Fig. 2). Although lake

surface area still had the highest individual explanation

power (35 %), short riparian vegetation (25 %) and shad-

ing (13 %) in combination contributed with slightly more

explanation power than lake size alone. All parameters

selected in the 2DAICc model in the restored landscape had

more than 5 % of individual explanation power compared

to the model for the un-restored landscape where only two

parameters (size and amount of riparian vegetation)

exceeded 5 % explanation power.

Both lake surface area and the proportion of riparian

vegetation had a positive effect on the presence of L.

Popul Ecol

123

62

pectoralis (Fig. 3). Examining the highest gain in the

likelihood of L. pectoralis’ presence at a probability level

of 0.5 in Fig. 3, differences between the two landscape

types appeared. In the restored landscape the highest gain

in probability of being present deriving from larger locality

size occurred in significantly smaller lakes and ponds

(0.05 ha [0.01; 0.20] (mean and 95 % CL)) compared to

the un-restored landscape (0.79 ha [0.30; 2.40]). The

influence of riparian vegetation on the presence of L.

pectoralis showed the same tendencies (i.e., that the

highest gain in probability occurred at a lower amount of

riparian vegetation in the un-restored landscape compared

to the restored). However, this difference was not signifi-

cant between landscape types (Fig. 3). The null-models did

not detect any random interaction effect between the

environmental variables and the two landscape types for

the presence of L. pectoralis in regard to lake surface area

(P = 0.45 [0.43; 0.46]) or proportion of riparian vegetation

(P = 0.57 [0.55; 0.58]). Thus, the differences between

landscapes in Figs. 2 and 3 are caused by differences in

habitat preferences of L. pectoralis in the two landscapes,

and are not artifacts of general differences between the un-

restored and restored landscape.

From the generated ROC-curves there was no difference

in the predictive power between the reduced model (AUC

= 0.84) and the 2DAICc selected model (AUC = 0.82) in

the un-restored landscape (z = -0.75, P = 0.45). In con-

trast, the 2DAICc selected model (AUC = 0.89) provided a

Table 1 Results of the stepwise

reduction of habitat variables in

the logistic regression models

accounting for the presence of

Leucorrhinia pecoralis in the

restored and un-restored

landscapes

Parameter Estimate 95 % CI Deviance AIC LRT Pr([Chi)

Un-restored landscapes

Intercept -5.97 (-9.14, -3.36) – – – –

log(surface area) 0.62 (0.32, 0.97) 93.88 97.88 19.6 <0.001

Floating vegetation – – 74.28 80.28 0.88 0.35

Submerged vegetation – – 73.4 81.4 1.12 0.29

Vegetation\ 1 m 0.03 (0.001, 0.06) 78.29 82.29 4.01 <0.05

Minimum edge slope – – 72.27 82.27 0.01 0.91

Restored landscape

Intercept -7.73 (-14.17, -3.22) – – – –

log(surface area) 1.11 (0.42, 2.13) 80.38 84.38 14.18 <0.001

Submerged vegetation – – 58.54 68.54 2.17 0.14

Vegetation\ 1 m 0.16 (0.05, 0.30) 75.44 79.44 9.24 <0.01

Shading – – 65.68 71.68 3.53 0.06

Minimum edge slope – – 62.16 70.16 3.62 0.06

Significant values are given in bold

Fig. 2 Independent

contributions (percentage of the

total independent explained

variance) of landscape predictor

variables in the 2DAICc

parameter selected models of

the restored landscape (white

bars) and un-restored

landscapes (black bars)

Popul Ecol

123

63

better prediction compared to the model containing only

the significant parameters (AUC = 0.80) in the restored

landscape (z = -1.98, P\ 0.05). This finding supports the

results of the hierarchical partitioning analysis showing

that the distribution of L. pectoralis in the restored land-

scape is much more dependent on additional habitat char-

acteristics than just habitat size compared to its

predominant role in the un-restored landscape.

Discussion

Our results revealed that the same two habitat variables—

lake size and the proportion of short riparian vegetation—

are significantly positive parameters determining the pres-

ence of L. pectoralis in lakes in both the un-restored and

the restored Estonian landscapes. However, the relative

importance of the different variables varied greatly

between the two landscape types. The species was found to

be much more habitat specific when the ecological niche

was derived from the restored landscape.

The general positive effect of lake area on the presence

of L. pectoralis probably reflects the underlying relation-

ships between habitat size and population size and stability

(Connor et al. 2000; Dunstan and Johnson 2006). Overall,

larger lakes are more temporally stable compared to smaller

water bodies and their age and turnover time between dry

and water-filled stages are longer (Sand-Jensen and Staehr

2009; Staehr et al. 2012). Therefore, the timeframe during

which a species of a given fundamental niche exists,

increases with lake size. The positive relationship to the

proportion of riparian vegetation is more directly linked to

the ecology of L. pectoralis, that is the resources needed to

complete its life cycle (Dennis et al. 2003). The shallow

vegetation growing along the lake perimeter acts both as

larval feeding habitat, larval emergence site, and resting and

egg-laying site for the adults (Wildermuth 1992; Sternberg

et al. 2000). Therefore, this vegetation type supports major

stages and activities of the species’ life cycle.

Differences in the ecological niche between the two

landscape types could indicate that a shift from the niche

defined in the un-restored to the restored landscape enables

the species to occupy more diverse microhabitats in the

restored landscape. Alternatively, it could indicate, that the

niche defined in the un-restored landscape is constrained by

the present landscape properties, while the actual niche is

better reflected in the restored landscape. Thus, the full

habitat spectra in which L. pectoralis can sustain a breed-

ing population are not present within the un-restored

landscape causing the observed differences in the ecolog-

ical niche between the two landscapes. We have no data on

the ecological niche of the species in our study area prior to

the agricultural shifts in the early 1990s. However, we

believe that the niche defined in the restored landscape

comes closer to the actual (and historical) ecological niche

of the species in the region. When niche shifts occur, the

adaptation period often requires much longer time spans

([25 years) than the one available in the present study

(Pearman et al. 2008). Also, differences in survival and

reproductive output between the present habitats (i.e.,

between lakes and small ponds) are known to favor niche

stasis (Holt and Gomulkiewicz 2004; Pearman et al. 2008).

Fig. 3 Probability of presence of L. pectoralis in relation to locality size (left) and percentage of riparian vegetation (right). The central tendency

is shown by the solid line and 95 % confidence limits are shown as dotted lines

Popul Ecol

123

64

Our finding that habitat dependencies of L. pectoralis

have lower explanation power in the un-restored than the

restored landscape (lower AICc values, Table S2 in ESM)

is evidence of delayed extinction, because past landscape

characteristics explain current species richness better than

present-day characteristics (Kuussaari et al. 2009). The

potential delayed extinction in the un-restored landscape

might cause an erroneous niche definition simply because

the distribution of the species is not in temporal equilib-

rium with the changing environment (Schurr et al. 2012).

Colwell and Rangel (2009) highlighted dispersal limi-

tation, lack of habitat characters in the contemporary bio-

tope, and species interactions as main reasons for the

mismatch between the realized and fundamental niche of a

species. In our study, parts of L. pectoralis’ fundamental

niche requirements are probably not available in the un-

restored landscape resulting in the mismatch between the

realized niche and the underlying fundamental niche. The

common dispersal distance of L. pectoralis is up to 27 km

(Jaeschke et al. 2013) compared to a maximum of 15 km

among study sites in the different study areas making it

unlikely that dispersal limitations should restrict the real-

ized niche. However, lower dispersal distances of the

individuals within the study sites are expected to accom-

pany afforestation because overview and flight possibility

are limited in forested areas (Chin and Taylor 2009).

This study does not define species interactions. They

could influence the realized niche of L. pectoralis, though

we do not suspect that species interactions are the main

reasons for the observed differences in the realized niche

between the two landscapes. Given that the same dragonfly

species communities are present across the study sites

(Martin 2013; L.L. Iversen, personal communication),

there is no reason to believe that there should be differ-

ences in interspecific dragonfly competition between the

two landscapes prior to habitat restoration of one of them.

This argument does not imply that there are no differences

in interspecific competition between the two landscapes

nowadays and that this could affect the realized niche

differently. However, it seems reasonable that these dif-

ferences would be closely linked to structural changes of

the restored landscape; i.e., changes in species interactions

caused by differences in habitat availability.

The conclusion based solely on the three un-restored

landscapes suggests that L. pectoralis becomes more of a

habitat generalist in Estonia, located in the central part of

its distribution range, than at the western European borders

of the range. However, the results obtained in the restored

Estonian landscape do suggest, that the species is much

more habitat specific here than in the un-restored land-

scape. The differences in preferences regarding locality

size (Fig. 3) and the dominant effect of size alone in un-

restored landscape contra the shared effect of size and

habitat parameters in the restored landscape in Estonia are

probably caused by the different history of land use

between the landscape types. Afforestation and abandon-

ment of extensive agricultural practices in the un-restored

landscape have probably lead to increased shading and

sedimentation caused by in-growth of emergent plants and

increased input of terrestrial material (Butler and Malanson

1995; Renwick 2006). Hence many ponds have become

unsuitable for L. pectoralis in the un-restored landscape

while the lakes still hold the species, simply due to their

larger size and slower rate of environmental and biological

deterioration. Reactivation of the old agricultural landscape

has created suitable habitat characteristics within the ponds

in the restored landscape, enabling L. pectoralis to re-oc-

cupy these breeding sites.

In summary our study demonstrates, firstly, the problems

of extracting relevant habitat components and defining a

species environmental niche from local species distribution,

and, secondly, the importance of incorporating landscape

history when defining the realized niche based on the

presence of a species according to local habitat character-

istics. If a species realized niche is derived from local

species distributions without incorporating landscape his-

tory it might lead to an erroneous niche definition. Ideally

ecological niches should be derived from preserved land-

scapes where the species distribution is in an equilibrium

state. However, such situations might not always be present,

especially not for declining or rare species restricted to a

few sites. One way to circumvent these problems is by using

landscape restorations to gain knowledge on the species

habitat dependencies before habitat degradation became

widespread. However, this can only be done if the restora-

tion mitigation changes the landscape towards its former

state and that the distribution of the species in focus have

been shaped by the former landscape.

Acknowledgments This study was supported by the Danish Min-

istry of Higher Education and Science (Project 0604-02230B), EU

LIFE? project DRAGONLIFE (LIFE08 NAT/EE/000257), the

Estonian Science Foundation (Grant 9051) and the Estonian Ministry

of Education and Science (Project SF0180012s09).

References

Agresti A (2010) Analysis of ordinal categorical data, 2nd edn. Wiley,

New Jersey

Baselga A, Lobo JM, Svenning JC, Aragon P, Araujo MB (2012)

Dispersal ability modulates the strength of the latitudinal

richness gradient in European beetles. Glob Ecol Biogeogr

21:1106–1113

Beale CM, Lennon JJ (2012) Incorporating uncertainty in predictive

species distribution modelling. Philos Trans B 367:247–258

Brauner O (2006) Einjahrige Entwicklung von Leucorrhinia pec-

toralis und Brachytron pratense in einem Kleingewasser

Popul Ecol

123

65

Nordostbrandenburgs. Libellula 25:61–75 (in German withEnglish abstract)

Brito JC, Acosta AL, Alvares F, Cuzin F (2009) Biogeography and

conservation of taxa from remote regions: an application of

ecological-niche based models and GIS to North-African Canids.

Biol Conserv 142:3020–3029

Burnham KP, Anderson DR (2002) Model selection and multimodel

inference: a practical information-theoretic approach, 2nd edn.

Springer, New York

Butler DR, Malanson GP (1995) Sedimentation rates and patterns in

beaver ponds in a mountain environment. Geomorphology

13:255–269

Chevan A, Sutherland M (1991) Hierarchical partitioning. Am Stat

45:90–96

Chin KS, Taylor PD (2009) Interactive effects of distance and matrix

on the movements of a peatland dragonfly. Ecography

32:715–722

Colwell RK, Rangel TF (2009) Hutchinson’s duality: the once and

future niche. P Natl Acad Sci USA 106:19651–19658

Connor EF, Courtney AC, Yoder JM (2000) Individuals–area

relationships: the relationship between animal population density

and area. Ecology 81:734–748

Corbet PS, Suhling F, Soendrerath D (2006) Voltinism of Odonata: a

review. Int J Odonatol 9:1–44

Council of the European Union (1992) Council directive 92/43/EEC

on the conservation of natural habitats and of wild fauna and

flora. Council of the European Union, Strasbourg

Cramer VA, Hobbs RJ, Standish RJ (2008) What’s new about old

fields? Land abandonment and ecosystem assembly. Trends Ecol

Evol 23:104–112

Davis AJ, Jenkinson LS, Lawton JH, Shorrocks B, Wood S (1998)

Making mistakes when predicting shifts in species range in

response to global warming. Nature 391:783–786

DeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing the

areas under two or more correlated receiver operating charac-

teristic curves: a nonparametric approach. Biometrics

44:837–845

Dennis RLH, Shreeve TG, Van Dyck H (2003) Towards a functional

resource-based concept for habitat: a butterfly biology view-

point. Oikos 102:417–426

Dennis RLH, Dapporto L, Dover JW (2014) Ten years of the

resource-based paradigm: the biotope-habitat issue and implica-

tions for conserving butterfly diversity. J Insect Biodivers 2:1–32

Dijkstra KD, Lewington R (2006) Fieldguide for identifying dragon-

flies of Great Britain and Europe. British Wildlife Publishing,

Dorset

Dingemanse NJ, Kalkman VJ (2008) Changing temperature regimes

have advanced the phenology of Odonata in the Netherlands.

Ecol Entomol 33:394–402

Dolny A, Harabis F (2012) Underground mining can contribute to

freshwater biodiversity conservation: allogenic succession forms

suitable habitats for dragonflies. Biol Conserv 145:109–117

Dunstan PK, Johnson CR (2006) Linking richness, community

variability, and invasion resistance with patch size. Ecology

87:2842–2850

Ferrier S (2002) Mapping spatial pattern in biodiversity for regional

conservation planning: where to from here? Syst Biol

51:331–363

Foster GN (1996) Leucorrhinia pectoralis (Charpentier, 1825). In:

van Helsdingen PJ, Willemse L, Speight MCD (eds) Background

information on invertebrates of the habitats directive and the

bern convention. Part 1—crustacea, coleoptera and lepidoptera.

European Invertebrate Survey, Leiden, pp 292–307

Foster DR, Swanson F, Aber J, Burke I, Brokaw N, Tilman D, Knapp

A (2003) The importance of land-use legacies to ecology and

conservation. Bioscience 53:77–88

Gotelli NJ, Graves CR, Rahbek C (2010) Macroecological signals of

species interactions in the Danish avifauna. P Natl Acad Sci

USA 107:5030–5035

Guisan A, Thuiller W (2005) Predicting species distribution: offering

more than simple habitat models. Ecol Lett 8:993–1009

Heikkinen RK, Luoto M, Kuussaari M, Toivonen T (2007) Modelling

the spatial distribution of a threatened butterfly: impacts of scale

and statistical technique. Landscape Urban Plan 79:347–357

Holt RD, Gomulkiewicz R (2004) Conservation implications of niche

conservatism and evolution in heterogeneous environments. In:

Ferriere R, Dieckmann U, Couvet D (eds) Evolutionary conser-

vation biology. Cambridge University Press, Cambridge,

pp 244–264

Iversen LL, Rannap R, Thomsen PF, Kielgast J, Sand-Jensen K

(2013) How do low dispersal species establish large range sizes?

The case of the water beetle Graphoderus bilineatus. Ecography

36:770–777

Jackson ST, Overpeck JT (2000) Responses of plant populations and

communities to environmental changes of the late quaternary.

Paleobiology 26:194–220

Jaeschke A, Bittner T, Reineking B, Beierkuhnlein C (2013) Can they

keep up with climate change? - Integrating specific dispersal

abilities of protected Odonata in species distribution modelling.

Insect Conserv Diver 6:93–103

Kuussaari M, Bommarco R, Heikkinen RK, Helm A, Krauss J,

Lindborg R, Ockinger E, Partel M, Pino J, Roda F, Stefanescu C,

Teder T, Zobel M, Steffan-Dewenter I (2009) Extinction debt: a

challenge for biodiversity conservation. Trends Ecol Evol

24:564–571

Lindborg R, Eriksson O (2004) Historical landscape connectivity

affects present plant species diversity. Ecology 85:1840–1845

Mac Nally R (2002) Multiple regression and inference in ecology and

conservation biology: further comments on identifying important

predictor variables. Biodivers Conserv 11:1397–1401

Mac Nally R, Walsh CJ (2004) Hierarchical partitioning public-

domain software. Biodivers Conserv 13:659–660

Margules CR, Pressey RL (2000) Systematic conservation planning.

Nature 405:243–253

Martin M (2013) Eesti kiilide maaraja. Tallinn, Keskkonnaminis-

teerium (in Estonian)Moilanen A, Hanski I (1998) Metapopulation dynamics: effects of

habitat quality and landscape structure. Ecology 79:2503–2515

Oliver T, Hill JK, Thomas CD, Brereton T, Roy DB (2009) Changes

in habitat specificity of species at their climatic range bound-

aries. Ecol Lett 12:1091–1102

Pearman PB, Guisan A, Broennimann O, Randin CF (2008) Niche

dynamics in space and time: species distribution modeling meets

phylogenetics. Trends Ecol Evol 23:149–158

Peterson U, Aunap R (1998) Changes in agricultural land use in

Estonia in the 1990s detected with multitemporal Landsat MSS

imagery. Landscape Urban Plan 41:193–201

Peterson AT, Soberon J, Pearson RG, Anderson RP, Martınez-Meyer

E, Nakamura M, Araujo MB (2011) Ecological niches and

geographic distributions. Princeton University Press, Princeton

Pulliam HR (1988) Sources, sinks, and population regulation. Am Nat

132:652–661

Pulliam HR (2000) On the relationship between niche and distribu-

tion. Ecol Lett 3:349–361

Rannap R, Lohmus A, Briggs L (2009) Restoring ponds for

amphibians: a success story. Hydrobiologia 634:87–95

Rannap R, Lohmus A, Tammaru T, Briggs L, de Vries W,

Biebelriether F (2012) Northern natterjack toads (Bufo calamita)

select breeding habitats that promot rapid development. Behav-

ior 149:737–754

Renwick WH (2006) Small artificial ponds in the United States:

impacts on sedimentation and carbon budge. In: Bernard JM,

Popul Ecol

123

66

Glysson D (eds) Proceedings of the 8th federal interagency

sedimentation conference, subcommittee on sedimentation,

Nevada, pp 738–744

Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC,

Muller M (2011) pROC: an open-source package for R and

S? to analyze and compare ROC curves. BMC Bioinformat

12:77

Sahlen G, Bernard R, Rivera AC, Ketelaar R, Suhling F (2004)

Critical species of Odonata in Europe. Int J Odonatol 7:385–398

Sand-Jensen K, Staehr PA (2009) Net heterotrophy in small Danish

lakes: a widespread feature over gradients in trophic state and

land cover. Ecosystems 12:336–348

Schorr M (1990) Grundlagen zu einem artenhilfsprogramm Libellen

der bundesrepublik Deutschland. Ursus Scientific Publishers,

Bilthoven (in German)Schurr FM, Pagel J, Cabral JS, Groeneveld J, Bykova O, O’ Hara RB,

Hartig F, Kissling WD, Linder HP, Midgley GF, Schroder B,

Singer A, Zimmermann NE (2012) How to understand species’

niches and range dynamics: a demographic research agenda for

biogeography. J Biogeogr 39:2146–2162

Soo T, Tullus A, Tullus H, Roosaluste E, Vares A (2009) Change

from agriculture to forestry: floristic diversity in young fast-

growing deciduous plantations on former agricultural land in

Estonia. Ann Bot Fenn 46:353–364

Southwood TRE (1977) Habitat, the templet for ecological strategies.

J Anim Ecol 46:337–365

Southwood TRE (1988) Tactics, strategies and templets. Oikos

52:3–18

Staehr PA, Baastrup-Spohr L, Sand-Jensen K, Stedmon C (2012)

Lake metabolism scales with lake morphometry and catchment

conditions. Aquat Sci 74:155–169

Stephens PA, Buskirk SW, del Rio CM (2007) Inference in ecology

and evolution. Trends Ecol Evol 22:192–197

Sternberg K, Schiel FJ, Buchwald R (2000) Leucorrhinia pectoralis.

In: Sternberg K, Buchwald R (eds) Die Libellen Baden-

Wurttembergs. Band 2 Großlibellen (Anisoptera). Literatur.

Verlag Eugen Ulmer, Stuttgart pp 415–427 (in German)Svenning JC, Skov F (2007) Ice age legacies in the geographical

distribution of tree species richness in Europe. Glob Ecol

Biogeogr 16:234–245

Symonds MRE, Moussalli A (2011) A brief guide to model selection,

multimodel inference and model averaging in behavioural

ecology using Akaike’s information criterion. Behav Ecol

Sociobiol 65:13–21

Thomaes A, Kervyn T, Maes D (2008) Applying species distribution

modelling for the conservation of the threatened saproxylic Stag

Beetle (Lucanus cervus). Biol Conserv 141:1400–1410

Toth S (1983) Libellen und ihre Biotope im Bakony-Gebirge. Folia

musei historico-naturalis Bakony 1983:45–54 (in German)Vares A, Jogiste K, Kull E (2001) Early growth of some deciduous

tree species on abandoned agricultural lands in Estonia. Balt For

7:52–58

Whittingham MJ, Stephens PA, Bradbury RB, Freckleton RP (2006)

Why do we still use stepwise modelling in ecology and

behaviour? J Anim Ecol 75:1182–1189

Wildermuth H (1992) Habitate und Habitatwahl der Großen

Moosjungfer (Leucorrhinia pectoralis) Charp. 1825 (Odonata,

Libellulidae). Z Okol Nat 1:3–22 (in German)Wildermuth H (1993) Populationsbiologie von Leucorrhinia pec-

toralis (Charpentier) (Anisoptera: Libellulidae). Libellula

12:269–275 (in German with English abstract)

Popul Ecol

123

67

68

Appendix - Chapter II

Fig. S1. Search time and accumulated dipnet records of L. pectoralis larvae from a study of 50 lakes.

Search time and accumulated dipnet records of L. pectoralis larvae from a study of 50 Swedish lakes using a

search time up to 90 min. All records of L. pectoralis were obtained within the first 40 min of searching (red

line).

69

Section S2:

Differences in monthly average temperatures within the timespan of the two generations studied (2009-

2011) were tested by two linear mixed models:

March, April and May:

“average temperature”= β0 + β1•”year"+ U1•”month"+ є

June and July

“average temperature”= β0 + β1•”year"+ U1•”month"+ є

Differences between years were tested by a likelihood ratio tests based on ML estimations and evaluated at a

5% significance level. We used temperature data from Southern Estonia (measured in Valga) and data was

acquired from the Estonian State Meteorological Institute (assessed the 16th of January 2015). We did not

find any differences in average temperature between the study years within the spring (p = 0.93) or summer

months (p = 0.12).

70

Table S1: The measured median, minimum and maximum values of parameters for restored and

unrestored landscapes.

Median (Min-Max)

1. Shading from surrounding trees (%)

Restored 2.5 (0-100)

Un-restored 2 (0-100)

2. Minimum edge slope (5, 10, 25, 45 or 90 °)

Restored 45 (10-90)

Un-restored 30 (5-90)

3. Submerged vegetation (%)

Restored 10 (0-100)

Un-restored 5 (0-100

4. Floating vegetation (%)

Restored 10 (0-100)

Un-restored 1 (0-75)

5. Riparian vegetation (%)

Restored 5 (0-25)

Un-restored 5 (0-75)

6. Vegetation higher than 1 m. (%)

Restored 0 (0-25)

Un-restored 1 (0-75)

7. Lake surface area (log(m2))

Restored 5.99 (4.38-12.91)

Un-restored 8.16 (3.91-13.66)

Table S2: AICC models from the initial set of seven habitat-parameters.

71

Unrestored landscape

(Intercept) Floating vegetation

Log(Surface area) Shadow

Min. edge slope

Submerged vegetation

Vegetation > 1m

Vegetation < 1m df logLik AICC ΔAICC

β - estimate

-5.97 0.62 0.03 3 -37.14 80.6 0

-5.44 0.58 -0.02 0.03 4 -36.29 81.2 0.55

-5.73 0.02 0.59 -0.03 0.04 5 -35.38 81.6 1.02

-6.22 0.01 0.63 0.03 4 -36.70 82 1.35

-5.97 0.64 -0.01 0.03 4 -36.70 82 1.36

-5.71 0.64 2 -39.15 82.5 1.84

-5.45 0.60 -0.02 -0.01 0.03 5 -35.83 82.5 1.93

-6.06 0.62 0.01 0.03 4 -37.08 82.7 2.12

-5.71 0.02 0.61 -0.03 -0.01 0.04 6 -34.79 82.8 2.21

-6.05 0.62 0.00 0.03 4 -37.14 82.8 2.23

-6.23 0.01 0.65 -0.01 0.03 5 -36.14 83.2 2.53

-5.51 0.58 -0.02 0.01 0.03 5 -36.22 83.3 2.7

-5.53 0.58 -0.02 0.00 0.03 5 -36.29 83.5 2.84

-5.31 0.61 -0.01 3 -38.69 83.7 3.1

-5.79 0.02 0.59 -0.03 0.01 0.04 6 -35.31 83.9 3.24

-5.91 0.02 0.60 -0.03 0.00 0.04 6 -35.36 84 3.35

-5.73 0.66 -0.01 3 -38.88 84.1 3.48

-5.87 0.01 0.64 3 -38.91 84.2 3.53

-6.32 0.01 0.63 0.01 0.03 5 -36.64 84.2 3.53

-6.07 0.65 -0.01 0.01 0.03 5 -36.65 84.2 3.56

-6.37 0.01 0.63 0.00 0.03 5 -36.68 84.2 3.63

-5.99 0.64 0.00 -0.01 0.03 5 -36.70 84.3 3.67

-5.37 0.62 0.00 3 -39.02 84.4 3.76

-5.76 0.64 0.00 3 -39.12 84.6 3.96

-5.53 0.60 -0.02 -0.01 0.01 0.03 6 -35.76 84.8 4.15

-5.48 0.60 -0.02 0.00 -0.01 0.03 6 -35.83 84.9 4.3

-6.20 0.63 0.00 0.01 0.03 5 -37.07 85 4.4

-5.45 0.01 0.61 -0.02 4 -38.26 85.1 4.47

-5.78 0.02 0.61 -0.03 -0.01 0.01 0.04 7 -34.72 85.1 4.52

-5.83 0.02 0.62 -0.03 0.00 -0.01 0.04 7 -34.78 85.3 4.63

-6.32 0.01 0.66 -0.01 0.01 0.03 6 -36.09 85.4 4.81

-5.32 0.63 -0.01 -0.01 4 -38.44 85.4 4.83

-6.32 0.01 0.66 0.00 -0.01 0.03 6 -36.13 85.5 4.89

-4.87 0.58 -0.01 -0.01 4 -38.50 85.6 4.97

-5.66 0.59 -0.02 0.00 0.01 0.03 6 -36.21 85.7 5.04

-5.89 0.01 0.66 -0.01 4 -38.58 85.7 5.11

-5.35 0.61 -0.01 0.01 4 -38.66 85.9 5.27

-5.31 0.64 -0.01 -0.01 4 -38.69 86 5.34

-5.55 0.01 0.63 0.00 4 -38.80 86.2 5.55

-6.03 0.02 0.60 -0.03 0.00 0.01 0.04 7 -35.28 86.2 5.63

-5.77 0.66 -0.01 0.00 4 -38.86 86.3 5.67

-5.92 0.01 0.64 0.00 4 -38.88 86.3 5.72

-6.53 0.01 0.64 0.00 0.01 0.03 6 -36.61 86.5 5.85

-6.15 0.65 0.00 -0.01 0.01 0.03 6 -36.65 86.5 5.92

-5.42 0.62 0.00 0.00 4 -39.00 86.6 5.96

-5.45 0.01 0.63 -0.02 -0.01 5 -37.95 86.8 6.16

-5.01 0.01 0.59 -0.02 -0.01 5 -38.08 87 6.41

72

-5.61 0.61 -0.02 0.00 -0.01 0.01 0.04 7 -35.76 87.2 6.59

-4.80 0.60 -0.01 -0.01 -0.01 5 -38.18 87.2 6.63

-5.49 0.01 0.61 -0.02 0.01 5 -38.23 87.3 6.71

-5.96 0.02 0.62 -0.03 0.00 -0.01 0.01 0.04 8 -34.70 87.6 7

-5.48 0.01 0.64 -0.01 -0.01 5 -38.41 87.7 7.08

-5.36 0.63 -0.01 -0.01 0.00 5 -38.41 87.7 7.08

-6.48 0.01 0.66 0.00 -0.01 0.01 0.03 7 -36.07 87.8 7.22

-4.91 0.59 -0.01 -0.01 0.00 5 -38.48 87.8 7.23

-5.93 0.01 0.67 -0.01 0.00 5 -38.56 88 7.38

-5.35 0.64 -0.01 -0.01 0.00 5 -38.68 88.2 7.62

-5.60 0.01 0.63 0.00 0.00 5 -38.78 88.4 7.82

-4.93 0.01 0.61 -0.02 -0.01 -0.01 6 -37.70 88.7 8.03

-5.49 0.01 0.63 -0.02 -0.01 0.00 6 -37.93 89.1 8.49

-5.06 0.01 0.59 -0.02 -0.01 0.00 6 -38.05 89.4 8.74

-4.84 0.60 -0.01 -0.01 -0.01 0.00 6 -38.17 89.6 8.97

-5.52 0.01 0.65 -0.01 -0.01 0.00 6 -38.40 90.1 9.43

-4.97 0.01 0.61 -0.02 -0.01 -0.01 0.00 7 -37.69 91.1 10.45

-0.35 -0.03 0.03 3 -45.98 98.3 17.67

-0.52 0.01 -0.03 0.04 4 -45.14 98.9 18.23

-0.43 -0.03 0.01 0.03 4 -45.79 100 19.52

-0.13 -0.03 0.00 0.03 4 -45.82 100 19.59

-0.27 -0.03 0.00 0.03 4 -45.86 100 19.66

-0.42 0.02 -0.03 -0.01 0.04 5 -44.91 101 20.05

-0.61 0.02 -0.03 0.01 0.04 5 -44.92 101 20.08

-0.34 0.01 -0.03 0.00 0.03 5 -45.05 101 20.33

-0.69 0.03 2 -48.65 102 20.84

-0.02 -0.03 -0.01 0.00 0.03 5 -45.67 102 21.58

-0.23 -0.03 0.00 0.01 0.03 5 -45.68 102 21.59

-0.35 -0.03 0.00 0.01 0.03 5 -45.69 102 21.61

-0.51 0.02 -0.03 -0.01 0.01 0.04 6 -44.72 103 22.03

-0.21 0.02 -0.03 0.00 -0.01 0.03 6 -44.79 103 22.17

-0.83 0.01 0.03 3 -48.27 103 22.26

-0.46 0.01 -0.03 0.00 0.01 0.03 6 -44.86 103 22.32

-0.46 -0.01 0.03 3 -48.46 103 22.63

-0.63 0.00 0.03 3 -48.57 104 22.86

-0.73 0.00 0.03 3 -48.61 104 22.93

0.07 -0.02 2 -49.75 104 23.03

0.49 -0.02 -0.01 3 -48.92 104 23.54

-0.12 -0.03 0.00 0.00 0.01 0.03 6 -45.55 104 23.7

-0.75 0.01 0.00 0.03 4 -48.13 105 24.2

-0.62 0.01 0.00 0.03 4 -48.14 105 24.21

-0.32 0.02 -0.03 0.00 -0.01 0.01 0.04 7 -44.64 105 24.3

-0.03 0.01 -0.03 3 -49.33 105 24.37

-0.87 0.01 0.00 0.03 4 -48.23 105 24.4

-0.37 -0.01 0.00 0.03 4 -48.36 105 24.66

-0.49 0.00 0.00 0.03 4 -48.44 105 24.83

0.02 -0.02 0.01 3 -49.63 106 24.96

-0.67 0.00 0.00 0.03 4 -48.55 106 25.03

0.39 0.01 -0.03 -0.01 4 -48.57 106 25.08

0.10 -0.02 0.00 3 -49.72 106 25.16

-0.26 1 -52.02 106 25.47

73

0.58 -0.02 -0.01 0.00 4 -48.85 106 25.63

0.44 -0.02 -0.01 0.01 4 -48.85 106 25.64

0.12 -0.01 2 -51.23 107 26

-0.52 0.01 0.00 0.00 0.03 5 -47.97 107 26.17

-0.09 0.01 -0.03 0.01 4 -49.19 107 26.32

-0.79 0.01 0.00 0.00 0.03 5 -48.10 107 26.43

-0.66 0.01 0.00 0.00 0.03 5 -48.12 107 26.47

0.03 0.01 -0.03 0.00 4 -49.27 107 26.48

-0.40 -0.01 0.00 0.00 0.03 5 -48.35 108 26.94

0.49 0.01 -0.03 -0.01 0.00 5 -48.45 108 27.13

0.05 -0.02 0.00 0.01 4 -49.61 108 27.17

0.33 0.01 -0.03 -0.01 0.01 5 -48.48 108 27.2

-0.34 0.01 2 -51.86 108 27.25

-0.29 0.00 2 -52.00 108 27.55

-0.25 0.00 2 -52.01 108 27.57

0.52 -0.02 -0.01 0.00 0.01 5 -48.79 108 27.82

0.05 0.00 -0.01 3 -51.12 109 27.95

0.18 -0.01 0.00 3 -51.19 109 28.09

0.12 -0.01 0.00 3 -51.23 109 28.17

-0.55 0.01 0.00 0.00 0.00 0.03 6 -47.96 109 28.51

-0.03 0.01 -0.03 0.00 0.01 5 -49.14 109 28.52

0.43 0.01 -0.03 -0.01 0.00 0.01 6 -48.38 110 29.36

-0.31 0.01 0.00 3 -51.84 110 29.38

-0.36 0.01 0.00 3 -51.84 110 29.38

-0.27 0.00 0.00 3 -52.00 110 29.71

0.12 0.01 -0.01 0.00 4 -51.06 111 30.06

0.04 0.00 -0.01 0.00 4 -51.12 111 30.17

0.18 -0.01 0.00 0.00 4 -51.19 111 30.32

-0.33 0.01 0.00 0.00 4 -51.82 112 31.58

0.12 0.01 -0.01 0.00 0.00 5 -51.06 113 32.35

Restored landscape

(Intercept) Floating vegetation

Log(Surface area) Shadow

Min. edge slope

Submerged vegetation

Vegetation > 1m

Vegetation < 1m df logLik AICC ΔAICC

β - estimate

-9.85 1.09 -0.06 0.06 0.19 5 -29.27 69.6 0

-10.34 1.09 -0.05 0.06 0.01 0.18 6 -28.18 69.9 0.28

-10.91 1.13 0.06 0.02 0.17 5 -29.83 70.7 1.11

-7.19 1.07 -0.06 0.17 4 -31.08 70.9 1.25

-7.82 1.10 -0.06 0.01 0.16 5 -30.15 71.4 1.76

-10.25 1.13 0.05 0.17 4 -31.43 71.5 1.94

-9.92 0.01 1.10 -0.06 0.06 0.18 6 -29.15 71.8 2.22

-9.86 1.09 -0.06 0.06 0.00 0.19 6 -29.27 72.1 2.45

-10.51 1.10 -0.05 0.06 0.02 -0.05 0.17 7 -28.00 72.1 2.46

-11.10 1.14 0.06 0.02 -0.08 0.16 6 -29.30 72.1 2.49

-8.39 1.15 0.01 0.16 4 -31.80 72.3 2.68

-10.34 0.00 1.09 -0.05 0.06 0.01 0.18 7 -28.18 72.4 2.83

-7.73 1.11 0.16 3 -33.10 72.6 3

-7.39 0.01 1.07 -0.06 0.16 5 -30.78 72.6 3.03

-10.79 0.00 1.11 0.06 0.02 0.17 6 -29.74 73 3.37

-7.24 1.07 -0.06 0.01 0.17 5 -31.06 73.2 3.59

74

-7.88 0.00 1.10 -0.06 0.01 0.16 6 -30.06 73.7 4.04

-7.82 1.10 -0.05 0.01 -0.03 0.16 6 -30.08 73.7 4.08

-10.19 1.13 0.06 -0.02 0.17 5 -31.39 73.8 4.23

-10.27 0.00 1.13 0.05 0.17 5 -31.43 73.9 4.3

-8.31 1.14 0.02 -0.06 0.15 5 -31.49 74 4.42

-11.06 -0.01 1.13 0.07 0.02 -0.09 0.17 7 -29.09 74.2 4.61

-9.93 0.01 1.10 -0.06 0.06 0.01 0.18 7 -29.15 74.4 4.76

-8.39 0.00 1.15 0.01 0.16 5 -31.80 74.7 5.04

-10.53 0.00 1.10 -0.05 0.06 0.02 -0.05 0.18 8 -27.99 74.7 5.08

-7.92 0.00 1.13 0.16 4 -33.05 74.8 5.18

-7.67 1.11 -0.01 0.16 4 -33.09 74.9 5.26

-7.47 0.01 1.08 -0.06 0.02 0.16 6 -30.75 75 5.42

-7.87 0.00 1.10 -0.05 0.01 -0.02 0.16 7 -30.02 76.1 6.51

-10.20 0.00 1.13 0.06 -0.02 0.17 6 -31.39 76.3 6.67

-8.27 0.00 1.14 0.02 -0.06 0.15 6 -31.48 76.5 6.85

-7.85 0.00 1.12 -0.01 0.15 5 -33.04 77.1 7.52

-6.99 0.88 -0.04 0.04 4 -34.64 78 8.36

-5.21 0.88 -0.04 3 -35.81 78 8.43

-7.83 0.91 -0.03 0.05 0.01 5 -33.57 78.2 8.61

-5.83 0.92 -0.04 0.01 4 -35.00 78.7 9.1

-5.75 0.01 0.92 -0.04 4 -35.06 78.8 9.21

-8.55 0.99 0.05 0.01 4 -35.18 79 9.43

-7.50 0.01 0.92 -0.04 0.04 5 -34.01 79.1 9.47

-9.17 1.04 0.05 0.02 -0.11 5 -34.06 79.2 9.56

-8.40 0.97 -0.03 0.05 0.02 -0.08 6 -33.00 79.5 9.91

-6.58 1.00 0.01 3 -36.59 79.6 9.98

-5.96 0.96 2 -37.72 79.6 10.03

-7.70 0.97 0.04 3 -36.62 79.7 10.04

-7.05 0.89 -0.04 0.04 -0.03 5 -34.54 80.1 10.53

-8.00 0.01 0.94 -0.04 0.05 0.01 6 -33.34 80.2 10.59

-5.21 0.89 -0.04 -0.02 4 -35.76 80.2 10.61

-6.11 0.01 0.94 -0.04 0.01 5 -34.59 80.3 10.65

-6.74 1.03 0.02 -0.09 4 -35.79 80.3 10.65

-6.05 0.96 -0.03 0.01 -0.07 5 -34.64 80.3 10.73

-1.60 -0.05 0.03 0.14 4 -35.99 80.7 11.07

-0.24 -0.05 0.12 3 -37.26 80.9 11.33

-5.73 0.01 0.92 -0.04 -0.01 5 -35.04 81.2 11.55

-7.72 0.97 0.04 -0.05 4 -36.32 81.3 11.72

-8.60 0.00 1.00 0.05 0.01 5 -35.17 81.4 11.79

-5.87 0.96 -0.05 3 -37.50 81.4 11.8

-7.52 0.01 0.93 -0.04 0.04 -0.02 6 -33.96 81.4 11.84

-6.37 0.01 1.00 3 -37.53 81.5 11.85

-9.12 0.00 1.04 0.06 0.02 -0.12 6 -34.05 81.6 11.98

-8.04 0.00 1.00 0.04 4 -36.50 81.7 12.08

-6.77 0.00 1.02 0.01 4 -36.53 81.8 12.15

-8.41 0.00 0.97 -0.03 0.05 0.02 -0.08 7 -32.92 81.9 12.3

-1.97 -0.05 0.03 0.01 0.13 5 -35.46 82 12.37

-6.20 0.01 0.96 -0.04 0.01 -0.05 6 -34.38 82.3 12.67

-0.45 -0.05 0.01 0.12 4 -36.92 82.5 12.93

-6.81 0.00 1.04 0.02 -0.09 5 -35.78 82.6 12.99

-1.61 0.00 -0.05 0.03 0.13 5 -35.98 83 13.42

75

-1.62 -0.05 0.03 0.01 0.14 5 -35.98 83 13.42

-0.31 0.00 -0.05 0.12 4 -37.19 83.1 13.48

-0.25 -0.05 0.01 0.12 4 -37.26 83.2 13.6

-6.24 0.01 1.00 -0.04 4 -37.33 83.4 13.75

-8.01 0.00 1.00 0.04 -0.05 5 -36.22 83.5 13.89

-1.99 0.00 -0.05 0.03 0.01 0.14 6 -35.44 84.4 14.79

-1.97 -0.05 0.03 0.01 -0.01 0.13 6 -35.44 84.4 14.79

-1.80 0.03 0.13 3 -39.05 84.5 14.89

-0.53 0.12 2 -40.19 84.6 14.97

-2.29 0.03 0.01 0.13 4 -38.10 84.9 15.28

-0.47 0.00 -0.05 0.01 0.12 5 -36.91 84.9 15.28

-0.44 -0.05 0.01 -0.01 0.12 5 -36.91 84.9 15.29

-0.33 0.00 -0.05 0.01 0.12 5 -37.18 85.4 15.82

-0.80 0.01 0.12 3 -39.52 85.5 15.85

-1.63 0.00 -0.05 0.03 0.01 0.13 6 -35.97 85.5 15.85

0.38 -0.04 2 -40.66 85.5 15.92

-0.72 -0.04 0.03 3 -39.76 85.9 16.33

-2.30 -0.01 0.04 0.01 0.14 5 -37.59 86.2 16.61

-1.76 -0.01 0.03 0.14 4 -38.82 86.3 16.71

-0.44 0.00 0.12 3 -40.09 86.6 16.98

-1.75 0.03 -0.02 0.13 4 -39.00 86.7 17.08

-0.49 -0.02 0.12 3 -40.14 86.7 17.08

-2.28 0.03 0.01 -0.05 0.13 5 -37.82 86.7 17.09

-1.99 0.00 -0.05 0.03 0.01 -0.02 0.14 7 -35.41 86.9 17.28

0.18 -0.04 0.01 3 -40.35 87.1 17.5

0.23 0.01 -0.05 3 -40.38 87.2 17.57

-0.69 -0.01 0.01 0.12 4 -39.28 87.3 17.65

-0.75 0.01 -0.05 0.12 4 -39.31 87.3 17.69

-1.04 -0.04 0.03 0.01 4 -39.31 87.3 17.71

-0.46 0.00 -0.05 0.01 -0.01 0.12 6 -36.91 87.3 17.73

0.39 -0.04 -0.01 3 -40.65 87.7 18.1

-0.79 0.01 -0.04 0.03 4 -39.58 87.9 18.25

-2.29 -0.01 0.04 0.02 -0.06 0.13 6 -37.23 88 18.35

-0.71 -0.04 0.03 -0.01 4 -39.75 88.2 18.58

-1.71 -0.01 0.03 -0.02 0.13 5 -38.76 88.6 18.97

-0.39 0.00 -0.02 0.12 4 -40.04 88.8 19.16

-0.62 -0.01 0.01 -0.05 0.12 5 -39.02 89.1 19.49

0.11 0.01 -0.04 0.00 4 -40.21 89.1 19.51

0.19 -0.04 0.01 -0.03 4 -40.26 89.2 19.61

0.10 1 -43.60 89.3 19.65

-1.05 -0.04 0.03 0.01 -0.04 5 -39.19 89.4 19.84

0.24 0.01 -0.05 -0.01 4 -40.38 89.5 19.85

-1.04 0.00 -0.04 0.03 0.01 5 -39.26 89.6 19.99

-0.92 0.02 2 -42.79 89.8 20.17

-0.78 0.01 -0.04 0.03 -0.01 5 -39.57 90.2 20.6

-0.14 0.01 2 -43.02 90.2 20.63

-1.32 0.03 0.01 3 -42.01 90.4 20.82

0.16 -0.04 2 -43.41 91 21.41

0.13 0.00 2 -43.59 91.4 21.76

0.13 0.00 -0.04 0.01 -0.02 5 -40.16 91.4 21.78

-0.09 0.01 -0.06 3 -42.56 91.5 21.92

76

-0.86 0.02 -0.04 3 -42.59 91.6 21.97

-1.34 0.03 0.01 -0.07 4 -41.47 91.6 22.03

-1.05 0.00 -0.04 0.03 0.01 -0.03 6 -39.17 91.9 22.25

-0.88 0.00 0.02 3 -42.74 91.9 22.29

-0.07 0.00 0.01 3 -42.95 92.3 22.69

-1.29 -0.01 0.03 0.01 4 -41.83 92.4 22.75

0.20 0.00 -0.04 3 -43.39 93.2 23.58

-1.30 -0.01 0.03 0.01 -0.08 5 -41.21 93.5 23.86

0.00 0.00 0.01 -0.07 4 -42.44 93.6 23.97

-0.82 0.00 0.03 -0.04 4 -42.53 93.8 24.14

77

78

79

Chapter III

Time restricted flight ability influences dispersal and colonization rates in a group

of freshwater beetles

80

Time restricted flight ability influences dispersal and colonization rates in a group of freshwater beetles

Lars Lønsmann Iversen1, 2, Riinu Rannap3, Lars Briggs2, Kaj Sand-

Jensen

1

1Freshwater Biological Laboratory, Biological Institute, University of

Copenhagen, Denmark. 2Amphi Consult ApS, International Science Park

Odense, Denmark. 3

Institute of Ecology and Earth Sciences, University of

Tartu, Estonia

Abstract Variation in the ability to fly or not is a key mechanism for differences in local

species occurrences. It is increasingly acknowledged that physiological or

behavioral mechanisms rather than morphological differences may drive flight

abilities. However, our knowledge on the seasonal variability and stressors

creating non-morphological differences in flight abilities and how it scales to

local and regional occurrences is very limited particular for small, short-lived

species such as insects.

Here we examine how flight ability might vary across seasons and between

two closely related genera of freshwater beetles with similar geographical

ranges, life histories and dispersal related morphology. By combining flight

experiments of > 1,100 specimens with colonization rates in a meta-

community of 54 ponds in Northern and Eastern Europe, we have analyzed

the relationship between flight ability and spatio-environmental distribution of

the study genera.

We find profound differences in flight ability between the two study genera

across seasons. High flight ability for Acilius (97 %) and low for Graphoderus

81

(14 %) corresponded to the different colonization rates of newly created ponds.

While Acilius dispersed throughout the season, flight activity in Graphoderus

was restricted to stressed situations immediately after the emergence of

adults. Regional colonization ability of Acilius was independent of spatial

connectivity and mass effect from propagule sources. In contrast,

Graphoderus species were closely related to high connectivity between ponds

in the landscape.

Our data suggest that different dispersal potential can account for different

local occurrences of Acilius and Graphoderus. In general our findings provide

some of the first insights into the understanding of seasonal restrictions in flight

patterns of aquatic beetles and their consequences for species distributions.

Introduction The ability of species to move between habitats is essential for their

distribution (Kokko and López-Sepulcre 2006, Gaston 2009). The spatial

occurrence and temporal persistence of species are closely linked to their

dispersal ability and dispersal strategy (Clobert et al. 2004, Bowler and Benton

2005). Ultimately, these properties can shape the geographical ranges of

species and determine continental diversity gradients (Svenning and Skov

2004, Geber 2011, Baselga et al. 2012). Thus, understanding how dispersal

evolves and persists among and within species in a changing environment is a

key element when evaluating both current and future drivers of biodiversity

change (Pereira et al. 2010, Urban 2015).

Although many processes such as environmental stability, kin competition and

inbreeding influence dispersal traits (Ronce 2007, Matthysen 2012, Starrfelt

and Kokko 2012), abiotic barriers are often considered to be the main

evolutionary drivers of dispersal (Baguette et al. 2013, Kubisch et al. 2014).

Usually, environmental disturbance increases dispersal provided that suitable

habitats are evenly distributed (Gadgil 1971, Levin et al. 1984, Poethke et al.

2003). However, high dispersal rates come at a cost. When habitat quality

varies spatially, a high level of dispersal can lead to emigration from optimal

habitats and may reduce the use of high-quality habitats (Holt 1985, North et

al. 2011, Bonte et al. 2012). In theory, this mechanism is further enhanced

82

when the mortality risk during dispersal is high, e.g. when island species have

to cross open waters or aquatic species move across hostile terrain (Bonte et

al. 2012, Kubisch et al. 2014). Spatial isolation and local habitat conditions

have been shown to alter intraspecific dispersal properties either as a

consequence of local habitat connectivity (Soons and Heil 2002, Hanski et al.

2004) or at the front edge of expanding range margins (Thomas et al. 2001,

Simmons and Thomas 2004).

Traditionally differences in dispersal properties within and between species

have been studied by analysis of different morphological structures related to

the dispersal organs of species such as wing structure or flight musculature

(Hargreaves and Eckert 2014). But morphological dispersal traits do not

necessarily explain species occurrences (e.g. Schulz et al. 2012, Grönroos et

al. 2013) and dispersal events driven by physiological or behavioral factors

with no relationship to morphological traits are well documented among larger

vertebrates (Bekoff 1977, Duckworth and Badyaev 2007). These physiological

or behavioral factors have also been proposed to be important for invertebrate

species (Hanski et al. 2004, Clobert et al. 2009) yet we know very little about

the seasonal variability and stressors generating non-morphological

differences in flight abilities and how this may influence local and regional

occurrences of invertebrate species (Bilton et al. 2001, Bilton 2014).

In this study we aimed at demonstrating how non-morphological differences in

flight abilities might vary across season and how these differences scale to

colonization rates of newly created habitats in five invertebrate species. We

demonstrated for a group of aquatic beetles distributed within distinctively

isolated habitats (lakes and ponds). As a consequence of the spatial isolation

and short geological lifetime of freshwaters, aquatic beetles have undergone

an evolutionary selection towards strong dispersal abilities during the invasion

of these habitats (Bilton et al. 2001, Arribas et al. 2012). Following the

invasion, subsequent reduction in dispersal has been reported for several

species (Jackson 1952, 1956), offering an opportunity for studying differences

in flight abilities both within and between species. Using experiments and field

observations we studied five species of European diving beetles from two

genera, Acilius and Graphoderus (Coleoptera: Dytiscidae), with similar

83

geographical ranges, food resources and life histories (Nilsson and Holmen

1995, Bergsten and Miller 2005), but contrasting regional occurrences in

Europe (Foster 1996, Foster and Bilton 2014). Both genera have well

developed wings and flight musculature (Bergsten and Miller 2005, Kehl and

Dettner 2007), but contrasting flight capabilities (Foster 1996, Kehl and Dettner

2007). All five study species are associated with richly vegetated standing

waters (Nilsson and Holmen 1995, Miller and Bergsten 2016) and the two

genera often co-occur together in the same lakes and ponds (Nilsson et al.

1994). Within the study region of interest the three Graphoderus species are

most often found in larger permanent habitats, contra the two Acilius species

which are omnipresent occupying both large and small habitats (Lundkvist et

al. 2001, Iversen et al. 2013).

Within this framework we firstly tested whether differences in flight ability are

species-specific or clustered within the two genera. During three time periods,

covering species phenology, we experimentally tested individual flight ability of

each species. If differences in dispersal have been the evolutionary

mechanism behind the split between Acilius and Graphoderus, we would

expect the seasonal patterns in flight abilitiies to be conditional on the two

genera studied and not species-specific. Secondly, we examined the

hypothesis that differences in dispersal potential scales to local occurrences by

detecting the colonization rates of our study species within 54 newly created or

restored ponds. If difference in flight ability does affect colonization probability

we would expect concordance between the observed experimental flight

abilities and in situ colonization rates.

Methods Flight experiments

In order to quantify the potential ability to fly within the five study species, we

performed ex-situ flight experiments. The setup consisted of four-litre

containers with only one possible exit point, either flying from a vertical or a

horizontal position (Fig. S1). The experiment measured the ability to fly or not

to fly and it did not record flight distance per se. Adult diving beetles were

84

collected and the experiments were conducted in late spring (end of May 2011

and 2015), midsummer (first half of July 2015) and early autumn (first half of

September 2011 and 2015). The sampling periods represented three different

phases of the species’ life history: 1. the period just after emergence from the

pupae (summer). 2. the pre-hibernation period when beetle activity is related to

foraging (autumn), and 3. the breeding period (spring). The containers were

placed in transparent plastic boxes which were moved between a greenhouse

and an outside area in order to keep the ambient daytime temperatures

between ~20 and 35 °C, and nocturnal temperatures above 10 °C. The

temperature range covers temperatures promoting dispersal activity in aquatic

beetles (Csabai et al. 2012). Individuals were collected at 16 different sites in

southern Sweden and eastern Denmark (Table S1). During the two years of

study, a total of 1,128 individuals were collected and tested (Table S1).

Following field sampling all animals were acclimatized for two to three days

prior to the experiments. High-quality tap water derived from groundwater

reservoirs was used during the acclimatization and experiments to exclude any

potential fledge caused by toxins or chemical traces from predators. The

experiments were conducted for 14 days with a constant density of 10

specimens per container. The trials were conducted separately for each

species and sex. The animals were not fed before or during the experiment

and any dead or dying animals were removed immediately from the

experiments and excluded from further analysis. All collected specimens of the

strictly protected G. bilineatus were released to their respective localities after

the experiment. Specimens of the other common species were killed in 96%

ethanol and checked for the presence of well-developed wings and flight

musculature. All of the specimens used in the experiments had well developed

wings and flight musculature.

In order to confirm that our setup was appropriate to determine differences in

flight ability between different species of diving beetles, we performed flight

experiments on 13 additional species (supplementary Table S2).

Colonization of newly created and restored ponds

85

The ability to colonize new habitats was tested within a network of three to

five-year old restored or newly created ponds in the Haanja landscape park,

Southern Estonia. New ponds were created and completely overgrown ponds

reactivated through conservation actions, e.g. by removing bushes, tall and

dense vegetation, and muddy sediment (Rannap et al. 2009). The land use in

the area is a mixture of open, extensively used farm- and grassland and mixed

forest. All five study species are common in the natural lakes, cattle ponds,

and beaver floods scattered throughout the landscape (L. L. Iversen personal

observations). In mid-June 2011 we recorded the presence of the two study

genera in 54 restored or newly created ponds. Due to the late sampling date,

adult beetle abundance was expected to be low and the presence of larvae

was used as the measure of colonization. Identification of larvae was restricted

to genus, because intraspecific morphological characters within both Acilius

and Graphoderus have not yet been described (Mogens Holmen personal

communication). A semi-standardized dipnetting method was used to detect

the presence or absence of the study organisms (Iversen et al. 2013). Diving

beetle larvae were actively searched for during a 45 minute period at each site

by sweeping a hand dipnet (40 × 40 cm frame) through the vegetation and

detrital material. Along with larvae occurrence we recorded vegetation cover,

structured in four density classes, and shading from surrounding trees (see

supplementary appendix for details).

Analysis of data

Applying linear contrast models to our data, differences in flight ability between

Acilius and Graphoderus were evaluated. When tested explicitly, P-values

related to the effect of explanatory variables were evaluated at a 5 %

significance level and, unless stated otherwise, correspond to a χ ² - test.

Reported confidence intervals correspond to 95 % likelihood confidence

intervals.

Using the observed ability to leave the experimental setup by flight as the

response variable, differences in flight ability were tested across individuals

with a linear logistic regression model via a logit-link function. Genus

(Graphoderus or Acilius), season (spring, summer or autumn) and the

86

interaction between these factors were defined as the explanatory variables.

Initial models including species, sampling year and site as explanatory

variables produced the same significance of genera and season as the models

reported here.Nor, were there any differences in flight ability between sexes

(same model but with sex of the individual as the only explanatory variable, p =

0.50). Within genera, differences in flight ability between species were

evaluated in a linear logistic regression model. The time spent in the

experimental setup before flying away was modelled by a gaussian linear

mixed model and tested by a likelihood ratio test; genus was used as a fixed

factor, species and sites as random factors.

Differences in colonization ability between Acilius and Graphoderus in the

newly created and restored ponds were evaluated by a mixed linear logistic

regression model via a logit-link function. Genus and restoration measure

(restored or newly created) were used as fixed factors, while site was used as

a random factor. Using an Outlying Mean Index analysis (OMI, Dolédec et al.

2000), we tested whether the observed colonization patterns were

independent of the habitat characteristics (within the studied localities). OMI

analysis provided a measure of niche position and niche breadth from a subset

of localities, compared to the environmental space present in a region. A

principal component analysis (PCA) was performed on the z-scores of the four

vegetation variables and the level of shading at each site. This PCA was used

as the sampling unit to calculate the OMI and related to the occurrence matrix

of the two genera. The observed OMI value was compared to the random OMI

value generated from randomly selected sites with the same frequency of

occurrence as Acilius and Graphoderus. Using a Monte-Carlo test generated

from 1000 null-model repetitions, P-values tested whether or not the observed

OMI value was greater (more niche specific) than expected by chance. We

created a local proximity index (Gustafson and Parker 1994) for each sampling

pond based on the size of and euclidian distance to all other lakes and ponds

within 5 km of the given pond. The index incorporates both isolation and the

size of potential propagules (mass effect (Leibold et al. 2004)) by dividing area

with the squared distances to the sampling pond and summing this for all lakes

and ponds within the search radius of each sampling pond (Whitcomb et al.

87

1981). We tested for differences in relationship between colonization

probability and site proximity index between Acilius and Graphoderus using

linear logistic regression model via a logit-link function, including the interaction

and fixed term of the proximity index and genera as explanatory variables and

the observed presence/non-presence as response variable.

Figure 1: Flight properties of Acilius (grey) and Graphoderus (black). The percentage of

individuals flying during the experimental trials. The experiments were conducted across three

different seasons; just after post-pupae state in July (summer), prior to hibernation in

September (autumn), and post

88

Results Flight experiments

Flight experiments showed profound differences in flight ability between the

two genera. Across seasons, flight ability was high for Acilius (97 %) and low

for Graphoderus (14 %) (Fig. 1) and there was a significant interaction

between genera and season (|χ ² |= 8.3, df (degrees of freedom) = 1121, p <

0.05). Between the two Acilius species there was no difference in flight ability

across species ((|χ ² |=1.1, df=400, p = 0.52), across seasons (p = |χ ² |=2.1,

df=402, 0.36), or for that matter between seasons across species (|χ ² |=2.3,

df=402, p=0.32). In contrast, the three Graphoderus species showed a

systematic change between the seasons for all species (|χ ² |=176.3, df=716, p

< 0.001), but no difference between species (|χ ² |=1.5, df=716, p = 0.47).

Graphoderus species had a distinct peak in flight in summer (dispersal rate 45

%) and almost no flight for the rest of the year (dispersal rate 3 %) (Fig. 2,

Table S1).

The time spent in the experimental setup differed significantly between the

genera (|LRT| (likelihood ratio test) = 376.6, df= 499, p < 0.001). Acilius flew

almost immediately (within 0.78 [0.74; 0.82] days; mean and 95 % CL) after

being introduced, while on average the dispersing individuals of Graphoderus

left the containers after 5.37 [4.86; 5.88] days (Fig. 2).

The setup documented flight events across a large range of additional species

without detecting any systematic restrictions (supplementary Table S2). Thus,

we are confident that the results reflect true differences in flight ability between

the two genera and similarities of species within genera

89

Figure 2: Accumulated flight events after experimental introduction across the three study

periods of Acilius (grey) and Graphoderus (black). The accumulated percentages refer to the

number of flying individuals per species. A. canaliculatus n = 239, A. sulcatus n = 159, G.

bilineatus n = 24, G. cinereus n = 61, and G. zonatus n = 21.

Colonization of newly created and restored ponds

Individuals of Acilius and Graphoderus had colonized 81 % and 31 % of the 54

“new” Estonian ponds, respectively. Acilius showed a significantly higher

likelihood of colonizing the sites than Graphoderus (|χ ² |= 28.9, df= 106, p <

0.001). The OMI estimation for the two genera showed no difference other

than expected by random distribution (Acilius p = 0.14 and Graphoderus p =

0.74). Hence, habitat availability and possible differences in habitat conditions

between sites did not influence colonization rates. There was a significant

difference in probability of occurrence along a connectivity gradient between

the two genera (|χ ² |= 5.9, df= 104, p < 0.05). Graphoderus showed a higher

likelihood of occurrence with an increase in the proximity index (relative

change in odds by a factor of 2.56 [1.21; 5.55], Fig.3), whereas Acilius

90

occurrence was independent of landscape structure (odds changed by a factor

of 0.79 [0.47; 1.40], Fig.3).

Figure 3: Probability of presence in 54 newly created Estonian ponds of Graphoderus (blue)

and Acilius (green) in relation to landscape proximity index. The central tendency is shown by

the solid line and 95 % confidence limits are shaded.

Discussion Our results document that after being placed in the experimental setup, the

two Acilius species flew immediately at all times, while the three Graphoderus

species only dispersed in the post-pupae state. High dispersal rates of Acilius

throughout the season resemble the dispersal patterns of other highly

91

dispersive diving beetles (Lundkvist et al. 2002, Miguélez and Valladares

2008, Boda and Csabai 2013). In contrast, the restricted dispersal events in

Graphoderus may be related to the oogenesis-flight syndrome (Dingle 1972).

Individuals disperse in the pre-mature state, then settle and allocate energy for

reproduction and thereby create the observed bimodal dispersal pattern

(Fronhofer et al. 2015). So far, very few studies have documented the

oogenesis-flight syndrome in aquatic beetles (Landin 1980), and our results

provide insights into the understanding of seasonal restrictions in flight

patterns of aquatic beetles (Bilton 2014).

Both the flight experiments and the colonization rates supported the

expectations of a mobility syndrome, where a high dispersal potential leads to

higher dispersal rates within populations (Cote et al. 2010, Ducatez et al.

2012). When dispersing, in any of the three study periods, Acilius showed a

markedly higher dispersal rate than Graphoderus and also a higher

colonization rate of the newly created Estonian ponds. During summer, only 45

% of the Graphoderus specimens left the experimental containers. This could

be interpreted as a conservative dispersal strategy according to which

populations are polymorphic and include both flying and non-flying individuals

(Vepsäläinen 1978, Landin 1980). Given the short timespan during which

dispersal is possible and the high mortality risk during emigration, having a

proportion of non-flying individuals might increase the likelihood of population

survival.

The permanent, well-developed wings and flight muscles of both Acilius and

Graphoderus suggest that the observed differences in the ability to fly was not

driven by morphological differences. The slow dispersal of all three

Graphoderus species in the experiments suggests that dispersal decisions are

a conservative choice within this genus. Other studies have highlighted that

variation in flight abilities could be a consequence of differences in

physiological tolerances, resources utilizations and local competition (Hanski

et al. 2004, King and Roff 2010, Baguette et al. 2011). Future work should aim

to clarify the underlying triggers leading to the different flight abilities between

Acilius and Grapoderus.

92

The correlation between colonization probability of Graphoderus and the

proximity index in the Estonian ponds suggest that both mass effect and

connectivity promote the occurrence of the genera (Iversen et al. 2013, Heino

et al. 2015). In, contrast the chance of Acilius colonizing any of the study

ponds was independent of the property of the surrounding landscape (at least

within the spatial scale of this study). This result could suggest that not only

are there differences in dispersal rates between the two species but also in

potential dispersal distances during flight. However, specific information on

dispersal distances from propagule sources would be needed in order to

differentiate the landscape dependent effect of potential dispersal distance

contra the frequency of flight events in Acilius and Graphoderus.

Our results can also explain the regional conservation status of the studied

species. All three Graphoderus species are threatened locally in many regions

and have even become extinct in some countries, while the two Acilius species

are common throughout their range. It is likely, that this rarity of the

Graphoderus species is a consequence of the high turnover rate and

destruction of freshwater habitats caused by anthropogenic activities in

Western Europe (Foster 1996, Foster and Bilton 2014). The restricted

dispersal of the Graphoderus species could limit their ability to track these

habitat changes, leading to decline in population numbers and extinction. In

contrast, the two Acilius species might be able to track these environmental

changes, partly due to a strong dispersal ability. Our results support previous

findings highlighting the importance of integrating landscape connectivity and

stability into conservation strategies of species like the three studied

Graphodersu species (Iversen et al. 2013) and suggest that such approaches

are important even for conservation on a local scale.

In summary, our study shows distinct differences and seasonal constraints in

flight ability within a group of diving beetles. Given the clear contrast between

genera and the remarkably similar dispersal properties among species of the

same genera, our data suggest that species flight ability is important for

profound differences in local occurrences between Acilius and Graphoderus.

The two Acilius species dispersed continuously throughout the season and

where independent of spatial connectivity and mass effect. In contrast, the

93

short window during which dispersal took place in all three Graphoderus

species scaled to differences in dispersal rates, within and across seasons,

and differences in colonization rates between the two genera.

Our findings provide some of the first insights into the understanding of

seasonal restrictions in flight patterns of aquatic beetles and invertebrates in

general.

Acknowledgements We thank Voldemar Rannap (Estonian Environmental Board) for practical

assistance during fieldwork. We also thank Lech Borowiec for providing photos

of the study species and Charlotte Nekman for commenting on this as well as

a previous version of the manuscript. All experimental work involving the

protected Graphoderus bilineatus was conducted under permit number 522-

2532-2015 of the County Administrative Board of Scania, Sweden, and 522-

366-2015 of the County Administrative Board of Blekinge, Sweden. This study

was supported by the Danish Ministry of Higher Education and Science

(Project 0604-02230B) and LIFE + DRAGONLIFE (LIFE08 NAT/EE/000257).

References • Arribas, P., J. Velasco, P. Abellán, D. Sánchez-Fernández, C. Andújar,

P. Calosi, A. Millán, I. Ribera, and D. T. Bilton. 2012. Dispersal ability

rather than ecological tolerance drives differences in range size

between lentic and lotic water beetles (Coleoptera: Hydrophilidae).

Journal of Biogeography 39:984-994.

• Baguette, M., S. Blanchet, D. Legrand, V. M. Stevens, and C. Turlure.

2013. Individual dispersal, landscape connectivity and ecological

networks. Biol Rev Camb Philos Soc 88:310-326.

• Baguette, M., J. Clobert, and N. Schtickzelle. 2011. Metapopulation

dynamics of the bog fritillary butterfly: experimental changes in habitat

quality induced negative density‐dependent dispersal. Ecography

34:170-176.

• Baselga, A., J. M. Lobo, J.-C. Svenning, P. Aragón, and M. B. Araújo.

2012. Dispersal ability modulates the strength of the latitudinal richness

94

gradient in European beetles. Global Ecology and Biogeography

21:1106-1113.

• Bekoff, M. 1977. Mammalian dispersal and the ontogeny of individual

behavioral phenotypes. American Naturalist:715-732.

• Bergsten, J., and K. B. Miller. 2005. Taxonomic revision of the Holarctic

diving beetle genus Acilius Leach (Coleoptera: Dytiscidae). Systematic

Entomology 31:145-197.

• Bilton, D. T. 2014. Dispersal in Dytiscidae. Pages 387-407 Ecology,

Systematics, and the Natural History of Predaceous Diving Beetles

(Coleoptera: Dytiscidae). Springer.

• Bilton, D. T., J. R. Freeland, and B. Okamura. 2001. Dispersal in

freshwater invertebrates. Annual review of ecology and

systematics:159-181.

• Boda, P., and Z. Csabai. 2013. When do beetles and bugs fly? A unified

scheme for describing seasonal flight behaviour of highly dispersing

primary aquatic insects. Hydrobiologia 703:133-147.

• Bonte, D., H. Van Dyck, J. M. Bullock, A. Coulon, M. Delgado, M.

Gibbs, V. Lehouck, E. Matthysen, K. Mustin, and M. Saastamoinen.

2012. Costs of dispersal. Biological Reviews 87:290-312.

• Bowler, D. E., and T. G. Benton. 2005. Causes and consequences of

animal dispersal strategies: relating individual behaviour to spatial

dynamics. Biological Reviews 80:205-225.

• Clobert, J., L. Galliard, J. Cote, S. Meylan, and M. Massot. 2009.

Informed dispersal, heterogeneity in animal dispersal syndromes and

the dynamics of spatially structured populations. Ecology Letters

12:197-209.

• Clobert, J., R. Ims, and F. Rousset. 2004. Causes, mechanisms and

consequences of dispersal. Ecology, genetics, and evolution of

metapopulations:307-335.

• Cote, J., J. Clobert, T. Brodin, S. Fogarty, and A. Sih. 2010. Personality-

dependent dispersal: characterization, ontogeny and consequences for

spatially structured populations. Philosophical Transactions of the Royal

Society B: Biological Sciences 365:4065-4076.

95

• Csabai, Z., Z. Kálmán, I. Szivák, and P. Boda. 2012. Diel flight

behaviour and dispersal patterns of aquatic Coleoptera and Heteroptera

species with special emphasis on the importance of seasons.

Naturwissenschaften 99:751-765.

• Dingle, H. 1972. Migration strategies of insects. Science 175:1327-

1335.

• Dolédec, S., D. Chessel, and C. Gimaret-Carpentier. 2000. Niche

separation in community analysis: a new method. Ecology 81:2914-

2927.

• Ducatez, S., D. Legrand, A. CHAPUT‐BARDY, V. M. Stevens, H.

Freville, and M. Baguette. 2012. Inter‐individual variation in movement:

is there a mobility syndrome in the large white butterfly Pieris

brassicae? Ecological Entomology 37:377-385.

• Duckworth, R. A., and A. V. Badyaev. 2007. Coupling of dispersal and

aggression facilitates the rapid range expansion of a passerine bird.

Proceedings of the National Academy of Sciences 104:15017-15022.

• Foster, G. 1996. Graphoderus bilineatus (DeGeer, 1774). Pages 40-48

in Background information on invertebrates of the Habitats Directive and

the Bern Convention. Part.

• Foster, G. N., and D. T. Bilton. 2014. The Conservation of Predaceous

Diving Beetles: Knowns, Unknowns and Anecdotes. Pages 437-462

Ecology, Systematics, and the Natural History of Predaceous Diving

Beetles (Coleoptera: Dytiscidae). Springer.

• Fronhofer, E. A., H. J. Poethke, and U. Dieckmann. 2015. Evolution of

dispersal distance: maternal investment leads to bimodal dispersal

kernels. Journal of theoretical biology 365:270-279.

• Gadgil, M. 1971. Dispersal: population consequences and evolution.

Ecology 52:253-261.

• Gaston, K. J. 2009. Geographic range limits: achieving synthesis.

Proceedings of the Royal Society of London B: Biological

Sciences:rspb. 2008.1480.

• Geber, M. A. 2011. Ecological and evolutionary limits to species

geographic ranges. Am Nat 178 Suppl 1:S1-5.

96

• Grönroos, M., J. Heino, T. Siqueira, V. L. Landeiro, J. Kotanen, and L.

M. Bini. 2013. Metacommunity structuring in stream networks: roles of

dispersal mode, distance type, and regional environmental context.

Ecology and Evolution 3:4473-4487.

• Gustafson, E. J., and G. R. Parker. 1994. Using an index of habitat

patch proximity for landscape design. Landscape and Urban Planning

29:117-130.

• Hanski, I., C. Erälahti, M. Kankare, O. Ovaskainen, and H. Sirén. 2004.

Variation in migration propensity among individuals maintained by

landscape structure. Ecology Letters 7:958-966.

• Hargreaves, A. L., and C. G. Eckert. 2014. Evolution of dispersal and

mating systems along geographic gradients: implications for shifting

ranges. Functional Ecology 28:5-21.

• Heino, J., A. S. Melo, T. Siqueira, J. Soininen, S. Valanko, and L. M.

Bini. 2015. Metacommunity organisation, spatial extent and dispersal in

aquatic systems: patterns, processes and prospects. Freshwater

Biology 60:845-869.

• Holt, R. D. 1985. Population dynamics in two-patch environments: some

anomalous consequences of an optimal habitat distribution. Theoretical

population biology 28:181-208.

• Iversen, L. L., R. Rannap, P. F. Thomsen, J. Kielgast, and K. Sand-

Jensen. 2013. How do low dispersal species establish large range

sizes? The case of the water beetleGraphoderus bilineatus. Ecography

36:770-777.

• Jackson, D. J. 1952. Observations on the capacity for flight of water

beetles. Pages 57-70 in Proceedings of the Royal Entomological

Society of London. Series A, General Entomology. Wiley Online Library.

• Jackson, D. J. 1956. Observations on flying and flightless water beetles.

Journal of the Linnean Society of London, Zoology 43:18-42.

• Kehl, S., and K. Dettner. 2007. Flugfähigkeit der in Deutschland

vorkommenden adephagen Wasserkäfer (Coleoptera, Hydradephaga).

Entomologie heute 19:141-161.

97

• King, E. G., and D. A. Roff. 2010. Modeling the Evolution of Phenotypic

Plasticity in Resource Allocation in Wing‐Dimorphic Insects. The

American Naturalist 175:702-716.

• Kokko, H., and A. López-Sepulcre. 2006. From individual dispersal to

species ranges: perspectives for a changing world. Science 313:789-

791.

• Kubisch, A., R. D. Holt, H.-J. Poethke, and E. A. Fronhofer. 2014.

Where am I and why? Synthesizing range biology and the eco-

evolutionary dynamics of dispersal. Oikos 123:5-22.

• Landin, J. 1980. Habitats, life histories, migration and dispersal by flight

of two water‐beetles Helophorus brevipalpis and H. strigifrons

(Hydrophilidae). Ecography 3:190-201.

• Leibold, M. A., M. Holyoak, N. Mouquet, P. Amarasekare, J. M. Chase,

M. F. Hoopes, R. D. Holt, J. B. Shurin, R. Law, and D. Tilman. 2004.

The metacommunity concept: a framework for multi‐scale community

ecology. Ecology Letters 7:601-613.

• Levin, S. A., D. Cohen, and A. Hastings. 1984. Dispersal strategies in

patchy environments. Theoretical population biology 26:165-191.

• Lundkvist, E., J. Landin, and F. Karlsson. 2002. Dispersing diving

beetles (Dytiscidae) in agricultural and urban landscapes in south-

eastern Sweden. Pages 109-123 in Annales Zoologici Fennici. JSTOR.

• Lundkvist, E., J. Landin, and P. Milberg. 2001. Diving beetle

(Dytiscidae) assemblages along environmental gradients in an

agricultural landscape in southeastern Sweden. Wetlands 21:48-58.

• Matthysen, E. 2012. Multicausality of dispersal: a review. Dispersal

ecology and evolution:3-12.

• Miguélez, D., and L. F. Valladares. 2008. Seasonal dispersal of water

beetles (Coleoptera) in an agricultural landscape: a study using

Moericke traps in northwest Spain. Pages 317-326 in Annales de la

Societe entomologique de France. Taylor & Francis.

• Miller, K. B., and J. Bergsten. 2016. Diving Beetles of the World:

Systematics and Biology of the Dytiscidae. JHU Press.

98

• Nilsson, A. N., J. Elmberg, and K. Sjoberg. 1994. Abundance and

species richness patterns of predaceous diving beetles (Coleoptera,

Dytiscidae) in Swedish lakes. Journal of Biogeography:197-206.

• Nilsson, A. N., and M. Holmen. 1995. The Aquatic Adephaga

(Coleoptera) of the Fennoscandia and Denmark. Ii. Dytiscidae: II-

Dytiscidea. Brill.

• North, A., S. Cornell, and O. Ovaskainen. 2011. Evolutionary responses

of dispersal distance to landscape structure and habitat loss. Evolution

65:1739-1751.

• Pereira, H. M., P. W. Leadley, V. Proença, R. Alkemade, J. P.

Scharlemann, J. F. Fernandez-Manjarrés, M. B. Araújo, P. Balvanera,

R. Biggs, and W. W. Cheung. 2010. Scenarios for global biodiversity in

the 21st century. Science 330:1496-1501.

• Poethke, H. J., T. Hovestadt, and O. Mitesser. 2003. Local extinction

and the evolution of dispersal rates: causes and correlations. The

American Naturalist 161:631-640.

• Rannap, R., A. Lõhmus, and L. Briggs. 2009. Restoring ponds for

amphibians: a success story. Hydrobiologia 634:87-95.

• Ronce, O. 2007. How does it feel to be like a rolling stone? Ten

questions about dispersal evolution. Annual Review of Ecology,

Evolution, and Systematics:231-253.

• Schulz, G., T. Siqueira, G. Stefan, and F. d. O. Roque. 2012. Passive

and active dispersers respond similarly to environmental and spatial

processes: an example from metacommunity dynamics of tree hole

invertebrates. Fundamental and Applied Limnology/Archiv für

Hydrobiologie 181:315-326.

• Simmons, A. D., and C. D. Thomas. 2004. Changes in dispersal during

species’ range expansions. The American Naturalist 164:378-395.

• Soons, M. B., and G. W. Heil. 2002. Reduced colonization capacity in

fragmented populations of wind‐dispersed grassland forbs. Journal of

Ecology 90:1033-1043.

• Starrfelt, J., and H. Kokko. 2012. The theory of dispersal under multiple.

Dispersal ecology and evolution:19.

99

• Svenning, J. C., and F. Skov. 2004. Limited filling of the potential range

in European tree species. Ecology Letters 7:565-573.

• Thomas, C. D., E. Bodsworth, R. J. Wilson, A. Simmons, Z. G. Davies,

M. Musche, and L. Conradt. 2001. Ecological and evolutionary

processes at expanding range margins. Nature 411:577-581.

• Urban, M. C. 2015. Accelerating extinction risk from climate change.

Science 348:571-573.

• Vepsäläinen, K. 1978. Wing dimorphism and diapause in Gerris:

determination and adaptive significance. Pages 218-253 Evolution of

insect migration and diapause. Springer.

• Whitcomb, R. F., C. S. Robbins, J. F. Lynch, B. L. Whitcomb, M. K.

Klimkiewicz, and D. Bystrak. 1981. Effects of forest fragmentation on

avifauna of the eastern deciduous forest. Pages 125-205 in R. L.

Burgess and D. M. Sharpe, editors. Forest Island Dynamics in Man-

Dominated Landscapes. Springer-Verlag, New York.

100

Appendix – Chapter III

Appendix S1: Habitat and spatial variables used in the study:

Habitat descriptions were conducted along with the collection of species data.

Instead of recording fluctuating water-chemistry variables, we focused on more

stable physical and biological variables. Variables assumed to influence

presence/absence of the different Aciliini species were selected based on

existing literature and results of preliminary fieldwork conducted on the group.

At each site, shading was defined as the percentage of the shoreline having

standing trees of more than two meter in height, present within three meters

from the shoreline. The shallow areas (distance from the banks with a depth <

30 cm) were estimated as the maximum extent for each of the four cardinal

banks. Following Iversen et al. (2013) larval and adult habitats were obtained

by visually estimating the percentage vegetation cover of the total lake surface

area. Vegetation structure were subdivided into 4 classes representing

submerged vegetation, floating vegetation, riparian vegetation < 1 meter in

height, and riparian vegetation > 1 meter in height. Surface area of ponds (<

0.5 ha ) was measured in the field, while lakes (> 0.5 ha ) were measured prior

to the fieldwork, using high resolution orthophotos. All ponds and lakes in the

study region was geolocated via high resolution orthophotos.

References

Iversen, L. L., Rannap, R., Thomsen, P. F., Kielgast, J., & Sand‐Jensen, K.

(2013). How do low dispersal species establish large range sizes? The case of

the water beetle Graphoderus bilineatus. Ecography, 36(7), 770-777.

101

Fig S1: Ex-situ flight experiments of the study species were conducted in 4 -

litre containers (a) with only one possible exit point (b), either flying from a

vertical or a horizontal position (c).

.

102

Table s1: Study sites and number of animals tested in the dispersal

experiments. The number in brackets refers to the amount of animals flying

during the experiment

Site name Lat, Long Country Sampling yearAcilius canaliculatus

Acilius sulcatus

Graphoderus bilineatus

Graphoderus cinereus

Graphoderus zonatus

BøllemosenN 55.827216, E 12.563542 Denmark 2011, 2015 2(2) 121(25)

GyngemosenN 55.723842, E 12.493528 Denmark 2015 8(8) 40(12)

AssarumN 56.227105, E 14.819313 Sweden 2015 1(1) 21(21) 2(1) 30(2)

GlimåkraN 56.327089, E 14.115704 Sweden 2011, 2015 19(17) 15(15) 13(1) 4(1) 11(1)

HemgylN 56.345193, E 14.774630 Sweden 2015 30(9) 11(4)

KarstorpN 56.314909, E 14.036817 Sweden 2011, 2015 18(18) 17(15) 14(0) 84(9) 67(9)

KnapsttorpaN 56.352765, E 14.113557 Sweden 2011, 2015 31(31) 28(0) 16(5)

Lönsboda AN 56.425885, E 14.385216 Sweden 2011, 2015 7(7) 26(25) 30(0) 14(1)

Lönsboda BN 56.394262, E 14.401792 Sweden 2011, 2015 1(1) 14(0) 9(0)

LångasjönN 56.351405, E 14.769773 Sweden 2015 15(14) 2(2) 25(9) 16(2)

NybygdasjönN 56.392431, E 13.922768 Sweden 2015 19(0)

RydN 56.459295, E 14.653355 Sweden 2011, 2015 68(66) 22(22) 28(2)

SkinnemyraN 56.422707, E 14.271922 Sweden 2011, 2015 33(32) 24(23) 14(0)

StockegyletN 56.338310, E 14.398565 Sweden 2011, 2015 29(29) 37(37) 5(0) 21(1) 16(7)

Stora KröksjönN 56.315356, E 14.436484 Sweden 2015 15(0)

VångasjönN 56.283016, E 14.629340 Sweden 2015, 2015 10(10) 25(5)

Total 242(236) 164(160) 232(24) 338(61) 152(21)

Spring 90(88) 82(81) 130(2) 118(5) 61(2)

Summer 55(54) 42(42) 50(22) 127(53) 32(16)

Autumn 97(94) 40(37) 52(0) 93(3) 59(3)

103

Table S2: Results from flight experiments conducted on 13 other species of

diving beetles. All trials were performed during the spring of 2011.

Species

Nr. of individuals

tested

Nr. of observed

flights

Colymbetes paykulli 18 18

Dytiscus lapponicus 14 0

Dytiscus marginalis 6 5

Hydaticus aruspex 6 5

Hydaticus seminiger 12 10

Hydaticus transversalis 2 1

Hyphydrus ovatus 4 1

Ilybius fuliginiosus 3 0

Rhantus exoletus 3 3

Rhantus frontalis 1 1

Rhantus grapii 8 6

Rhantus notaticollis 6 6

Rhantus suturalis 3 2

104

105

Chapter IV

Community composition of diving beetles

across a human altered microhabitat

gradient in a raised bog system

106

Community composition of diving beetles across a human altered microhabitat gradient in a raised bog system

Lars Lønsmann Iversen1,2, Lars Briggs2, Johannes Edvardsson3,

Leonas Jarašius4, Emil Kristensen1, Jūratė Sendžikaitė4, Kaj Sand-

Jensen1 1Freshwater Biological Laboratory, Biological Institute, University of

Copenhagen, Denmark. 2Amphi Consult ApS, International Science Park

Odense, Denmark. 3Dendrolab.ch, Institute of Geological Sciences, University

of Bern, Switzerland. 4

Institute of Botany, Nature Research Centre, Vilnius,

Lithuania

Abstract Species living in pools on European peat bogs face immediate extinction

threats as a consequence of human disturbance and climate change. Thus,

understanding the environmental factors affecting these communities is

essential if bog pool species are to be efficiently protected. We studied the

composition of diving beetles in 50 bog pools along an afforestation gradient in

an extensive Lithuanian raised bog ecosystem. From the core to the edge of

the study area, species associated to bog pool habitats decreased in individual

abundance. In contrast, species associated to the surroundings landscape

increased in abundance. Bog pool species were negatively affected by the loss

of natural hydrology and the establishment of pine trees leading to increasing

brownification. Although our conclusions are drawn from a single bog system

we show that when undergoing environmental change, species communities in

open bog pool habitats are influenced by the same core-edge gradients and

107

face the same conservation challenges as other isolated habitats. Our study

highlights rewetting schemes and tree removal as potential restoration

measures for threatened diving beetles in bog pools

Introduction Freshwater habitats in ombrotrophic bogs are among the most vulnerable and

threatened aquatic habitats in Europe (Joosten et al. 2012, Beadle et al. 2015).

Peat excavation, drainage, and increasing atmospheric nutrient deposition are

continuously causing a decline in ombrotrophic bog habitats across the

continent (Aerts et al. 1992, Rydin and Jeglum 2013). Given the limited

distribution of ombrotrophic bogs surrounded by highly contrasting landscape

types, the regional occurrence of species in bogs is by nature patchy and

isolated (Spitzer and Danks 2006, Buchholz 2016). These conditions have

placed bog species and habitats on the top list of threatened biodiversity and

ecosystems in Europe (Evans 2006, Spitzer and Danks 2006) Furthermore,

increasing temperatures due to climate change are expected to enhance the

loss of water and open bog habitats (Edvardsson et al. 2015b, Holmgren et al.

2015).

Tree colonisation on peat bogs is one such degradation threat, and reduced

biodiversity and carbon sequestration are expected if drier conditions turn

moss-dominated bogs to tree-covered habitats (Limpens et al. 2014, Holmgren

et al. 2015). At present there is growing evidence of accelerating tree

colonization in northern peatlands (Linderholm and Leine 2004, Edvardsson et

al. 2015b).

A decrease in species associated to bog pools is an expected response to the

increasing habitat degradation of peat bogs (Mazerolle 2003, van Kleef et al.

2012), while an increase in abundance and richness of non-specific species is

anticipated (Scudder 1987, Mieczan et al. 2014). The increase of non-specific

species is often explained as a consequence of invasion of species from the

areas surrounding ombotrophic bogs (Antonović et al. 2012, Sushko 2016).

These patterns are not unique for raised bogs and can be extended to blanket

bogs and other Sphagnum dominated habitats (Swengel and Swengel 2011,

Drinan et al. 2013). Although these general patterns have been demonstrated ,

108

we know surprisingly little about the underlying environmental conditions and

processes causing changes in community compositions along these core-

perimeter gradients in peatlands in general and ombrotrophic bogs in particular

(Spitzer and Danks 2006, Kreutzweiser et al. 2013).

The present field study aimed at elucidating the relative role of general and

bog specific environmental factors in structuring species communities of

ombrotrophic bog pools. We hypothesize that communities in bog pools are

affected by specific environmental processes for this habitat and processes

affecting community assemblages in general. We predict that this leads to

differences between the abundance of individuals and environmental variables

depending on the level of association of species guilds to pristine bog pool

habitats. We further hypothesize that these different dependencies to bog pool

environments lead to different gradients in individual abundance between

communities along a habitat degradation gradient (Ries et al. 2004). Species

guilds associated with pristine bog pools should exhibit a decrease of

abundance from the edge to the core of bog complexes (Davies et al. 2001). In

contrast, species mainly associated to surrounding habitats should penetrate

across the edge of the raised bog creating a decrease in abundance towards

the core of the bog area (Ewers and Didham 2007).

We address these questions by studying communities of diving beetles

(Coleoptera: Dytiscidae) in 50 bog pools sampled in a large but partly drained

bog system in Western Lithuania. We identify abundance gradients in two

guilds of diving beetles resembling species preferring Sphagnum-rich habitats

(from the genus Ilybius) and species preferring forest ponds and groundwater

fed ponds (from the tribe Aciliini) (Nilsson and Holmen 1995, Duff 2012, Miller

and Bergsten 2016). The two guilds represent large diving beetles sharing the

same phenological attributes, hibernating as adults and breeding during the

following spring (Nilsson 1986, Nilsson and Holmen 1995). Thus, we argue

that differences in habitat preferences will be one of the main determinants of

community assemblage in the two study groups allowing us to test our target

hypothesis.

Methods

109

Study area

The Aukštumala is a 3702 ha bog complex situated in Western Lithuania

(55°24′ N, 21°20′ E). The central and western parts of the bog remain

protected from anthropogenic activities, whereas the eastern part has

undergone drainage and peat-mining (Fig.1). The pristine parts of the area

have an open bog surface with isolated small birch trees and about 380 small

dystrophic bog pools. In the Nemunas delta region, large-scale land

reclamation activities were implemented already in the 19th

century to regulate

the water regime in fertile lands of the delta which could be readily used for

agriculture. In 1882, in the south-eastern part of Aukštumala raised bog, a peat

mining factory was built, launching manual peat excavation activities (Purvinas

2006). The greatest changes in the raised bog took place in 1968, following

installation of protective embankments, water pumping stations, roads and

ditches. The intensive reclamation resulted in drainage of about two-thirds of

the total bog area, which was subsequently designated for industrial peat-

cutting (Jarašius et al. 2014). The rest of the bog territory (1017 ha) was

declared a Telmological Reserve in 1995 (Jarašius et al. 2015). As a

consequence of the intensive long term drainage in the adjacent areas of the

Reserve, the open raised bog vegetation in Aukštumala has been replaced by

tree and shrub vegetation (Jarašius et al. 2014, Edvardsson et al. 2015b).

Data collection

Sampling of diving beetles took place in the latter half of May 2014 during the

peak reproductive period of the two species groups (Nilsson 1986, Nilsson and

Holmen 1995). A total of 50 pools were studied following a consistent strategy.

At the start of each day of fieldwork one locality within the Aukštumala raised

bog was selected at random. The chosen locality was sampled, and

subsequently the nearest neighboring pool was sampled, and so forth. If

localities were directly interconnected by waterways, only one of them was

randomly selected for sampling.

110

Fig.1: Geographical locations of the 50 bog pools studied in the Aukštumala raised bog. The

extent of the active peat excavation within Aukštumala is seen the eastern part of the map.

(Bottom left) A pristine bog pool with no surrounding pine tree vegetation, and (bottom right) a

forested bog pool in the eastern part of the study area.

Before sampling a given pool, a schematic overview of the microhabitats along

the shores of the pool was constructed. We categorized microhabitats based

on difference in vegetation structure (Table S1) because diving beetle

assemblages are closer related to structure than specific plant species (Gioria

2014, Vamosi and Wohlfahrt 2014). A vegetation structure was considered to

be present if there was a minimum of one meter of coherent vegetation along

the banks and accumulated 10 meters of a given structure at a site. The

111

presence of different vegetation structures was established based on the

census of two field person’s independent evaluation.

Diving beetle sampling was adopted from the sampling protocol for aquatic

Coleoptera of (Oertli et al. 2005). A hand net (50*50 cm, mesh size 0.25 mm)

was continuously swept through the habitat (i.e. vegetation and debris) 10

times for each sample. For each identified microhabitat structure 20 samples

were collected scattered across the range of the microhabitat. The species

identification followed (Nilsson and Holmen 1995). All specimens from the

Aciliini tribe were identified and counted in the field while species of Ilybius

were collected and later identified in the laboratory under the microscope.

For each study site we determined five explanatory environmental variables.

Instead of recording fluctuating water-chemistry variables, we focused on more

stable physical and biological variables known to affect the structure of water

beetle communities or key descriptors of raised bog gradients.

Water samples were collected at the end of the sampling period for each of the

study sites. The samples were filtered through a GF/C filter (1.2 µm) and the

absorbance of colored dissolved organic matter (CDOM, m-1

We estimated the proportion of tree cover by Pinus sylvestris within 10 m of

the border of each study site. Tree cover was obtained from existing forest

maps (Edvardsson et al. 2015) updated using high definition orthophotos from

2014. We expect tree cover to structure local micro-temperatures and thereby,

potentially, affecting diving beetle communities (e.g. (Eyre et al. 2006, Schäfer

et al. 2006). Additionally we used tree cover as a descriptor of core-perimeter

status of the raised bog (Rydin and Jeglum 2013).

), were measured

at 440 nm using a Shimadzu UV-1800 spectrophotometer. An increase in

CDOM, also referred to as brownification due to peat degradation (Price 2003,

Gažovič et al. 2013) will increase light attenuation, reduce lake productivity

(Karlsson et al. 2009, Kritzberg et al. 2014) and alter community structure of

organism in small nutrient-poor ponds (Hansson et al. 2013, Lindholm et al.

2016). As a consequence of these changes increased brownification has been

shown to affect the structure of bog pool communities of diving beetles along

bog-forestry gradients (Drinan et al. 2013) and species communities in general

(Solomon et al. 2015; Williamson et al. 2015).

112

Pool area and the diversity of microhabitats are parameters expected to

promote both species richness and abundance of diving beetles in general

(Nilsson et al. 1994, Fairchild et al. 2000) and raised bog species in particular

(Verberk et al. 2006, van Duinen 2013). Microhabitat diversity within a site was

the total number of different microhabitats (Table S1) and pool area was

measured on high resolution orthophotos.

The water table in ombrotrophic bogs is thought to be a primary factor in bog

afforestation dynamics. Lower water tables, a large acrotelm zone and a

deeper unsaturated zone at the bog surface promote tree establishment and

increase radial tree growth (Boggie 1972, Limpens et al. 2014, Edvardsson et

al. 2015b). For each study site we estimated the water level, defined as the

distance between the bog surface and the water-saturated peat. Water level

predictions were established using a random forests model (Breiman 2001)

fitted via water levels measured as the mean water table from April to October

in 2007-2014 at 12 stations scattered across the study area (Jarašius et al.

2014). In the model spatial placement and elevation (at 0.5 m resolution) of a

given site were used to predict the depth of the water table.

Data analyses

Changes in diving beetle abundance along the afforestation gradient were

analysed by a logistic regression model. On a log-scale the accumulated count

of sampled individuals was used as the response variable and the percentage

of tree cover as the explanatory variable. The effect of tree cover on beetle

abundance was evaluated at a 5 % significance level by a χ ² -test. We

adjusted for differences in sampling intensity between pools (20 samples per

microhabitat type present) by adding an offset variable in the models

containing the logarithm to the number of microhabitats present at a given site

(Zuur et al. 2009).

Following (Baselga 2013), we partitioned the changes in species abundance

into two different processes: balanced variation in abundance (BC-bal),

whereby species at one site are substituted by different species at another site

but the number of individuals remains constant; and abundance gradients (BC-

gra), whereby individuals are lost from one site to the other. Derived from site

113

level Bray-Curtis dissimilarities these indices were estimated using the

betapart package in R (Baselga and Orme 2012). In order to adjust for

differences in sample intensity we computed the dissimilarities using 20

random samples from each study site. This step was repeated 1000 times and

average distance matrixes were calculated for BC-bal and BC-gra respectively

(Baselga 2010). For species within both Ilybius and Aciliini we compared site

specific differences in abundance turnover with differences in tree cover

between sites using a mantel test in the R package vegan (Oksanen et al.

2016).

To evaluate the effects of different bog degradation factors on diving beetle

abundance in a given microhabitat, we used structural equation modelling

(SEM). In SEM causal links between variables of interest are defined and

evaluated in the form of interconnected equations (Shipley 2009, Grace et al.

2012). Recognising that different microhabitat types might have different

offsets in diving beetle abundance the SEM models were implemented within a

mixed-effect framework using microhabitat type as a random intercept. We

created a hypothetical path diagram based on our expected connections

between parameters (Fig. S1). Local linear models were then fitted according

to the structure of the data. The abundance data were modelled following a

Poisson distribution in a mixed logistic regression model. All other endogenous

variables were modelled using Gaussian mixed effect models. It should be

noted that in non-linear mixed models inference between models becomes

conditional to the random effects (Stroup 2012). In our case this restricts the

parameter interpretation to be within and not between each of the five

microhabitat types present in the study area.

Model fit and model evaluation in the SEM were based on local estimations

(Lefcheck 2015). We evaluated the SEM model using Fisher’s C statistic

(Shipley 2000) and AIC values derived from Fisher’s C (Shipley 2013). Starting

from the initial network we added all missing pathways lowering the AIC value

of the entire SEM. Following (Lefcheck 2015) we then performed a stepwise

removal of all parameters with a p-value > 0.05 in each of the local models.

Finally the Fisher’s C statistic was checked once again in order to ensure that

the hypothesized relationships in the SEM were consistent with the data. The

114

level of variance explained by each component model for all response

variables was given as the conditional R2 (or pseudo R2

Given the ongoing peat excavation along the eastern border of the study area

(Fig. 1), we expected a strong directional pattern of the drainage related

processes in Aukštumala. Thus, two types of spatial autocorrelation could be

present in the study area. One related to the longitudinal drainage gradient

(captured by the measured and unmeasured factors associated to the

drainage of the bog). Another potential source could be a random correlation

between the drainage gradient and single parameters acting independently of

the drainage gradient. If the effect of an explanatory variable on a response

variable is a consequence of the observed drainage gradient and not just

random associations, the effect size should be independent of the spatial

placement of the study sites and solely explained by the variation of the

explanatory variable itself. In each of the component models in the final SEMs

we tested for spatial independence of the effect size for a target explanatory

variable by adding an interaction between the given variable and the

longitudinal placement of the study sites. The effect modification of longitude

on the explanatory variables was evaluated at a 5 % significance level by a χ ²

-test. If the effect size of a variable significantly varied spatially, and could not

be explained alone by the variable itself, we included the variable in the final

model as an adjusting variable but excluded all direct interpretations from the

variable in our model inference.

) based on the

variance of both the fixed and random effects (Nakagawa and Schielzeth

2013). In order to enable model conversion and compare effect sizes across

variables all environmental variables were scaled to the mean of zero and the

standard deviation of one prior to any model fitting.

All analyses were conducted in R ver. 3.1.0 or later (www.r-project.org).

Results A total of 15 diving beetle species (six Aciliini and nine Ilybius species, Table

S2) was recorded in the Aukštumala raised bog. There was a positive

relationship between local species richness and abundance in bog pools for

both beetle groups (Fig S2). The abundance of Aciliini (P < 0.001) and Ilybius

115

(P< 0.001) changed significantly along the afforestation gradient, but Aciliini

showed a positive relationship between abundance and forest cover, while

Ilybius abundance showed a negative relationship with forest cover (Fig. 2).

The two species groups also differed with respect to the processes causing

changes in community compositions along the afforestation gradient. Number

of individuals within Aciliini increased significantly with forest cover (BC-gra vs.

differences in tree cover, Mantel statistic r = 0.09, P< 0.05) with no

replacement of species along the gradient (BC-bal vs. differences in tree

cover, Mantel statistic r = -0.02, P = 0.685). In contrast, number of individuals

of Ilybius decreased slightly with higher forest cover (BC-gra vs. differences in

tree cover, Mantel statistic r = 0.07, P = 0.09) but with a profound replacement

of species along the afforestation gradient (BC-bal vs. differences in tree

cover, Mantel statistic r = 0.15, P < 0.01)

Fig. 2: Average individual diving beetle abundance per microhabitat in relation to pine tree

cover among the 50 studied bog pools. The left panel represents core species from the Ilybius

group and the right panel edge species from the Aciliini group. The central tendency is shown

by the solid line and 95 % confidence limits are shown as shaded areas.

The structural equation models showed different pathways regarding the

causal effect of tree cover on beetle abundance for the two study groups

(Table 1). Among the 18 unique paths in the models only the effect size of

water level on microhabitat density and the effect size of tree cover on pool

area weres patially independent (Table S4). In Aciliini there was a direct

116

positive relationship between tree cover and beetle abundance (Fig.3a). In

contrast, tree cover had an indirect negative effect on Ilybius abundance by

increasing brownification (Fig 3b). Pool area alone had a positive effect on

Aciliini abundance (β=0.13) but a negative effect on Ilybius abundance (β=-

0.29). The diversity of microhabitats did not have any effect on either of the

two groups when pool area had been adjusted for. The structural equation

models also suggested that there was a positive effect of water level on diving

beetle abundance of both groups, although this effect was to some extent

counter balanced for the Aciliini species by the decrease in tree cover with

higher water level (Fig 3b, Table 1).

117

Fig.3: Structural equation models (SEM) of diving beetle abundance in Aciliini (a) and Ilybius

(b) incorporating a random effect of the sampled microhabitat type. Boxes represent measured

variables and the Rc2 is based on the variance of both the fixed and random effects. Arrows

represent unidirectional relationships among variables. Blue arrows denote positive

118

relationships, and red arrows negatives ones. Arrows for non-significant paths (P ≥ 0·05) are

dashed. Path coefficients give the strength of the direct, indirect and total effect on abundance

(= standardized β – coefficients, se main text).

Table 1: Direct, indirect and combined standardized effects of parameters affecting diving

beetle abundance in the structural equation models developed for diving beetle species in 50

pools in a Lithuanian raised bog. The standardized effects can vary from – 1 (perfect negative

effect), over 0 (no effect) to 1 (perfect positive effect).

Tree

cover

Water

level

Pool

area Brownification

Aciliini

Direct effect 0.44 0.25 0.13 0

Indirect effect 0 -0.16 0 0

Total effect 0.44 0.09 0.13 0

Ilybius

Direct effect 0 0.29 -0.29 -0.2

Indirect effect -0.08 -0.07 0 0

Total effect -0.08 0.22 -0.29 -0.2

Discussion The change in diving beetle abundance along the afforestation gradient was in

accordance with the expected pattern for both the Aciliini and Ilybius species

group. The decrease in Aciliini abundance in bog pools with falling forest cover

from the edge towards the open central parts of the raised bog resembles an

abundance gradient of species penetrating into a unique habitat from the

surrounding contrasting landscape (Ewers and Didham 2007). In contrast the

Ilybius species strongly suggested a representation of bog specialist preferring

the pristine and open parts of the bog and decreasing as the edge effect and

the forest coved increased (Davies et al. 2001, Barbosa and Marquet 2002).

These relationships were also supported by the different processes structuring

the community change in species abundance. Thus, the gradual decrease of

specimens but static composition of Aciliini species from the perimeter to the

119

core of the Aukštumala bog suggest an intrusion gradient from the surrounding

landscape with all Aciliini species declining as the environmental conditions

change towards open and pristine bog conditions (Ewers and Didham 2008,

Seabloom et al. 2013). Among the Ilybius species, the decrease in total

abundance and replacement of species from the open pools in the center to

the pools surrounded by trees at the periphery of the bog system suggest a

change from the most demanding to the least demanding Ilybius species for

pristine bog pool conditions (Davies et al. 2001, Haddad et al. 2015).

Effects of the environmental variables highlighted that both general and bog

specific factors structured the abundance of the two beetle groups. The

positive effect of water level on abundance of both Aciliini- and Ilybius-species

likely reflects a positive association with the extent of shallow zones at the

perimeter of the pools (Gioria 2014). Likewise, the positive influence of pool

area and forest cover on Aciliini abundance was in agreement with the general

preference for larger forested ponds and lakes of species in this taxonomic

group (Lundkvist et al. 2001, Bergsten and Miller 2006).

We did not find any effect of microhabitat diversity on diving beetle abundance

in the individual microhabitats for either Aciliini or Ilybius. This finding

contradicts previous work demonstrating a positive link between species

richness or abundance and habitat heterogeneity for diving beetles (Nilsson et

al. 1994, Picazo et al. 2010) and bog macro-invertebrates (Verberk et al.

2006). Within the sampled microhabitats there was no accumulated effect of

the number of habitat types present in a given bog pool on the abundance of

diving beetles in a given microhabitat. Thus, in our case study any potential

effect of microhabitat heterogeneity on the size of the local species pool was

not created by habitat diversity at the bog pool level.

The Ilybius species also contradicted the general notion that pool area

enhanced beetle abundance, although higher water level did increase total

habitat area (Fig 3). It has been hypothesized that predation from predators

such as dragonfly larva in large bog pools restricts the presence of diving

beetles to small predator-free habitats (Larson et al. 1990). Given that smaller

bog pools often has fewer large macroinvertebrate species (and predatory

dragonfly larvae) (Verberk et al. 2008, Hannigan and Kelly-Quinn 2012), the

120

observed decrease in Ilybius abundance with an increase in bog pool area

could be the result of a restriction of Ilybius-species to small predator-free bog

pools.

The observed link between water level, tree cover and brownification is

regarded as a common chain of reactions in ombrotrophic bogs and especially

in those exposed to human impacts (Laine and Vanha-Majamaa 1992,

Haapalehto et al. 2011). As the water table decreases tree establishment is

promoted and radial tree growth increased. This development accelerates peat

degradation and subsequently increases brownification in the bog pools.

According to our predictions the Ilybius group showed a negative relationship

to these habitat traits with the increase of forest cover towards the perimeter of

the raised bog. The effect of gain in tree cover and decrease in water level had

an indirect negative effect on Ilybius abundance by increasing the level of

brownification at a given site.

Over the past century the establishment of pine trees, sometimes at

accelerating rates, has been noted both on Lithuanian (Edvardsson et al.

2015a, Edvardsson et al. 2015b) and Swedish peat bogs (Linderholm and

Leine 2004, Edvardsson and Hansson 2015). At Aukštumala, about 75% of the

trees have established since the mid-1990s, which is believed to be mainly

related to the interaction between the current and historical peat cuttings and

warmer and drier summer conditions (Edvardsson et al. 2015b). Projections of

drier summers and increased temperatures are expected to enhance the

observed environmental gradients in the study area by an increasing peat

degradation, lowering the water table and promote tree establishment (Holden

et al. 2007). These expected changes should advance the observed gradients

in diving beetle abundance and the effect of the environmental variables in the

study area. Furthermore, increasing temperatures do not only enhance the

ongoing environmental changes in the European raised bogs, they also create

synergetic effects with other factors such as brownification. Thus it has been

shown that the combination of higher temperatures and brownification has an

negative effect in lakes and ponds by altering phytoplankton species

composition (Hansson et al. 2013), lowering zooplankton abundance and

shortening food chains (Urrutia-Cordero et al. 2016).

121

Our results highlight potential conservation mitigation which could facilitate and

preserve natural diving beetles and associated aquatic communities in bog

pool habitats. In agreement with other studies rewetting schemes aimed at

raising local water tables and thereby dampening the rate of tree

establishment and peat degradation would accommodate bog specific diving

beetles. In the Aukštumala raised bog the blocking of drainage ditches and

outlets has already been shown to reduce radial tree growth (Jarašius et al.

2015). Rewetting actions are also known to support the abundance of

ombotrophic bog invertebrates and buffer potential negative effects expected

from climate change scenarios (Carroll et al. 2011, Brown et al. 2016).

Furthermore, removal of already established pine trees is also expected to

have a positive effect on bog pool macroinvertebrates (Drinan et al. 2013). In

our study area, removal of pine trees might reduce brownification and its

negative influence on bog pool beetles.

Even though our conclusion are drawn from a single bog system we show that

when undergoing environmental changes, species communities associated

with open bog pools are affected by a multitude of factors leading to a gradual

decrease of the abundance of individuals. Our results demonstrate that bog

pool communities associated with pristine conditions near the center of the

raised bog decrease in abundance as the edge effect becomes stronger

species associated with the surrounding areas take over. This finding suggests

that bog pool habitats are influenced by the same core-edge gradients and

faces the same conservation challenges as species in other isolated unique

habitats. The increasing pressure from human habitat alterations, such as peat

cutting, together with current and future climate changes pose an immediate

threat for diving beetle communities in pristine bog pools.

Acknowledgements We thank Nerijus Zableckis, Lietuvos gamtos fondas, for providing

accommodation and practical assistance during the fieldwork. This study was

supported by the Danish Ministry of Higher Education and Science (Project

0604-02230B) and EU LIFE project LIFEAukstumala (LIFE12

NAT/LT/000965).

122

References • Aerts, R., B. Wallen, and N. Malmer. 1992. Growth-limiting nutrients in

Sphagnum-dominated bogs subject to low and high atmospheric nitrogen

supply. Journal of Ecology:131-140.

• Antonović, I., A. Brigić, Z. Sedlar, J. Bedek, and R. Šoštarić. 2012.

Terrestrial isopod community as indicator of succession in a peat bog.

Zookeys:171.

• Barbosa, O., and P. A. Marquet. 2002. Effects of forest fragmentation on

the beetle assemblage at the relict forest of Fray Jorge, Chile. Oecologia

132:296-306.

• Baselga, A. 2010. Partitioning the turnover and nestedness components of

beta diversity. Global Ecology and Biogeography 19:134-143.

• Baselga, A. 2013. Separating the two components of abundance‐based

dissimilarity: balanced changes in abundance vs. abundance gradients.

Methods in Ecology and Evolution 4:552-557.

• Baselga, A., and C. D. L. Orme. 2012. betapart: an R package for the

study of beta diversity. Methods in Ecology and Evolution 3:808-812.

• Beadle, J. M., L. E. Brown, and J. Holden. 2015. Biodiversity and

ecosystem functioning in natural bog pools and those created by rewetting

schemes. Wiley Interdisciplinary Reviews: Water 2:65-84.

• Bergsten, J., and K. B. Miller. 2006. Taxonomic revision of the Holarctic

diving beetle genus Acilius Leach (Coleoptera: Dytiscidae). Systematic

Entomology 31:145-197.

• Boggie, R. 1972. Effect of water-table height on root development of Pinus

contorta on deep peat in Scotland. Oikos:304-312.

• Breiman, L. 2001. Statistical modeling: The two cultures (with comments

and a rejoinder by the author). Statistical Science 16:199-231.

• Brown, L. E., S. J. Ramchunder, J. M. Beadle, and J. Holden. 2016.

Macroinvertebrate community assembly in pools created during peatland

restoration. Science of the Total Environment 569:361-372.

• Buchholz, S. 2016. Natural peat bog remnants promote distinct spider

assemblages and habitat specific traits. Ecological Indicators 60:774-780.

123

• Carroll, M. J., P. Dennis, J. W. PEARCE‐HIGGINS, and C. D. Thomas.

2011. Maintaining northern peatland ecosystems in a changing climate:

effects of soil moisture, drainage and drain blocking on craneflies. Global

Change Biology 17:2991-3001.

• Davies, K. F., B. A. Melbourne, and C. R. Margules. 2001. Effects of

within‐and between‐patch processes on community dynamics in a

fragmentation experiment. Ecology 82:1830-1846.

• Drinan, T. J., G. N. Foster, B. H. Nelson, J. O’Halloran, and S. S. C.

Harrison. 2013. Macroinvertebrate assemblages of peatland lakes:

Assessment of conservation value with respect to anthropogenic land-

cover change. Biological Conservation 158:175-187.

• Duff, A. 2012. Beetles of Britain and Ireland. Vol. I. Sphaeriusidae to

Silphidae. AG Duff Publishing, West Runton, Norfolk.

• Edvardsson, J., and A. Hansson. 2015. Multiannual hydrological

responses in Scots pine radial growth within raised bogs in southern

Sweden. Silva Fennica 49:1354.

• Edvardsson, J., E. Rimkus, C. Corona, R. Šimanauskienė, J. Kažys, and

M. Stoffel. 2015a. Exploring the impact of regional climate and local

hydrology on Pinus sylvestris L. growth variability–A comparison between

pine populations growing on peat soils and mineral soils in Lithuania. Plant

and Soil 392:345-356.

• Edvardsson, J., R. Simanauskiene, J. Taminskas, I. Bauziene, and M.

Stoffel. 2015b. Increased tree establishment in Lithuanian peat bogs--

insights from field and remotely sensed approaches. Sci Total Environ

505:113-120.

• Evans, D. 2006. The habitats of the European Union habitats directive.

Pages 167-173 in Biology and Environment: Proceedings of the Royal

Irish Academy. JSTOR.

• Ewers, R. M., and R. K. Didham. 2007. The effect of fragment shape and

species' sensitivity to habitat edges on animal population size.

Conservation Biology 21:926-936.

124

• Ewers, R. M., and R. K. Didham. 2008. Pervasive impact of large-scale

edge effects on a beetle community. Proceedings of the National Academy

of Sciences 105:5426-5429.

• Eyre, M., G. Foster, M. Luff, and S. Rushton. 2006. The definition of British

water beetle species pools (Coleoptera) and their relationship to altitude,

temperature, precipitation and land cover variables. Hydrobiologia

560:121-131.

• Fairchild, G. W., A. M. Faulds, and J. F. Matta. 2000. Beetle assemblages

in ponds: effects of habitat and site age. Freshwater Biology 44:523-534.

• Gažovič, M., I. Forbrich, D. F. Jager, L. Kutzbach, C. Wille, and M.

Wilmking. 2013. Hydrology-driven ecosystem respiration determines the

carbon balance of a boreal peatland. Science of the Total Environment

463:675-682.

• Gioria, M. 2014. Habitats. Pages 307-362 Ecology, Systematics, and the

Natural History of Predaceous Diving Beetles (Coleoptera: Dytiscidae).

Springer.

• Grace, J. B., D. R. Schoolmaster, G. R. Guntenspergen, A. M. Little, B. R.

Mitchell, K. M. Miller, and E. W. Schweiger. 2012. Guidelines for a

graph‐theoretic implementation of structural equation modeling. Ecosphere

3:1-44.

• Haddad, N. M., L. A. Brudvig, J. Clobert, K. F. Davies, A. Gonzalez, R. D.

Holt, T. E. Lovejoy, J. O. Sexton, M. P. Austin, and C. D. Collins. 2015.

Habitat fragmentation and its lasting impact on Earth’s ecosystems.

Science Advances 1:e1500052.

• Hannigan, E., and M. Kelly-Quinn. 2012. Composition and structure of

macroinvertebrate communities in contrasting open-water habitats in Irish

peatlands: implications for biodiversity conservation. Hydrobiologia

692:19-28.

• Hansson, L.-A., A. Nicolle, W. Granéli, P. Hallgren, E. Kritzberg, A.

Persson, J. Björk, P. A. Nilsson, and C. Brönmark. 2013. Food-chain

length alters community responses to global change in aquatic systems.

Nature Climate Change 3:228-233.

125

• Holden, J., L. Shotbolt, A. Bonn, T. Burt, P. Chapman, A. Dougill, E.

Fraser, K. Hubacek, B. Irvine, and M. Kirkby. 2007. Environmental change

in moorland landscapes. Earth-Science Reviews 82:75-100.

• Holmgren, M., C. Y. Lin, J. E. Murillo, A. Nieuwenhuis, J. Penninkhof, N.

Sanders, T. Bart, H. Veen, H. Vasander, and M. E. Vollebregt. 2015.

Positive shrub–tree interactions facilitate woody encroachment in boreal

peatlands. Journal of Ecology 103:58-66.

• Haapalehto, T. O., H. Vasander, S. Jauhiainen, T. Tahvanainen, and J. S.

Kotiaho. 2011. The Effects of Peatland Restoration on Water‐Table Depth,

Elemental Concentrations, and Vegetation: 10 Years of Changes.

Restoration Ecology 19:587-598.

• Jarašius, L., V. Lygis, J. Sendžikaitė, and R. Pakalnis. 2015. Effect of

Different Hydrological Restoration Measures in Aukðtumala Raised Bog

Damaged by Peat Harvesting Activities. Botanica Lithuanica 21:150-159.

• Jarašius, L., D. Matulevičiūtė, R. Pakalnis, J. Sendžikaitė, and V. Lygis.

2014. Drainage impact on plant cover and hydrology of Aukštumala raised

bog (western Lithuania). Botanica Lithuanica 20:109-120.

• Joosten, H., M.-L. Tapio-Biström, and S. Tol. 2012. Peatlands: guidance

for climate change mitigation through conservation, rehabilitation and

sustainable use. Food and Agriculture Organization of the United Nations.

• Karlsson, J., P. Byström, J. Ask, P. Ask, L. Persson, and M. Jansson.

2009. Light limitation of nutrient-poor lake ecosystems. Nature 460:506-

509.

• Kreutzweiser, D., F. Beall, K. Webster, D. Thompson, and I. Creed. 2013.

Impacts and prognosis of natural resource development on aquatic

biodiversity in Canada’s boreal zone 1. Environmental Reviews 21:227-

259.

• Kritzberg, E. S., W. Granéli, J. Björk, C. Brönmark, P. Hallgren, A. Nicolle,

A. Persson, and L. A. Hansson. 2014. Warming and browning of lakes:

consequences for pelagic carbon metabolism and sediment delivery.

Freshwater Biology 59:325-336.

126

• Laine, J., and I. Vanha-Majamaa. 1992. Vegetation ecology along a

trophic gradient on drained pine mires in southern Finland. Pages 213-233

in Annales Botanici Fennici. JSTOR.

• Larson, D., R. Roughley, and R. Aiken. 1990. Odonate predation as a

factor influencing dytiscid beetle distribution and community structure.

Quaestiones Entomologicae 26:151-162.

• Lefcheck, J. S. 2015. piecewiseSEM: Piecewise structural equation

modelling in r for ecology, evolution, and systematics. Methods in Ecology

and Evolution.

• Limpens, J., M. Holmgren, C. M. Jacobs, S. E. Van der Zee, E. Karofeld,

and F. Berendse. 2014. How Does Tree Density Affect Water Loss of

Peatlands? A Mesocosm Experiment. PLoS One 9:e91748.

• Linderholm, H. W., and M. Leine. 2004. An assessment of twentieth

century tree-cover changes on a southern Swedish peatland combining

dendrochronoloy and aerial photograph analysis. Wetlands 24:357-363.

• Lindholm, M., R. Wolf, A. Finstad, and D. O. Hessen. 2016. Water

browning mediates predatory decimation of the Arctic fairy shrimp

Branchinecta paludosa. Freshwater Biology.

• Lundkvist, E., J. Landin, and P. Milberg. 2001. Diving beetle (Dytiscidae)

assemblages along environmental gradients in an agricultural landscape in

southeastern Sweden. Wetlands 21:48-58.

• Mazerolle, M. J. 2003. Detrimental effects of peat mining on amphibian

abundance and species richness in bogs. Biological Conservation

113:215-223.

• Mieczan, T., M. Tarkowska-Kukuryk, W. Płaska, and J. Rechulicz. 2014.

Abiotic predictors of faunal communities in an ombrotrophic peatland lagg

and an open peat bog. Israel Journal of Ecology & Evolution 60:62-74.

• Miller, K. B., and J. Bergsten. 2016. Diving Beetles of the World:

Systematics and Biology of the Dytiscidae. JHU Press.

• Nakagawa, S., and H. Schielzeth. 2013. A general and simple method for

obtaining R2 from generalized linear mixed‐effects models. Methods in

Ecology and Evolution 4:133-142.

127

• Nilsson, A. 1986. Life cycles and habitats of the northern European

Agabini(Coleoptera, Dytiscidae). Entomologica Basiliensia. 1986.

• Nilsson, A. N., J. Elmberg, and K. Sjoberg. 1994. Abundance and species

richness patterns of predaceous diving beetles (Coleoptera, Dytiscidae) in

Swedish lakes. Journal of Biogeography:197-206.

• Nilsson, A. N., and M. Holmen. 1995. The Aquatic Adephaga (Coleoptera)

of the Fennoscandia and Denmark. Ii. Dytiscidae: II-Dytiscidea. Brill.

• Oertli, B., D. Auderset Joye, E. Castella, R. Juge, A. Lehmann, and J.-B.

Lachavanne. 2005. PLOCH: a standardized method for sampling and

assessing the biodiversity in ponds. Aquatic Conservation: Marine and

Freshwater Ecosystems 15:665-679.

• Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, G. L.

Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner. 2016. Package

‘vegan’ v. 2.3-5.

• Picazo, F., J. Moreno, and A. Millán. 2010. The contribution of standing

waters to aquatic biodiversity: the case of water beetles in southeastern

Iberia. Aquatic ecology 44:205-216.

• Price, J. S. 2003. Role and character of seasonal peat soil deformation on

the hydrology of undisturbed and cutover peatlands. Water Resources

Research 39.

• Purvinas M. 2006. Pelkininkų kaimai ir kolonijos Šilutės apylinkėse:

tradicinės gyvensenos bruožai ir gyvenviečių raida. – Liaudies kultūra

4:10–20.

• Ries, L., R. J. Fletcher Jr, J. Battin, and T. D. Sisk. 2004. Ecological

responses to habitat edges: mechanisms, models, and variability

explained. Annual Review of Ecology, Evolution, and Systematics:491-

522.

• Rydin, H., and J. K. Jeglum. 2013. The Biology of Peatlands. OUP Oxford.

• Schäfer, M. L., E. Lundkvist, J. Landin, T. Z. Persson, and J. O.

Lundström. 2006. Influence of landscape structure on mosquitoes

(Diptera: Culicidae) and dytiscids (Coleoptera: Dytiscidae) at five spatial

scales in Swedish wetlands. Wetlands 26:57-68.

128

• Scudder, G. 1987. Aquatic and semiaquatic Hemiptera of peatlands and

marshes in Canada. Memoirs of the Entomological Society of Canada

119:65-98.

• Seabloom, E. W., E. T. Borer, Y. Buckley, E. E. Cleland, K. Davies, J. Firn,

W. S. Harpole, Y. Hautier, E. Lind, and A. Macdougall. 2013. Predicting

invasion in grassland ecosystems: is exotic dominance the real

embarrassment of richness? Global Change Biology 19:3677-3687.

• Shipley, B. 2000. A new inferential test for path models based on directed

acyclic graphs. Structural Equation Modeling 7:206-218.

• Shipley, B. 2009. Confirmatory path analysis in a generalized multilevel

context. Ecology 90:363-368.

• Shipley, B. 2013. The AIC model selection method applied to path analytic

models compared using ad‐separation test. Ecology 94:560-564.

• Spitzer, K., and H. Danks. 2006. Insect biodiversity of boreal peat bogs.

Annual review of entomology 51:137.

• Stroup, W. W. 2012. Generalized linear mixed models: modern concepts,

methods and applications. CRC press.

• Sushko, G. G. 2016. Succession changes in diversity and assemblages

composition of planthoppers and leafhoppers in natural ancient peat bogs

in Belarus. Biodiversity and Conservation:1-17.

• Swengel, A. B., and S. R. Swengel. 2011. High and dry or sunk and

dunked: lessons for tallgrass prairies from quaking bogs. Journal of Insect

Conservation 15:165-178.

• Urrutia-Cordero, P., M. K. Ekvall, and L.-A. Hansson. 2016. Local food

web management increases resilience and buffers against global change

effects on freshwaters. Scientific reports 6.

• Vamosi, S. M., and B. Wohlfahrt. 2014. Community Patterns in Dytiscids.

Pages 409-436 Ecology, Systematics, and the Natural History of

Predaceous Diving Beetles (Coleoptera: Dytiscidae). Springer.

• van Duinen, G. A. 2013. Rehabilitation of aquatic invertebrate communities

in raised bog landscapes. Nijmegen: Radboud Universiteit Nijmegen.

• van Kleef, H. H., G.-J. A. van Duinen, W. C. Verberk, R. S. Leuven, G. van

der Velde, and H. Esselink. 2012. Moorland pools as refugia for

129

endangered species characteristic of raised bog gradients. Journal for

Nature Conservation 20:255-263.

• Verberk, W., G. Van Duinen, A. Brock, R. Leuven, H. Siepel, P.

Verdonschot, G. Van der Velde, and H. Esselink. 2006. Importance of

landscape heterogeneity for the conservation of aquatic macroinvertebrate

diversity in bog landscapes. Journal for Nature Conservation 14:78-90.

• Verberk, W. C., H. Siepel, and H. Esselink. 2008. Applying life‐history

strategies for freshwater macroinvertebrates to lentic waters. Freshwater

Biology 53:1739-1753.

• Zuur, A., E. N. Ieno, N. Walker, A. A. Saveliev, and G. M. Smith. 2009.

Mixed Effects Models and Extensions in Ecology with R. Springer Science

& Business Media.

130

Appendix – Chapter IV

Table S1: List of microhabitat types used in the study.

Habitats occurring at the shoreline (land–water interface)

Acronym Dominating plant structure

H1 Covered by Sphagnum spp but > 25 % of

the cover consist of Pinus detrius

H2 Cover dominated by Sphagnum spp with

an open vegetation of short stemmed

graminoids (e.g.Trichophorum cespitosum

or Rhynchosphora alba)

H3 Cover dominated by short stemmed

graminoids (e.g.Trichophorum cespitosum

or Scheuchzeria palustris). Sphagnum

spp occupying less than 25 % of the

cover.

H4 Cover dominated by Carex rostrata

H5 Cover dominated by Typha sp

131

Table S2: Studied diving beetles species observed in bog pools of the

Aukštumala raised bog.

Aciliini

Percentage of sites occupied

by the species

Acilius canaliculatus 30

Acilius sulcatus 82

Graphoderus austriacus 4

Graphoderus bilineatus 24

Graphoderus cinereus 40

Graphoderus zonatus 52

Ilybius

Ilybius aenescens 74

Ilybius ater 10

Ilybius crassus 8

Ilybius fenestratus 28

Ilybius fuliginosus 14

Ilybius guttiger 44

Ilybius quadriguttatus 10

Ilybius similis 4

Ilybius subaeneus 14

132

Table S3: Ficher’s C statistic of the final SEM (depicted in Fig.3) and each of

the paths not included in the structural model.

Aciliini

Missing path (Conditional variables have

been omitted from output table for

clarity) estimate

std.

error df

critical

value

P-

value

Microhabitat diversity ~ water level + … -0.16 0.10 74 -1.58 0.12

Brownification ~water level + … 0.04 0.13 76 0.34 0.73

Microhabitat diversity ~ Aciliini

abundance + … 0.02 0.02 73 0.72 0.48

Brownification ~ Aciliini abundance + … 0.01 0.03 74 0.31 0.75

Brownification ~ pool area + … -0.19 0.11 75 -1.69 0.10

Fisher’s C statistic: df

P-

value

11.62 10 0.311

Ilybius

Missing path (Conditional variables have

been omitted from output table for

clarity) estimate

std.

error df

critical

value

P-

value

Microhabitat diversity ~ water level + … -0.16 0.10 74 -1.58 0.12

Brownification ~water level + … 0.04 0.13 76 0.34 0.73

Ilybius abundance ~ tree cover + … 0.16 0.11 75 1.47 0.14

Brownification ~ pool area + … -0.19 0.11 75 -1.69 0.10

Fisher’s C statistic: df

P-

value

13.48 8 0.096

133

Table S4: Spatial correlation of effect sizes

Original model

Parameter interaction with longitude

β estimate (standard deviation)

P-value

Ilybius abundance ~ water level

+ pool area + brownification +

microhabitat diversity + Tree

cover

water level -0.08 (0.11) 0.53

pool area -0.22 (0.35) 0.46

brownification -0.03 (0.05) 0.59

microhabitat

diversity 0.06 (0.17) 0.74

tree cover 0.13 (0.14) 0.34

Aciliini abundance ~ water level

+ pool area + brownification +

microhabitat diversity + Tree

cover

water level 0.42 (0.29) 0.16

pool area 0.21 (0.12) 0.07

brownification 0.09 (0.05) 0.08

microhabitat

diversity -0.22 (0.16) 0.17

tree cover 0.22 (0.13) 0.09

Brownification ~ water level +

tree cover

water level -0.05 (0.16) 0.92

tree cover -0.23 (0.4) 0.56

Microhabitat diversity ~tree

cover + water level + pool area

tree cover 0.19 (0.18) 0.27

pool area 0 (0.16) 1.00

water level -1.38 (0.27)

<

0.001

Pool area ~ tree cover + water

level tree cover -0.56 (0.23)

< 0.05

water level 0.46 (0.4) 0.24

Tree cover ~ water level water level -0.5 (0.31) 0.11

134

Fig S1: A conceptual framework with the assumed influence of the selected

environmental variables on diving beetle abundance.

135

Fig S2: The relationship between species richness and abundance of all study

diving beetle species

136