UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

A house of cards: Patterns of aquatic invertebrate diversity in agricultural ditches

Whatley, M.H.

Publication date2014Document VersionFinal published version

Link to publication

Citation for published version (APA):Whatley, M. H. (2014). A house of cards: Patterns of aquatic invertebrate diversity inagricultural ditches.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an opencontent license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, pleaselet the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the materialinaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letterto: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. Youwill be contacted as soon as possible.

Download date:27 Jul 2021

201518-os-Whatley.indd 1 02-04-14 09:22

A House of Cards

Patterns of aquatic invertebrate diversity in agricultural ditches

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus

prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties ingestelde commissie,

in het openbaar te verdedigen in de Agnietenkapel op dinsdag 13 mei 2014, te 14:00 uur

door Merrin Hazel Whatley

Geboren te Auckland, Nieuw-Zeeland

Promotie commissie Promotor: Prof. dr. W. Admiraal Co-promotor: Dr. H.G. van der Geest Overige leden: Dr. P.D. Armitage

Prof. dr. K. Kalbitz Dr. W.C.E.P. Verberk

Prof. dr. ir. P.F.M. Verdonschot Prof. dr. J.T.A. Verhoeven Faculteit der Natuurwetenschappen, Wiskunde en Informatica

_____________________________________________________________ This research was funded by Hoogheemraadschap Hollands Noorderkwartier (HHNK) and Stichting Waterproef, Edam. The study was conducted at the department of Aquatic Ecology and Ecotoxicology (AEE), Institute for Biodiversity and Ecosystem Dynamics (IBED), University of Amsterdam (UvA). Cover design by Nigel Upchurch and Merrin Whatley Printed by Ipskamp Drukkers BV

Copyright © 2014 by Merrin H. Whatley

ISBN: 978-94-6259-158-5

“It is to be emphasized that although patterns may underlie the rich and varied tapestry of the natural world, there is no single simple pattern.” Robert M. May 1974

Contents

GENERAL INTRODUCTION................................................................................ 7

MACROPHYTE LOSS DRIVES DECADAL CHANGE IN BENTHIC INVERTEBRATES IN PEATLAND DRAINAGE DITCHES ..................... 21

THE ROLE OF EMERGENT VEGETATION IN STRUCTURING AQUATIC INSECT COMMUNITIES IN PEATLAND DRAINAGE DITCHES .................................................................................................................... 43

LINKAGES BETWEEN BENTHIC MICROBIAL AND FRESHWATER INSECT COMMUNITIES IN DEGRADED PEATLAND DITCHES ....... 69

TEMPORAL ABIOTIC VARIABILITY STRUCTURES INVERTEBRATE COMMUNITIES IN AGRICULTURAL DRAINAGE DITCHES ................ 93

SYNTHESIS .............................................................................................................. 125

CONCLUSIONS ....................................................................................................... 135 REFERENCES ......................................................................................................... 137 SUMMARY ............................................................................................................... 150 SAMENVATTING ................................................................................................... 153 ACKNOWLEDGMENTS ......................................................................................... 157

Ilybius fenestratus

Chapter 1

General Introduction

Chapter 1

8

A house of cards is an unstable structure, which is at risk of collapsing if either knocked by an external force or if an essential element of the house is lost. By making the analogy between a house of cards and the diversity of macroinvertebrate communities occupying North Holland’s agricultural drainage ditches I ask the reader to consider the stability of an ecosystem which has lost many of its elements. Herein arise two important questions, 1) what are the key environmental drivers which structure the house of cards, and 2) can the stability of the structure be improved in order to keep it standing? To address these questions this thesis reviews long-term and contemporary provincial monitoring data in combination with data collected during field experiments and from analytical analysis to disentangle the key environment divers and underlying mechanism of aquatic invertebrate diversity in the province. In low-lying agricultural areas, drainage ditches are ubiquitous features of the landscape and provide valuable habitat to a range of aquatic species, including macroinvertebrates (Beltman 1983; van der Hammen 1992; Armitage et al. 2003; Herzon & Helenius 2008; Verdonschot 2012). Yet, in the province of North Holland, The Netherlands, an increase in land use intensification and the associated loss of habitat has caused a significant decline in species diversity (Kleijn et al. 2004; van Eerden et al. 2010). Aquatic invertebrates are diverse both in form and in the habitats they occupy. Invertebrates feed on primary producers, organic material and other invertebrates making nutrients available for larger animals, such as fish and birds, which prey on invertebrates (Fig. 1.1). In addition, invertebrates are involved in fundamental ecosystem functions, including the recycling of organic matter. Invertebrates stimulate microbial communities (fungi and bacteria) through grazing, breaking down Coarse Particulate Organic Matter (CPOM) and reworking sediments, thereby affecting sediment structure and redox conditions (Hunting et al. 2012). Thus, invertebrates are highly interconnected with other components of the ditch ecosystem and are responsive to changes in landscape heterogeneity, vegetation and habitat structure, nutrient enrichment and macro-ions (Scheffer et al. 1984; Higler & Verdonschot 1989; Armitage et al. 2001; Vlek, et al. 2004; Verberk et al. 2006; O’Toole et al. 2008) (Fig. 1.1).

Chapter 1

9

Patterns of invertebrate species diversity vis-á-vis community composition are underpinned by biological and abiotic conditions and the interactions between them. Broadly speaking, community assemblages can be considered at three different levels, regional diversity (γ diversity), local diversity (α diversity) and the divergence in species among local communities (β diversity) (Fig. 1.2). Key biological and abiotic conditions include the total regional pool of species, the ability of species to colonize habitats, competition and predation, degree of connectivity between suitable habitats and environmental filtering (MacArthur & Wilson 1967; Conrad et al. 1999; Bilton, Freeland & Okamura 2001; Chase 2003; Donald & Evans 2006; Winemiller et al. 2010).

Fig. 1.1. Invertebrate trophic position and sediment reworking in drainage ditches. Environmental filtering essentially influences community composition by filtering out those species that are poorly adapted to the conditions of an environment, while selecting for species that possess traits which increase their chance of survival and reproductive success (Poff 1997; Chase 2003; Verberk, van Noordwijk & Hildrew 2013). For example, the larvae of the rare dragonfly

Chapter 1

10

(Aeshna viridis) may be excluded from waters due to absence of the water soldier plant (Stratiotes aloides) which provides essential refugia and habitat for the larvae, thus the absence of S. aloides functions as a filter for Aeshna viridis (Rantala et al. 2004; Suhonen et al. 2013). Indeed many aquatic invertebrate species are dependent upon vegetation, especially in linear drainage ditches where submerged and emergent vegetation provide the primary structural habitats (Scheffer et al. 1984; Higler & Verdonschot 1989; Verdonschot, et al. 2012a). Consequently, present day patterns of species composition are the result of interactions between the environment and the organisms living within it and as such can be site specific (e.g. Suren et al. 2008; Verdonschot et al. 2012a). Under these circumstances the underlying environmental drivers can be difficult to ascertain from taxonomic data alone, yet the analysis of species traits offers the opportunity to investigate the causal mechanisms underlying the relationship between environmental drivers and taxonomic composition (see Verberk et al. 2013 and references therein). Moreover, aquatic invertebrates are one of the few groups for which a wide variety of biological traits have been described (Usseglio-Polatera et al. 2000; Tachet et al. 2002; Bonada et al. 2006; Verberk et al. 2008a). Spatial and temporal environmental features as proposed by Southwood (1977) may be viewed as a habitat templet, upon which taxonomic and functional community composition are structured through natural selection and evolutionary events (Townsend & Hildrew 1994). Therefore, there is a tendency for biological traits (i.e. reproductive behaviour, body size and dispersal potential) to be taxonomically linked by evolutionary processes, leading to sets of traits being associated with one another (Giller et al. 2004). For example species which have synchronized life-cycles often support resting stages such as diapause (Mousseau & Dingle 1991; Gonzalez et al. 2001; Lytle & Poff 2004; Beche et al. 2006). The chance of success under a synchronized life-cycle strategy is thus greatly increased in predictable environments as these species have evolved to wait out poor conditions so individuals will encounter favourable conditions on emergence. Moreover, traits are not isolated from one another and it is preferable to consider them as assemblages bound together through individual species and organised by temporal selection events (Stearns 1976). Yet, as a single species hold numerous traits, practical problems often arise in determining relationships between traits and how best to statistically analyse trait expression. Thus, to gain mechanistic insights from trait analysis species traits must be combined in a meaningful way, life-history strategies are one such combination (Stearns 1976; Poff 1997; Verberk et al. 2013). A species life-history strategy is shaped (in-part) by environmental filtering processes. Therefore the representation of life-history strategies within a community can shed light on how species are influenced by their environment and have evolved to overcome particular obstacles (Poff 1997; Lytle & Poff 2004; Verberk et al. 2008a; b, 2013). For example, invertebrate

Chapter 1

11

species may experience the environment as stable and predictable, changeable but in a regular and predictable manner or unpredictable. In theory, these different environmental conditions will filter species, resulting in the predictable expression of species life-history strategies (see, Fig. 1.2). Agricultural intensification European agricultural landscapes have been in existence for thousands of years and the initial patchy clearance of dense broad leaf forests opened up the landscape, creating a mosaic of habitats. Yet, with growing concern over future land requirements for food and biofuel production many governments have implemented production based incentives. The subsequent agricultural intensification relies heavily on inputs of fertilizer, herbicides and pesticides, and caused a reduction in vegetation diversity in these landscapes, leading to the homogenization of habitats (Busch 2006). Consequently, many species are now in decline in agricultural landscapes with those habitats surrounded by agricultural land becoming increasingly isolated (Donald & Evans 2006; Hendrickx et al. 2007). Of all environments, aquatic ecosystems are the most at risk, with current agricultural demands estimated to account for 70% of all global freshwater withdrawals (Molden 2007). Increasing agricultural production will heighten the demand for water and simultaneously increase the diffuse release of nutrients, sediments, pesticides and herbicides to water bodies, exposing aquatic life to multiple stressors associated with eutrophication, increased turbidity and toxicity (Wood & Armitage 1997; Camargo et al. 2005). Nowhere in Europe has agricultural intensification been more absolute than in the low-lying landscapes of The Netherlands (van Dam 2001). The effects on aquatic ecosystems in these landscapes have been particularly significant over the last hundred years, which has greatly reduced the ecosystems carrying capacity of wetlands for birds (van Eerden et al. 2010). Therefore, understanding the mechanisms by which environmental filters are structuring aquatic communities in agricultural landscapes is a fundamental step towards reducing the impact on biodiversity and protecting the same freshwater resources so essential to agriculture.

Chapter 1

12

Fig 1.2. Environmental filtering of species and their respective life-history strategies. Letters denote individual species and icons represent different life-history strategies. Species are filtered from a regional pool of species (γ diversity) resulting in local community (α diversity), and the difference between two local communities, species turnover (β diversity). This figure is adapted from Southwood (1977); Chase (2003) and Verberk et al. (2013). Study site The province of North Holland (The Netherlands) was historically part of a vast delta system which contained an expanse of raised bog wetlands. Yet much of these original peat wetlands have been drained for agricultural production and the peat cut away and sold for fuel (van Dam 2001; Lamers et al. 2002). The drainage of these peatlands resulted in wide-scale land subsidence and subsequently the province, which was then exposed to the open sea (the Zuiderzee), was frequently flooded during high tides and storm surges (van Dam 2001). This flooding resulted in a unique flora and fauna in the area which included both freshwater species alongside those adapted to brackish conditions. Marine flooding also deposited macro-ions, particularly sulphates and iron on the land which can still be found at relatively high concentrations in the soils today (Pons 1992). In 1932 the situation changed with the construction of the Afsluitdijk which closed of the Zuiderzee creating the largest manmade lake in the world, the IJsselmeer. Essentially this created a

Chapter 1

13

freshwater environment overnight and the waters and soils in the province have been gradually desalinising ever since (van der Hammen 1992; van Dam 2009). Today the peat areas account for less than a third of their original extent (Fig. 1.3). The province of North Holland is characterized by an interlacing network of drainage ditches, which are the dominant water type in this landscape and underlie a unique hydrology. The majority of North Holland is situated between 1 and 2 meters below sea-level and the water levels are strictly managed. This has led to the development of hydrological terminology which is quintessentially Dutch. The term polder is commonly used and refers to an area of low-lying land and the ditches encompassed within it. Polders are protected by a series of dykes and their water levels are strictly managed by a system of pumping stations and inlets to achieve optimal conditions for agricultural production. The main inputs of water to polder ditches are rain, inlet of river derived waters, agricultural run-off and upward seepage, while water is removed from the system by pumping, evaporation and downward seepage.

Fig 1.3. The extent of peatlands in the province of North Holland in a) 850 AD (based on van Eerden et al. 2010), and b) present day extent. The water let into polders originates from the river Rhine and is rich in minerals (notably, Cl-, HCO3- and SO42-) (Roelofs 1991; Lamers et al. 2002). The inlet of these waters can alter chemical conditions in the aquatic environment. High nutrient concentrations and period anoxia in sediments stimulates anaerobic breakdown of organic matter with nitrates, sulphates and iron-oxides serving as electron acceptors. Iron-sulphides can be produced as a by-product of sulphate and iron-oxide reduction and compete for sorption sites with sediment bound phosphorus, which is then released back into the water

Chapter 1

14

(Lamers et al. 2002; Smolders et al. 2006). Moreover, mineralization processes are stimulated by the addition of bicarbonate which buffers the water, neutralizing decay-inhibiting acids. This combination of factors, in addition to artificially low water tables, can be particularly damaging in peatlands, leading to rapid and wide-spread peat soil degradation (Roelofs 1991; Janse & Van Puijenbroek 1998; Lamers et al. 2002; Smolders et al. 2006). The hydrological network incorporates a series of large arterial canals that move mineral rich water in and around the province which is then mixed with the rainwater that is pumped out of the polders. This network can be viewed upon as a nested structure with polders nested within larger canals, networks of ditches nested within polders and individual ditches within the network containing their own micro-habitats, e.g. benthic sediments, macrophytes and open water (Fig. 1.4). Monitoring, vegetation removal and dredging The European Water Framework Directive (2000/60/EC; WFD) was introduced by the European Parliament and Council in the year 2000 with the goal of achieving a “good chemical and ecological status of waters” for inland, coastal and groundwater(s) within all European Union member states with the initial date set for 2015 (http://ec.europa.eu/environment/water/water-framework). The status of a water body is assessed against chemical and biological criteria. These criteria are set lower for artificial or highly modified waters, in which case reaching a “good ecological potential” is the goal. Determining what the “ecological potential” of their artificial and modified waters is one of the major challenges for water managers in the province of North Holland (as for the entirety of The Netherlands) with all of North Holland’s inland waters falling into this category (Ligtvoet et al. 2008). Since the introduction of the WFD there have been further incentives for Dutch water managers to assess the state of their waters and find the means of improving the chemical and ecological conditions of these waters. Being the most densely populated country in Western Europe and supporting one of the most intensive agricultural systems, eutrophication is one of the main issues affecting water quality in The Netherlands which has implications for both chemical and biological conditions. To address this, measures have been taken over recent years to reduce the amount of nutrient entering surface waters, predominantly by improving wastewater treatment processes (Ligtvoet et al. 2008; Junier & Mostert 2012). Yet, despite these efforts there has been little change in nutrient levels in surface waters because the majority of the nutrients entering surface waters originate from agricultural production (Ligtvoet et al. 2008; Junier & Mostert 2012). In addition, the WFD is based on the concept of

Chap

ter 1 15

Fi

g. 1.

4. A

n ex

ample

of a

nest

ed h

ydrol

ogica

l syst

em in

Nort

h H

ollan

d’s p

older

lands

cape

a) a

erial

photo

graph

of a

pea

tland

pold

er (W

orme

r en

Jispe

rveld)

alo

ngsid

e a cl

ay p

older

(Wijd

ewor

mer),

b) a

sche

matic

repr

esent

ation

of th

e nest

ed h

ydro

logica

l syst

em.

Chapter 1

16

river basins and the quality of drainage ditches do not need to be reported to the EU Council, despite being the dominant water type in The Netherlands. Consequently, ditches are viewed primarily as infrastructure and the biodiversity value of these waterways are largely overlooked by water managers (Verdonschot 2012). Yet, many ecological studies have been conducted in drainage ditch systems, especially in the Netherlands (Scheffer et al. 1984; Higler & Verdonschot 1989; van der Hammen 1992; Armitage et al. 2003; Verdonschot 2012; van Zuidam 2013). Moreover, the availability of long-term monitoring data, collected by the water authority, makes it possible to track changes that have occurred in these ecosystems over decades of intensive land use, as well as during the more recent measures to improve water quality (van der Hammen 1992; van Dam 2009). As one of the main functions of ditches is to drain water from the surrounding land there is a long tradition of managing vegetation growing in and around North Holland’s drainage ditches to prevent flooding. Vegetation used to be removed by hand annually, yet it is now done mechanically. Machinery efficiently removes virtually all vegetation growing within and along the banksides of the ditches. Moreover, the frequency and type of machinery used greatly influences both composition and growth of vegetation and thus influences the invertebrate community which depend upon it for vital habitat (Twisk et al. 2000; Lamers et al. 2002; van Zuidam & Peeters 2012). In addition to vegetation removal sediment dredging is performed, predominantly in degraded peatlands in which ditches quickly accumulate fine, particulate, organically rich sediments. Dredging can improve water quality by removing nutrient rich sediments and improve water clarity, however, the benefits are only short-term if the wide-scale problems of eutrophication and desiccation of peat soils are not addressed (Lamers et al., 2002; Verberk et al., 2007). Moreover, the act of dredging is very disruptive and can negatively impact aquatic invertebrates, particularly if undisturbed patches of vegetation are not left as refugia for invertebrates from which recolonization can take place following dredging activities (Twisk et al., 2003; Verberk et al., 2007). Thus, the availability of long-term and contemporary monitoring data for the province of North Holland offers a unique opportunity to study biodiversity patterns in a historical, yet rapidly modernizing and intensively managed agricultural environment. Scope of this thesis This thesis seeks to elucidate how aquatic invertebrate communities are structured by environmental drivers (i.e. filters) in intensively managed agricultural drainage ditches in the province of North Holland, The Netherlands. As the greatest density of drainage ditches occur in peatland

Chapter 1

17

regions which are also remnants of the original wetlands, peatlands are the focal point of this thesis. To this end the following questions are defined:

What are the key environmental drivers determining aquatic invertebrate community composition in agricultural ditches?

What are the mechanisms underlying the response of aquatic invertebrates to these environmental drivers?

How can management practices be adjusted to improve ecological conditions in these agricultural ditches?

Thesis outline Conditions have changed in the agricultural landscape of North Holland over recent decades. Despite efforts made to reduce nutrient loading in the province, predominately by improving waste water treatment, there has been a dramatic decline in the diversity of submerged macrophytes in the province. Yet, it is not clear how the decline in submerged macrophytes relates to abiotic conditions or how this has impacted the invertebrate community, therefore Chapter two covers the analysis of monitoring data collected in peatland ditches over a twenty two year period from 1985 till 2007. In combination with this review a field experiment was run using artificial plastic plants which were sampled alongside natural habitats to reveal the habitat preferences of invertebrate species occupying peatland drainage ditches (Chapter two). Although many ditches in the province contain few or no species of submerged plant they can still support extensive emergent reed beds dominated by Phragmites australis and Typha angustifolia. Yet, this vegetation is mechanically cleared away annually to prevent clogging and maintain the drainage function of ditches. In Chapter three the importance of emergent vegetation structure on the taxonomic and life-history strategy composition of aquatic insects was investigated in both a eutrophic peatland, containing no submerged macrophytes and in a restored mesotrophic peatland with submerged vegetation. The chemical environment in ditches is likely to be strongly affected by the inlet of mineral rich river waters; however linking abiotic parameters directly with invertebrate data in the field may give unclear results, partly due to the strong correlations between abiotic parameters. Furthermore, ditches are predominantly shallow environments with a high level of interaction between benthic sediments and littoral environments. Therefore, in Chapter four microbial phospholipid profiles were analysed in conjunction with data on abiotic parameters, macrophytes and aquatic insects to determine how abiotic conditions were influencing the ditch community and if parameters indicative of mineral rich waters could be identified as drivers. Abiotic variability is expected to be high in the province due to the influence of mineral rich water, yet to date there have been no studies

Chapter 1

18

investigating the role this plays in structuring the invertebrate community. Moreover, determining how this may impact the distribution of species, in conjunction with average concentrations of nutrients and macro-ions, can provide essential information on how to promote biodiversity in the landscape. Therefore, Chapter five investigates the relationship between abiotic conditions (variability and average conditions), invertebrate species-turnover and the representation of insect life-history strategies. The analysis includes data collected from drainage ditches over a 7 month period at the provincial wide scale, covering the three dominant soil types (sand, clay and peat) to determine if diversity patterns are unique to different areas. In Chapter six the culmination of findings and ideas presented throughout the thesis are discussed. The implications of these findings for the management of invertebrate communities and the overall biodiversity in this landscape are considered and some recommendations for monitoring and managing this unique landscape are offered.

Chapter 1

19

Chapter 1

Elodea nuttallii

Chapter 1

Chapter 2

Macrophyte loss drives decadal change in benthic invertebrates in peatland drainage ditches

Published: M.H. Whatley, van Loon E.E., van Dam H., Vonk J.A., van der Geest H.G. and Admiraal W. (2014)

Freshwater Biology 59, 114-126.

Chapter 2

22

Abstract Agricultural peatlands and their associated drainage systems are often highly managed and exposed to anthropogenic pressures, such as eutrophication and stable water tables, maintained via drainage during periods of high rainfall and inlet of, alkaline rich, waters during dry periods. These pressures promote peat degradation, resulting in the accumulation of fine degraded peat particles which dramatically alter aquatic habitats by smothering surfaces and decreasing water quality. Consequential effects on benthic communities are expected but have not been investigated so far. We hypothesized that peat degradation can lead to the decline of submerged macrophytes, which are of critical importance to sustaining biodiversity of benthic invertebrate communities. To investigate this we analysed decadal (1985 – 2007) changes in benthic species richness in 29 peat ditches in The Netherlands and, to determine patterns of macroinvertebrate habitat occupancy, carried out a complementary field experiment with submerged artificial macrophytes, natural sediments and emergent bank vegetation. Results from long-term monitoring indicate that chemical conditions in agricultural peat ditches have improved slightly over the last decades; however there has been a simultaneous decline in benthic invertebrate species richness and densities corresponding to a decline in the numbers of submerged macrophytes. The apparent dependence of macroinvertebrates on macrophytes was reinforced by our field experiment which revealed that invertebrate density was highest in submerged artificial plants, while invertebrate species richness was highest in natural emergent vegetation. Conversely, degraded peat sediments supported extremely few invertebrates. Our results clearly illustrate the strong influence of submerged macrophyte loss on macroinvertebrate assemblages in peatland waters. Furthermore, this suggests that improvements in water quality alone will not benefit invertebrates in the absence of suitable vegetative habitats. Introduction Peatlands are valuable ecosystems, recognised for their natural, social and economic resources, but these habitats have been greatly reduced. Centuries of human induced modifications including peat extraction, intensive agriculture and nutrient inputs have resulted in the degradation of peatlands. It has been estimated that peatlands once covered approximately 24% of The Netherlands (850AD), while today they account for less than 2% of its area (van Eerden et al. 2010). All natural peat ecosystems have been lost and those remaining have been drained for agriculture and are, to some extent, degraded (van Dam 2001; Lamers et al. 2002; van Eerden et al. 2010). Agricultural peatlands are often intersected by networks of drainage ditches, which can provide valuable habitat for aquatic organisms (Painter 1999;

Chapter 2

23

Armitage et al. 2003; Herzon & Helenius 2008; Verdonschot et al. 2011). Smaller ditches (< 20 m wide) in particular, being the most numerous and spatially heterogeneous aquatic habitats in these peatlands, facilitate the occurrence of a wide range of benthic invertebrate species. For example (Williams et al. 2003) noted that, despite having low local invertebrate species richness, the ditches supported more rare species than rivers and streams. The communities of small ditches, however, are often exposed to eutrophication, due to the leaching of nutrients from the adjacent agricultural soils (Janse & van Puijenbroek 1998) and changes in water quality prompted by hydrological management (Roelofs 1991; Lamers et al. 2002; Smolders et al. 2006). Water tables in Dutch peatland areas are maintained within strict limits by pumping excess water out during wet periods and the inlet of external waters, originating from the River Rhine via Lake IJsselmeer, during dry periods. Inlet waters have higher concentrations of sulphate and carbonate than area specific peatland waters. This can increase alkalinity of surface waters and cause degradation of peat soils and subsequent release of fine organic particles into the aquatic environment (Roelofs, 1991). Abiotic conditions in the resulting sludge layer cause the release of nutrients from organic material previously conserved in peat soils; i.e. internal eutrophication (Fig. 2.1). Subsequent algae blooms and growth of floating species (Lemna and Azolla), coupled with the aforementioned sludge accumulation, increase turbidity and light attenuation in ditches, consequently submerged macrophyte coverage declines dramatically (Wood & Armitage 1997; Lamers et al. 2002; Harrison et al. 2007; Verberk et al. 2007). Changes in water quality including pH, dissolved oxygen, alkalinity and macro-ions (e.g. Cl-, K+, Mg2+ and Ca2+), can strongly influence invertebrate assemblages (van der Hammen 1992; Verbruggen et al. 2011), while nutrient enrichment predominantly influences invertebrates indirectly via a reduction in dissolved oxygen and loss of macrophyte habitat (Verbruggen et al., 2011).

Fig. 2.1. Relationship between physicochemical and biological components in agricultural peatland ditches under degradation. Schematic adapted from Verberk et al. (2007).

Chapter 2

24

Submerged macrophytes provide essential habitat for macro-invertebrates (Scheffer et al. 1984; Kovalenko et al. 2010), since they offer protection from fish predation (Goyke & Hershey 1992) and provide a range of food sources (Underwood & Thomas 1990; Newman 1991), therefore macrophyte decline can negatively impact invertebrate assemblages. Although it is known that invertebrate communities are influenced by deterioration in water quality the impact of submerged macrophyte loss resulting from peat degradation has not been widely studied. We tested the role of macrophytes in driving benthic invertebrate community composition in North Holland peatland drainage ditches. To investigate the relationship between submerged macrophyte communities and benthic invertebrates we reviewed long-term trends in benthic biodiversity and abiotic characteristics from regional monitoring data. In support of this analysis, a field experiment was run in agricultural peatland ditches in which we sampled macroinvertebrates from natural and artificial substrata (macrophytes) to determine patterns of invertebrate occupancy in different habitats. We expect to observe a strong influence of habitat structure on invertebrates and a higher species richness and density in vegetated habitats (both natural and artificial) compared to bare sediments. Methods Long-term trends in benthic richness –regional monitoring data We investigated temporal trends in water quality, submerged macrophytes and macroinvertebrates by reviewing monitoring data collected over 22 years (1985 to 2007) by the North Holland Water Authority, Hoogheemraadschap Hollands Noorderkwartier (HHNK). All monitoring sites were ditches (width < 16 m, water depth < 1.2 m) situated in peat areas. A complete overview of the HHNK monitoring data covering this period and sample collection details can be found in van Dam (2009). Monitoring data, including benthic invertebrates, submerged macrophytes and abiotic variables were reviewed. The majority of monitoring locations were visited at least 3 times during the 22 year period with some locations sampled more than others (Fig. 2.2a). Observations for invertebrates were available for more locations than for macrophytes while abiotic variables (total nitrogen (mg L-1), total phosphorus (mg L-1), chloride (mg L-1), chlorophyll-a (μg L-1), total sulphate (SO42-, mg L-1) and water transparency (cm)) were available for all locations but not during all time periods. The invertebrate dataset comprised of 198 samples derived from 29 locations and 17 sampling years between 1985 and 2007. Macroinvertebrates (adult body length > 1 mm) were sampled in either spring or late summer/autumn and were identified to species, with the exception of worms

Chapter 2

25

(Oligochaeta) which were not identified beyond this subclass and were included only for density analysis. Samples were collected with a dip-net using a multi-habitat sampling technique whereby different habitats (i.e. submerged macrophytes, sediments, open water and emergent vegetation) were subsampled and combined to make one sample. The combined sampled length of the ditch covered approximately 5 m, equivalent to an area of 1.5 m-2 within a 100 m section. Invertebrate abundance was standardized to density (individuals m-2). The macrophyte dataset consisted of 72 samples derived from 22 locations and 11 sampling years between 1990 and 2007. Macrophyte species were sampled over a 50 m transect covering the width of the entire ditch in late summer/autumn and included all floating and submerged aquatic plants. Invertebrate habitat occupancy – 2010 field experiment To test the responses of macroinvertebrates to the presence of submerged macrophytes, artificial substrata were deployed within the Wormer and Jisperveld (52º30’N, 4º50’E), The Netherlands (Fig. 2.2b). The Wormer and Jisperveld is a low-lying peat meadow of approximately 2500 ha, comprising low intensity agricultural land intersected by drainage ditches (Janssen et al. 2005). Submerged macrophytes have completely disappeared due to rapid peat degradation and accumulation of extremely fine (< 200 μm) peat particles, which has created a thick layer of amorphous mud and very turbid conditions.

Fig. 2.2. Map showing the position of HHNK monitoring locations and 2010 sampling locations. The small insert map indicates the position of the sample locations within The Netherlands. a) Monitoring locations were sampled from 1985 to 2007; symbol size represents the number of samples collected for invertebrates (triangles) and macrophytes (circles) at each location. b) Position of the three ditches in the Wormer and Jisperveld (North Holland, the Netherlands) where artificial plants were placed and dip-net sampling undertaken in 2010. Image adapted from Janssen et al. (2005). Sampling was undertaken over a three week period from late July to early August 2010. Three separate ditches with a similar morphology (width <

Chapter 2

26

17 m, water depth < 1 m) were selected for invertebrate sampling and water quality measurements (Fig. 2.2b). Morphological parameters were measured in situ and, due to the highly mobile nature of the degraded peat material, sediment (degraded peat sediments) and water depths were averaged from 18 measurements taken in each ditch. Samples of undisturbed overlying water were analysed three times over the sampling period. Measurements of conductivity, turbidity, dissolved oxygen (DO, measured between 10 am – 2 pm, 5 cm below the surface and at the sediment-water Interface), and pH were taken in the field. Conductivity was measured with a WTW LF 92 meter and Tetracon 96 cell (Weilheim, Germany) and oxygen (DO) and pH were measured with a WTW pH/Oxi 340i/set meter (Weilheim, Germany). Turbidity was measured with a WTW TURB 350 IR meter (Weilheim, Germany) and subsequently converted to water transparency (cm) using the conversion table developed by Kevin Fermanich (University of Wisconsin 2010). Analysis of nitrogen (TN), phosphorus (TP and PO43-), sulphate (SO42-), total iron (Fe), carbonate (CO32-), chloride (Cl-) and chlorophyll-a were carried out in the laboratory using standardised national protocols accredited by the Netherlands Standards Institute (NEN). To determine patterns of invertebrate habitat occupancy, macroinvertebrates (adult body length > 1 mm) were sampled from three different habitats within each ditch; emergent bank-side vegetation, bare sediments and submerged artificial plants. From this point forward the habitats will be referred to as Bank, Sediment and Plastic Plant, respectively. Bank vegetation was dominated by reed species characteristic of ditches in the region (Phragmites australis and Typha angustifolia).

Fig. 2.3. Individual components (left) and the assemblage of the artificial substrata (right). Substrata consisted of a Plastic Plant mounted on a non-buoyant plastic base, placed on plant fibre (Hessian cloth) to simulate root structure and housed inside a galvanized steel cage. Artificial substrata, consisting of a Plastic Plant (resembling the common water plant hornwort, Ceratophyllum demersum) mounted on a non-

Chapter 2

27

buoyant plastic base and placed on plant fibre (Hessian cloth) to simulate root mass, were housed in galvanized steel cages (mesh 1 cm, base 15 x 15 cm, height 25 cm) (Fig. 2.3). In each ditch four cages were deployed adjacent to the bank and approximately 5 cm above the sediment-water Interface (Fig.2 4). Substrata were left in the field for 56 days prior to sampling to allow time for colonization (Wise & Molles 1979; Higler & Verdonschot 1989). Invertebrates were collected over a 3 week period from late July to early August, 2010. This sampling period is in accordance with current and long-term monitoring protocols (Bijkerk 2010). Bank and Sediment samples were collected with a dip-net (mesh, frame, bag depth; 900 μm, 25 x 29 cm, 25 cm) by sweeping the net continuously along a 5 m transect. This sampling method was adopted to maintain distinct samples for each habitat and is comparable to standard monitoring practices in The Netherlands. Plastic Plants were sampled by placing a dip-net under the substrata before removing them from the water, to reduce invertebrate loss. Invertebrate density from Plastic Plant substrata were standardized assuming a 5% macrophyte coverage area, equivalent to the median macrophyte coverage recorded during the early monitoring period (1990 – 1991).

Fig. 2.4. Ditch sampling scheme for 2010 macroinvertebrate field experiment, showing the positioning of Plastic Plants and where Bank and Sediment samples were collected with a dip-net along a continuous 5 m length taken parallel to the shoreline. Four replicate invertebrate samples were taken in each habitat type (Bank, Sediment and Plastic Plants) in each ditch (Fig. 2.4), giving a total of 12 samples per ditch (with the exception of ditch 3 for which one Sediment sample was lost). Invertebrate samples were taken back to the laboratory and sorted live and subsequently preserved in 70% ethanol for later determination. Where possible, invertebrates were identified to species, with the exception of Chironomid larvae (Diptera), which were identified to subfamily. The following taxonomic groups were identified with the corresponding keys: Ephemeroptera

Chapter 2

28

(Elliott & Humpesch 2010), Heteroptera (Savage 1989), case-bearing Trichoptera (Wallace et al. 2003), caseless Trichoptera (Edington & Hildrew 1995), Hirudinea (Elliott & Mann 1979), Gastropoda (Macan 1977), Coleoptera and Odonata (Nilsson 2005) and Diptera (Nilsson 2005). Amphipoda, Isopoda and Mysida were each represented by a single common species, and Bivalvia were grouped by genus. Mites (Arachnida) and worms (Oligochaeta) were not identified to species but abundance was recorded for density calculations. Statistical Analysis Long-term monitoring data (1985 – 2007) were collected repeatedly from 29 locations. Due to the monitoring design, sequential observations of abiotic variables, invertebrates and macrophytes taken from the same location were not statistically independent from one another. For this reason Generalized Estimating Equations (GEE) were used to analyse the long-term trends and correlations between variables (Liang & Zeger 1986). The basic model was defined with sampling location as the clustering variable, season (in case of invertebrates) as the predictor variable, location as the factor and either time (year), macrophyte species richness or abiotic variables as covariables. Changes in invertebrate species richness over time were investigated at the local scale (α diversity), equivalent to the average species richness per sample, the regional scale (γ diversity), equivalent to the regional species pool over a two year period, and by calculating an additive model of beta diversity (βAdd), equivalent to γ – α, to determine the variation among communities during the monitoring period (Lande 1996). In this study we used additive βAdd diversity because it had the advantage of being in the same units as α and γ diversity and, thus, straightforward comparisons between these different measures of diversity were possible (Lande 1996; Anderson et al. 2011). GEE model predictions were plotted with observed species richness (i.e. α diversity) of macrophytes and invertebrates over time to determine temporal trends and between invertebrate richness against macrophyte richness to ascertain the relationship between these two groups.

Chapter 2

29

The predictions were calculated by the following equations: yj = β0,j + β1x (1a) β0,j = uj + β0 (1b) where: y = species richness at location j; β0,j = model intercept for location j; β0 = component of model intercept which is independent of location; uj = location-specific component of intercept; β1 = slope of x; x = value of the covariable, either year, macrophyte species richness or abiotic variable; In this study we will report the form of this model which is aggregated over the different locations: y = β0 + β1x (2a) β0 = Σj (nj β0,j/N) (2b) where the summation is a weighted average using the relative number of observations per location (nj/N) to determine the relative importance of each intercept β0,j . The general form of a GEE (Equation 2) can then be used to predict at locations different from those where measurements were collected. By doing so, the term uj disappears from the model and the variance of the values uj adds to the model error. Models run on invertebrate species richness and macrophyte richness were tested against all covariables, while invertebrate densities (individuals m-2) were analysed over time. We applied an autoregressive correlation structure to our model to correct for correlations between observations from the same location in close temporal succession to one another (Quinn & Keough 2002). The distributions of species richness and abiotic variables were normal with an identity link function, i.e. the dependent variable was not transformed within the GEE model. Densities of dominant invertebrate taxa displayed skewed residual distributions, therefore a gamma distribution with a log link function was applied in the GEE model to analyse temporal trends of individual density. GEE models were run in IBM SPSS Statistics (v. 20). Data collected during the 2010 field experiment were analysed to compare species richness and densities of benthic invertebrates between the three ditch habitats. Invertebrate abundances were converted to density (individuals m-2). As two different sampling methods were used during the field study (artificial substrata and net samples of naturally occurring substrata)

Chapter 2

30

species richness was rarefied against invertebrate abundance to standardise samples (Gotelli & Colwell 2001) using EcoSim version 7.72 (Gotelli & Entsminger 2011). Since the three different habitats were located within each ditch, nested-ANOVAs (habitat nested within ditch) were run in IBM SPSS Statistics (v. 20) to test for significant differences in species richness and invertebrate density between ditches and habitats. In the case of a significant test result, a Tukey HSD post hoc tests was run. Normality of both the monitoring and 2010 field experiment data were checked with a Shapiro-Wilk test and QQ-plots were used to assess homogeneity of variances. If these assumptions were not met data were log10 transformed or, in the case of density data, log10 (x + 1) transformed prior to statistical analysis.

Fig. 2.5. Overview of invertebrate species richness during the monitoring period (1985 – 2007). Three diversity indices were derived: γ = regional diversity, i.e. total number of species collected over two years; α = local diversity, i.e. mean species richness of a single sample (± 95% CI); βAdd = γ - α, i.e. difference in species richness between regional and local scales. Grey bars indicate the number of samples collected within each time period. No diversity indices were calculated for 1987 – 1989 since only one sample per year was available for this period. Results Long-term trends –regional monitoring data There was a significant reduction in total nitrogen (β1 = -0.006, S.E. = 0.002, P = 0.005, r2 = 0.97, n = 98), total phosphorus (β1 = -0.011, S.E. = 0.005, P = 0.038, r2 = 0.75, n = 193), total sulphate (β1 = -0.015, S.E. = 0.002, P < 0.001, r2 = 0.71, n = 181) and chloride (β1 = -0.008, S.E. = 0.002, P < 0.001, r2 = 0.76, n = 193) in surface waters during the monitoring period. Although water transparency declined over time (β1 = -1.83, S.E. = 0.77, P = 0.018, r2 = 0.81, n

Chapter 2

31

= 112), no trend was observed for chlorophyll-a concentrations (data not shown). Transparency was not significantly related to surface water chlorophyll-a concentration, due in part to temporal variation, but was negatively correlated to total phosphorus (β1 = -26.68, S.E. = 11.73, P = 0.023, r2 = 0.80, n = 112). Total sulphate covaried with several other variables, chlorophyll-a (β1 = -0.397, S.E. = 0.189, P = 0.036, r2 = 98, n = 88) total nitrogen (β1 = -0.20, S.E. = 0.090, P = 0.026, r2 = 98 , n = 88), chloride (β1 = 0.599, S.E. = 0.083, P < 0.001, r2 = 0.63, n = 181) and transparency (β1 = 73.38, S.E. = 20.18, P < 0.001, r2 = 0.86, n = 112). Further interactions between abiotic parameters relating to seasonal dynamics and complex non-linear relationships were also probable but such analysis was beyond the scope of this study. All measures of macroinvertebrate diversity (α, γ and βAdd) decreased over time and βAdd diversity was positively affected by the number of samples collected in each time period (Fig. 2.5). Therefore, α diversity (species richness from this point forward) was considered to be a representative proxy of invertebrate diversity within the monitoring dataset. Between 1990 and 2007 macrophyte species richness declined significantly in 83% of monitoring locations (GEE, β1 = -0.38, S.E. = 0.06, P < 0.001, r2 = 0.68, n = 72) with an overall loss of eight macrophyte species during this period (Fig. 2.6a). Macrophyte species richness was not related to nutrients, however it was negatively correlated with chlorophyll-a concentration (GEE, β1 = -31.71, S.E. = 15.21, P = 0.023, r2 = 0.82, n = 33) and positively correlated with water transparency, although this was on the boundaries of statistical significance (GEE, β1 = 0.15, S.E. = 0.08, P = 0.050, r2 = 0.84, n = 31). Invertebrate species richness declined simultaneously with macrophyte richness, a significant decline was observed in 79% of monitoring locations between 1985 and 2007 (GEE, β1 = -0.014, S.E. = 0.002, P < 0.001, r2 = 0.43, n = 198), with an overall loss of 43 species over this period (Fig. 2.6b). Invertebrate species richness was not affected by sampling season or any measured abiotic parameter but was significantly positively correlated with macrophyte richness (GEE, β1 = 0.024, S.E. = 0.006, P < 0.001, r2 = 0.35, n = 72) (Fig. 2.7).

Chapter 2

32

Fig. 2.6. Species richness of a) macrophytes and b) invertebrates plotted over time. Grey dots represent field observations and the black line is the GEE prediction, calculated from the model formula. Dashed lines show 95% confidence belts for the GEE prediction.

Fig. 2.7. Invertebrate species richness plotted against macrophyte species richness. Grey dots represent observations in the field and the black line is the GEE prediction, calculated from the model formula. Dashed lines show 95% confidence belts for the GEE prediction.

Chap

ter 2

33

Tabl

e 2.1

. Med

ian d

ensit

ies (N

o. m-2

) at t

he b

eginn

ing a

nd en

d of

the 2

2 yea

r mon

itorin

g peri

od fo

r dom

inant

inver

tebra

te ta

xa in

each

orde

r with

at l

east

24

obser

vatio

ns o

r 12%

of a

ll sa

mples

and

Gen

eraliz

ed E

stima

ting

Equ

ation

mod

el res

ults o

f den

sities

aga

inst t

ime.

Rang

es in

den

sity

of ea

ch ta

xon

are g

iven

in pa

renth

eses.

Valu

es we

re ca

lculat

ed fo

r 200

6 be

caus

e only

one o

bserv

ation

was

ava

ilable

for 2

007.

Tax

a are

listed

in or

der o

f lea

st to

most

numb

er of

obser

vatio

ns,

N/A

– to

o few

obser

vatio

ns fo

r stat

istica

l ana

lysis,

n.s.

– n

ot sig

nific

ant.

Med

ian d

ensit

y (ra

nge)

G

EE

mod

el ou

tput

Ta

xon

Ord

er

1985

20

06

β1

S.E

. P

r2 N

o. O

bs.

Oect

is fur

va

Trich

opte

ra

7 (2

- 11

) N

/A

-0.0

7 0.

01

< 0

.001

0.

62

24

Ischn

ura

elega

ns

Odo

nata

7

(3 -

10)

N/A

-0

.07

0.02

<

0.0

01

0.39

43

H

aliplu

s rufi

collis

Co

leop

tera

18

(17

- 18)

N

/A

-0.0

5 0.

04

n.s.

n.s.

45

Siali

s lut

aria

Meg

alopt

era

4 (1

- 41

) 1

(1 -

3)

-0.0

4 0.

02

n.s.

n.s.

55

Neom

ysis i

ntegr

a M

ysid

a 2

(1 -

70)

N/A

-0

.09

0.03

0.

009

0.43

65

O

ligoc

haet

a

Olig

ocha

eta

730

(17

- 841

) 11

(5 -

16)

-0.0

8 0.

04

0.02

5 0.

55

100

Helo

bdell

a sta

gnali

s H

irudi

nea

15 (1

2 - 1

8)

4 (1

- 15

) 0.

05

0.02

0.

008

0.30

10

7 Pl

anor

bis p

lanorb

is G

astro

poda

44

(38

- 50)

12

(5 -

43)

-0.0

4 0.

02

n.s.

n.s.

108

Caen

is sp

. E

phem

erop

tera

8

(1 -

25)

1 (1

- 23

) -0

.01

0.03

n.

s. n.

s. 11

2 A

sellus

sp.

Isop

oda

4 (3

- 4)

1

(1 -

5)

-0.0

2 0.

02

n.s.

n.s.

125

Chiro

nomu

s sp.

D

ipte

ra

207

(1 -

300)

3

(1 -

14)

-0.1

1 0.

02

< 0

.001

0.

28

145

Gam

maru

s tigr

inus

A

mph

ipod

a 8

(2 -

62)

6 (1

- 13

) 0.

02

0.02

n.

s. n.

s. 15

4 A

rrenu

rus s

p.

Ara

chni

da

4 (1

- 31

9)

3 (1

- 9)

-0

.08

0.03

0.

002

0.18

16

3 Si

gara

sp.

Het

erop

tera

62

(10

- 70)

2

(1 -

22)

-0.0

6 0.

02

0.00

2 0.

32

165

All

taxa

-

1412

(128

4 - 2

084)

11

4 (3

3 - 1

79)

-0.0

3 0.

01

< 0

.001

0.

39

198

Chapter 2

34

Total invertebrate density (number of individuals m-2) declined significantly over time. Changes in the median densities and the ranges of the dominant taxa in each order are shown alongside GEE model outputs reflecting temporal trends in the densities of taxa (Table 2.1). The majority of taxonomic groups declined significantly, including worms (subclass: Oligochaeta), Arrenurus sp. (Arachnida), Oectis furva (Trichoptera), Ischnura elegans (Odonata) and Sigara sp. (Heteroptera). However, the greatest rate of decline was observed in nonbiting midges Chironomus sp. (Diptera) followed closely by the crustacean Neomysis integer (Mysida). The predatory leech Helobdella stagnalis (Family: Glossiphoniidae) was the only species to increase significantly in density over the 22 year monitoring period. Invertebrate habitat occupancy – 2010 field experiment The three ditches in the Wormer and Jisperveld were comparable in salinity (chloride), water transparency, pH and nutrients (Table 2.2). Furthermore, all ditches had a steep oxygen gradient, with surface water DO ranging between 12.9 – 4.6 mg L-1, while DO at the sediment-water interface ranged between 0.1 – 0.6 mg L-1. Ditch 1 had the highest concentrations of TN and TP and chlorophyll-a, although these differences were not significant.

Fig. 2.8. Non-rarefied species richness (Non-rar.) and cumulative species richness of habitats rarefied against 20 (Rar. 20) and 100 individuals (Rar. 100). The results were pooled from 12 samples of each habitat over the three ditches (with the exception of the habitat Sediment which comprised of 11 samples). The rarefaction process was computed repeatedly 1000 times. Error bars represent ± 95% CI.

Chapter 2

35

Table 2.2. Morphological and chemical characteristics of the three ditches sampled in the Wormer and Jisperveld in 2010. All values are calculated from the mean of three measurements (unless otherwise noted) ± 95% CI. Measurements

1

Ditch 2

3

Ditch width (m) 7 ± 3 10 ± 1 5 ± 1

Water depth (cm)a 30.3 ± 4.8 23.0 ± 1.8 63.0 ± 9.9

Sediment depth (cm)a 14.8 ± 4.1 16.5 ± 2.5 30.6 ± 4.1

DO 5cm water depth (mg L-1)b 9.3 ± 2.6 10.1 ± 2.8 5.8 ± 1.2

DO SWI (mg L-1)c 0.3 ± 0.3 0.3 ± 0.2 0.3 ± 0.2

Turbidity (NTU) 38 ± 4 29 ± 7 45 ± 14

Transparency (cm)d 22 ± 2 28 ± 5 20 ± 5

Chlorophyll-a (μg L-1) 147 ± 160 24 ± 7 91 ± 68

pH 8.7 ± 0.9 8.8 ± 1.4 7.9 ± 0.3

Conductivity (μS cm-1) 814 ± 123 1061 ± 107 1014 ± 35

Chloride (mg Cl- L-1) 140 ± 30 177 ± 31 170 ± 23

Carbonate (mg CO32- L-1) 157 ± 45 200 ± 30 190 ± 20

Sulphate (mg SO42- L-1) 50.0 ± 8.6 85.3 ± 4.3 83.7 ± 5.4

Total Iron (μg Fe L-1) 183 ± 17 123 ± 69 56 ± 7

Total P (mg P L-1) 0.35 ± 0.12 0.17 ± 0.08 0.2 ± 0.07

Orthophosphate (mg P L-1) N/A 0.03 ± 0.02 0.01 ± 0.01

Total N (mg N L-1) 5.3 ± 1.5 3.5 ± 0.5 3.2 ± 0.7 aAveraged from 18 measurements per ditch bDissolved Oxygen (DO) cSediment-Water Interface (SWI) dConverted from turbidity according to University of Wisconsin (2010) The 2010 field experiment revealed that non-rarefied invertebrate species richness was significantly higher in Plastic Plants compared to Sediment samples (nested ANOVA, F(5,3) = 34, P < 0.001). However, the greatest number of species were recorded in Bank samples, which were significantly higher than Plastic Plants, rarefied against 200 individuals (nested ANOVA, F(5,3) = 49, P < 0.001) and Sediment samples, non-rarefied data (nested ANOVA, F(5,3) = 54, P < 0.001) (Fig. 2.8). The majority of species found in Plastic Plants (non-rarefied) were from the taxonomic groups Gastropoda, Hirudinea and Trichoptera (6 – 10

Chapter 2

36

sp.), while emergent Bank vegetation had the highest number of species in the Coleoptera, Diptera and Gastropoda (11 – 15 sp.) followed by Hirudinea, Heteroptera and Trichoptera (6 – 10 sp.). The number of species was low in all taxonomic groups in Sediments (1 – 5 sp.) (Fig. 2.9). Total invertebrate density (individuals m-2) differed significantly between habitats (nested ANOVA, F(8,6) = 196, P < 0.001), being highest in Plastic Plants (mean density ± 95% CI, 816 ± 171), intermediate in Bank vegetation (303 ± 100) and lowest in Sediments (34 ± 11). In Plastic Plants the most numerous taxa were G. tigrinus (Amphipoda) and Chironomus larvae (Diptera) (> 100 m-2), followed by S. lutaria (Megaloptera) (51 – 100 m-2) then Asellus aquaticus (Isopoda), Bithynia tentaculata (Gastropoda), H. stagnalis (Hirudinea), Caenis sp. (Ephemeroptera), Cyrnus flavidus (Trichoptera) and Oligochaeta (21 – 50 m-2). Emergent Bank vegetation supported the highest density of N. integer (Mysida) (< 100 m-2) followed by G. tigrinus (Amphipoda) and Chironomidae larvae (Diptera) (51 – 100 m-2) and, finally, Ischnura elegans (Odonata) (21 – 50 m-2) (Fig. 2.9). Invertebrate density was exceptionally low in sediments, with over 50% of all taxonomic groups present at a density < 1 m-2

and predominantly consisted of the most common taxa, namely Chironomus larvae, G. tigrinus and N. integer. Discussion The aim of this study was to determine the role of submerged macrophyte habitats in driving benthic macroinvertebrate community composition in agricultural peatland waters. The combined analysis of long-term monitoring data and the complementary 2010 field experiment data draws attention to the strong relationship between submerged macrophytes and invertebrate species richness. Over the 22 year monitoring period invertebrate and macrophyte species richness declined synchronously. Moreover, the prominent colonization of artificial submerged macrophytes (Plastic Plants) within two months after placement in the field experiment demonstrated the rapid response of invertebrate communities to macrophyte habitat availability. Invertebrate densities recorded in Plastic Plant substrata were similar to natural densities recorded at the beginning of the HHNK monitoring period (1985), before macrophytes had been lost from these peat ditches. These changes were highlighted by a shift from clear water conditions, in which a range of macrophyte species were present, to a turbid state characterised by high concentrations of suspended particles and dominated by algae. The driver for this shift was probably long-term degradation of peat soils prompted by oxygenation and the inlet of sulphate rich alkaline waters, causing subsequent particle accumulation in the aquatic environment (Roelofs 1991; Lamers et al. 2002).

Chapter 2

37

Fig. 2.9. Invertebrate density (individuals m-2) and species richness represented in each taxonomic group within each habitat. The total number of species in each taxonomic group is indicated between brackets. Values are calculated from the average of 12 replicate samples, except for the habitat Sediment. N/A indicates no data available. Physicochemical processes underlying declining benthic species richness in North Holland peatlands In recent years, measures have been taken to reduce nutrient inputs into North Holland peatland ditches although, while TN, TP and SO42- concentrations have declined over time, concentrations remain high and the majority of waters are still eutrophic to hypertrophic. The effects of reducing external nutrient loadings to surface waters may have been masked by nutrient recycling in these historically enriched peatlands. Under anoxic conditions, microbial breakdown of organic material and the further release of nutrients is exacerbated by high concentrations of ions like sulphate and nitrate which can act as electron acceptors (Jones 1979; Caraco, Cole & Likens 1989), promoting bacterial degradation of peat. Sulphate reduced to sulphide can bind to reduced iron, generating iron sulphides (FeS and FeS2) (Holmer & Storkholm 2001). This chemical competition for iron causes iron bound phosphates (Fe-PO4) to be released into the water column (Smolders et al. 2006). Roelofs (1991) demonstrated that this process was further accelerated in Dutch peatlands under alkaline conditions, causing a positive feedback of continued peat degradation and high nutrient availability in ditches. The monitoring data indicates that high nutrient concentrations have persisted for several decades, coincident with a reduction in water transparency in North Holland’s agricultural peatlands, compromising submerged macrophyte growth and favouring algae and floating macrophytes (such as Lemna sp.) (Barko et al. 1991; Søndergaard et al. 2003; Geurts et al. 2009). The

Chapter 2

38

loss of submerged macrophytes in these shallow, soft sediment ditches can lead to the establishment of an alternative stable state via mechanisms similar to those found in productive shallow lakes (Scheffer et al. 1993) characterized by turbid conditions under which light attenuation increases, retarding the growth of submerged vegetation (Scheffer 1990). The decline in water clarity due to fine sediments combined with high nutrient availability initiates macrophyte decline while growth of algal and floating plants, alongside unstable sediments, constrains macrophyte re-establishment (Janse & van Puijenbroek 1998; Schutten et al. 2005). Consequences of submerged macrophyte loss for benthic invertebrates The importance of macrophyte habitat structure is well established and, in line with our study, has been identified previously as an important driver of the spatial distribution of invertebrates within ditches (Scheffer et al. 1984). Macrophytes provide habitats required for various invertebrate life stages e.g. oviposition and emergence (McLaughlin & Harris 1990; Orr & Resh 1992), and hence their structure and complexity are central to promoting invertebrate diversity and overall density (Jeffries; Higler & Verdonschot 1989; Lucena-Moya & Duggan 2011). For example, Kovalenko et al. (2010) and Hansen et al. (2011) reported a positive relationship between macrophyte complexity and invertebrate abundance. In a similar study conducted in eutrophic agricultural ditches, Hinojosa-Garro et al. (2010) found that increased vegetation complexity supported more invertebrate species and higher densities of predators (Coleoptera, Hemiptera and Odonata) and grazers (Gastropoda). Comparable results were obtained in our field study in which Plastic Plants also supported higher densities of grazing snails and the predatory alderfly S. lutaria (Megaloptera) and caddisfly Cyrnus flavidus (Trichoptera). Sediments composed of degraded peat have a semi-liquid structure providing no support for infauna (e.g. Chironomidae and Oligochaeta) and no refuge from predation by benthivorous fish (Moss & Timms 1989). Moreover, the positive influence of macrophytes on sediment oxygen conditions (Carpenter & Lodge 1986) and stability (Moss & Timms 1989) results in feedback mechanisms, whereby the loss of vegetation exacerbates physical and chemical conditions within degraded peat sediments. We observed both long-term negative effects of declining macrophyte habitat on invertebrate species richness and density, as well as short-term positive effects of introduced artificial macrophytes on the invertebrate community. The decline in invertebrate richness over the monitoring period was observed over both local and regional scales, as reflected by changes in γ (gamma) and βAdd (beta) diversity (Fig. 2.5). These findings suggest that a reduction in macrophyte habitat, driven by the aforementioned peat degradation process, underlies the regional decline in invertebrate richness over

Chapter 2

39

the last two decades. During the monitoring period densities of Chironomus sp. (Diptera), Oectis furva (Trichoptera) and Oligochaeta declined significantly while, in our 2010 field study, densities of these same taxa were greater in submerged artificial macrophytes than in natural vegetation. Observed differences in species richness between Plastic Plants and natural Bank vegetation could be explained by the greater habitat complexity provided by natural vegetation, containing a range of species compared to Plastic Plants, which represented a single plant species. Aside from providing a physical habitat for organisms, differences in invertebrate densities between artificial and natural substrata could be partly explained by increased food availability to grazers via growth of epiphytic algae (Carpenter & Lodge 1986; Underwood & Thomas 1990) or by protection of prey species from benthivorous fish, known to be present in the area (Hofman 2007). Invertebrate densities in artificial plants, however, were comparable to densities seen in North Holland ditches in 1985 (when macrophytes were present) and within the ranges of those found in other lentic freshwaters with submerged macrophytes (Viljoen, Cyrus & Wepener 2001; Storey 2007; Verdonschot et al. 2011). Additionally, it takes time for periphyton communities to develop on Plastic Plants and although the artificial substrata were left in the field for more than 50 days before sampling, we assume this food source is more readily available on natural vegetation. This suggests that food availability and predation were not the predominant factors determining the extensive and rapid colonisation of artificial macrophytes by invertebrates. Benthic macroinvertebrate assemblages in degraded peatland ditches are stressed by hypoxia (caused by eutrophication), fine particle accumulation and habitat loss. The continued decline in benthic species richness and water transparency in recent years, despite decreases in external nutrient loading, suggests that sediment resuspension and internal nutrient remobilization, exacerbated by peat degradation, are underlying this trend. In our study we observed the regional shift from macrophyte to algae dominated systems caused by eutrophication and sediment loading. The results of our long-term monitoring data analysis and field experiment highlights the importance of macrophyte habitat structure to invertebrate communities. The underlying negative effects of peat degradation on the benthic invertebrate community therefore appeared to be largely indirect through the resulting loss of macrophytes. Management implications It is evident that benthic invertebrate species richness would improve in agricultural peat ditches if submerged macrophytes were re-established. In waters devoid of submerged plants invertebrates are concentrated along zones of emergent bank vegetation. Thus, importance should be placed on

Chapter 2

40

maintaining and expanding the existing vegetation. The combination of fine particle accumulation, eutrophication and increased alkalinity causes multiple stressor effects on both invertebrates and submerged vegetation (Lamers et al. 2002; Verberk et al. 2007). Due to the potentially high rate of internal nutrient release in agricultural peat areas, peat degradation needs to be addressed by removing nutrient-rich sediment (e.g. by dredging) in combination with a reduction of external nutrient loading and the inlet of alkaline and SO42- enriched waters (Lamers et al. 2002; Smolders et al. 2006). This could be facilitated by allowing more flexible water tables, moving towards an integrative terrestrial-aquatic management approach (Janssen et al. 2005) aimed at reducing overall peat oxidation and mineralization. Ultimately a clear-water phase is necessary to allow submerged macrophytes to re-establish. To achieve this re-establishment, nutrients and particularly light attenuation need to be reduced to values lower than when the system switched initially (van Nes et al. 2002). To facilitate an increase in transparency sediment resuspension must be reduced. This may be achieved through biomanipulation to reduce the numbers of benthic and planktivorous fish, in combination with dredging to remove nutrient-rich fine sediments and to reduce sediment resuspension (Lamers et al. 2002; Verberk et al. 2007). In a field study run in several hydrologically isolated ditches in the Wormer and Jisperveld, Hofman (2007) found this combination of dredging and fish removal increased zooplankton numbers, improved water clarity and led to the re-establishment of submerged vegetation. This demonstrates that rehabilitation of smaller, isolated ditches is achievable, although applying such measures to the whole peat area would be very costly and may not give the same results, particularly in larger water bodies that are more exposed to the wind. In conclusion, this study demonstrates the strong influence of submerged macrophyte loss and introduction of artificial macrophytes on benthic invertebrate assemblages. High nutrient concentrations and oxidation of peatlands has resulted in the degradation of peaty soils, leading to the release of fine particles into the water. Amorphous sediments composed of degraded peat particles provide unsuitable habitats for benthic invertebrates while eutrophication, coupled with increased alkalinity and low dissolved oxygen, have additional detrimental effects on the invertebrates. The accumulation of degraded peat particles in the aquatic environment predominantly has an indirect negative effect on invertebrates, by triggering the loss of macrophytes, underlying the need to address both physicochemical and biological components in the management and restoration of agricultural peatlands.

Chapter 2

41

Acknowledgments We would like to thank Gert van Ee, Emile Nat and Ron van Leuken for their constructive advice throughout this project, Ivo Roessink, José van Diggelen and Fons Smolders for orientating us at the start of the project, Nigel Upchurch and Thijs de Boer for their help with the maps and diagrams, Ellard Hunting for his feedback on the manuscript, Pim Koelma, Coen Wagner and Alejandra Goldenberg for their assistance during fieldwork and Andre Timmer and Ed Zijp of Natuurmonumenten, for allowing us to access the Wormer and Jisperveld. This research was funded by Stichting Waterproef and Hoogheemraadschap Hollands Noorderkwartier.

Chapter 2

42

Phragmites australis

Chapter 2

43

Chapter 3

The Role of Emergent Vegetation in Structuring Aquatic Insect Communities in Peatland Drainage Ditches

Under revision: M.H. Whatley, van Loon E.E., Vonk J.A., van der Geest H.G. and Admiraal W.

(Aquatic Ecology)

Chapter 3

44

Abstract Availability of macrophyte habitat is recognized as an important driver of aquatic insect communities in peatland drainage ditches; however, eutrophication can lead to the decline of submerged vegetation. While emergent vegetation is able to persist in eutrophicated ditches vegetation removal, carried out during ditch maintenance, can reduce the availability of this habitat. In this study we applied the landscape filtering approach to determine if the absence of emergent vegetation is a habitat filter which structures aquatic insect communities in peatland drainage ditches under different trophic conditions. To this end, a field study was carried out in one mesotrophic (NM) and one eutrophic (WJ) peatland in the province of North Holland, The Netherlands. We assigned life-history strategies to insect species and applied linear mixed models and Redundancy Analyses (RDA) to taxonomic and functional aquatic insect community data. Our results indicate that while differences between peatlands primarily determine the species pool within each wetland, emergent vegetation acted as a secondary filter by structuring functional community composition within ditches. The eutrophic peatland was dominated by insects adapted to abiotic extremes, while species with good dispersal abilities were strongly related to emergent vegetation cover. This study demonstrates the applicability of life-history strategies to provide insight on the filtering of species due to availability of emergent macrophyte habitat. To ensure greater diversity of insect communities in ditch habitats it is recommended that some vegetation be spared during maintenance to leave patches from which insect recolonization can occur. Introduction Peatland drainage ditches are productive environments that potentially support a range of aquatic plant and insect species (Armitage et al. 2003; Twisk et al. 2003; Herzon & Helenius 2008; Verdonschot et al. 2011). In The Netherlands many peatlands are managed as agricultural systems and intersected by ditches to maintain constant water tables. Regular vegetation management is carried out to maintain the drainage function and prevent the possible terrestrialisation of ditches (Twisk et al. 2000; Lamers et al. 2002). But efficient vegetation removal reduces the habitat available for aquatic insects as well as other organisms, such as birds, fish and periphytic algae. In addition, intensive land use has caused many Dutch peatlands to become eutrophicated, predominantly due to mineralization of peat soils leading to internal eutrophication in combination with external eutrophication caused by run-off of nutrient rich water from surrounding (agricultural) lands and the inlet of sulphur rich, alkaline water (Sinke et al. 1990; Lamers et al. 2002).

Chapter 3

45

In the province of North Holland, water originating from the River Rhine is supplied to peatlands during dry periods. River Rhine water increases carbonate and sulphate concentrations in peatlands, which in turn increases peat mineralization and nutrient release (Lamers et al. 2002; Smolders et al. 2006). Hydrological management regimes in The Netherlands result in different peatlands being subjected to varying levels of nutrient enrichment. Eutrophication can negatively affect aquatic insects in various ways by reducing oxygen levels (Verdonschot 2012) and inducing declines in submerged vegetation, resulting in the loss of habitat and reduced availability of periphytic food (Varga 2003; Hinojosa-Garro et al. 2010). Furthermore, eutrophication can lead to the accumulation of degraded peat sediments, which form an anoxic, amorphous layer of mud, excluding species that are ill-equipped to cope with these harsh conditions (Verberk et al. 2007; 2008b).

Submerged vegetation has declined in North Holland peatlands over recent decades (Whatley et al. 2014). This is partly due to eutrophic conditions causing light attenuation, by increasing water turbidity and excessive growth of algae and floating plants (i.e. Lemna and Azolla sp.), which outcompete submerged plants (Janse & van Puijenbroek 1998). However, emergent helophyte species (e.g. Phragmites australis and Typha angustifolia) can tolerate poor water quality and are able to form extensive stands if left unmanaged. Emergent vegetation can support a variety of aquatic insects (Murkin et al. 1992; Radomski & Goeman 2001), their aerial structures serve to orientate flying insects and underwater stems provide essential habitat for oviposition and emerging nymphs and larvae (de Szalay & Resh 2000). While aquatic insects are negatively influenced by the loss of submerged vegetation (Keast 1984; Hinojosa-Garro et al. 2010; Lucena-Moya & Duggan 2011; Whatley et al. 2014), the presence of emergent vegetation stands may serve as a suitable surrogate habitat for some species. Therefore, the absence of emergent macrophyte habitat could operate as a habitat filter and structure aquatic insect communities in peatland ditches.

Landscape filtering, as proposed for lotic systems by Poff (1997), is essentially the sorting of species via a set of hierarchical habitat filters (i.e. operating at different scales from microhabitat to catchment characteristics). At each level species may either pass through or be blocked by a habitat filter, their ability to pass through a filter will depend on whether or not they are equipped with the suitable traits. This provides a theoretical framework for understanding the mechanisms underlying insect community composition under different environmental conditions. Habitat filters can be considered as the environmental processes operating within the temporal spatial dimensions of Southwood's (1977) habitat template. Here, Southwood describes how the habitat provides the templet which governs the evolutionary development of species traits. In this way species traits are intrinsically linked to their environment. Unlike taxonomic measures, which are often site specific, analysis

Chapter 3

46

of trait (i.e. functional) composition allows the comparison of spatially separated populations (Menezes et al. 2010; Verberk et al. 2013).

Functional composition of aquatic insect communities can be determined by categorizing species by their life-history strategies (Verberk et al. 2008a; b). Life-history strategies (LHS) are based on specific traits relating to reproductive behaviour, development, dispersal, developmental trade-offs to invest in other traits and the interrelationship between these traits (Stearns 1976). They therefore represent an integrated response to the environment and may give direct information about how a particular environment is experienced by the species inhabiting it (Stearns 1976; Verberk et al. 2008b; Verberk et al. 2013). Investigation of insect life-history strategy composition (termed functional composition from this point forward) under different environmental conditions can provide insight into how environmental drivers are underlying the present community composition. For example eutrophic waters that experience periodic anoxia, as is the case for the ditches sampled in this study, are likely to support insect species which have life-history strategies that allow them to tolerate low oxygen concentrations, while stable environments or those that experience a predictable level of disturbance, may support more species with synchronized life-cycles and relatively long juvenile development times (Verberk et al. 2008a; b).

The aim of this study is to evaluate if emergent vegetation structure operates as a habitat filter on aquatic insect communities in peatland drainage ditches under different trophic conditions. We investigated insect community functional composition alongside taxonomic composition in emergent reed stands of variable dimensions in two peatlands, one eutrophic degraded system, the Wormer and Jisperveld (WJ) and one rehabilitated mesotrophic system, the Naardermeer (NM). We expected taxonomic composition to differ between the two peatlands and analysis of functional composition to show that species adapted to abiotic extremes prevailed in the eutrophic peatland. Furthermore, we anticipated that ditches with a large emergent vegetation area would be more taxonomically diverse and support a greater abundance of taxa adapted to stable environmental conditions than ditches with little or no emergent vegetation. Materials and methods Study sites and sampling design Two peatlands with different trophic states were selected for this study: a eutrophic peatland, the Wormer and Jisperveld (WJ), and a mesotrophic peatland in the Naardermeer reserve (NM). Both peatlands are situated in the province of North Holland, The Netherlands and provide typical examples of a nutrient enriched peatland and a restored peatland in this landscape (Fig. 3.1).

Chapter 3

47

The Wormer and Jisperveld (52º30’N, 4º50’E) is a low-lying, highly modified eutrophic peatland covering approximately 2500 ha. The peatland comprises many drainage ditches and is managed as low intensity pastoral meadows to maintain good nesting habitat for meadow birds (Janssen et al. 2005). The area is surrounded by intensively farmed agricultural land and small towns and the water table is maintained at an artificially stable level via an inlet and an outlet, which are linked to an external hydrological system. To maintain a constant water table, mineral rich (alkaline) River Rhine derived water is supplied to the peatland during periods of low rainfall, thereby altering the chemistry of the surface water. High nutrient levels in the peatland, combined with stimulated peat oxidation, cause peat degradation and create unstable sediments and turbid waters. Consequently, submerged macrophyte richness has declined in the Wormer and Jisperveld over the last two decades (Whatley et al. 2014). The ditches in this peatland are dredged every 7 years and the emergent vegetation is cut back annually. The ditches sampled for this study had not been dredged in the past 6 years.

Fig. 3.1. Position of a) the Wormer and Jisperveld and b) Naardermeer reserve in The Netherlands and the location of the 6 ditches sampled in 2011. The Naardermeer reserve (52º17’N, 5º06’E) was established in 1906 and is the oldest protected nature reserve in The Netherlands. The reserve contains approximately 1,077 ha marshland, of which 677 ha is natural (mineral

Chapter 3

48

poor) peatland (Wassen et al. 1989; Bootsma et al. 1999) in which our sampling sites were situated. The peatland is surrounded by 400 ha of historic agricultural land, designated as a buffer zone and kept free of livestock since 1997 to protect the enclosed wetlands. The Naardermeer water table is characterized by an infiltration zone in the east and seepage in the west, which discharges from the reserve on to low-lying farmland (Wassen et al. 1989). Historically the peatland’s hydrology was unmanaged, allowing natural fluctuations of the water table, however, as water abstraction increased in the surrounding area the Naardermeer peatland became increasingly dry, with subsequent oxidation of the peat soils. To abate this process a dephosphatation plant was built in 1984 and River Rhine derived water was pumped to the reserve, via the plant, to increase water table levels during dry periods (Bootsma et al. 1999). These measures maintained peat soils since water tables were increased and no additional nutrients were introduced to the wetland from external waters. The reserve is managed for wildlife preservation and is surrounded by intensively farmed agricultural land and the town of Bussum to the southeast. The ditches in NM are not dredged and removal of emergent reed vegetation takes place annually in the reserve where the sampled ditches were situated.

Six ditches (width ≤ 12.2 m, water depth ≤ 1.2 m), three in each peatland, were sampled following a nested factorial design with peatland type (i.e. trophic state) as a two level factor and vegetation and insect measurements nested within ditch (Fig. 3.2). Ditches were visited over a 3 week period from mid-August to early September 2011 to sample insects and to record vegetation measurements. Sampling during this time period has been shown to provide representative samples of insect communities in Dutch drainage ditches (Beltman 1983; Bijkerk 2010). Furthermore, emergent stands are well established by this time and annual vegetation maintenance (mowing and clearing) hasn’t yet taken place. Water chemistry was measured twice in each ditch, once five months prior to and once during the sampling period. Surface water conductivity, turbidity and pH were measured in situ in undisturbed overlying water with a WTW LF 92 conductivity meter and Tetracon 96 cell, a WTW TURB 350 IR turbidity meter and a WTW pH/Oxi 340i/set meter, respectively. Analysis of surface water concentrations of total nitrogen (TN), nitrate (NO3-), nitrite (NO2-), ammonium (NH4+), total phosphorus (TP), orthophosphate (PO43-), sulphate (SO42-), total sulphide (S2-), total iron (Fe), carbonate (CO32-) and chloride (Cl-) were determined in the laboratory using standardised national protocols accredited by the Netherlands Standards Institute (NEN).

Chapter 3

49

Fig. 3.2. Sampling design; emergent vegetation area and height were measured in 6 ditches located in two peatlands of differing trophic status. Peatlands and ditches were selected for differences in nutrient and emergent vegetation area to achieve a nested factorial sampling design. Four measurements of emergent vegetation area and height above water were collected in four different locations within each ditch. Emergent vegetation surveys Emergent vegetation area and height were determined for each ditch from four vegetation surveys. Emergent vegetation area (m2) is equivalent to the vegetation cover measured from the bankside towards the centre of the ditch along a 5 meter stretch of bank. Emergent vegetation height (cm) was measured above the surface of the water. The six ditches were selected to obtain an equal number with small (≤ 1 m2) and large (≥ 5 m2) vegetation area. Emergent vegetation stands in the sampled ditches were all dominated by characteristic helophyte species (Phragmites australis and Typha angustifolia). These emergent species are common in nutrient rich peatlands and because they have the same vertical growth form, vegetation measurements collected in different ditches were comparable to one another in our study. In addition to emergent vegetation measurements, total submerged macrophyte species richness was also recorded at each ditch, for qualitative purposes.

Chapter 3

50

Table 3.1. Aquatic invertebrate life-history strategies (LHS), the defining traits of each strategy and the environmental conditions each strategy is adapted to. Illustrations of taxa for each strategy are representative examples only (see Appendix 3.1 for full list). This table summarizes information published in Verberk et al. (2008a; b)

LHS Defining traits Environmental conditions

Example species

D1 Strong dispersal, rapid juvenile development, long-lived adults, high voltinism.

Predictable variation & fragmented.

Triaenodes bicolor Trichoptera

D2 Strong dispersal, slow juvenile development, large clutch size.

Stable & fragmented.

Ischnura elegans Odonata

D3 Moderate dispersal, rapid reproductive development, high voltinism, short-lived adults, (mainly semelparous).

Briefly suitable & fragmented.

Gerris argentatus Heteroptera

T1 Tolerant to abiotic extremes, small

body size, elongated development time, weak dispersal.

Stable & harsh. Chironomus sp. Diptera

S1 Synchronized emergence, long juvenile development, short-lived adults.

Predictable & stable. Caenis horaria Ephemeroptera

S2 Synchronized emergence, iteroparous, has resistant stages.

Predictable & changeable.

Ilybius fenestratus Coleoptera

S3 Synchronized emergence, rapid juvenile development, long-lived adults, seasonally iteroparous.

Predictable & moderately stable.

Ecnomus tenellus Trichoptera

R1 Protracted oviposition, rapid juvenile development, relatively long-lived adults, moderate dispersal.

Briefly suitable & fragmented.

Hydrometra sp. Heteroptera

Chapter 3

51

Insect sampling and identification Aquatic insects were collected between August and September in accordance with current Netherlands national monitoring protocols (Bijkerk 2010). Aquatic insects were sampled adjacent to the bank with a dip-net (mesh 900 μm, frame 25 x 29 cm, bag depth 25 cm) by sweeping the net continuously along a 5 meter length, approximating to a sampling area of 1.5 m2. Care was taken to collect insects from the emergent vegetation zone and not from the sediment. One insect sample was collected, prior to vegetation measurements, in the same location where emergent vegetation dimensions were recorded, giving a total of four insect samples per ditch. Insect samples were sorted into groups in the laboratory and preserved in 70% ethanol for further identification.

Where possible, insects were identified to species with the exception of Chironomidae larvae [Diptera], which were identified to either subclass or genus. The following taxonomic groups were identified with the corresponding keys: Ephemeroptera (Elliott & Humpesch 2010), Heteroptera (Savage 1989), case-bearing Trichoptera (Wallace et al. 2003), caseless Trichoptera (Edington & Hildrew 1995), Coleoptera and Odonata (Nilsson 2005) and Diptera (Nilsson 2005; Orendt & Spies 2010). Life-History Strategies (LHS) Aquatic insect species were classified by their life-history strategies following the work of Verberk et al. (2008a). Species that were not already categorized in the studies of Verberk et al. (2008b; 2010) were categorized on the basis of the defining traits, which were derived for each genus using the trait database of Tachet et al. (2002). Specific traits from this database relating to dispersal, reproduction, life-cycle duration, potential number of reproductive cycles per year, resistance forms and trophic level were applied using the assignment methods developed by Verberk et al. (2008a). Individuals belonging to eight life-history strategies (LHS) were relevant to this study (Table 3.1). These represented the four trait domains, being related to dispersal (D1, D2 and D3), synchronization (S1, S2 and S3), reproduction (R1) and development, i.e. species that are able to tolerate environmental stress because of developmental trade-offs and other specific traits (T1). The S2 and S3 synchronized strategist were very low in abundance and their defining traits were similar to that of S1 (see, Table 3.1), thus the three ‘S’ strategists were combined into one group for statistical analysis. A full list of insect species encountered during this study and their respective life-history strategies is provided in Appendix 3.1.

.

Chapter 3

52

Statistical analysis Insects and ditch ecological characteristics were sampled following a nested, factorial design with emergent vegetation measurements and peatland type assigned as fixed factors. Fixed factors were nested in ditch, which was assigned as a random factor in the statistical model (see Fig. 3.2). The two peatlands were selected for their differences in trophic conditions with WJ being eutrophic and NM being mesotrophic. Differences in abiotic characteristics between peatlands were analysed with linear mixed models. Prior to analyses residual plots and QQ-plots were used to assess normality and homogeneity of variances of the abiotic and vegetation parameters. If these assumptions were not met data were log10 transformed. Insect abundance was standardized to density (number of individuals m-2) and log10 (x +1) transformed prior to statistical analysis.

Ordination techniques were used to examine the relationship between aquatic insect taxonomic or functional community composition and emergent vegetation dimensions and peatland type. Detrended Correspondence Analyses (DCA), with 26 segments and down-weighted rare species, were used to determine gradient length of insect taxonomic and functional composition. Gradient length was 2.0 for taxonomic composition and 1.0 for functional composition, indicating that a direct ordination technique based on linear combinations of insect variables (in this case Euclidean distance) were suitable (Ter Braak & Smilauer 2002). Subsequently, Redundancy Analyses (RDA) were carried out to examine the relationship between aquatic insect taxonomic or functional community composition and emergent vegetation dimensions and peatland trophic status. Vegetation area and height were each included as continuous variables and peatland was defined as a fixed factor in the analysis with significance based on the proximity to the centroid of all samples collected in ditches from the same peatland.

To investigate the relationship between insect taxonomic and functional indices and environmental variables (emergent vegetation structure and peatland type), three plausible candidate linear mixed models were specified a priori and subsequently tested (Burnham & Anderson 2002). Models were constructed with insect response variables (y), emergent vegetation area (VA) and height (VH) as continuous fixed predictors, peatland type (PT) as a binary fixed factor and ditch (Ditch) as a random factor to capture inherent differences between ditches (i.e. sites). Vegetation area and height and peatland type were nested in ditch to account for intrinsic differences between ditches (e.g. food availability or toxicants). The full model (model I) contained both emergent vegetation and peatland type as predictors and was tested against reduced models which contained only peatland type (model II) or emergent vegetation (model III) predictors.

Chapter 3

53

Models: I (Full model) Response = vegetation + peatland type nested in ditch yij = αVAi + αVHi + βPTi ~1| bi + εi II (Peatland model) Response = peatland type nested in ditch yij = βPTi ~1| bj + εi III (Vegetation model) Response = vegetation nested in ditch yij = αVAi + αVHi ~1| bj + εi

yij is the response data for each ditch i and location j within each ditch, βPTi has two values: WJ and NM. Furthermore the random intercept bj is location within a ditch, which is assumed to follow a normal distribution with zero mean and a variance υloc. The model error εi is assumed to be normally distributed with zero mean and a variance of υditch. bj ~ N(0, υloc) εi ~ N(0, υditch)

The Akaike Information Criterion corrected for small (n – K < 40)

sample sizes (AICc) and Akaike weights (wi) were used to determine the support for each model by the observations (Burnham & Anderson 2002). The wi are normalized relative likelihoods for each model and can be interpreted as the probability or the performance of each candidate model in relation to the other models in the set. Models with wi > 0.7 were considered to be strongly supported, between 0.4 – 0.7 as moderately supported and with wi < 0.4 as minimally supported by the data (Burnham & Anderson 2002). Only models with significant (P < 0.05) parameters and uncorrelated Gaussian residuals were considered adequate. Variance inflation values were examined for each predictor to check for influence of collinearity with results showing that none of the predictors need be excluded from the models. Emergent vegetation variables were not correlated with peatland (r ≤ 0.02) and only slightly correlated to each other (Pearson correlation, r = 0.53). All statistical analyses were run in R for Windows using the packages nlme, car and biology (version 2.15, R Development Core Team 2012).

Chapter 3

54

Results Water chemistry and vegetation As anticipated, nutrients concentrations differed between the two peatlands with WJ ditches having significantly higher concentrations (linear mixed model, P < 0.01) of TP and TN than NM ditches (Table 3.2). There was some difference between ditches within the WJ as ditch 2 was semi-isolated hydrologically and had lower concentrations in TP, TN, carbonate, pH and chloride than other WJ ditches. Yet, the abiotic conditions of ditch 2 were more similar to the other WJ ditches than to NM ditches. The trophic status of each peatland was reflected by the plant community with five aquatic species recorded in mesotrophic NM, including two truly submerged species (Chara globularis and Potamogeton natans), two submerged/emergent species (Sparganium emersum and Alisma plantago-aquatica) and one floating leaf species, the yellow water lily (Nuphar lutea). Only one species of submerged plant (Lemna trisulca) was recorded in WJ, in ditch 2. Conversely, the average height and area of emergent vegetation stands were comparable across both peatlands. Insect Community Composition and Life-History Strategies A total of 70 insect taxa were recorded over both peatlands with a greater number of taxa recorded in NM ditches (60 species) compared to the more eutrophic WJ ditches (40 species). Conversely, average insect density (individuals m-2 ± 95% CI) was slightly greater in WJ ditches (187 ± 55) than NM ditches (80 ± 37) (linear mixed model, P = 0.11). The difference in density was largely due to a greater abundance of Chironomidae larvae (dominated by Chironomus sp.) found in the WJ compared to NM.

The highest density and highest taxonomic richness of all life-history strategies (LHS) were found within the T1 strategy (see Table 3.1 on insect LHSs), with a total of 21 species and an average total density of 72 individuals m-2 per ditch (predominantly Chironomus sp.). The D1 strategy was the second most abundant group and was dominated by the trichopteran Triaenodes bicolor, followed by the S1 strategy, which was dominated by the ephemeropteran Caenis sp. and the damselfly Coenagrion sp.. The lowest densities were recorded for S2 and S3 strategists, each comprising 4 species and an average density of only 1 individual m-2. The number of LHSs did not differ between peatlands but were significantly positively related to species richness (linear mixed model, P = 0.005) and the number of Ephemeroptera, Odonata and Trichoptera (EOT) species (P = 0.03). The density of R1 strategies accounted for less than 10% of the total insect density in each ditch and was not statistically analysed.

Chap

ter 3

55

Tabl

e 3.2

. Ditc

h mo

rpho

logica

l var

iables

, sur

face w

ater

chem

istry

and

emerg

ent v

egetat

ion m

easu

remen

ts. V

alues

show

n ar

e aver

ages

± 9

5% C

I, n

= n

umbe

r of

replic

ates

per p

eatla

nd. S

ignifi

cant

diff

erenc

es be

tween

peat

lands

, as d

eterm

ined

by l

inea

r mix

ed m

odels

, are

indica

ted a

s **

P <

0.01

and

***

P <

0.0

01.

Ditc

h W

orm

er a

nd Ji

sper

veld

Naa

rder

mee

r

1 2

3

4 5

6

Abio

tic co

nditi

ons n

= 6

M

ax d

itch

wid

th (m

) 6

6 5

7

8 12

Max

wat

er d

epth

(cm

) 70

70

90

83

68

77

pH

9.

2 ±

0.2

6.

7 ±

0

8.7

± 0

.1

7.

1 ±

0.4

8.

1 ±

1.6

7.

8 ±

1.9

Turb

idity

(NTU

) 32

.6 ±

3.2

3.

1 ±

1

36.4

± 6

.9

2.

8 ±

3.7

3.

6 ±

0.6

4.

4 ±

0.4

Carb

onat

e (m

g CO

32- L

-1)

140

± 20

65

± 4

9 14

0 ±

20

10

5 ±

49

105

± 49

10

5 ±

49

Ch

lorid

e (m

g Cl

- L-1

) 14

5 ±

69

74 ±

29

145

± 69

125

± 30

99

± 2

2 10

0 ±

20

TN

(mg

N L

-1)

4.0

± 1

2.

5 ±

0.2

4.

5 ±

1.3

1.0

± 0

.5

1.3

± 0

1.

3 ±

0.2

**

* N

O3- (

mg

N L

-1) a

<

0.1

<0.

1 <

0.1

<

0.1

<0.

1 <

0.1

N

O2- (

mg

N L

-1) a

<

0.01

<

0.01

<

0.01

<0.

01

<0.

01

<0.

01

N

H4

<0.

02

0.04

<

0.02

<0.

02

<0.

02

<0.

02

TP

(mg

P L-

1 ) 0.

30 ±

0.0

1 0.

11 ±

0.0

2 0.

34 ±

0.1

0.04

± 0

.03

0.05

± 0

.01

0.04

± 0

.03

**

Orth

o-P

(mg

P L-

1 )a

<0.

005

0.01

6 <

0.00

5

<0.

005

<0.

005

<0.

005

Ir

on (μ

g Fe

L-1

) 16

5 ±

127

17

5 ±

69

210

± 13

7

160

± 1

57

90 ±

0

100

± 20

Sulp

hate

(mg

SO42-

L-1)

67 ±

2.9

34

± 0

44

± 4

1.1

77

± 4

4.1

81 ±

56.

8 53

± 4

.9

Su

lphi

de (m

g S2

- L-1

)b 1.

7 2.

6 1.

8

0.7

0.6

0.7

a Be

low d

etecti

on li

mit i

n all

loca

tions

Chap

ter 3

56

Tabl

e 3.2.

cont

inue

d.

Wor

mer

and

Jisp

erve

ld

N

aard

erm

eer

D

itch

1 2

3

4 5

6

Eme

rgent

veget

ation

n

= 12

Heig

ht (c

m)

68 ±

31

145

± 9

213

± 19

259

± 28

23

2 ±

30

35 ±

13

A

rea

(m-2

) 1

± 1

19 ±

3

1 ±

0

7 ±

1 5

± 2

1 ±

0

Subm

erge

d pl

ant s

p.

0 1

0

3 3

3

NO

TE -

Chem

ical a

nd m

orpho

logica

l par

amete

rs we

re me

asur

ed tw

ice in

each

ditch

, onc

e 3 m

onth

s prio

r to a

nd on

ce du

ring t

he in

sect s

ampli

ng p

eriod

. Vege

tatio

n me

asur

emen

ts we

re tak

en in

4 lo

catio

ns w

ithin

each

ditch

and i

nclud

e aver

age h

eight

abov

e wat

er an

d av

erage

vegeta

tion

area

(m2 ).

Chapter 3

57

Fig. 3.3. Biplots of the first and second constrained axis from Redundancy Analysis (RDA) run with environmental variables, peatland and a) insect taxonomic community composition, or b) insect functional composition. Emergent vegetation area and height were included as continues factors and peatland type as the nesting factor with two levels, NM or WJ. The proportion of variance explained by the first and second constrained axes are shown. In both RDAs peatland type was significant (P ≤ 0.02) and emergent vegetation parameters were also significant (P ≤ 0.01).

Chapter 3

58

Eutrophication, emergent vegetation structure and insect community composition Insect community composition was significantly correlated to emergent vegetation structure and peatland type. Redundancy Analysis (RDA) explained a significant proportion of variation in insect taxonomic composition (P = 0.005) (Fig. 3.3a). Partitioning of variances indicated that the three constrained RDA axes accounted for 55.8% of the taxonomic variation with the first and second constrained axes accounting for 32% and 18% of the variance. Peatland type was strongly correlated to the first RDA axis (Pearson correlation coefficient, r = 0.98) and explained 31% of the taxonomic variation (P = 0.005). Emergent vegetation area was strongly correlated to the second RDA axis (r = 0.97) and explained 16% of the taxonomic variation (P = 0.005) and vegetation height was strongly correlated to the third RDA axis (r = 0.91) and explained 8% of the taxonomic variation (P = 0.01). Redundancy analysis also explained a significant proportion of functional variation in the insect community (P = 0.005) accounting for 55.7% of total variation in the dataset with the first and second axes explaining 37% and 13% of the variance, respectively (Fig. 3.3b). Insect functional composition was most strongly related to emergent vegetation area which was strongly correlated to the first RDA axis (r = 0.98) and explained 26% of the variation (P = 0.005), while vegetation height was negatively correlated to the second RDA axis (r = – 0.60) and explained 18% of the variation in functional composition (P = 0.005). Peatland type was positively correlated to the second RDA axis (r = 0.76) and explained 12% of the functional variation in the insect community (P = 0.02).

Analysis of the relationship between individual habitat filters and insects showed that EOT richness was negatively associated with WJ, the eutrophic peatland (Fig. 3.4a and b). Conversely, total insect density appeared to be positively associated with the eutrophic peatland which was essentially due to a greater density of tolerant (T1) strategists in WJ ditches. The density of EOT taxa and dispersal strategists (D1 and D2) were positively associated with emergent vegetation cover (Fig. 3.4c and d). Total insect richness, EOT richness and density of taxa with synchronized life cycles (S) also appeared to be positively associated with emergent vegetation cover, while the density of T1 taxa appeared to be negatively associated with vegetation area.

Chapter 3

59

Table 3.3. Linear mixed model results with insect response variables and emergent vegetation and peatland eutrophication predictor variables nested in ditch. Number of predictor variables (K), Akaike information criterion corrected for small sample size (AICc) and Akaike weights (wi) are shown for each model. The model which performed the best of the three candidate models is in bold, significant models are indicated as * P < 0.05, ** P < 0.01 and *** P < 0.001, n = 4 per ditch. Insect response Model Log

likelihood K AICc wi

(I) Full Model -60.52 4 131.14 0.81 Sp. richness (II) Eutrophication -65.66 2 135.90 0.07 (III) Vegetation -63.88 3 134.97 0.12 (I) Full Model -120.32 4 250.74 1.00 Total density (II) Eutrophication -131.03 2 266.63 0.00 (III) Vegetation -127.17 3 261.54 0.00 (I) Full Model -44.42 4 98.95 0.88 ** EOT richness (II) Eutrophication -49.16 2 102.88 0.12 * (III) Vegetation -49.77 3 106.74 0.00 (I) Full Model -8.74 4 30.81 0.18 ** EOT density (II) Eutrophication -10.49 2 28.18 0.00 (III) Vegetation -8.86 3 27.83 0.82 ** (I) Full Model -15.60 4 41.31 0.06 S density (II) Eutrophication -15.78 2 36.14 0.83 (III) Vegetation -16.52 3 40.24 0.11 (I) Full Model -10.95 4 32.01 0.35 *** D1 density (II) Eutrophication -11.54 2 27.65 0.00 (III) Vegetation -11.77 3 30.74 0.65 * (I) Full Model -6.00 4 22.10 0.34 *** D2 density (II) Eutrophication -8.88 2 22.33 0.00 (III) Vegetation -6.77 3 20.74 0.66 * (I) Full Model -11.43 4 32.97 0.08 D3 density (II) Eutrophication -12.18 2 28.94 0.59 (III) Vegetation -11.43 3 30.06 0.33 (I) Full Model -5.17 4 20.44 0.06 *** T1 density (II) Eutrophication -5.22 2 15.01 0.94 * (III) Vegetation -8.30 3 23.81 0.00

Chapter 3

60

Results of the significant (P < 0.05) linear mixed models showed that model I, the full model containing both peatland type and emergent vegetation predictors, explained the most variation in EOT richness (P = 0.01, wi = 0.88) (Table 3.3). Model II, the reduced model which represented peatland type, explained much of the variation in the density of T1 strategists (P = 0.02, wi = 0.94). Model III, the reduced model representing emergent vegetation dimensions, explained much of the variation in EOT density (P = 0.009, wi = 0.82) and density of D1 (P = 0.02, wi = 0.65) and D2 dispersal strategists (P = 0.02, wi = 0.66).

Discussion The role of emergent vegetation as a habitat filter of aquatic insects under different trophic conditions In this study we investigated the role of emergent vegetation structure as a habitat filter which shapes aquatic insect communities in peatland drainage ditches under different trophic conditions. We found that both taxonomic and functional community composition were influenced by this habitat filter. In a hierarchical context, peatland type primarily determined the pool of species in each wetland, while emergent vegetation was a secondary filter operating at the level of individual ditches. We acknowledge that conclusions on the effect of eutrophication based solely on the results from a single sampling event in two peatlands are limited. Ideally more peatlands covering a range of spatial and temporal nutrient conditions should be investigated in order to make generalizations outside our study areas. Yet, concentrations of nutrients (TP and TN) were the only measured variables that differed significantly between the two wetlands in our study, thus the observed differences in nutrients are likely to be underlying, at least in-part, the differences between the insect communities of these two peatlands. Moreover, nutrients and eutrophication have been shown by others to significantly impact aquatic invertebrate communities (van der Hammen 1992; Smith et al. 2007; O’Toole et al. 2008; Verdonschot 2012).

Chapter 3

61

Fig. 3.4. Average values ± 95% CI for ditches in different peatlands a) total number of insect taxa and EOT taxa b) total insect density, EOT and LHS density and ditches with high and low average emergent vegetation area c) total number of insect taxa and EOT taxa d) total insect density, EOT and LHS density. Significant results from linear mixed model are indicated as * P < 0.05, ** P < 0.01 and *** P < 0.001, n = 4 per ditch.

The influence of eutrophication on aquatic insects is generally

considered to be indirect via changes in oxygen concentrations and macrophyte communities, which can lead to loss of habitat and changes in food source (Janse & van Puijenbroek 1998; Varga 2003; Hinojosa-Garro et al. 2010; Verdonschot 2012). In our study the most eutrophic peatland had a high density of taxa that exhibited traits which increase tolerance of low oxygen concentrations (i.e. T1 strategists dominated by Chironomus sp. larvae). Dominance by tolerant T1 taxa in eutrophic ditches supports our expectation that nutrient enrichment creates inhospitable conditions for species without the necessary adaptations (i.e. air breathing or increased haemoglobin production). In this way eutrophication acts as a habitat filter by excluding species ill-equipped to deal with oxygen depletion related stress, allowing T1 taxa to persist and proliferate. However, in our study peatland type and vegetation were, to some extent uncoupled, allowing us to investigate if the presence of emergent macrophyte habitat mediated against the negative effects of eutrophication.

Chapter 3

62

The density of T1 strategists was slightly lower in well vegetated ditches, while we found greater densities of species with good dispersal abilities (D1 and D2) and Ephemeroptera, Odonata and Trichoptera (EOT) species in these same ditches. This provides some evidence that the presence of emergent vegetation helps to support a diverse insect community. Emergent vegetation stands can positively influence aquatic insects by providing suitable oviposition and emergence sites (Batzer & Wissinger 1996; de Szalay & Resh 2000; Foote & Hornung 2005), refugia from predatory fish (Warfe & Barmuta 2004) and by increasing niche availability (Southwood 1977; Giller & McNeill 1981). Consequently, this is an invaluable habitat for insects and the absence of emergent macrophyte habitat may lead to the exclusion of certain species (e.g. Caenis sp., Coenagrion sp. and Triaenodes bicolor) in peatland ditches.

It is recognised that taxonomic composition can be highly site specific (Suren et al. 2008; Menezes et al. 2010) and as anticipated, insect taxonomic composition was characteristic to each peatland in our study. In particular, the number of EOT species was indicative of the lower nutrient peatland (NM). In Californian wetlands EOT species richness and abundance were found to decrease under increasing disturbance, associated with eutrophication and catchment urbanization (Lunde & Resh 2012). Some EOT species are recognised as being sensitive to eutrophication and are associated with oligotrophic or mesotrophic conditions (Twisk et al. 2000; Yuan 2004; O’Toole et al. 2008). However, several EOT species found in our study are also common in nutrient rich waters (i.e. Caenis sp., Ischnura elegans and Agrypnia obsoleta) (Lenat 1993; Smith et al. 2007; Lunde & Resh 2012). We found a greater density of these species in the ditches with more emergent macrophyte habitat and a similar relationship was reported for Odonata by Foote and Hornung (2005). These results demonstrate that EOT taxa are good indicators of both structural and abiotic habitat conditions in peatland ditches. Mechanisms underlying aquatic insect community composition The use of insect life-history strategies in this study allowed us to investigate the possible mechanisms underlying the observed community composition. Furthermore, this allowed us to compare the spatially separated communities of the two peatlands. Although taxonomic composition was distinctive between each peatland, functional community composition did not conform principally to peatland type and was more strongly influenced by emergent vegetation structure. This finding supports the theory that landscape filters operate primarily by preventing species which do not possess the appropriate traits from pass through the filters (Poff 1997), thereby influencing community trait composition.

Tolerant strategists (T1) dominated in the most eutrophic sites recorded in our study and were found at a much lower density in highly

Chapter 3

63

vegetated ditches. The costs involved with adapting to harsh abiotic conditions, like haemoglobin production (Moller Pillot 2009), osmoregulation and decreased metabolism, results in the individual having less resources to invest in other traits, which may lead to reductions in growth rates, dispersal or biomass (Stearns 1976). This has consequences for the competitive ability of T1 species under benign environmental conditions (Verberk et al. 2008a, b). Conversely, species with good dispersal abilities (e.g. Triaenodes bicolor, Enallagma cyathigerum and Hesperocorixa linnaei) can colonize suitable habitats more readily than dispersal limited species. The strong positive relationship between dispersal strategies and emergent vegetation structure could reflect the importance of macrophyte oviposition sites for this group. Another mechanism could relate to D2 strategists exhibiting longer juvenile development, and for juveniles to reach maturity the environment must be stable for a long enough period of time (Stearns 1976; Southwood 1977). This suggests that emergent vegetation stands may mediate against eutrophication by providing a stable environment. For these reasons insects with strong dispersal abilities may also perceive emergent vegetation as an indicator of good habitat and may readily colonize these sites and avoid sites with little emergent vegetation. Thus, lack of emergent vegetation may operate as a filter through the preferences of individual species and exclude species which require vegetation cover for refuge, food supply and life-cycle completion.

Species which have invested in synchronizing their life-cycles (e.g. Caenis sp. and Ecnomus tenellus) are not necessarily strong dispersers and they are better adapted to predictable conditions (Stearns 1976; Verberk et al. 2008a). Thus, we expected that the density of S strategists would be greater in highly vegetated ditches and in the mesotrophic peatland (NM) in our study. Although there was some evidence of this (Fig. 3.4b and d) the trend was not significant. This was likely due to the large degree of variability in the density of S strategists between ditches. In general the ditches with low emergent vegetation cover and situated in the eutrophic peatland (WJ) contained very low numbers of S strategists, while ditches with good water quality and high vegetation cover supported greater numbers of these strategists. This supports the theory that S strategists benefit from stable conditions but also highlights the need for additional studies in this field. In summary both taxonomic and functional insect composition reflected emergent vegetation structure and inherent differences between peatlands including trophic state. While peatland type primarily determined the taxonomic composition within each wetland, emergent vegetation structure operated as a secondary filter driving functional community composition within ditches. Large-scale mechanical vegetation clearance will likely cause a reduction in aquatic insect richness. To mitigate the potential negative impacts of vegetation management it is, therefore, advisable to retain small patches of undisturbed emergent vegetation, to provide refugia which act as sources of

Chapter 3

64

individuals to facilitate recolonization ( see also Painter 1999 and Verberk et al. 2007). Furthermore, as insects have seasonal reproductive cycles with many taxa emerging into flying adults (i.e. terrestrial phases), the timing of vegetation removal can be critical to reducing the potential negative impacts, as recommended by Twisk et al. (2000). Acknowledgements We thank Gert van Ee, Emile Nat, Ron van Leuken and Herman van Dam for their help throughout this project, Pim Koelma, Coen Wagner and Alejandra Goldenberg for their assistance during fieldwork, Nigel Upchurch for the insect illustrations, Wilco Verberk for his comments on this paper, two anonymous reviewers for their constructive criticism of this paper and Annemieke Ouwehand, Andre Timmer and Ed Zijp of Natuurmonumenten, for allowing us to access the Wormer and Jisperveld and Naardermeer reserves. This research was funded by Stichting Waterproef and Hoogheemraadschap Hollands Noorderkwartier.

Chapter 3

65

Appendix 3.1. List of insect taxa and their respective life-history strategies (LHS). The assignment of life-history strategies is based of published data listed as reference. Taxa Order LHS Reference

Athripsodes aterrimus Trichoptera D1 1 Anopheles maculipennis complex Diptera D1 1 Agraylea multipunctata a Trichoptera D3 2

Agrypnia obsoleta a Trichoptera T1 3

Agrypnia pagetana a Trichoptera T1 2

Brachytron pratense Odonata T1 1 Cloeon dipterum Ephemeroptera D2 3 Cyrnus flavidus a Trichoptera S1 2

Caenis sp. a Ephemeroptera S1 2

Corixa panzeri a Heteroptera D1 3

Coquillettidia richiardii a Diptera S1 2

Ceratopogonidae a Diptera T1 3

Chaoborus sp. a Diptera S1 3

Chironomus sp. a Diptera T1 3

Coenagrion sp. a Odonata S1 3

Dixella sp. a Diptera D1 3

Enallagma cyathigerum Odonata D2 1 Erythromma najas a Odonata D2 2

Ecnomus tenellus a Trichoptera S3 2

Enochrus sp. a Coleoptera S3 3

Gerris argentatus a Heteroptera D3 3

Gerris odontogater a Heteroptera D3 3

Graphoderus bilineatus a Coleoptera D2 2

Glossosoma boltoni a Trichoptera S2 2

Gyrinus marinus a Coleoptera S3 2

Gyrinus paykulli a Coleoptera S3 2

Chapter 3

66

Appendix 3.1. continued Taxa Order LHS Reference

Graptodytes pictus Coleoptera R1 3 Haliplus confinis a Coleoptera R1 3

Haliplus fulvicollis Coleoptera T1 3 Haliplus laminatus a Coleoptera R1 3

Haliplus lineolatus a Coleoptera R1 3

Haliplus ruficollis Coleoptera R1 3 Holocentropus dubius Trichoptera T1 3 Holocentropus picicornis Trichoptera T1 3

Hydrobius fuscipes Coleoptera R1 3 Hydrometra gracilenta Heteroptera R1 3 Hygrotus inaequalis Coleoptera T1 3 Hesperocorixa linnaei Heteroptera D1 3 Hyphydrus ovatus Coleoptera T1 3 Helius sp. a Diptera T1 2

Hydrovatus cuspidatus a Coleoptera D1 2

Ischnura elegans Odonata D2 3 Ilybius fenestratus Coleoptera S2 3 Ilyocoris cimicoides Heteroptera T1 3 Liopterus haemorrhoidalis Coleoptera D1 1 Limnephilidae sp. a Trichoptera S2 3

Molanna angustata a Trichoptera D2 2

Mesovelia furcata a Heteroptera D3 2

Mystacides longicornis a Trichoptera S1 2

Microvelia reticulata Heteroptera R1 3 Nepa cinerea a Heteroptera T1 2

Noterus sp. Coleoptera T1 3 Notonecta glauca Heteroptera D1 3 Notonecta viridis Heteroptera D1 3 Notiphila sp. a Diptera T1 2

Chapter 3

67

Appendix 3.1. continued Taxa Order LHS Reference

Orthetrum cancellatum a Odonata D2 2

Oecetis furva a Trichoptera D2 2

Orthocladiinae a Diptera D1 3

Plea minutissima Heteroptera T1 3 Porhydrus lineatus Coleoptera D1 3 Pyrrhosoma nymphula Odonata T1 3 Phryganea sp. a Trichoptera T1 2

Ranatra linearis a Heteroptera R1 2

Rhantus suturalis Coleoptera D1 3 Sigara falleni a Heteroptera D3 3

Sigara striata Heteroptera D3 3

Sialis lutaria Megaloptera T1 1 Triaenodes bicolor Trichoptera D1 3 Tanytarsini sp. a Diptera T1 3

Tanypodinae a Diptera D1 3

References: 1Verberk et al. (2010), 2Tachet et al. (2002), 3Verberk et al. (2008a) aLife-history strategy assigned to genus or subclass

Chapter 3

Freshwater diatoms

Chapter 3

Chapter 4

Linkages between benthic microbial and freshwater insect communities in degraded peatland ditches

Manuscript: M. H. Whatley, E. E. van Loon, C. Cerli, J. A. Vonk, H. G. van der Geest and W. Admiraal

Chapter 4

70

Abstract Many wetlands are heavily modified and identifying the environmental drivers of indicator groups like aquatic insects is complicated by multiple stressors and co-varying environmental factors. Yet, incorporating data from other biological groups, such as microbial communities, potentially reveals which environmental factors are underpinning insect community composition. In the present study we investigated the application of benthic microbial community composition, as determined by phospholipid fatty acid (PLFA) analysis, alongside aquatic insect data in 25 peatland ditches in the province of North Holland, The Netherlands. We applied clustering and Principal Component Analysis to a matrix of 26 PLFAs to organise ditches by the microbial community. Generalized Linear Models were used to examine correlations between microbial PLFAs, insects, vegetation (emergent and submerged) and abiotic factors. The ratio of heterotrophic (e.g. sulphate reducing bacteria) to autotrophic (e.g. algae and cyanobacteria) derived PLFAs could be estimated by the ratio between Saturated and Branched to Monounsaturated and Polyunsaturated fatty acids (SB/MP). SB/MP was correlated with insect community composition, differences in water chemistry (in particular bicarbonate, sulphate and nutrients) and vegetation cover in the ditches. Moreover, ditches distinguished by their microbial communities differed in the number of insects they supported with differences most pronounced for Odonata, Trichoptera and Chironomus larvae. This study demonstrates that integrating microbial and aquatic insect community data provides insight into key environmental drivers in modified aquatic ecosystems and facilitates the development of remediation strategies for degraded wetlands. Introduction Wetlands are some of the most heavily impacted environments in the world. In Europe alone, wetland loss is estimated in excess of 50% of the original land area while in New Zealand it is thought to be as high as 90% (see Moser et al. 1996 and references therein). The productivity of wetland soils combined with a reliable water supply has resulted in the drainage and conversion of wetlands to agricultural land. Many of these wetlands contain large numbers of drainage ditches which potentially support a range of aquatic biota reminiscent of the natural wetland environment. However, determining the ecological conditions of such heavily modified environments is complicated by the presence of multiple stressors and a lack of suitable reference conditions for comparative ecological assessments. Aquatic invertebrates are widely used as indicators of wetland ecosystem status because they reflect conditions relating to nutrients, micro-ions, salinity and habitat structure, among other factors (Scheffer et al. 1984; van der Hammen 1992; Smith et al. 2007; O’Toole et al. 2008; Lunde &

Chapter 4

71

Resh 2012; Verdonschot, et al. 2012a). Within the invertebrate community insects are particularly good indicators of the general ecosystem conditions as the majority of insects have a terrestrial phase in their life-cycle (Crichton et al. 1978; Nilsson 2005; Elliott & Humpesch 2010). However, in heavily modified wetlands, relating patterns of insect diversity to abiotic conditions is complicated by interaction effects. For example, increased nutrients may supply food (i.e. by stimulating epiphytic algae growth) to insect species, but concomitantly result in stress by causing diurnal fluctuations in oxygen concentrations, thus limiting the distribution of these same species. Analysis of other groups of organisms, such as microbes, may reveal additional information on environmental factors that are involved in driving shifts in insect diversity. Determination of benthic microbial composition, by phospholipid fatty acid (PLFA) analysis is a well-established in situ measure of microbial biomass and community composition (Piotrowska-Seget & Mrozik 2003; Kaur et al. 2005). Yet, ecosystem studies linking aquatic insects and microbes are rare outside of food web studies (e.g. Goedkoop et al. 1998; Vos et al. 2002; Peeters et al. 2004). Microbial lipids and Polyunsaturated Fatty Acids (PUFAs) in particular, are essential dietary requirements of insects. Sediment dwelling insects in-turn can influence microbial communities through grazing (Goedkoop et al. 1997; Traunspurger et al. 1997), sediment mixing and detrital processing (Hunting et al. 2012). Moreover, aquatic insects in lentic environments depend on vegetation for habitat and microbes (bacteria and algae) influence aquatic macrophytes by mineralizing organic matter and competing for resources (e.g. light and nutrients). Rooted vegetation in-turn influences microbes by altering sediment conditions, water chemistry and detritus composition (see Fig 4.1). Benthic microbial community composition analysed in relation to environmental drivers, such as the degree of vegetation cover, hydrological regime, nutrients, pH and edaphic conditions, can reveal the role of these drivers in underpinning ecosystem state (Bååth et al. 1995; Boon, Virtue & Nichols 1996; Gao et al. 2005). Moreover, integration of microbial and insect data can be applied to get a better overview of wetland health, particularly in modified landscapes which lack suitable reference conditions. An advantage of PLFA analysis is that it provides an accurate measure of the living and active microbial community because PLFAs are quickly hydrolysed in dead cells (Harvey et al. 1986; Findlay et al. 1989; Sundh et al. 1997). Furthermore, PLFAs can be used to identify the presence of different microbial groups because they differ in their fatty acid compositions (Kaur et al. 2005). Comparison of different biological communities, such as benthic microbial communities and aquatic insects, provides information on how environmental filtering is affecting both benthic and littoral communities, giving a wider overview of factors which affect the entire aquatic community.

Chapter 4

72

Fig. 4.1. Interrelationships between benthic microbial community (bacteria and algae), aquatic insects and aquatic vegetation. The drainage ditches of North Holland’s peatlands are some of the most degraded wetland environments in Western Europe. These wetlands are remnants of a once vast system of river deltas and raised bogs that covered much of The Netherlands (van Dam 2001; van Eerden et al. 2010). Regular vegetation clearance and sediment dredging is performed to maintain drainage ditches and a strict hydrological regime is applied to ensure stable water tables and increased productivity of the surrounding agricultural land. External inputs of River Rhine waters contribute carbonate and sulphate to the peatlands, which in turn increases peat mineralization and the release of nutrients (e.g. by facilitating the anaerobic mineralization of peat and release of sediment bound phosphates) (Lamers et al. 2002; Smolders et al. 2006). The inlet of these mineral rich waters, in combination with oxidation arising from wetland drainage and the diffuse release of nutrients from the surrounding agricultural land, are underlying the eutrophication of North Holland’s peatlands (Sinke et al. 1990; Lamers et al. 2002). In the present study we assess if microbial data can reveal habitat suitability for aquatic insects and help to identify key environmental drivers of ditch communities in North Holland’s peatlands. We used monitoring data (collected by the local Water Authority as part of standard monitoring) available for insects, water chemistry and submerged and emergent vegetation. In addition we collected sediment samples for microbial PLFA analysis and measurements to estimate emergent vegetation cover. Using clustering and Principal Component Analysis (PCA) we classified 25 peatland ditches, based on microbial community composition, and investigated how this classification related to insect communities, submerged and emergent vegetation and abiotic conditions in peatland ditches. We then employed Generalized Linear Models

Chapter 4

73

(GLMs) and Canonical Correspondence Analysis (CCA) to determine the variation in insect community composition explained by microbial and environmental (abiotic and vegetation) predictor variables. We expect differences in vegetation and abiotic conditions between ditches to be reflected by microbial community composition and anticipate that microbial PLFAs will provide insight into factors underlying aquatic insect community composition in North Holland’s peatlands. Methods Study sites and environmental data collection Monitoring locations were sampled for macroinvertebrates, aquatic vegetation and abiotic variables in 2011 as part of the annual monitoring of surface waters undertaking by the North Holland Water Authority, Hoogheemraadschap Holland’s Noorderkwartier (HHNK). Based on the dimensions of the water body, soil type and salinity 25 locations were selected from a larger dataset to obtain a set of samples from similar habitats. The locations were all ditches, situated in peat areas with average chloride levels less than 1000 mg Cl- L-1 and representing a range in water quality (e.g. nutrients, sulphate and bicarbonate), from nature reserves with good ecological and chemical status (according to the European Water Framework Directive) to intensively farmed agricultural land with poor status (Fig 4.2). Measurements of conductivity, water transparency and pH were taken in the field at monthly intervals from February to July. Monthly samples of undisturbed overlying water were also collected during the same period and analysis of nitrogen (Kjeldahl nitrogen (KN), NO3-, NO2- and NH4+), total phosphate (TP) and PO43-, sulphate (SO42-), total iron (Fe), bicarbonate (HCO3-) and chloride (Cl-) were carried out in the laboratory using standardized national protocols accredited by the Netherlands Standards Institute (NEN). Vegetation measurements Vegetation surveys were undertaken in all ditches in June and July, as part of monitoring carried out by HHNK. Coverage and species richness of emergent, submerged and floating vegetation were assessed by visual inspection along a 50 m stretch of ditch. In addition to this monitoring data, we took measurements of emergent vegetation extension (i.e. the distance the vegetation was growing from the bank towards the centre of the ditch) to test an alternative way of estimating emergent vegetation cover over a 50 m ditch length. To this end 5 measurements of vegetation extension were taken along a 5 meter bank length, which was representative of emergent vegetation cover over a 50 m ditch length. The average extension (± 95% CI) was 45 ± 12 cm in

Chapter 4

74

the sampled ditches, so in this case we measured vegetation up to 1 meter away from the bank. The average extension was then calculated from these 5 measurements, giving one value between 0 – 1 m for each ditch. Emergent vegetation was then entered as a two level factor in statistical models with ditches having a value > 0.5 m categorized as having high emergent vegetation coverage and those with a value < 0.5 m as having low emergent vegetation coverage. An initial check of the data indicated that monitoring data on the number of submerged and emergent macrophyte species and their respective coverage were not significantly correlated to insects or microbial PLFAs. However, the presence/absence of submerged vegetation and the emergent vegetation extension measurements were significant correlated to insect abundance and PLFA content (|r| > 0.7, P < 0.05, Bonferroni-Holm) and where therefore applied in subsequent statistical analyses.

Fig. 4.2. Position of the 25 agricultural peatland ditches sampled in 2011 in North Holland. Ditches were sampled for aquatic insects, microbial phospholipids and abiotic parameters. Measurements of ditch dimensions, submerged vegetation and emergent vegetation were also taken in each ditch. Ditches sampled for insects in spring (April – May) are in bold and underlined.

Chapter 4

75

Insect sampling and identification Aquatic insects were collected once from each ditch between April and July 2011, in accordance with current Netherlands national monitoring protocols (Bijkerk 2010). Five ditches were sampled between 18th April and 26th May and twenty ditches between 6th June and 27th July. Although seasonal emergence can influence community composition, insects collected during the mid-spring to summer period have been shown to express low temporal variation in Dutch peatland ditches (Beltman 1983). Insects were collected within a 50 m stretch of ditch in which 5 – 10 meters of habitat were sampled with a dip-net (mesh-size 900 μm, width 30 cm). A multi-habitat sampling technique was employed, whereby different habitats (submerged macrophytes, sediments, open water and emergent vegetation) were subsampled proportionally and combined. Insect samples were taken back to the laboratory, sorted into groups and preserved in 70% ethanol for further identification. Where possible, insects were identified to species with the exception of Chironomidae larvae [Diptera], which were identified to either subclass or genus. Sediment sample collection, preparation and microbial phospholipid extraction Four sediment samples were collected in June in the same 50 m stretch of ditch that vegetation assessments were carried out and insects were collected. The four sediment samples were pooled to make one representative sample for each ditch. The majority of ditches had sediments composed of amorphous organically rich muds (i.e. degraded peat sediments), making it impossible to sample with a core sampler, therefore surficial sediments were gently collected with a dip-net. Sediment samples were freeze dried (Coolsafe model, 55-4) and subsequently stored at -20°C. Sediment organic matter (OM) content was determined in two replicate sub-samples for each location by loss-on-ignition (LOI) at 550°C for 2.5 hours in a furnace (Carbolite model, ELF 11/14B). Triplicate sediment subsamples were analysed for 6 locations to determine the representativeness of a single subsample. The locations included sites with high and low vegetation cover in eutrophic and mesotrophic ditches. The results showed that PLFA profiles were all strongly correlated between subsamples from the same ditch (rmedian = 0.97, P < 0.001), while the PLFA content among all 25 ditches expressed a wider range of correlations (rmedian = 0.60, P = 0.4). Therefore, a single 0.3 g (dry weight) sediment subsample from each ditch was analysed to determine the benthic microbial PLFA composition. The dried peaty sediments formed a natural powder which did not require sieving and care was taken not to include any large particles (sticks, stones, empty shells or animals) in the subsamples. PLFAs were extracted from freeze-dried sediment samples based on the method developed by Bligh and Dyer

Chapter 4

76

(1959) with modifications by White et al. (1979); Frostegård et al. (1991) and Dickson et al. (2009). Briefly, lipids were extracted from about 0.3 g dried sediment with a solution (1:2:0.8) of chloroform-methanol-citric acid buffer (0.15M, pH 4.0). Prior to extraction L-alpha-phosphatidiylcholin-1, 2-dinonadecanoyl (1.05 mg ml-1) was added to each sample, to serve as internal standard. Total lipid extracts were fractionated into neutral lipids, glycolipids and phospholipids on silica gel columns by elution with chloroform-acetic acid (100:1), acetone and methanol, respectively. PLFA extracts were first hydrolysed to fatty acids with NaOH in MeOH (0.5M, 100°C for 10 min), they were then methyl esterified with BF3 in MeOH (12.5% in water free methanol, 80°C for 15 min) and the excess BF3 was hydrolysed by saturated NaCl water solution. The fatty acids were then collected by liquid-liquid extraction with hexane, dried under N2, re-dissolved in trimethylpentane-nonane solution (9:1) and stored at -20°C prior to GC-MS analysis. Gas Chromatography and GC-Mass Spectrometry Sediment PLFAs were detected by gas chromatography mass spectrometry (GC-MS) using a Finnigan Trace Gas Chromatographic Mass Spectrophotometry (122-5562) interfaced to a Finnigan Trace MS on a SP-2560 (100 m X 0.25 mm X 0.2 μm) capillary column. Samples (1 μL) were injected at a split ratio of 50, on column (SP™ -2560 Capillary GC Column, L × I.D. 100 m × 0.25 mm) with helium as the carrier gas, at a constant flow rate of 0.8 ml/min. Initial oven temperature was 40°C, held for 1 min, subsequently increased to 160°C at a rate of 80°C/min, then raised to 210°C at a rate of 10°C/min, then to 250 at a rate of 1°C/min, then to 300 at rate of 25°C/min and held at 300°C for 10 min. The MS was in electron impact (EI+) mode with a full scan mass range of m/z 40–450. Lipids were identified by chromatographic retention time and mass spectral comparison with fatty acid methyl ester (FAME) mix and bacterial acid methyl ester (BAME) mix standards by Supelco®. Both standard mixtures were analysed in 5 different concentrations several times in between real samples. These were used as calibration curves to calculate the content of each identified compound in the samples. Peak areas for individual lipids were quantified using the Xcalibur program (version 1.0.0.1). Lipid concentrations were standardized by sediment OM content (dry weight) and losses were corrected for by using the internal standard recovery ratio.

Chapter 4

77

Fatty acid biomarkers The assignment of PLFAs to microbial groups was based on published literature where PLFA profiles had been determined either in field studies in conjunction with genetic analysis or from pure cultures (see Table 4.1 and references therein). In conjunction with the literature, the correlation matrix between fatty was analysed to assess associations between individual fatty acids within our dataset (see Appendix 4.1 for the summary of sediment PLFA content found during this study). Thereby PLFAs that were strongly correlated to one another (Pearson r > 0.8, P < 0.001, Bonferroni-Holm correction), and were sighted as being derived from the same microbial groups in the literature, were considered to originate from the same microbial group. Statistical analysis Residual and QQ-plots were used to assess normality and homogeneity of variances of abiotic, microbial PLFA and insect data. The sequential Bonferroni–Holm method was applied to correlations to correct for family-wise errors associated with multiple comparisons (Quinn & Keough 2002). If variables did not comply with the necessary model assumptions log10

transformation was applied prior to analysis. Two Principal Component Analyses were run. To explore the differences in water quality and dimensions between the ditches a PCA was run with all continuous abiotic variables, ditch width and water depth. To determine patterns of fatty acid distribution between ditches and analyse the data in multidimensional space, a separate PCA was run on microbial PLFA concentrations. These PCAs will be referred to as environmental (PCAE) and microbial (PCAM) hereafter. To categorize ditches by their microbial PLFA profiles we ran the relative proportion of microbial phospholipids in a single-linkage hierarchal cluster analysis, based on Euclidean distance. After consideration of the distribution and grouping of lipids, ascertained during cluster analysis and the PCAM, a ratio indicative of microbial heterotrophic to autotrophic PLFAs, was subsequently calculated as: SB/MP = Saturated FA + Branched FA / Monounsaturated FA +Polyunsaturated FA

Chapter 4

78

Table 4.1. Benthic microbial fatty acids and the microbial groups they are associated with. PLFAs were assigned to taxonomic groups after investigating correlations between PLFAs in our samples and based on published references. Fatty acids which were strongly correlated to one another (Pearson coefficients r > 0.7, P < 0.001) were considered to originate from the same taxonomic group.

Lipid group Fatty acids Microbial groups References Symbolsa

Saturated, Branched and Methyl branched fatty acids

14:0, 15:0, 16:0, 17:0, 18:0, 20:0, a15:0, i15:0, i16:0, i17:0, cy17:0

Bacteria, mainly gram positive anaerobic

1, 2, 3 ◊

10Me16, i17:1ω7c

Sulphate reducing bacteria

4, 5

♦

Monounsaturated and Polyunsaturated fatty acids

16:1ω7, 18:1ω9c, 18:1ω9t, 20:1ω9c 18:2ω6c, 18:2ω6t, 18:3ω3, 18:3ω6c, 20:4ω6, 20:5ω3, 22:6ω3

Of mixed origin including gram negative bacteria but mainly autotrophs, e.g. green algae and diatoms

3, 6, 7, 8, 9, 10, 11

□

16:1ω9, 18:2(9,12)

Cyanobacteria

8

■

References: 1Vestal & White (1989), 2Piotrowska-Seget & Mrozik (2003), 3Kaur et al. (2005), 4Taylor & Parkes (1983), 5Edlund et al. (1985), 6Potts et al. (1987), 7Léveillé et al. (1997), 8Siegenthaler & Murata (1998), 9Smoot & Findlay (2001), 10Dijkman & Kromkamp (2006), 11Findlay et al. (2008). a Symbols correspond with those in Figure 4.3b.

Chapter 4

79

Generalized linear models (GLMs) were run to assess the relationship between SB/MP and environmental variables. A set of GLMs were firstly run to determine which individual water quality variables were the most influential to microbial lipids (model set 1). Following this selection three candidate models (model set 2) were run to test a full model, containing both water quality and vegetation variables (i.e. presence or absence of submerged vegetation and degree of emergent vegetation coverage1) against two reduced models (II and III) which contained either water quality (PCAE axis 1) or vegetation. The following candidate models were specified a priori and subsequently tested: SB/MP ~ WQ + V SB/MP ~ WQ SB/MP ~ V Where: WQ = water quality (PCAE axis 1) V = as single variable representing emergent vegetation cover and presence or absence of submerged vegetation The Akaike information criterion, corrected for small sample size (AICc) and Akaike weights (wi) were used to determine the support for each model by the observations (Burnham & Anderson 2002). The wi are normalized relative likelihoods for each model and can be interpreted as the probability or the performance of each candidate model in relation to the other models in the set. Models with wi > 0.7 were considered to be strongly supported, between 0.4 – 0.7 as moderately supported and with wi < 0.4 as minimally supported by the data (Burnham & Anderson 2002). Only models with significant (P < 0.05) parameters and uncorrelated Gaussian residuals were considered adequate. To determine differences in aquatic insect abundance between ditches clustered by microbial fatty acids GLMs were run using the quasi-Poisson error model. Differences in insect species richness between clusters were tested with GLMs using either a Gaussian or Poisson error model. The total number and abundance of Ephemeroptera, Odonata and Trichoptera (EOT) taxa were used as an index of ditch habitat quality. Differences in vegetation and water quality parameters between clusters were tested using one-way ANOVA.

1 Represented by a single factor with 4 levels, where 0 = absence of submerged vegetation with small emergent vegetation cover, 1 = absence of submerged vegetation with large emergent vegetation cover, 2 = presence of submerged vegetation with small emergent vegetation cover and 3 = presence of submerged vegetation with large emergent vegetation cover

Chapter 4

80

A multivariate direct ordination technique was carried out on log10 transformed insect abundance data to determine insect community variability in relation to vegetation and abiotic conditions. Detrended correspondence analysis (DCA) was initially performed with down weighting of rare species to investigate the gradient length for aquatic insect community composition. The lengths of the first and second DCA axes were 2.84 and 2.69, respectively indicating that a unimodal response model was appropriate (Ter Braak & Smilauer 2002). Canonical correspondence analysis (CCA) was then run with insect abundances and environmental variables. Because a large number of abiotic (i.e. surface water chemistry) variables were correlated to one another, variance inflation factors (VIFs) were inspected and preliminary models run to determine which abiotic variable(s) were most strongly correlated with the insect community. Vegetation and benthic microbial variables were then tested in a similar fashion. The final CCA contained only significant environmental variables that were not strongly correlated to one another (VIFs < 2.0). All statistical analysis was run in R for windows using functions from the vegan, car and MuMIn packages (R Development Core Team 2012). Results Ditch characteristics Average values of surface water chemistry, ditch dimensions, emergent vegetation area and proportion of ditches supporting submerged vegetation are shown in Table 4.2. Results of the environmental Principal Component Analysis (PCAE) and Pearson correlation coefficients between the first and second PCAE axes and environmental data indicated that water quality and ditch dimensions were the dominant environmental gradients in the sampled ditches. Water quality was represented by the first axis, which was correlated to all abiotic variables and ditch dimensions were represented by the second axis, which was strongly correlated to ditch width and depth. The first and second PCA axes explaining 55% and 14% of variance between the 25 ditches, respectively. Emergent vegetation coverage was high in 14 ditches (56% of all ditches) and submerged vegetation was present in 14 ditches. Six ditches (24%) supported both high emergent vegetation coverage and submerged vegetation. Aquatic insect composition A total of 74 insect taxa were collected during this study and the average (± 95% CI) taxonomic richness was 13 ± 3 per ditch sample. Coleoptera were the most taxonomically rich group represented by 22 taxa but few were caught with an average abundance of < 1 individual collected per ditch. The average abundance of insect taxa was 95 ± 24 individuals collected in each sample, with

Chapter 4

81

29% of taxa recorded in this study being singletons (taxa recorded once). Diptera were the most abundant taxon group consisting of 8 genera/subclasses and Chironomidae larvae were the most abundant of all with Glyptotendipes and Chironomus recorded in > 70% of all locations with an average abundance of 10 ± 7 and 11 ± 6, respectively. Megaloptera was the least abundant taxon group, represented by one species (Sialis lutaria) with an abundance of 1 ± 1. Table 4.2. Average values and ranges of surface water parameters, ditch dimensions and vegetation collected in 25 ditches in 2011. Water chemistry values are taken from half year averages and ditch dimensions were measured simultaneously. Vegetation parameters were recorded in June 2011, n = 25. Pearson correlation coefficients of continuous variables with environmental principal component (PCAE) axes 1 and 2 ** good (|r| ≥ 0.7, P < 0.001) and * moderate (|r| ≥ 0.6, P < 0.05).

Parameters Average ± 95% CI

Range PCAE-1 PCAE-2

Abiotic variables:

Ammonium (mg NH4+ L-1) 0.47 ± 0.24 0.02 - 2.66 -0.92** -0.22

Kjeldahl nitrogen (mg N L-1) 3.8 ± 0.8 1.2 - 9.1 -0.70** -0.27

Nitrite (mg N L-1) 0.02 ± 0.01 0.01 - 0.07 -0.87** -0.07

Nitrate (mg N L-1) 0.29 ± 0.07 0.09 - 0.72 -0.89** 0.10

Total phosphorus (mg P L-1) 0.48 ± 0.14 0.05 - 1.27 -0.88** -0.10

Orthophosphate (mg PO43- L-1)

0.21 ± 0.09 0 - 0.8 -0.90** -0.05

Sulphate (mg SO42- L-1) 84 ± 12 23 - 133 -0.81** 0.08

Chloride (mg Cl- L-1) 265 ± 62 89 - 790 -0.78** 0.38

Total iron (μg Fe L-1) 823 ± 345 80 - 4192 -0.87** -0.28

Bicarbonate (mg HCO3- L-1) 198 ± 30 80 - 368 -0.85** 0.39

pH 8.0 ± 0.2 6.4 - 8.9 -0.54 0.39

Conductivity (mS m-1) 154 ± 23 70 - 329 -0.41 0.37

Transparency (cm) 36 ± 6 14 - 65 0.60* 0.22 Sediment organic matter (%) 40 ± 6 13 - 72 0.56 -0.05 Ditch dimensions: Width (m) 11 ± 3 2 - 35 -0.05 0.83** Water depth (cm) 68 ± 10 7 - 119 0.12 0.86** Vegetation: Ditches with large emergent vegetation coverage (%)

60 0.41 0.21

Ditches with submerged vegetation (%)

56

0.09

-0.34

Chapter 4

82

A total of 27 Ephemeroptera, Odonata and Trichoptera (EOT) taxa were recorded with an average abundance of 17 ± 9. Three Ephemeroptera taxa were recorded with an average abundance of 8 ± 7 and eighteen Trichoptera taxa with an abundance of 5 ± 3. Odonata had an average abundance of 4 ± 4 and was represented by 6 species with Pyrrhosoma nymphula being the least abundant, recorded only once in this study and Coenagrion spp. (pulchellum/puella) being the most abundant genus. Benthic microbial phospholipids and environmental characteristics A total of 26 PLFAs, ranging from 14:0 to 22:6ω3, were recorded in this study including established PLFA biomarkers for a range of microbial groups (see Table 4.1 and Appendix 4.1 in supplementary information). Cluster analysis, run on the relative proportion of benthic microbial PLFAs, grouped 24 out of 25 ditches into two clusters with a single ditch (number 18) positioned outside the two clusters (Fig. 4.3a). The average total PLFA content, based on sediment dry weight, was similar between the two main clusters at 15.6 ± 3.5 mg g-1 C (± 95% CI) in cluster 1 and 16.7 ± 6.8 mg g-1 C in cluster 2 ditches. However, there was a clear difference in the proportions of each fatty acid group with SB/MP (heterotrophic to autotrophic PLFAs) being greater in cluster 2 ditches (Fig. 4.3a). Vegetation cover also differed significantly between the ditches of each cluster. Cluster 2 ditches concomitantly supported submerged vegetation and greater emergent vegetation cover (ANOVA, F(1,22)= 11 P = 0.003). Furthermore, water quality (reflected by PCAE axis 1) was significantly positively related to cluster, i.e. cluster 2 ditches had better water quality (ANOVA, F(1,22) = 8.32, P = 0.009). There was no difference between ditch dimensions in the two clusters. The first two axes of the Principal Component Analysis (PCAM), run with benthic microbial phospholipid content, explained 57% of the variation between ditches with 34% of the variation explained by the first axis and 23% by the second axis (Fig. 4.3b). Saturated fatty acids (FAs) and branched FAs were clustered together in the centre of the biplot as they were ubiquitous to all ditches (see Fig 4.3b). These two lipid groups were predominantly associated with gram negative bacteria and anaerobic gram positive bacteria such as sulphate reducing bacteria (i.e. 10Me16:0, i17:1ω7 and cy17:0) (see Table 4.1 and references therein). Conversely, polyunsaturated FAs and monounsaturated FAs, were dispersed away from the centre of the biplot with the exception of two lipid biomarkers for cyanobacteria (16:1 ω 9 and 18:2(9,12)), which were situated in the centre of the biplot (Fig 4.3b).The ratio of saturated and branched FAs to monounsaturated and polyunsaturated FAs (i.e. SB/MP) was positively related to both vegetation and water quality and indicated a gradient from greater to lower water quality and greater to lower vegetation cover in the

Chapter 4

83

ditches (Table 4.3). Only abiotic parameters which were significantly correlated to SB/MP are shown in Table 4.3 and the Akaike weights indicated SB/MP was most strongly correlated to HCO3- , pH and SO42-. Table 4.3. Generalized linear model (GLM) results showing the response of microbial PLFAsa to the presence of submerged and emergent vegetation and abiotic factors. Model set 1 tests the response of SB/MP against individual water quality parameters. Model set 2 tests the response of SB/MP against a full model containing both vegetation and water quality (PCAE axis 1) and the two reduced models (II and III). Model outputs are shown with degrees of freedom (df), Akaike information criterion corrected for small sample size (AICc) and Akaike weights (wi), n = 25. Predictors Coefficients df Log likelihood AICc wi

Model set 1: HCO3- -2.16 3 -14.18 35.5 0.67 pH -0.62 3 -15.60 38.3 0.16 SO42- -2.03 3 -16.29 39.7 0.08 NO3- -1.29 3 -16.72 40.6 0.05 PO43- -0.45 3 -17.35 41.8 0.03 Transparency 0.021 3 -18.72 44.6 0.01 Model set 2: (I) Full Model 4 -10.22 30.4 0.95 (II) Water quality 0.47 3 -14.65 36.4 0.05 (III) Vegetation 0.36 3 -17.36 41.9 < 0.01 a PLFA data are expressed as the ratio of saturated and branched to monounsaturated and polyunsaturated fatty acids (SB/MP). Microbial and aquatic insect community composition There were clear differences in species richness and the abundance of insect taxa between the ditches grouped in the microbial PLFA cluster analysis (Fig. 4.3c). The average total insect abundance and abundance of Chironomus sp., Odonata and Trichoptera taxa were all significantly higher in cluster 2 ditches (GLM, P < 0.05). Moreover, total insect richness and EOT richness were also greater in cluster 2 ditches (GLM, P < 0.05).

Chap

ter 4

84

Fi

g. 4

.3. S

umma

ry of

micro

bial p

hosp

holip

id fat

ty ac

id (P

LFA

) and

aqu

atic

insec

t dat

a fro

m 2

5 pe

atlan

d dit

ches;

a) E

uclid

ean

dista

nce b

ased

den

drog

ram

based

on

PLFA

con

tent.

Mac

rophy

tes su

rveyed

in ea

ch d

itch,

subm

erged

(▲ =

pres

ence

and ∆

= ab

sence)

and

emerg

ent v

egetat

ion ex

tent (▲

= la

rge a

nd ∆

= sm

all).

Total

aver

age f

atty

acid

(FA

) con

tent a

nd th

e aver

age r

atio

of sa

tura

ted F

A +

bra

nche

d FA

/ p

olyun

satu

rated

FA

+ m

onou

nsatu

rated

FA

(SB/

MP)

in ea

ch cl

uster

; b) B

iplot

of th

e prin

cipal

compo

nent

an

alysis

(PCA

M) r

un w

ith dr

y weig

hts o

f ind

ividu

al PL

FAs.

The p

lot sh

ows t

he va

rianc

e in

PLFA

data

expla

ined

by ax

es 1

and

2 of

the P

CAM

; c) A

verag

e abu

ndan

ce of

insec

t gro

ups b

etween

ditc

hes i

n ea

ch cl

uster

. Sign

ifica

nt d

iffere

nces

(GLM

) are

indic

ated

as *

P <

0.0

5 an

d **

P <

0.0

1.

Chapter 4

85

The CCA explained a significant proportion of the variation in aquatic insect community composition (P ≤ 0.001) and showed that variation in the insect community was significantly correlated to the microbial lipid ratio SB/MP (P < 0.001), (i.e. the ratio of heterotrophic to autotrophic PLFAs), the presence of submerged vegetation (P < 0.001) and total phosphorus (TP) (P = 0.01) (Fig 4.4). The first constrained axis explained 19% of the variation in the insect community and was positively related to SB/MP and negatively related to TP. The second constrained axis explained 14% of the variation and was negatively related to the presence of submerged vegetation. All abiotic parameters were initially tested in the CCA and the final selection (shown in the biplot) consisted of those parameters which expressed the strongest relationship with variation in the insect community (Fig 4.4). During the variable selection procedure ammonium, orthophosphate, nitrite and Kjeldahl nitrogen were also significant correlated to variability in the insect community, however none of these variables were as strongly correlated to the insect community as TP. Moreover, TP was positively correlated to all of these aforementioned abiotic parameters (r > 0.66, P < 0.05), so TP was indicative of overall eutrophication in the ditches. Discussion Determining the underlying drivers of insect composition can be difficult in modified environments, particularly when suitable reference conditions are not available. The present study shows that combining data on microbes together with data on insects and vegetation lead to the identification of key divers affecting the interacting community components. This observation expands on a trend observed before. For example Feio et al. (2007) reported that diatoms and invertebrates provided complementary information on environmental conditions in streams with diatoms being more sensitive to water chemistry and invertebrates being more indicative of morphological characteristics. Similarly, Johnson and Ringler (2014) found that fish and invertebrate indices were indicative of habitat and water quality in streams, but the two groups were not correlated to one another. In the present study we integrated benthic microbial and insect data to determine if variation in these two communities were correlated with one another and identify environmental factors which structure these two communities. Microbial fatty acids and invertebrates have been incorporated typically in studies on food web dynamics. For example Vos et al. (2002) and Peeters et al. (2004) demonstrated that concentrations in sedimentary bacterial fatty acids and total PUFAs (including non PLFAs) positively influenced benthic invertebrates (e.g. Chironomidae larvae and other detritivorous taxa) by providing a high quality source of food. In our study, bacterial fatty acid and

Chapter 4

86

total PUFA contents were not related to the abundance of sediment dwelling Chironomidae larvae. This may be because degraded peatland sediments are rich in organic matter (accounting for on average 40% of the total sediment dry weight) in conjunction with PUFA content being over 1000 times greater than those recorded by Vos et al. (2002) and Peeters et al. (2004). Consequently, food quality and availability are not likely to be limiting and, therefore, are probably not an important factor structuring the insect community in nutrient rich peatland ditches. Another confounding factor could be the sample method used in our study which included littoral and open water insect species as well as benthic fauna. This type of sampling may reduce the sensitivity of statistical techniques to determine relationships between benthic microbes and insects collected from a variety of ditch substrates. However, Chironomidae larvae comprise of many sediment dwelling taxa (particularly Chironomus sp. and Glyptotendipes sp.) therefore the relationship between these taxa and food quality parameters would likely have been detected if it was prominent.

Fig. 4.4. First and second constrained axes of canonical correspondence analysis run with insect taxa and environmental variables. Environmental variables are presence or absences of submerged vegetation, total phosphorus (TP) and benthic microbial lipids expressed as the ratio of Saturated and Branched to Monounsaturated and Polyunsaturated fatty acids (SB/MP). Arrows represent continuous variables and labels in boxes show the centroids for submerged vegetation levels, all variables were significantly correlated with insect data (P ≤ 0.01 ). Circles represent ditches and the cluster they are associated with (○ = cluster 1, = cluster 2). Dominant taxa are represented by abbreviations: CHIR= Chironomus sp., COE = Coenagrion spp., C.DIP = Cloeon dipterum, C.HOR = Caenis horaria, C.ROB = Caenis robusta, ENDO = Endochironomus sp., GLY = Glyptotendipes sp., G.ODO = Gerris odontogaster, HAL = Haliplus sp., HEP = Helophorus sp., H.DUB = Holocentropus dubius, H.INA = Hygrotus inaequalis, I.CIM = Ilyocoris cimicoides, I.ELE = Ischnura elegans, N.CLA = Noterus clavicornis, ORT = Orthocladiinae, P.MIN = Plea minutissima, SIG = Sigara sp., S.LUT = Sialis lutaria, TAN = Tanytarsus sp., TANY = Tanypodinae, T.BIC = Triaenodes bicolor.

Chapter 4

87

In our study we assessed if microbial data was indicative of key environmental drivers of aquatic insects in North Holland’s peatlands. Variation in aquatic insect community composition was strongly correlated to benthic microbial phospholipids and the presence of submerged and emergent vegetation in North Holland’s peat land ditches (Fig. 4.3 and 4.4). The microbial community (as represented by clustered ditches and the SB/MP ratio) were indicative of water quality/chemical conditions and vegetation, while invertebrates were more indicative of vegetation/habitat structure. The initial clustering of ditches by microbial PLFAs resulted in two main clusters, with ditches situated in the second cluster characterised by more vegetation and supporting a greater abundance of insect taxa, predominantly Odonata, Trichoptera and Chironomus larvae. These taxa are of particular interest as they have shown a marked decline in North Holland’s peatland ditches over recent decades (van Dam 2009; Whatley et al. 2014). The ditches with the most diverse insect communities were associated with cluster 2 and were more likely to support submerged vegetation and have greater emergent vegetation cover compared with cluster 1 ditches. Similarly, Boon et al. (1996) reported measurable differences in microbial PLFA composition in response to the presence of vegetation in an Australian wetland. Vegetation influences microbial communities by stabilizing and adding organic material to sediments and creating pockets of oxygen around their roots zones (Sand-Jensen et al. 1982; Palmer et al. 2000; Colmer 2003). Furthermore, the expansion of oxic/anoxic boundary zones in sediments acts to alter potential electron donors and redox conditions exploited by microbes. However in the present degraded peatlands sediment deposits are substantial, very unstable and not consolidated by plant roots. Our observations may actually be affected by a decoupling of vegetation and sediment microbial markers. Nevertheless published findings highlight the role of vegetation as ecosystem engineers in these aquatic systems. Emergent and submerged vegetation influenced both benthic microbial communities and aquatic insects, with many species dependent on vegetation for habitat and periphytic food (Scheffer et al. 1984; De Szalay & Resh 1997; Twisk et al. 2003; Verdonschot et al. 2012a). Environmental conditions in many of North Holland’s peatlands are dominated by the process of peat degradation which leads to an accumulation of fine amorphous organic particles (Lamers et al. 2002). These particles smother surfaces and regular resuspension of particles causes light attenuation and subsequent decline in submerged vegetation, resulting in habitat loss (Scheffer et al. 1993; Whatley et al. 2014). The supply and concentrations of nutrients is yet another important factor influencing aquatic vegetation distribution in drainage ditches (De Lange 1972; Rip, et al. 2006; van Zuidam 2013). Under hypertrophic conditions submerged vegetation is often lost and replaced by algae and floating macrophytes (Janse & Van Puijenbroek 1998). In our study, the ratio of heterotrophic to autotrophic PLFAs was strongly

Chapter 4

88

associated with chemical conditions, specifically nutrients, sulphate, bicarbonate and pH. These findings indicate a greater abundance of autotrophs in sediment samples of more eutrophic ditches, likely arising from the deposition of planktonic algal and cyanobacteria on surficial sediments. The assignment of PLFAs to taxonomic groups, such as algae and anaerobic bacteria, are not always evident as the same PLFAs can be present in a number of different organisms (Frostegård et al. 2011). We found strong correlations between certain PLFAs. For example, fatty acids 16:0, 10Me16:0, 16:1ω9 and cy17:0 are indicative of sulphate reducing bacteria Desulfobacter sp. (Taylor & Parkes 1983) and were perfectly correlated to each other (Pearson r = 1). While, fatty acids 18:2ω6, 18:3ω3 and 18:3ω6 may originate from algal, cyanobacteria, fungi or higher plants (Siegenthaler & Murata 1998; Piotrowska-Seget & Mrozik 2003; Dijkman & Kromkamp 2006). These three PUFAs (18:2ω6, 18:3ω3 and 18:3ω6) were strongly correlated to one another in our study (r ≥ 0.8) and were most abundant in the more eutrophic locations. Considering the association of 18:2ω6, 18:3ω3 and 18:3ω6 with an increase in nutrient concentrations, particularly dissolved inorganic nutrients (NH4+, NO3- and PO43-) and the absence of submerged vegetation, we concluded that they were likely to have originated from algae and cyanobacteria. Aquatic insects are predominately indirectly impacted by nutrient loading with diversity/species richness generally decreasing under eutrophic conditions. The inlet of external waters into Dutch peatlands has been shown to trigger internal eutrophication, thereby negatively influencing peatland ecosystems (Roelofs 1991; Lamers et al. 2002). Neither bicarbonate nor sulphate were directly correlated to variability in insect community composition in the CCA. Yet the ratio of heterotrophic to autotrophic PLFAs (SB/MP) was negatively associated with these macro-ions and positively associated with vegetation. This indicates that ditches with high bicarbonate and sulphate concentrations are more likely to be dominated by algae and are less likely to support macrophytes. Thus high HCO3- and SO42- are indicative of reduced habitat quality for aquatic insects. These findings support our expectation that integrating benthic microbial and aquatic insect data can clarify the drivers of (insect) communities in these degraded wetlands. In the case of North Holland the drivers of insect communities include nutrient enrichment, the inlet of mineral rich water and the availability of suitable macrophyte habitat, particularly submerged vegetation.

Chapter 4

89

Acknowledgements We thank Gert van Ee, Emile Nat, Ron van Leuken and Herman van Dam for their help throughout this project, Pim Koelma, Coen Wagner and Alejandra Goldenberg for their assistance during fieldwork, Nigel Upchurch for the map. Thanks to Boris, Linda and Soraya whose work inspired this paper. We extend thanks to Annemieke Ouwehand, Andre Timmer and Ed Zijp of Natuurmonumenten, for supporting our work in the Wormer and Jisperveld and Naardermeer reserves. We also thank two anonymous reviewers, Wilco Verberk and Patrick Armitage for their helpful comments on this paper. This research was funded by Stichting Waterproef and Hoogheemraadschap Holland’s Noorderkwartier.

Chapter 4

90

Appendix 4.1. Phospholipid Fatty Acid (PLFA) content of sediments collected from 25 ditches in June 2011. Average values are shown ± 95% CI alongside the full range of values. PLFA content (mg g-1 C) Average ± 95% CI Range i15:0 0.63 ± 0.16 0.19 - 1.89 a15:0 1.00 ± 0.26 0.35 - 3.55 i16:0 0.40 ± 0.10 0.12 - 1.09 i17:0 2.56 ± 0.90 0.46 - 11.68 Branched 5.37 ± 2.98 2.45 - 16.13 C14:0 0.62 ± 0.21 0.16 - 2.58 C15:0 0.62 ± 0.21 0 - 0.49 C16:0 1.40 ± 0.43 0.33 - 5.33 10Methyl16:0 0.26 ± 0.06 0.06 - 0.67 cy17:0 0.19 ± 0.06 0.03 - 0.71 C17:0 0.64 ± 0.32 0.07 - 3.90 C18:0 0.28 ± 0.12 0 - 1.14 C20:0 0.21 ± 0.07 0 - 0.89 Saturated 1.56 ± 0.77 0.93 - 3.96 C16:1 2.76 ± 0.83 0.62 - 10.24 C16:1n7 1.72 ± 0.80 0 - 7.81 C17:1n7 0.17 ± 0.05 0 - 0.50 C18:1n9t 0.13 ± 0.09 0 - 0.91 C18:1n9c 0.43 ± 0.17 0 - 1.55 C20:1n9 0.12 ± 0.07 0 - 0.82 Monounsaturated 3.97 ± 1.98 2.38 - 10.11 C18:2 0.49 ± 0.14 0.1 - 1.59 C18:2n6t 0.14 ± 0.06 0 - 0.57 C18:2n6c 0.59 ± 0.3 0 - 2.65 C18:3n3 0.13 ± 0.09 0 - 0.99 C18:3n6 0.47 ± 0.21 0 - 1.92 C20:4n6 0.02 ± 0.02 0 - 0.19 C20:5n3 0.39 ± 0.25 0 - 2.65 C22:6n3 0.22 ± 0.17 0 - 1.77 Polyunsaturated 1.38 ± 0.64 0.49 - 3.48 Total fatty acids 21.09 ± 8.12 8.82 - 48.64

Chapter 4

91

Chapter 4

92

Ischnura elegans

Chapter 4

93

Chapter 5

Temporal abiotic variability structures invertebrate communities in agricultural drainage ditches

Manuscript: M. H. Whatley, J. A. Vonk,

H. G. van der Geest and W. Admiraal

Chapter 5

94

Abstract Abiotic variability is known to structure lotic invertebrate communities, yet its influence on lentic invertebrates is not clear. This study investigates the role of variability of nutrients and macro-ions in structuring macroinvertebrate communities in agricultural drainage ditches. Many low-lying agricultural areas contain drainage ditches which potentially provide suitable habitat for aquatic invertebrates. Hydrological management in the province of North Holland (The Netherlands) involves seasonal variability of water quality with the inlet of mineral rich, river derived water in summer. This temporal variability was analysed from monitoring data, collected over a 7 month period (February till August) and covering 84 ditches in three soil regions; sand, clay and peat. We ran linear regressions to determine correlations between invertebrate diversity and average abiotic conditions and abiotic variability (expressed as the Median Absolute Deviation). Invertebrate diversity was determined as local (α diversity), regional (γ diversity) and species-turnover (β diversity). In addition, functional community composition was examined by analysing insect life-history strategies. This study demonstrates that abiotic variability is as important as average abiotic conditions in structuring lentic invertebrate communities. Moreover, analysis of insect life-history strategies (LHS) in the different soil regions suggests that abiotic variability influences species-turnover by triggering the removal of some dominant species, thereby creating space for other species. Colonization and the community’s ability to recover from disturbance will depend on the proximity of disturbed sites to potential source populations. We conclude that aquatic invertebrates in agricultural ditches are being impoverished by wide-spread eutrophication and highly variable conditions. Introduction Agricultural landscapes cover approximately 38% of the Earth’s total land area (Clay 2004), thus it’s important that we understand how management activities underlie biodiversity patterns in these landscapes. In The Netherlands the management intensity of agricultural landscapes has increased over recent decades (Kleijn et al. 2004). In the low-lying provinces the landscape is characterized by networks of drainage ditches, which potentially provide suitable habitat to a range of aquatic organisms including invertebrates (Beltman 1983; van der Hammen 1992; Herzon & Helenius 2008; Verdonschot 2012). However, the diversity of invertebrates has been in decline over recent decades (van Dam 2009; Whatley et al. 2014). Individual hydrological systems in North Holland, known as polders, include plots of land separated by ditches and surrounded by the larger arterial

Chapter 5

95

canal system. Polders are hydrologically managed to maintain artificially stable water tables suitable for agriculture. To this end, water is pumped out of polders during wet periods and mineral rich river water is let into polders during dry periods. This inlet of mineral rich water can have an especially strong effect on abiotic conditions by altering the pH and buffering capacity of waters and increasing sulphate concentrations, which can trigger the release of sediment bound phosphates leading to internal eutrophication (Roelofs 1991; Lamers et al. 2002). Therefore, abiotic conditions in these agricultural ditches exhibit a high degree of variability, related to the inlet of external waters as well as maintenance (i.e. vegetation removal and dredging) and nutrient rich run-off from surrounding agricultural land (Twisk et al. 2000; Lamers et al. 2002; Verberk et al. 2007; Verdonschot 2012). Although abiotic variability is expected to influence biodiversity patterns in these areas, as has been demonstrated for invertebrates in lotic environments (McCabe & Gotelli 2000), the importance of abiotic variability has not been investigated for lentic invertebrate communities in North Holland’s heavily impacted environment. Invertebrate community composition can be influenced by many factors, including patch size and habitat connectivity (i.e. the ability of individuals to colonize new habitats) as well as environmental filtering processes (Chase 2003; Verdonschot et al. 2012b). More specifically, the stability and predictability of an environment can influence the persistence of species based on their specific traits relating to dispersal, tolerances and reproductive behaviour. While an inherent degree of (temporal) variability exists in nature, unpredictable events like those described above in agricultural ditches can be a disturbance for biological communities. Disturbance is defined here as an event which creates temporarily unsuitable abiotic conditions, such as anoxia, osmotic stress or decreased water transparency and ultimately leads to the local reduction in invertebrate abundance or the complete removal of certain species. Disturbance events can cause a range of community responses and may promote diversity at intermediate levels by removing competitive and dominate species (i.e. the intermediate disturbance hypothesis (Connell 1978)), but the greater the frequency and intensity of the disturbance the greater the chance it will lead to a decline in diversity. A stable equilibrium may not be reached in environments which experience regular disturbances and contain a range of species adapted to colonize newly available habitats (Connell 1978; Huston 1979, 1994). The rate at which a community can recover from a disturbance relates to habitat connectivity, the size of the available species pool (γ diversity) and species dispersal abilities. Therefore, it is important to consider multiple measures of diversity to understand how disturbance events are structuring communities in a given environment (McCabe & Gotelli 2000; Chase 2003). Moreover, community dynamics are linked to the specific adaptations of individual species to cope with disturbances. Species life-history strategies reflect a species

Chapter 5

96

integrated response to their environment because temporal and spatial environmental factors can filter out species which do not have suitable life-history strategies (Stearns 1976; Southwood 1977; Poff 1997; Verberk, van Noordwijk & Hildrew 2013). The representation of specific life-history strategies may, therefore, reveal additional information about how environmental conditions are driving community composition (Verberk et al. 2008b; Verberk et al. 2010; Verberk et al. 2013). Aside from environmental filtering, habitat size and geographical distance between suitable patches of habitat will also determine the ability of a population to recover from a disturbance (MacArthur & Wilson 1967). Indeed, invertebrate diversity in Dutch agricultural systems often varies between ditches and is likely to be associated with changeable environmental conditions (Verdonschot 2012). Since the size and number of aquatic habitats are in decline as a result of the increased agricultural intensity, it is important to understand the role of abiotic disturbance in structuring species diversity. A comprehensive study carried out over 20 years ago in the province of North Holland indicated that patterns in macroinvertebrate community composition were related to soil region, among other factors (van der Hammen 1992). These patterns were likely associated with differences in land use and hydrology in each soil region (van der Hammen 1992). However, land use has intensified over the past 20 years and it is not clear if these original patterns in diversity still hold. In this study we investigated the role of abiotic variability in underlying patterns of diversity in freshwater macroinvertebrates in an intensively managed low-lying agricultural landscape. To this end we examined monitoring data collected in the province of North Holland from 84 drainage ditches situated in the three dominant soil regions (sand, clay and peat). We investigated correlations between abiotic variability, average abiotic conditions and macroinvertebrate species-turnover to determine the importance of abiotic variability in driving invertebrate communities. In addition, the representation of specific insect life-history strategies, (relating to dispersal, synchronized life-cycles, tolerance to abiotic extremes and reproductive strategies) were investigated in different soil regions to provide further insight on insect community structure. Relationships between the geographical distance between ditches, abiotic conditions and species-turnover were also assessed to determine if scale was an important factor structuring invertebrate communities within this landscape.

Chapter 5

97

Materials and Methods Data collection and preparation Abiotic parameters Drainage ditches were sampled over a period of four years (2008-2011) as part of the standard monitoring of waters in the province of North Holland, The Netherlands (Fig. 5.1). From the available data 84 drainage ditches (17 sand, 24 peat and 43 clay soil ditches) were selected from a wider set of monitoring sites based on water body size (width range 2 – 40 m, depth range 0.1 – 1.8 m), salinity (average chloride concentrations < 1000 mg L-1) and sampling season of the invertebrate communities (from mid to late summer, i.e. June to August). In the same ditches abiotic parameters were measured monthly, between February and August during the same year that invertebrates were sampled. Transparency and pH were measured in the field (by Secchi disk and a WTW pH/Oxi 340i/set meter, respectively) and undisturbed, overlying water was collected for laboratory analyses of nutrients (nitrogen KN, NO3-, NO2- and NH4+), phosphates (TP and PO43-)) and macro-ions (sulphate (SO42-), total iron (Fe), bicarbonate (HCO3-) chloride (Cl-), sodium (Na+), magnesium (Mg2+), calcium (Ca2+) and potassium (K+)), using protocols accredited by the Netherlands National Standards Institute (NEN). Invertebrates Invertebrates were surveyed once in the 84 ditches, between June and August in either 2008, 2010 or 2011, in accordance with the Netherlands national monitoring protocols (Bijkerk 2010). Macroinvertebrates (> 1 mm length) were collected with a dip-net (mesh-size 900 μm, width 30 cm) using a multi-habitat sampling technique whereby 5 to 10 meters of the available habitats (submerged macrophytes, sediments, open water and emergent vegetation) were subsampled proportionally over a 50 m stretch of ditch. Invertebrate subsamples from the same ditch were combined and sorted into groups in the laboratory and preserved in 70% ethanol for further identification. Where possible, invertebrates were identified to lowest possible taxonomic level, usually species, using identification keys recommended by Bijkerk (2010). Although we measured taxonomic richness in this study we refer to species richness from this point forward.

Chapter 5

98

Fig 5.1. Position of ditches sampled in the three soil regions. Invertebrates were collected once in the summer (June to August) from each ditch in 2008, 2010 or 2011. Abiotic variables were collected from all ditches at monthly intervals from February till August in the same year invertebrates were collected. Macroinvertebrate diversity We determined diversity as species richness (α diversity), species-turnover (β diversity) and species richness per soil region (γ diversity). Species-turnover was calculated using the Jaccard’s dissimilarity index (Cd) with binary data. Bray-Curtis dissimilarities were also calculated to determine differences in species abundance. Abundance data were square root transformed and subsequently standardized using Wisconsin double standardization (Gauch Jr. et al. 1977; Oksanen 2013) prior to analysis. The Jaccard’s index is predominantly referred to throughout the study, it is a metric measure of dissimilarity that can be used in conjunction with linear models (Quinn & Keough 2002) and is calculated by:

Chapter 5

99

Where b and c are the number of unique species present in each community and a is the number of species shared between two communities (Quinn & Keough 2002). Dissimilarity values for the two indices range between 1, where no species are the same and 0, where species composition is identical between samples. Species accumulation curves were constructed to compare the cumulative number of invertebrate species collected with every additional sample and compare α, β and γ diversity among soil regions. Insect life-history strategies Insect life history strategies (LHS) were assigned to each species following the methods published by Verberk et al. (2008a) and Verberk et al. (2008b). Species that were not in these studies were categorized on the basis of the defining traits, which were derived for each genus using the trait database of Tachet et al. (2002). Specific traits from this database relating to dispersal, reproduction, life-cycle duration, potential number of reproductive cycles per year, resistance forms and trophic level were applied. Individuals belonging to eight LHS were relevant to this study (for details see Table 3.1 and Verberk et al. (2008a)). These represented the four trait domains, being related to dispersal (D1, D2 and D3), synchronization (S1, S2 and S3), reproduction (R1) and development, i.e. species that are able to tolerate environmental stress because of developmental trade-offs and other specific traits (T1). The traits and adaptations characteristic of each LHS are briefly described as follows. D1 – strong dispersal, rapid juvenile development, long-lived adults and high voltinism; adapted to fragmented habitats with a predictable level of variation. D2 – strong dispersal, slow juvenile development and large clutch size; adapted to stable and fragmented habitats. D3 – moderate dispersal, rapid reproductive development, high voltinism, short-lived adults (mainly semelparous); adapted to briefly suitable and fragmented habitats. T1 - tolerant to abiotic extremes by making trade-offs resulting in small body size, elongated development time or weak dispersal; adapted to stable and harsh environments. R1 – protracted oviposition, rapid juvenile development, relatively long-lived adults and moderate dispersal; adapted to briefly suitable and fragmented habitats. S1 - synchronized emergence, long juvenile development and short-lived adults; adapted to predictable and stable environments. S2 – synchronized emergence, iteroparous and resistant stages; adapted to predictable but changeable environments. S3 – synchronized emergence, rapid juvenile development, long-lived adults and seasonally iteroparous; adapted to predictable and moderately stable environments. The S2 and S3 synchronized strategists were very low in abundance and their defining

Chapter 5

100

traits were similar to that of S1, thus the three ‘S’ strategists were combined into one group for statistical analysis. Statistical analysis Abiotic data Median absolute deviations (MAD) were calculated for nutrients, macro-ions, pH and water transparency to quantify abiotic variability over the sampling period. Combined abiotic variability within each soil region was determined by running Principal Component Analyses with log10 (x +1) transformed MAD values (PCAV). Likewise, average abiotic conditions were determined in each soil region by running separate PCAs with log10 transformed average values (PCAA). In this way two PCAs were produced for each soil region giving a total of six PCAs. A matrix of Euclidean distances between ditches was calculated with geographic distance to determine if the physical separation between ditches was related to abiotic conditions or invertebrate richness in each soil region. Abiotic parameters were compared between ditches in different soil regions using one-way ANOVAs. Homogeneity of variances and normality of residuals were checked for each parameter by checking box and whisker plots, Q-Q plots and running Levene’s test of homogeneity. ANOVAs were considered significant at P < 0.01 to avoid acceptance of marginal results, which may have been the result of unequal sample sizes between the different soil regions (Quinn & Keough 2002). The sequential Bonferroni–Holm method was applied to correlations and linear models to correct for family-wise errors associated with multiple comparisons (Quinn & Keough 2002). Invertebrate data Mantel tests were employed to determine the relationship between invertebrate β diversity, geographical distance, and dissimilarity in abiotic conditions among ditches (Quinn & Keough 2002). Four matrices were entered into Mantel tests, namely Jaccard’s dissimilarity index (Cd), the dissimilarity in ditch geographical coordinates (G), the dissimilarity in abiotic variation (V) and the dissimilarity in average abiotic conditions (A). In addition, partial Mantel tests were run to control for the influence of spatial correlation in abiotic factors, whereby a third matrix was entered as a controlling factor (i.e. to test for the influence of dissimilarity in abiotic variation, while controlling for geographical location). Mantel test results were based on Pearson’s correlation coefficients (r) and P values were determined from randomization tests over 10,000 permutations. Linear regressions were run to assess the relationship between invertebrate diversity (β and γ) and abiotic stability (PCAV axis 1) in the three soil regions. In

Chapter 5

101

this case the average β diversity was calculated from pair-wise comparisons between groups of three ditches, with the exception of one group of 2, in the sand region, and one group of 4, in the clay region. The pair-wise comparisons were made between ditches which shared similar levels of abiotic variation (PCAV – axis 1). Similarly, γ diversity was determined as the combined species assemblage of the same ditch groups from which the above mentioned β diversities were calculated from. With the exception of PCAA – axis 2 in sand ditches, the first two PCA axes (for PCAA and PCAV) for each soil region were positively correlated to the abiotic parameters (see Appendix 5.1). To determine the influence of abiotic conditions on invertebrate and insect α diversity and insect functional (LHS) composition a set of three candidate models (in this case linear regressions with Gaussian residual distributions) were run. A full model (I), containing both abiotic variation (PCAV axis 1 and 2) and average abiotic conditions (PCAA axis 1 and 2), was tested against two reduced models which contained either abiotic variation (II) or average abiotic values (III). The following candidate models were therefore specified a priori: α ~ V + A α ~ V α ~ A Where: V = Variation in abiotic parameters (PCAV axes 1 and 2) A = Average values of abiotic parameters (PCAA axes 1 and 2)2 The Akaike weights (wi) were calculated from the Akaike information criterion, which were corrected for small sample size (AICc), and used to determine the support for each model by the observations (Burnham & Anderson 2002). The wi are normalized relative likelihoods for each model and can be interpreted as the probability or the performance of each candidate model in relation to the other models in the set. Models with wi > 0.7 were considered to be strongly supported, between 0.4 – 0.7 as moderately supported and with wi < 0.4 as marginally supported by the data (Burnham & Anderson 2002). Only models with significant (P < 0.05) parameters and uncorrelated Gaussian residuals were considered adequate. All statistical analysis was run in R for Windows using functions provided in the vegan, car and MuMIn packages (R Development Core Team 2012).

2 PCAA axis 2 (sand) was not included in the models because it was not significantly correlated to any abiotic parameter; therefore this axis was not ecologically meaningful.

Chapter 5

102

Results Characteristics of agricultural ditches Abiotic conditions Average abiotic conditions differed significantly (ANOVA, P < 0.01) between ditches in the three soil regions (Table 5.1). Sand ditches had significantly lower concentration of SO42- and Mg2+ and the highest concentration of phosphates (TP and PO43-). Clay ditches had the highest concentration of Ca2+. Peat ditches had the highest concentration of Cl-, total Fe, Na+ and nitrogen (KN and NO3-), while having the lowest concentration of HCO3- and the lowest water transparency. The degree of abiotic variation (MAD) also differed significantly between soil regions. Sand ditches had the lowest variation in SO42- but the highest variation in PO43-. Clay ditches had the highest variation in HCO3- and Ca2+ and peat ditches exhibited the highest variation in Cl- and Na+. The six PCAs (two for each soil region) explained a large proportion of the variation in abiotic conditions over the three regions (see Appendix 5.1). The first and second axes of the PCAs relating to average abiotic conditions (PCAA) explained 85% and 7% of the variation in sand ditches. PCAA axis1 (PCAA-1) was significantly positively correlated to concentrations of macro-ions (Cl-, HCO3-, SO42-, Ca2+, K+, Mg2+ and Na+) and nutrients (KN, TP and PO43-) and negatively correlated to concentrations of total Fe. PCAA axis2 (PCAA-2) did not correlate to any abiotic parameter in sand ditches. In clay ditches PCAA-1 and PCAA-2 explained 57% and 21% of the variation in average abiotic conditions. PCAA-1 was positively correlated with concentrations of macro-ions (Cl-, total Fe, HCO3-, Ca2+, Mg2+ and Na+) and nutrients (KN and TP). PCAA-2 was positively correlated to SO42-, Ca2+ and negatively associated with water transparency in clay ditches. In peat ditches PCAA-1 and PCAA-2 explained 40% and 37% of the variation in average abiotic conditions. PCAA-1 was positively correlated with concentrations of total Fe, K+ and nitrogen (NH4+, KN and NO3-) and PCAA-2 was positively correlated to concentrations of macro-ions (Cl-, K+, Mg2+ and Na+) in peat ditches.

Chapter 5

103

Table 5.1. Abiotic parameters in ditches from each soil region. Values shown are averages and variability determined as the Median Absolute Deviation (MAD) calculated from monthly measurements. Ditches were sampled between February to August in the same year invertebrates were sampled, n = 7. Significant differences between soil regions are indicated in bold (ANOVA, P < 0.01, Bonferroni-Holm correction).

Average Variation (MAD) Number of ditches

Sand 17

Clay 43

Peat 24

Sand 17

Clay 43

Peat 24

Macro-ions Chloride (mg Cl- L-1) 218 203 361 25 20 68 Total iron (μg Fe L-1) 776 350 1025 100 70 163 Bicarbonate (mg HCO3- L-1) 303 373 245 20 50 30 Total sulphate (mg SO42- L-1) 73 121 109 7 16 20 Calcium (mg Ca2+ L-1) 94 123 74 8 25 7 Potassium (mg K+ L-1) 14.2 11.5 15.9 1 1 2 Magnesium (mg Mg2+ L-1) 23.9 31.7 37.9 3 3 5 Sodium (mg Na+ L-1) 137 130 223 20 15 35 Nutrients Ammonium (mg N L-1) 0.18 0.23 0.47 0.06 0.06 0.07 Kjeldahl Nitrogen (mg N L-1) 2.4 2.3 4.2 0.5 0.3 0.5 Nitrite (mg N L-1) 0.03 0.02 0.03 0.01 0.01 0.01 Nitrate (mg N L-1) 0.26 0.16 0.33 0 0 0 Total phosphorus (mg P L-1) 1.43 0.86 0.69 0.33 0.17 0.14 Orthophosphate (mg P L-1) 1.15 0.60 0.37 0.32 0.14 0.12 Field measurements pH 8.0 8.1 8.0 0.2 0.2 0.2 Transparency (cm) 45 43 34 10 8 5

The first and second axes of the three PCAs run on variation in abiotic parameters, represented by MAD values (PCAV) explained 44% and 22% of the variation in sand ditches. PCAV-1 was positively correlated to variation in Cl-, Mg2+ and Na+ and PCAV-2 was positively correlated to total Fe in sand ditches. In clay ditches, PCAV-1 accounted for 39% the variation and was positively correlated to variation in HCO3-, SO42-, Ca2+ and negatively to variation in Na+. PCAV-2 accounted for 17% of the variation in clay ditches and was positively correlated to Cl-, total Fe, K+ and Na+. In peat ditches PCAV-1 accounted for 47% of the variation and was positively correlated to variation Cl-, K+, Mg2+ and Na+ and PCAV-2 accounted for 20% of the variation in peat ditches and

Chapter 5

104

was positively correlated to total Fe and KN. The Pearson’s correlation coefficients (r) between abiotic variables and axes 1 and 2 of the six PCAs are presented in supporting information (Appendix 5.1). Invertebrate community composition A total of 159 invertebrate species, including 90 insect species, were encountered during this study. There was no difference in the cumulative number of invertebrate species between ditches in the three soil regions (Fig. 5.2). All six species accumulation curves indicated a low α diversity (average number of species in one sample), ranging between 25 to 26 invertebrate species and 11 to 13 insect species, while γ diversity was relatively high, ranging between 111 and 134 invertebrate species and 58 and 72 insect species. The most numerous taxa collected within the ditches of the three soil regions were Gastropods (Anisus vortex and Bithynia tentaculata), followed by Isopods (Asellus aquaticus), Amphipods (Gammarus sp.), Arachnida (Limnesia undulata), Diptera (Chironomus sp. and Glyptotendipes sp.) and Heteroptera (Sigara sp.). The three most abundant species were Asellus aquaticus, Anisus vortex and Gammarus sp., in sand ditches and Gammarus sp., Anisus vortex and Sigara sp. in clay and peat ditches. A large proportion of species were recorded at low numbers, with 43 species having less than 6 observations over all ditches, including 15 singleton species (i.e. observed once during this study). Within this group of the 43 least numerous species, insects made up the largest proportion with 15 species of Coleoptera (e.g. Coelostoma orbiculare), 6 species of Diptera (e.g. Dicrotendipes lobiger), 4 species of Heteroptera (e.g. Gerris odontogaster) and 3 species of Trichoptera (e.g. Cyrnus flavidus) having less than 6 observations over all ditches. The Arachnida followed the insects as the second least abundant group with 7 species having less than 6 observations (e.g. Neumania vernalis and Arrenurus cuspidifer).

Chapter 5

105

Fig. 5.2. Cumulative number of species sampled from ditches in each soil region. Solid lines represent invertebrate species richness and dashed lines represent insect richness. Invertebrate diversity in relation to geographical distance and environmental conditions Invertebrate species-turnover (β diversity) was high among individual ditches in different soil regions (i.e. the difference in α diversity among soil regions) (Fig. 5.3a). Dissimilarities were comparable when calculated by Jaccard’s index (i.e. presence of species) and Bray-Curtis index (i.e. relative abundance of species). Invertebrate species-turnover between soil regions (i.e. the difference in γ diversity) was low between ditches in the three soil regions. Dissimilarities were slightly higher between peat and sand ditches and lowest between clay and sand ditches. Although Bray-Curtis dissimilarity were slightly greater than Jaccard’s, indicating that community composition was similar yet the relative abundance of species changed between regions, the two indices were, however more or less analogous (Fig. 5.3b). These results indicate that despite the similarity in species composition between the three soil regions, invertebrate species composition varied considerably between individual ditches. Furthermore, the ditches of neighbouring soil regions shared more species than regions situated further apart (i.e. peat and clay shared more species than sand and peat).

Chapter 5

106

Fig 5.3. Invertebrate β diversity between ditches and soil type regions. Dissimilarities were calculated with binary (Jaccard), and with standardized abundance data (Bray-Curtis), a) β diversity between ditches (α diversity, averages + 95% CI), and b) β diversity between soil type regions (γ diversity). Geographical distance The results of Mantel tests between invertebrate community dissimilarity and geographical distance indicated that invertebrate β diversity (Jaccard’s index) was positively correlated with geographic distance (G) in peat and clay ditches but not in sand ditches (Table 5.2). Dissimilarity in abiotic variability (V) and average abiotic conditions (A) were positively correlated to invertebrate β diversity in all three soil regions. Yet, partial Mantel tests showed that when controlling for geographical distances, dissimilarity in abiotic variability between ditches were only significantly correlated to β diversity in sand and clay ditches.

Chapter 5

107

While the dissimilarity in average abiotic conditions were only positively correlated to β diversity in sand ditches, when controlled for geographical distance between ditches (Table 5.2). Table 5.2. Correlation coefficients (r) derived from simple and partial Mantel tests for ditches in different soil regions. Invertebrate β diversity was calculated with Jaccard’s index (Cd) and tested against Euclidean distances for geographical distance between ditches (G), abiotic variability (V) and average abiotic conditions (A). Abiotic measures were derived from PCA axes 1 and 2. In partial Mantel tests the correlation between matrices Cd and A, V or G were conditioned by a third matrix (G, V or A). Significant results are indicated in bold as * P < 0.05, ** P < 0.01, *** P < 0.001. Mantel test Partial Mantel test Cd.G Cd.V Cd.A Cd.G|V Cd.G|A Cd.V|G Cd.A|G Sand 0.10 0.27** 0.30** 0.11 0.12 0.27** 0.31** Clay 0.28** 0.36*** 0.17* 0.19** 0.23*** 0.30*** < 0.01 Peat 0.22** 0.18* 0.17* 0.20** 0.20** 0.15 0.14

Environmental factors Invertebrate β diversity (Jaccard’s index, based on pair-wise comparisons between ditch groups) was positively related to abiotic variability (PCAV 1) in peat and clay ditches (linear regression, r2 = 0.62, P = 0.01 and r2 = 0.48, P = 0.007, respectively) (Fig. 5.4a and b). While there was no relationship between invertebrate β diversity and abiotic variability in sand ditches (Fig. 5.4c). Regional diversity (γ, the total species richness of the same three ditches from which the above mentioned β diversities were calculated from) was not related to abiotic variability (PCAV 1) nor gradients in the average abiotic conditions (PCAA 1) in any of the three soil regions. Similarly, neither abiotic variability nor average abiotic conditions were significantly related to invertebrate α diversity in any of the soil regions (Table 5.3). However, in peat ditches insect α diversity was positively correlated to PCAV 2 (r2 = 0.15, P = 0.04), which was indicative of variability in total Fe and KN in peat ditches. There was no relationship between insect α diversity and abiotic conditions in sand and clay ditches. The representation of the different insect life-history strategies (LHS) was similar between clay and peat ditches but sand ditches had slightly fewer dispersal strategists (D1, D2 and D3). Considering results over all soil regions abiotic variation was significantly related to the representation of life-history strategies (D1, D3 and T1) (Table 5.3). Similarly, average abiotic conditions were significantly related to differences in the proportions of all LHS applied in this study (D1, D2, D3, T1 and R1). However, relationships between abiotic conditions and invertebrates differed between soil regions.

Chapter 5

108

Abiotic variability was both negatively and positively associated with the representation of different insect life-history strategies. In sand ditches T1 taxa were negatively correlated with PCAV-1 (r2 = 0.57, P < 0.001), which was indicative of variability in Cl-, Na+ and Mg2+. While, in clay ditches D1taxa were positively correlated to PCAV-1 (r2 = 0.13, P = 0.03), which represented variability in specific macro-ions (HCO3-, SO42-, Ca2+ and Na2+). Similarly, in peat ditches, insect α diversity and the proportion of D1 taxa were both positively associated with PCAV-2 (r2 = 0.15, P = 0.04) and (r2 = 0.37, P = 0.005), respectively. In peat ditches PCAV-2 was indicative of variability in Cl-, Na+ and Mg2+. Conversely, also in peat ditches, D3 taxa were negatively associated with PCAV-1 (r2 = 0.20, P = 0.04), which was indicative of variability in Fe and KN (see also supplementary data in Appendices 5.1 and 5.2). The majority of life-history strategies appeared to be negatively associated with average abiotic conditions. In sand ditches D1 taxa were negative correlated to PCAA-1 (r2 = 0.30, P = 0.01), which was indicative of macro-ion and nutrient concentrations. Similarly, in clay ditches D2 and R1 taxa were both negatively associated with PCAA-1 (r2 = 0.15, P = 0.01) and (r2 = 0.25, P < 0.001), respectively. In clay ditches PCAA-1 was indicative of greater concentrations in macro-ions (Cl-, Fe, HCO3-, K+, Mg2+, Na+), nutrients (KN, TP) and lower transparency. Conversely, T1 taxa were weakly positively correlated to PCAA-1 (r2 = 0.10, P = 0.03) in clay ditches. In peat ditches D3 were concomitantly positively associated with PCAA-1 and negatively associated with PCAA-2 (r2 = 0.38, P = 0.009). In peat ditches PCAA-1 was indicative of concentrations of total Fe, NH4+ and KN, while PCAA-2 was indicative of concentrations of Cl-, Na+ and Mg2+ (full model outputs are supplied in supplementary information Appendix 5.2, Tables S1 – S3).

Chapter 5

109

Fig. 5.4. Invertebrate β diversity (Jaccard’s index) plotted against abiotic variation (PCAV axis 1), a) sand ditches, b) clay ditches and c) peat ditches. Significance levels were determined by linear regression.

Chap

ter 5

110

Tabl

e 5.

3. A

kaik

e weig

hts c

alcula

ted fr

om li

near

regre

ssion

s. A

set o

f thr

ee reg

ressio

ns w

ere ru

n to

test t

he re

spon

se of

inver

tebra

te an

d in

sect a

lpha

divers

ity

(aver

age n

umbe

r of s

pecie

s per

samp

le) a

nd th

e perc

enta

ge of

differ

ent i

nsect

life-

histo

ry str

ategi

es, a

gain

st ab

iotic

varia

tion

(PCA

V a

xes 1

and

2) a

nd a

verag

e ab

iotic

condit

ions (

PCA

A a

xes 1

and

2).

Sign

ifica

nt re

sults

(P <

0.0

5) a

re sh

own

in bo

ld. A

verag

e perc

enta

ges of

each

insec

t life

-hist

ory st

rateg

y with

in ea

ch so

il reg

ion a

re un

derli

ned.

Degr

ee of

Ada

ptatio

n su

mmar

izes

the

degre

e to

which

insec

ts wi

th d

iffere

nt li

fe-hi

story

strat

egies

are

adap

ted to

spe

cific

envir

onme

ntal

condit

ions a

nd in

dicat

ed a

s Wea

kly

, Mod

eratel

y a

nd W

ell a

dapte

d . P

CA re

sults

and

full

mode

l out

puts

are p

rovid

ed in

App

endix

5.1

and 5

.2.

In

sect

life

-hist

ory

stra

tegi

es

α

inve

rtebr

ates

α

inse

cts

D1

D2

D3

S T1

R1

N

o. o

f tax

a 15

9 90

17

6

4 16

34

9

Degr

ee of

Ada

ptatio

n

E

nviro

nmen

tal v

ariab

ility

H

abita

t fra

gmen

tatio

n

H

arsh

abi

otic

cond

ition

s

Sa

nd

11%

3%

18

%

14%

48

%

7%

(I) F

ull M

odel

0.08

0.

19

0.23

0.

10

0.09

0.

15

0.16

0.

17

(II) V

ariat

ion

0.48

0.

70

0.06

0.

54

0.52

0.

73

0.84

0.

19

(III)

Ave

rage

0.

44

0.11

0.

71

0.36

0.

39

0.12

<

0.01

0.

65

Clay

5%

21

%

13%

9%

45

%

7%

(I) F

ull M

odel

0.34

0.

51

0.35

0.

48

0.27

0.

29

0.73

0.

44

(II) V

ariat

ion

0.39

0.

21

0.63

0.

05

0.69

0.

11

0.04

0.

00

(III)

Ave

rage

0.

27

0.28

0.

03

0.46

0.

04

0.60

0.

23

0.56

Pe

at

9%

6%

21%

12

%

47%

5%

(I)

Ful

l Mod

el 0.

17

0.35

0.

32

0.17

0.

56

0.15

0.

26

0.25

(II

) Var

iatio

n 0.

58

0.55

0.

37

0.32

0.

05

0.29

0.

32

0.37

(II

I) A

vera

ge

0.25

0.

09

0.32

0.

51

0.39

0.

56

0.43

0.

39

Chapter 5

111

Discussion The network of drainage ditches covered in this study spanned across approximately 650 km2 of agricultural land and three soil regions. All soil regions had high concentrations of macro-ions and nutrients, which are characteristic of the agricultural polder landscape in The Netherlands (van der Hammen 1992; van Dam 2009; Verdonschot et al. 2012b). Although species richness (α, β and γ diversity) was consistent among soil regions, there were differences in the representation of insect life-history strategies and importantly there were differences in abiotic conditions between soil regions. Individual polders have developed distinctive abiotic conditions, for example in relation to land use and edaphic factors like organic matter content and the capacity of soils to bind nutrients. This is particularly relevant for clay and peat soils (Roelofs 1991; Janse & van Puijenbroek 1998; Lamers et al. 2002; Verberk et al. 2007). Moreover, the eutrophicated and hydrologically managed peat areas are prone to releasing large quantities of fine organically rich sediments as a result of peat mineralization and internal eutrophication processes (Roelofs 1991; Smolders et al. 2006; Verberk et al. 2007). This was reflected in our study by the low average water transparency recorded in peat soil ditches. The majority of ditches are subject to eutrophication in conjunction with a decline in submerged macrophyte diversity. Consequently the availability of suitable habitats is limited in these environments (Whatley et al. 2014). Moreover, it is likely that habitat fragmentation and dispersal processes strongly influence population dynamics in drainage ditches, as has been observed for lentic invertebrates in the western Mediterranean region (Arribas et al. 2012). High β diversity (species-turnover) alongside low α diversity (local diversity) were identified for invertebrate communities in this study. These findings are in agreement with those for ditch communities in other parts of the Netherlands (Verdonschot 2012). In addition, invertebrate species-turnover was correlated to geographical distance between ditches. Sand covered the smallest sample area and peat covered the largest sample area in this study. Accordingly, the correlation between abiotic parameters and invertebrate species-turnover was not related to geographical distance in sand ditches, but was moderately related in clay ditches, whilst geographical distance and abiotic conditions were strongly interconnected in the peat environment. A major factor which may be underlying this trend is the degree of connectivity between ditches. While the province of North Holland is crisscrossed by interlacing ditches, the degree of hydrological connectivity between them is strongly influenced by management (e.g. damming, pumping and inlet of mineral rich water) and may vary seasonally depending on rainfall and water use. Thus water chemistry may be weakly correlated to geographical distance due to seasonal changes in the pumping and inlet of water.

Chapter 5

112

Aside from the effects of eutrophication and spatial factors, our study identified abiotic variability as an important driver of invertebrate diversity in agricultural drainage ditches. These findings are in agreement with relationships documented between environmental variation and invertebrate community composition in wetlands, lakes, estuaries and river systems, predominantly as a result of seasonal changes in vegetation growth and hydrological dynamics in these environments (Laprise & Dodson 1993; De Szalay & Resh 2000; Bunn & Arthington 2002; Malmqvist 2002). In lotic systems the impacts of sporadic pulses of suspended sediments and nutrients have been reported as having particularly strong effects on aquatic invertebrates (Wood & Armitage 1997; Dolédec et al. 2006; Wagenhoff et al. 2011). Moreover, regular abiotic variation could have long-term effects on communities that may not be detectable for weeks or even generations after the event. For example, Verdonschot & Verdonschot (2013) reported that the effects of experimental shading in combination with anoxic conditions on macroinvertebrate communities in ditches were not evident until several weeks following the event. In this case changes in taxonomic composition appeared to relate to species mobility and breathing method. Less mobile species, which were unable to avoid undesirable conditions and those species which obtained oxygen from the water showed the greatest rates of decline (Verdonschot & Verdonschot 2013). The occurrence of abiotic variability can be expected to increase with land use intensification as inputs of nutrients and the requirement for water will go up, placing more pressure on aquatic environments. Therefore in situ studies, as presented here, provide valuable information on how communities cope with environmental variability. Spatial and temporal environmental filters structure aquatic invertebrate communities and in the present study there was a clear relationship between invertebrate species-turnover (β diversity) and temporal abiotic variability in clay and peat ditches. Conversely, when we analysed the relationship between abiotic variability and invertebrate diversity at local (α) and regional (γ) scales no relationship was detected. Yet, this was not the case when considering insect (α) diversity, which was positively associated with abiotic variability in peat ditches. Insect species richness was generally low in all soil regions and these findings may suggest that abiotic variability is actually promoting the diversity of invertebrate communities in North Holland’s peatland drainage ditches. To investigate the possible mechanisms underlying the positive relationship with insect α diversity we look at the relationship between insect life-history strategies and abiotic conditions in the three soil regions. The two most prevalent insect life-history strategies (LHS) found in this study were those associated with dispersal (D1, D2 and D3) and development of trade-offs to invest in adaptations to harsh environmental conditions (T1). Dominance of these strategies suggests that good dispersal

Chapter 5

113

ability and tolerance are necessary traits for invertebrates living in nutrient rich agricultural ditches. T1 species are better equipped to deal with long-term abiotic disturbance induced by eutrophication (e.g. periodic anoxia) since these species invest in traits like increased haemoglobin production and saprobic tolerance (Verberk et al. 2008a,b). In this way they are able to deal with harsh abiotic conditions and can persist where other species cannot. The percentage of D3 (which encompassed only 4 species) was higher than the percentage of D1 (encompassing 17 species). Notably, D3 was dominated by a single genus of Heteropteran, Sigara. Adult Sigara are active fliers and can evacuate sites which have unfavourable conditions in search of new habitats (Savage 1989; Nilsson 2005). Consequently, a reduction in the relative abundance of D3 taxa could indicate emigration of Sigara adults, although gravid females may lose condition in their wing muscles, preventing them from undertaking aerial evacuation. Several of the more abundant D1 species, namely Corixa sp. and Hesperocorixa sahlbergi, occupy similar niches to Sigara. These three taxa predominantly feed on detrital material, microalgae and microinvertebrates and have preference for habitats comprised of macrophytes and detrital material (see Tachet et al. 2002). Thus, abiotic variability may promote a short term increase in local diversity by reducing competition (Huston 1994) and creating habitat space for other species. Yet it must not be overlooked that no relationship was found between regional diversity (γ diversity) and abiotic variability. Environmental variability can both stimulate and supress diversity by promoting temporal niche partitioning or by excluding species and causing stochastic extinctions (Shurin et al. 2010). In the case of North Holland, abiotic variability predominately arises from fluctuations in macro-ions and appears to stimulate species turn-over but concomitantly supresses taxa with particular LHS, namely D3 and T1. Although informative, field studies such as described here, are based on observations and remain descriptive. Additional controlled and long-term experiments are needed to move towards unravelling the underlying mechanisms of species responses to abiotic variability and further improve the state of knowledge in this field. In his seminal research on invertebrates in North Holland, van der Hammen (1992) analysed macroinvertebrate, abiotic and macrophyte data collected between 1979 and 1985, and classified the province (including the districts south of Amsterdam) into eight hydro-biological regions. These regions predominantly related to patterns of invertebrate taxonomic community composition, geomorphological features and chloride concentrations (van der Hammen 1992). In the present study we covered fragments of three hydro-biological regions, namely the dune region (i.e. sand), the dune fringe region (transition from sand to clay) and the polder region (clay and peat). The polder region accounts for the largest land area. Historically the polder region was the richest in nutrients (average TP and KN; 1.4 mg L-1 and

Chapter 5

114

4.8 mg L-1), while the dune fringe region was intermediate (1.2 mg L-1 and 3.6 mg L-1) and the dune region had the lowest concentrations of nutrients (0.24 mg L-1 and 1.9 mg L-1) (van der Hammen 1992). However data collected between 2008 and 2011 indicated that sand ditches have similar concentrations of nitrogen (KN, NH4+, NO3- and NO2-) as clay and peat ditches and the highest concentrations and variability in phosphates (PO43- and TP). Nowadays, the three regions appear to be more similar in nutrient concentrations than they were some 20 years ago. The implications of this for invertebrate diversity may include the loss of relatively stable habitats in sand ditches which could contribute to a loss of diversity in the province, a trend already reported by for other taxa in agricultural landscapes (Kleijn et al. 2004; Donald et al. 2006). In conclusion this study shows that abiotic variability, particularly of macro-ions, is playing a significant role in structuring macroinvertebrate diversity in North Holland’s drainage ditches. Furthermore, correlations between insect life-history strategies and abiotic conditions in the different soil regions illustrates that the response of different species is dependent upon their functional attributes and suggests that the balance in competition between groups for habitat and other resources is influenced by abiotic variability. Variability in macro-ions most likely arises from the inlet of mineral rich water, while high average nutrient conditions are associated with intensive land use. These findings have implications for agricultural landscapes outside North Holland as agricultural intensification is forecast to increase globally in the coming decades. Under these circumstances abiotic variability and its associated impacts on biota is likely to increase within aquatic environments. Acknowledgements We thank Gert van Ee, Emile Nat, Ron van Leuken and Herman van Dam for their help throughout this project, Nigel Upchurch for the map, Piet Verdonschot for early discussions on this paper and Wilco Verberk and Patrick Armitage for their comments. This research was funded by Stichting Waterproef and Hoogheemraadschap Holland’s Noorderkwartier. .

Chap

ter 5

115

Appe

ndix

5.1.

Pear

son co

rrelat

ion co

effici

ents

(r) b

etween

aver

age a

biotic

cond

ition

s and

the f

irst a

nd se

cond c

ompo

nent

s of P

rincip

al Co

mpon

ent A

nalys

es (P

CA) r

un w

ith a

verag

e valu

es (P

CAA),

and b

etween

abio

tic va

riabil

itya an

d the

first

and

secon

d axe

s of t

he P

CA ru

n wi

th a

biotic

varia

bility

(PCA

V).

Si

gnifi

cant

corre

lation

s are

indica

ted a

s goo

d (P

< 0

.01,

bold

) and

mod

erate

(P <

0.0

5, it

alic).

Sign

ifica

nce l

evels

were

adju

sting

follo

wing

the B

onfer

roni

-Holm

cor

rectio

n me

thod

.

Ave

rage

V

ariat

ion

(MA

D)

Sand

Clay

Peat

Sa

nd

Cl

ay

Pe

at

N

umbe

r of

ditc

hes

17

43

24

17

43

24

PC

AA1

PCA

A2

PCA

A1

PCA

A2

PCA

A1

PCA

A2

PC

AV1

PCA

V2

PCA

V1

PCA

V2

PCA

V1

PCA

V2

Var

iance

ex

plain

ed

85

%

7%

57%

21%

40%

37%

44

%

22

%

39

%

17

%

47

%

20

%

Mac

ro-io

ns

Cl

- (m

g L-

1 ) 0.

89

0.14

0.

83

-0.4

2 0.

40

0.86

0.83

0.

35

-0.4

4 0.

54

0.89

0.

16

Tota

l Fe

(μ

g L-

1 ) -0

.75

0.53

0.

75

0.53

0.

71

-0.4

7

-0.3

3 0.

75

0.27

0.

67

-0.2

9 0.

66

HCO

3- (m

g L-

1 ) 0.

91

0.15

0.

77

0.17

0.

41

0.24

0.63

-0

.11

0.75

-0

.04

0.08

0.

09

SO42-

(mg

L-1 )

0.94

0.

06

0.29

0.

61

0.11

0.

24

0.

34

-0.4

9 0.

60

0.43

0.

58

0.25

Ca

2+ (m

g L-

1 ) 0.

92

0.19

0.

34

0.48

-0

.15

0.38

0.44

-0

.03

0.89

-0

.17

0.21

0.

07

K+ (m

g L-

1 ) 0.

96

0.17

0.

89

-0.3

1 0.

61

0.69

0.67

0.

27

-0.0

4 0.

58

0.85

0.

12

Mg2

+ (m

g L-

1 ) 0.

93

0.14

0.

91

0.02

0.

59

0.74

0.76

-0

.35

0.12

0.

39

0.85

0.

08

Na+

(mg

L-1 )

0.

93

0.12

0.

87

-0.4

0 0.

46

0.85

0.80

-0

.07

-0.5

9 0.

51

0.91

0.

15

a Calcu

lated

as t

he M

edian

Abs

olute

Devi

ation

(MA

D) f

or in

dividu

al ab

iotic

para

meter

s

Chap

ter 5

116

Appe

ndix

5.1.

contin

ued.

A

vera

ge

V

ariat

ion

(MA

D)

Sa

nd

Clay

Pe

at

Sa

nd

Clay

Pe

at

Num

ber o

f di

tche

s 17

43

24

17

43

24

PC

AA1

PCA

A2

PCA

A1

PCA

A2

PCA

A1

PCA

A2

PC

AV1

PCA

V2

PCA

V1

PCA

V2

PCA

V1

PCA

V2

Var

iance

ex

plain

ed

85%

7%

57

%

21%

40

%

37%

44%

22

%

39%

17

%

47%

20

%

Nut

rient

s

NH

4+

(mg

N L

-1)

0.53

0.

22

0.30

-0

.20

0.82

-0

.22

0.

38

-0.1

6 -0

.23

-0.0

6 0.

06

0.43

KN

(mg

N L

-1)

0.85

0.

20

0.63

0.

09

0.82

-0

.10

0.

57

0.07

0.

37

0.04

-0

.16

0.63

N

O2-

(mg

N L

-1)

0.56

0.

09

0.45

-0

.08

0.68

0.

05

0.

53

-0.1

1 -0

.23

-0.0

2 0.

17

0.38

NO

3- (m

g N

L-1

) 0.

38

-0.1

2 0.

33

-0.1

7 0.

45

0.38

0.38

-0

.11

-0.0

4 0.

28

0.32

0.

12

TP (m

g P

L-1 )

0.73

0.

18

0.57

0.

01

0.58

-0

.17

0.

42

-0.2

0 0.

17

0.10

-0

.08

0.38

PO

43-

(mg

P L-

1 ) 0.

70

0.16

0.

53

-0.1

7 0.

46

-0.1

7

0.34

-0

.24

0.02

0.

01

-0.1

5 0.

33

Field

mea

surem

ents

pH

0.

88

-0.0

6 0.

40

-0.3

8 -0

.38

0.43

0.27

0.

08

0.17

0.

21

-0.1

2 -0

.34

Tran

spar

ency

(c

m)

0.63

-0

.42

-0.6

1 -0

.47

-0.2

4 0.

45

0.

49

-0.2

3 0.

26

-0.0

5 -0

.02

0.18

Chap

ter 5

117

Appe

ndix

5.2.

Lin

ear r

egress

ion ou

tputs

in ea

ch so

il typ

e test

ing p

attern

s in

invert

ebra

te α

divers

ity a

nd in

sect l

ife-h

istory

stra

tegy c

ompo

sition

aga

inst

three

mo

dels,

(I) t

he co

mbin

ation

of a

biotic

varia

bility

(PCA

V) a

nd a

verag

e con

dition

s (PC

AA),

(II) a

biotic

varia

bility

, and

(III)

aver

age c

ondit

ions.

Aka

ike

infor

matio

n cri

teria

contro

lled f

or sm

all sa

mple

size (

AIC

c) an

d A

kaik

e weig

hts (

w i) a

re sh

own

along

side P

value

s, ad

justed

with

the s

eque

ntial

Bon

ferro

ni-H

olm

correc

tion

meth

od, n

.s. =

not

signif

icant

. Ta

ble S

1. Sa

nd d

itche

s

Sand

n

= 1

7 A

vera

ge

Var

iatio

n (M

AD

)

PCA

A1

PCA

V1

PCA

V2

df

Log

likeli

hood

A

ICc

w i

P α

Mod

el I

-0.5

1 -2

.76

4

-60.

69

132.

7 0.

08

n.s.

inve

rtebr

ates

M

odel

II

-3

.11

3

-60.

69

129.

2 0.

48

n.s.

M

odel

III

-3.2

2

3

-60.

77

129.

4 0.

44

n.s.

α M

odel

I -2

.18

-5

.9

4 -7

0.15

15

0.4

0.19

n.s

. In

sect

s M

odel

II

-4.7

4 3

-70.

31

147.

8 0.

7 n.s

.

Mod

el II

I 3.

44

3 -7

2.08

15

1.4

0.11

n.s

. D

1 M

odel

I -1

.11

0.39

4 -8

.68

28.7

0.

23

n.s.

M

odel

II

-0

.36

3

-11.

8 31

.4

0.06

n.s

.

Mod

el II

I -0

.72

3 -9

.33

26.5

0.

7 0.

01

D2

Mod

el I

-0.1

1 0.

3

4 -7

.27

25.9

0.

1 n.s

.

Mod

el II

0.22

3 -7

.32

22.5

0.

54

n.s.

M

odel

III

0.18

3

-7.7

3 23

.3

0.36

n.s

. D

3 M

odel

I 0.

17

0.39

4 -1

4.3

39.9

0.

09

n.s.

M

odel

II

0.

5

3 -1

4.34

36

.5

0.52

n.s

.

Mod

el II

I 0.

55

3 -1

4.65

37

.1

0.39

n.s

.

Chap

ter 5

118

Appe

ndix

5.2.

Tabl

e S1.

Sand

ditc

hes c

ontin

ued.

Sand

n

= 1

7 A

vera

ge

Var

iatio

n (M

AD

)

PCA

A1

PCA

V1

PCA

V2

df

Log

likeli

hood

A

ICc

w i

P S

Mod

el I

0.14

0.48

4

-10.

75

32.8

0.

15

n.s.

M

odel

II

0.44

3

-10.

88

29.6

0.

73

n.s.

M

odel

III

-0.0

6

3

-12.

66

33.2

0.

12

n.s.

T1

Mod

el I

-3.6

8

-36.

64

4 -6

9.9

151.

1 0.

16

n.s.

M

odel

II

-35.

58

3 -6

9.98

14

7.8

0.84

<

0.00

1

Mod

el II

I 11

.79

3 -7

7.26

16

2.4

0 n.s

. R1

M

odel

I -0

.89

0.35

4 -1

1.11

33

.6

0.17

n.s

.

Mod

el II

0.

25

3 -1

2.76

33

.4

0.19

n.s

.

Mod

el II

I -0

.54

3 -1

1.52

30

.9

0.65

n.s

.

Chap

ter 5

119

Appe

ndix

5.2.

Tabl

e S2.

Clay

ditch

es

Clay

n

= 43

A

vera

ge

Var

iatio

n (M

AD

)

PCA

A-1

PC

AA-2

PC

AV-1

PA

C V-2

df

Lo

g lik

eliho

od

AIC

c w i

P

α M

odel

I -6

.034

4.

946

4 -1

39.5

21

288.

1 0.

34

n.s.

inve

rtebr

ates

M

odel

II

2.

953

3 -1

40.6

02

287.

8 0.

39

n.s.

M

odel

III

-2.7

6

3 -1

40.9

64

288.

5 0.

27

n.s.

α M

odel

I -4

.92

3.24

8 4

-115

.751

24

0.6

0.51

n.s

. In

sect

s M

odel

II

1.

622

3 -1

17.8

71

242.

4 0.

21

n.s.

M

odel

III

-2.7

7

3 -1

17.6

13

241.

8 0.

28

n.s.

D1

Mod

el I

0.

2854

0.

5602

4 -2

2.04

7 53

.1

0.35

n.s

.

Mod

el II

0.

4655

3 -2

2.62

3 51

.9

0.63

0.

03

M

odel

III

0.20

84

3

-25.

867

58.3

0.

03

n.s.

D2

Mod

el I

-1.0

29

0.59

07

0.

2065

5

-30.

775

73.2

0.

48

n.s.

M

odel

II

-0.4

379

3

-35.

469

77.6

0.

05

n.s.

M

odel

III

-0.8

919

3

-33.

34

73.3

0.

46

0.01

D

3 M

odel

I 0.

5343

-0.6

927

4

-35.

466

80

0.27

n.s

.

Mod

el II

-0

.541

7 0.

4608

4

-34.

531

78.1

0.

69

n.s.

M

odel

III

0.

3984

3

-38.

534

83.7

0.

04

n.s.

Chap

ter 5

120

Appe

ndix

5.2.

Tabl

e S2.

Clay

ditch

es con

tinue

d.

Clay

n

= 43

A

vera

ge

Var

iatio

n (M

AD

)

PCA

A-1

PC

AA-2

PC

AV-1

PA

C V-2

df

Lo

g lik

eliho

od

AIC

c w i

P

S M

odel

I

0.59

34

0.19

05

4

-24.

529

58.1

0.

29

n.s.

M

odel

II

-0.0

0623

3

3 -2

6.67

60

0.

11

n.s.

M

odel

III

0.

4668

3

-24.

989

56.6

0.

60

n.s.

T1

Mod

el I

0.43

9 -0

.349

4

-0.1

266

5 -0

.25

12.1

0.

73

n.s.

M

odel

II

0.14

16

3

-5.6

8 18

0.

04

n.s.

M

odel

III

0.35

52

3

-3.9

2 14

.4

0.23

0.

03

R1

Mod

el I

-1.1

17

-0.4

389

0.

2345

5

-21.

666

55

0.44

n.s

.

Mod

el II

-0

.289

7

3 -2

9.51

4 65

.6

0 n.s

.

Mod

el II

I -0

.961

5

3 -2

3.92

9 54

.5

0.56

<

0.0

01

Chap

ter 5

121

Appe

ndix

5.2.

Tabl

e S3.

Peat

ditch

es.

Peat

n

= 2

4 A

vera

ge

Var

iatio

n (M

AD

)

PCA

A-1

PC

AA-2

PC

AV-1

PC

AV-2

df

Lo

g lik

eliho

od

AIC

c w i

P

α M

odel

I

-3.9

6

6.21

4

-86.

4 18

2.9

0.17

n.s

. in

verte

brat

es

Mod

el II

6.7

3 -8

6.66

18

0.5

0.58

n.s

.

Mod

el II

I

-5.3

5

3

-87.

52

182.

2 0.

25

n.s.

α M

odel

I

-15.

31

10.2

3

4 -6

9.32

14

8.7

0.35

n.s

. In

sect

s M

odel

II

4.

74

3 -7

0.31

14

7.8

0.55

0.

04

M

odel

III

3.44

3 -7

2.08

15

1.4

0.09

n.s

. D

1 M

odel

I 0.

51

0.4

4 -9

.34

28.8

0.

32

n.s.

M

odel

II

0.

67

3 -1

0.64

28

.5

0.37

0.

005

M

odel

III

0.89

3 -1

0.81

28

.8

0.32

n.s

. D

2 M

odel

I -0

.39

0.

18

4

-17.

41

44.9

0.

17

n.s.

M

odel

II

0.14

3 -1

8.18

43

.6

0.32

n.s

.

Mod

el II

I -0

.35

3

-17.

73

42.7

0.

51

n.s.

D3

Mod

el I

0.89

-0.7

9

4 -1

6.37

42

.8

0.56

0.

008

M

odel

II

-0.6

8

3 -2

0.3

47.8

0.

05

0.04

Mod

el II

I 0.

73

-1.0

2

4

-16.

69

43.5

0.

39

0.00

9

Chap

ter 5

122

Appe

ndix

5.2.

Tabl

e S3.

Peat

ditch

es con

tinue

d. Pe

at

n =

24

Ave

rage

V

aria

tion

(MA

D)

PC

AA-1

PC

AA-2

PC

AV-1

PC

AV-2

df

Lo

g lik

eliho

od

AIC

c w i

P

S M

odel

I -0

.74

0.66

0.48

5

-14.

78

42.9

0.

15

n.s.

M

odel

II

0.32

3 -1

7.16

41

.5

0.29

n.s

.

Mod

el II

I

0.55

3

-16.

5 40

.2

0.56

n.s

. T1

M

odel

I

0.49

-0

.25

4

6.88

-3

.7

0.26

n.s

.

Mod

el II

-0.1

4 3

5.65

-4

.1

0.32

n.s

.

Mod

el II

I -0

.2

0.2

4 7.

42

-4.7

0.

43

n.s.

R1

Mod

el I

-0

.8

0.62

4 -1

2.37

34

.8

0.25

n.s

.

Mod

el II

0.15

3

-13.

38

34

0.37

n.s

.

Mod

el II

I 0.

21

3

-13.

35

33.9

0.

39

n.s.

Chapter 5

123

Chapter 5

124

Gyraulus crista

Chapter 5

125

Chapter 6

Synthesis

Chapter 6

126

The evidence presented in this thesis illustrates the potential consequences of intensive agricultural management on aquatic ecosystems. North Holland’s agricultural ditches are ecologically degraded and this is reflected by the invertebrate, macrophyte and microbial communities. Moreover, there has been a substantial decline in the number of species supported in these ditches over recent decades. At the beginning of this thesis I compared the ecosystems of intensively managed agricultural wetlands to a house of cards and posed two questions:

1. What are the key environmental drivers which structure the house of cards, and

2. Can the stability of the structure be improved in order to keep it standing?

I will now review the evidence provided in this thesis in view of these two questions. Habitat fragmentation and environmental instability in agricultural landscapes The degradation of habitats in agricultural landscapes, brought about by habitat fragmentation, reduced habitat size and environmental instability poses major threats to biodiversity (Sala et al. 2000; Donald & Evans 2006; Hendrickx et al. 2007; Griffen & Drake 2008). Fragmentation and regular disturbance events have repercussions for population dynamics, breeding success, individual fitness and ecosystem functioning (Laurance 2002, 2008; Fahrig 2003; Donald & Evans 2006; Hendrickx et al. 2007). This thesis shows how aquatic invertebrates in North Holland’s drainage ditches are structured by these factors. Patterns of macrophyte growth provide patches of desirable habitat, amongst open bare sediment, while environmental conditions in drainage ditches exhibit temporal patchiness with individual ditches having high nutrient concentrations and fluxes in micro-ions, thus the two dominant environmental filters acting on aquatic invertebrates in the province are: 1) a decline in macrophyte habitat (emergent and submerged vegetation) and 2) widespread eutrophication in combination with abiotic instability (Chapters 2, 3 and 5). Invertebrates in North Holland’s drainage ditches were strongly related to the presence of vegetation (Chapters 2 and 3). Moreover, ditches which contained submerged macrophytes in conjunction with high emergent vegetation coverage were distinctive in the microbial and invertebrate communities they supported (Chapter 4). Being linear water bodies the main structural habitat in ditches is associated with bank slope and vegetation (Higler & Verdonschot 1989; Verdonschot et al. 2012a). In general, North Holland’s agricultural ditches are U shaped in cross-section with straight banks and edges. In comparison, organically rich peatland soils provide soft banks, with some having gentle slopes that provide larger areas of habitat. Furthermore, under

Chapter 6

127

the right abiotic conditions they can support diverse macrophyte communities (Lamers et al. 2002; Armitage et al. 2003; Herzon & Helenius 2008). The habitats provided by drainage ditches, particularly in peatland environments, can potentially support a diverse invertebrate community, including rare species (Painter 1999; Williams et al. 2003; Nijboer & Verdonschot 2004; Verdonschot et al. 2011). In the Netherlands drainage ditches are cleared of vegetation on an annual basis, this creates a high level of disturbance which is tolerated only by fast growing, pioneer species like Phragmites australis and Typha angustifolia which can recover quickly. Submerged vegetation is negatively impacted by regular cutting and vegetation clearance is contributing to plant communities in ditches becoming dominated by algae and floating macrophytes (van Zuidam & Peeters 2012). Painter (1999) reported that the number of rare invertebrate species increased with the quantity of detritus, bank slope, the age of the ditch, macrophyte coverage and the time since the ditch was last cleaned. In this study Painter found that ditches cleared every 4 years along alternating 50 m stretches supported the most diverse plant communities. Thus the annual clearance of vegetation seen in Dutch drainage ditches is contributing to the loss of macrophyte species by excluding those unable to recover from regular cutting and ultimately causing the homogenization of macrophyte habitats. Moreover, vegetation removal may result in the loss of invertebrate eggs and larvae collected and removed along with plant material, causing a reduction in the recruitment of invertebrates. For example, Twisk et al. (2000) found that the presence of caddisfly larvae was related to the frequency, timing and machinery used to dredge sediment and clear vegetation. The second environmental filter affecting aquatic invertebrates in the province is wide-spread eutrophication in combination with abiotic variability. The most eutrophic drainage ditches are dominated by a few invertebrate taxa (Chironomidae larvae, Gammaridae and Gastropods) which have invested in traits that allow them to tolerate harsh abiotic concentrations, such as air breathing, haemoglobin production and resistance to sulphide and nitrate toxicity (Chapters 3 and 5). The pressures of eutrophication are also reflected by epiphytic diatom communities in the province, with a reduction in species-turnover evident under increasing eutrophication in peatland environments (Goldenberg et al. 2014). Because agricultural drainage ditches are hydrologically managed, with mineral rich waters let-in during dry periods, the stability of abiotic conditions is related to hydrological connectivity. In chapter 3 we see evidence of this in the conditions of a semi-isolated ditch which contained a rich invertebrate fauna, submerged vegetation and lower concentrations of macro-ions and nutrients in comparison to the surrounding ditches (Chapter 3). This ditch was located in the middle of the Wormer and Jisperveld, a degraded peatland, where the majority of the ditches contain no submerged vegetation at all. It would be of great interest to sample more of these isolated and semi-

Chapter 6

128

isolated ditches in the province, to determine if these could be supporting source populations from which recruitment takes place, to recolonize ditches which experience regular abiotic disturbance events. If so, hydrologically isolated ditches could play an integral role in invertebrate population dynamics, as the loss of species from a ditch may only be transient if dispersal is not a limiting factor and the regional pool of species is large enough (Giller et al. 2004). Active dispersal is generally rare in aquatic invertebrates. This is likely to be related to the reproductive trade-offs commonly associated with flight (Bilton et al. 2001). Alternatively, many species employ passive dispersal methods such as wind dispersal, phoresy or crawling (Bilton et al. 2001).Yet, in North Holland there is a strong presence of insects exhibiting good dispersal abilities (Chapters 3 and 5). This observation, in combination with the high species-turnover between individual ditches (Chapter 5), suggests that species community composition is in a state of perpetual turn-over, with species occurrence in any given ditch dependent upon the stability and habitat quality at that time. Mechanisms underlying aquatic invertebrate distribution Spatial and temporal features in the environment may be considered as a template upon which biodiversity is determined from a selection of species with the best adaptations (Stearns 1976; Townsend & Hildrew 1994). As a species life-history strategy (LHS) reflects its evolutionary response to environmental conditions, so can analysis of LHS composition reveal how the environment is being experienced by the invertebrate community (Stearns 1976; Verberk et al. 2008a b, 2013). The expression of insect LHS in the province of North Holland reflects the role of strong dispersal abilities and tolerance to eutrophication and abiotic disturbance in determining a species chance of survival and reproduction (Chapters 3 and 5). In agricultural landscapes biodiversity is often impacted by multiple stressors which act as the key environmental filters in this context. For example, eutrophication, increased water abstraction and annual vegetation removal are common stressors to aquatic communities, which occur together in agricultural systems. These act simultaneously on an organism causing either a synergistic effect (i.e. when the combined effects is larger) or an antagonistic effect (i.e. when the cumulative impact is smaller) (Folt et al. 1999). Most crucial to a community’s ability to deal with multiple stressors is the existence of species co-tolerances, that is – the alignment of a species tolerance to multiple stressors which occur together in the environment (Vinebrooke et al. 2004). The most abundant species collected in North Holland ditches were Chironomidae larvae (specifically, Chironomus and Glyptotendipes spp.) which persisted under high concentrations of nutrients and macro-ions. Moreover, their relative

Chapter 6

129

abundance increased under harsh conditions as a result of other species declining in number. This illustrates the species adaptation to this harsh environment, possibly as a consequence of having co-tolerances for eutrophication and high concentrations of micro-ions, such as bicarbonates and sulphates (Chapter 5). Chapter 2 covers the serious decline in submerged macrophyte species richness shown in peatlands in recent decades. Decline of macrophyte communities in these environments is related to sediment bound P concentrations, disturbance from regular vegetation cutting, low sediment stability and lack of seed stock (Barko et al. 1991; van Zuidam & Peeters 2012; van Zuidam et al. 2012). Indeed North Holland’s remnant peatlands have particularly high sediment P concentrations, unstable sediments and are consequently dredged regularly. In addition, emergent vegetation is cut back every year, often to ground level. The decline of these plant communities signifies the overall health of the ditch ecosystem and the combination of eutrophication, and regular disturbance from abiotic variability and vegetation removal are clearly underlying both invertebrate and microbial community structure (Chapters 2, 4 and 5). In this thesis several environmental filters are identified, namely loss of macrophyte habitat (Chapters 2, 3 and 4), widespread eutrophication (Chapters 3 and 4), variation in abiotic conditions (Chapter 5) and habitat fragmentation. There are few species who appear to do well under such conditions and the lower the number of species the greater the likelihood that ecosystem functioning will be affected (Loreau 1998; Hooper et al. 2005). The likely implications of biodiversity loss on the functioning of agricultural ecosystems will now be discussed. The potential consequences of biodiversity loss in North Holland’s drainage ditches Biodiversity is declining at an alarming rate globally and land use intensification is considered to be the main contributor to past and predicted future biodiversity loss, particularly in freshwater environments (Sala et al. 2000; van Eerden et al. 2010). Over recent decades there has been a dramatic decline in macroinvertebrate diversity in North Holland’s agricultural peatlands, which appears to be coupled with the decline in submerged macrophyte diversity and associated habitat loss (Chapter 2). These peatlands represent a habitat which was once dominant in the region, supporting a diversity of plant and animal life. Yet the wetlands have been heavily exploited and land use intensity has increased dramatically over recently decades (Kleijn et al. 2004; Busch 2006; Donald & Evans 2006; van Eerden et al. 2010). The response of a community to a changing environment is subject to a variety of interactions such as co-tolerances to multiple stressors, trophic interactions and the sequence and identity of species loss (Chase 2003; Petchey

Chapter 6

130

et al. 2004; Vinebrooke et al. 2004). Moreover, fundamental ecosystem functions can be affected. Cardinale et al. (2002) provide evidence of how a reduction in biodiversity can influences functions like organic matter processing through interspecific facilitation. In addition, regional processes like dispersal and disturbance can influence the degree to which biodiversity affects ecosystem functioning (Cardinale et al. 2004). In truth, it’s very hard to ascertain what the overall consequences of biodiversity loss on ecosystem function will be because the intricacies of species-environment interactions often result in case specific conditions (Gessner et al. 2004). Yet, what is certain is that if biodiversity in agricultural landscapes continues to decline, ecosystem processes, like organic matter processing, nutrient cycling and the breakdown of contaminants will be affected (Hooper et al. 2005; Balvanera et al. 2006). Fundamentally, in the case of North Holland the reduction in macrophyte growth has implications for water quality as loss of plants reduces sediment stability and nutrient dynamics (Barko et al. 1991; Palmer et al. 2000; Horppila & Nurminen 2001). The stabilization of macrophyte communities could then serve to ameliorate the negative impacts associated with increased land use intensity. But perhaps one of the most important implications of biodiversity loss is that a reduction in diversity is associated with a decline in overall ecosystem stability and resilience to environmental disturbances (Pimm 1984; Folke et al. 2004; Balvanera et al. 2006). Increased frequency and intensity of disturbances are anticipated as a result of land use intensification and climate change (Sala et al. 2000). Ultimately, if too much biodiversity is lost from North Holland there will be a greater chance of further extinctions with potential consequences for key ecosystem process. Considerations for the management of agricultural wetlands There is an evident decline in the biodiversity of agricultural landscapes seen in many terrestrial and aquatic species and this is largely associated with increased land use intensity (Hendrickx et al. 2007; van Eerden et al. 2010) (Chapter 2). Invertebrates can be considered as the ‘canaries’ of freshwater systems, due to their intermediate trophic position, largely resident life-style and relatively fast life-cycles. Thus invertebrates provide information on ecosystem health that is not perceivable from abiotic parameters and can indicate changes in the ecosystem much faster than vegetation. The European Water Framework Directive (2000/60/EC; WFD) principally seeks to 1) define and set environmental objectives for the improvement of chemical and ecological conditions in fresh and coastal waters, and 2) develop methods to implement and achieve these objectives (Junier & Mostert 2012). The review of ecological conditions of North Holland’s waterways that has come about through the WFD has revealed that current

Chapter 6

131

nutrient levels, particularly phosphorus, in waterways pose a major obstacle to achieving these goals. Van der Bolt et al. (2003) came to the striking conclusion that agricultural production in The Netherlands would need to be reduced by two-thirds to achieve the goals of the WFD. Certainly the current rates of manure application are leading to an increase in soil phosphorus levels, yet there appear to be no plans to take measures to reduce agricultural nutrient inputs (Ligtvoet et al. 2008; Junier & Mostert 2012). Alternatively, to achieve higher ecological quality scores the focus has been on design measures to restore structural habitat such as the development of nature friendly banks (natuurvriendelijke oevers, in Dutch). Agri-environment schemes (AES), another measure which focuses on the creation of habitat, have been introduced by the EU (under EU regulation 1257/1997) and some North American states in response to the observed decline in diversity in birds, terrestrial invertebrates and aquatic invertebrates. In fact over 1.6 billion euros is estimated to be spent annually on AES in Europe alone (Donald & Evans 2006). Yet, there appears to be a lack of reliable data available to determine the effectiveness of AES, while mixed reports on their apparent success likely arises from the ad-hoc manner in which they are often implemented (Kleijn et al. 2004). An important factor is likely to be the size of the habitats created by AES, as many protected areas consist of small marginal strips of land. This is likely to be why they have been reported to benefit insects but not necessarily birds (often the target group), which require larger habitats and so regularly leave AES areas (Kleijn et al. 2004). Thus, for management strategies to successfully increase biodiversity in agricultural landscapes the requirements of habitat size and quality for biota must be met. Moreover, it is of fundamental importance that the location of existing habitats and the dispersal abilities of species be considered when designing new habitat, for example nature friendly banks (natuurvriendelijke oevers) should be accessible from established habitats. Moreover, if we are to get the most out of such measures sound monitoring data is essential to determine where measures have been successful and where they have not. Without adequate monitoring data there is no certainty on the effectiveness of habitat improvement schemes, nor is it possible to adjust measures to achieve the most desirable results. Monitoring Some of the most frequently asked questions by water managers to aquatic scientists include, what should be measured? Where and how many samples should be taken? And how often should these samples be taken? Although, the practical relevance of these questions is clear, when it comes to monitoring there is no one index which can serve as the Holy Grail or any magic number of samples or frequency of sampling which will provide the answers. One of the most important aspects of designing a monitoring programme relates to the

Chapter 6

132

question(s) that are being asked or what problems need to be solved? For example, water managers are often interested in determining the long-term biodiversity trends in their district. In this case consistency in the monitoring locations and the methods used when collecting data are essential. Naturally, over time methods may change as new information and techniques come to light. This is not necessarily a problem as long as any changes are clearly documented, and ideally the new methods should produce results which can be compared to the original methods. Likewise, the number of monitoring locations may change over time. Again this is not an insurmountable problem, as long as the consistency between historical and contemporary locations are taken into consideration. For these reasons it is highly advisable that water managers seek scientific advice when adjusting their monitoring plans, for receiving good advice should save losses in valuable data and time. In addition to standard monitoring questions circumstances arise when water managers need additional information, for example to determine the effectiveness of habitat improvement measures such as the establishment of nature friendly banks. In this case sampling will need to be more intensively focussed. And in addition it may be advisable to alter the sampling methods, say by sampling different habitats separately or using different sampling methods, such as activity traps (Verdonschot 2010). Moreover, additional information on the response of specific taxa such as mayflies (Ephemeroptera), dragonflies (Odonata) and caddisflies (Trichoptera) (i.e. EOT species), invertebrate life-history strategies or microbial community structure can provide valuable insight into the processes regulating invertebrate community composition (Chapters 3, 4 and 5). Although there are arguments against using trait-based approaches, such as life-history strategies (LHS), these arguments are generally based on the premise that information is lacking and that a high level of expertise is required to correctly assign traits. Yet, freshwater invertebrates are one of the few groups where traits have been described for thousands of species and they have been very well researched (see, Usseglio-Polatera et al. 2000 and Bonada et al. 2006 and references therein). Furthermore, the approach of combining traits into LHS or other relevant categories (i.e. to determine community sensitivity to a particular stressor) has provided promising results (Chapters 3 and 5) (see also Verberk et al. 2008a; b, 2010, 2013 and references therein). Because functional composition is not strictly associated with species composition it allows for the comparison of geographically separate populations (Chapter 3). This offers the potential for information relating to environmental drivers of agricultural communities to be shared between different regions, countries or even continents. Secondly, the premise that a high level of taxonomic expertise is required for trait-based approaches is not really an obstacle when considering that a certain level of expertise is already necessary to identify invertebrate

Chapter 6

133

species. Moreover, there is a high level of trait detail available in existing trait databases for aquatic invertebrates. However, care is certainly needed when selecting the sets of traits to analyse and determining which are most meaningful to a given situation. For example, species sensitivities to specific pesticides may be used if the impact of pesticide application rates are of interest. Such investigatory analysis is integral in determining the impact of further land use intensification and habitat loss on aquatic biota. Moreover, the interconnectedness of aquatic communities necessitates an understanding of how stressors influence different fractions of the community e.g. vegetation, microbial communities, diatoms, invertebrates, fish and birds. Thus conclusions on the ecological consequences of land use intensity should be based on multiple components of the biological community to facilitate the development of successful management strategies. Habitat restoration Invertebrates in drainage ditches are intrinsically linked to the availability of plant habitat (Chapters 2 and 3). Many ditches in the province contain little in the way of submerged vegetation and are often dominated by free floating species, such as duckweed (Lemna sp.) or the aquatic fern (Azolla sp.). Van Zuidam and Peeters (2012) demonstrated that cutting negatively impacted two common species of submerged macrophyte (Potamogeton lucens and P. compressus) and showed that regular cutting stunted below ground growth of plants. Furthermore, the closer to the sediment layer plants were cut and the frequency of cutting increased the incidence of plant mortality (van Zuidam & Peeters 2012). Clearly taking measures to reduce the frequency of vegetation cutting would help restore submerged vegetation and overall biodiversity. This could be approached in different ways, for example in ditches where clearance is necessary vegetation could be cut at varying the depths, in ditches where the clearance is not so important removal could be rotating to occur every second year and most significantly some ditches could be left unmanaged to undergo a process of natural terrestrialisation. Vegetation removal is clearly affecting plant communities in the province but other factors are important too. Eutrophication and the inlet of mineral rich riverine waters are known to alter abiotic conditions in polders, this is particularly true in the case of peat environments (Roelofs 1991; Lamers et al. 2002; Smolders et al. 2006; Geurts et al. 2009). In agricultural wetlands peat soils are rapidly degraded, resulting not just in loss of habitat but also creating an expensive management problem, as the waterways require regular dredging to remove accumulated organic sludge. While the effects of eutrophication on invertebrates may be largely indirect (via changes in macrophyte habitat and epiphytic food quality), eutrophication can lead to harsh conditions, including periodic anoxia and the production of toxic compounds like sulphides. These

Chapter 6

134

conditions can exclude some species, thereby reducing invertebrate diversity (Chapters 3 and 4). Moreover, increasing the productive output of land requires a higher input of water, consequently more mineral rich river water is let into agricultural polders. Extraction of groundwater resources in the Netherlands has altered the hydrology dramatically; leading to the problem of desiccation (termed ‘verdroging’, in Dutch) which has been recognised as a serious environmental problem facing the country since the 1980’s (van Ek et al., 2000). While the influence of eutrophication on North Holland’s aquatic invertebrate community has been further exacerbated by abiotic conditions becoming more similar between soil type regions (Chapter 5), where once differences had existed (van der Hammen 1992). Wetland management Under the Ramsar convention a wetland is defined as“ areas of marsh, fen, peatland or water, whether natural or artificial, permanent or temporary, with water that is static or flowing, fresh, brackish or salt, including areas of marine water the depth of which at low tide does not exceed six meters” (http://ramsar.wetlands.org). Presently the peatlands of North Holland are managed under either low or high intensity agriculture and their water levels are kept artificially low, therefore they are not being managed as wetland environments. Consequently, there is a high rate of peat degradation occurring in these areas (Chapter 1). Recognizing the value of these wetland environments as remnants of what was once the dominant soil type in the province is a critical first step. To abate further degradation of the peat soil these areas need to be managed as wetland systems with flexible water tables, moving towards increased internalization of hydrology and decreasing the inlet of mineral rich river water. Such measures offer the best chance of reducing further peat soil degradation, stabilising the biological communities and increase the ability of these wetlands to perform valuable services including nutrient removal and water retention (Janssen et al. 2005). The value of hydrologically isolated ditches This thesis has reviewed some of the various environmental drivers responsible for structuring an unstable ‘house of cards’, drainage ditch ecosystem. The inevitable question then arises, what measures can be taken to stabilize this ecosystem? During the course of this research it became clear that the majority of monitoring sites in North Holland’s network of drainage ditches are relatively large, easily assessable waterways. These larger ditches have a high degree of hydrological connectivity and are more likely to experience abiotic variations as a result of the inlet of mineral rich river water. Thus much of the monitoring data is collected from disturbed sites in the system and the true regional diversity of North Holland may currently be underestimated. The

Chapter 6

135

habitat value of ditches which are either semi-isolated or completely isolated hydrologically must be recognized by water managers. These waters provide protected habitats which are more likely to have relatively stable abiotic conditions and a well-developed macrophyte community. Furthermore, the highest density of these ditches can be found in peat environments and different pockets of habitat will differ slightly in abiotic and vegetation characteristics and thus invertebrate species composition. If managed correctly, a network of small habitats can be protected which all together support a greater diversity of fauna, similar to what has been reported for ponds and small lakes (i.e. Williams et al. 2003; Scheffer & van Geest 2006). Semi-isolated and isolated ditches need to be protected and monitored more closely to determine the true pool of species in the region. This information will be invaluable in assessing the role of isolated habitats in providing source populations from which species emigrate and re-colonize ditches which experience regular abiotic and mechanical disturbances. Ultimately, the importance of vegetation and the inlet of mineral rich river water are reflected in both invertebrate and benthic microbial communities (Chapter 4). These are fundamental drivers of drainage ditch ecosystems, particularly in remnant peatlands, and must be managed accordingly. In addition, determining the processes of dispersal and immigration (i.e. meta-community dynamics) are fundamental to the effective management of aquatic biodiversity in North Holland’s agricultural landscape. Conclusions This thesis provides an overview of the interrelationships between different fractions of the drainage ditch ecosystem in North Holland’s intensively managed agricultural wetlands. A substantial decline in invertebrate diversity was described, concomitant to the decline in submerged macrophyte diversity in the remnant peatlands of the province over recent decades. The accumulation of degraded peat sediments is the most likely cause of this collapse. This has demonstrated that there is a strong influence of submerged macrophyte loss on benthic invertebrate assemblages. Furthermore, both taxonomic and functional insect composition reflected variations in emergent vegetation structure and peatland eutrophication. Likewise, benthic microbial phospholipids were correlated with aquatic insect composition and patterns of submerged and emergent vegetation growth and, more importantly, provided additional information on specific environmental factors driving the biodiversity in ditches, namely bicarbonate and sulphate concentrations in peatlands. The review of regional monitoring data in the three main soil type regions (sand, clay and peat) illustrates the significant influence of abiotic variability in structuring macroinvertebrate taxonomic and functional composition in the landscape. Thus eutrophication is operating as the primary

Chapter 6

136

environmental filter on invertebrate communities, while habitat structure (vegetation) operates as a secondary filter, driving taxonomic and functional community composition within ditches. A high degree of hydrological management in combination with intensive land use are underlying biodiversity patterns in the province. Fortunately, with the large number of ditches and canals in North Holland’s landscape there is good potential to provide the habitat necessary to support a diverse range of aquatic biota.

References

137

References Anderson M.J., Crist T.O., Chase J.M., Vellend M., Inouye B.D., Freestone A.L., et al.

(2011) Navigating the multiple meanings of B diversity: a roadmap for the practicing ecologist. Ecology Letters 14, 19–28.

Armitage P.D., Lattmann K., Kneebone N. & Harris I. (2001) Bank profile and structure as determinants of macroinvertebrate assemblages - seasonal changes and management. Regulated Rivers: Research & Management 17, 543–556.

Armitage P.D., Szoszkiewicz K., Blackburn J.H. & Nesbitt I. (2003) Ditch communities: a major contributor to floodplain biodiversity. Aquatic Conservation: Marine and Freshwater Ecosystems 13, 165–185.

Arribas P., Velasco J., Abellán P., Sánchez-Fernández D., Andújar C., Calosi P., et al. (2012) Dispersal ability rather than ecological tolerance drives differences in range size between lentic and lotic water beetles (Coleoptera: Hydrophilidae): dispersal ability and range size in water beetles. Journal of Biogeography 39, 984–994.

Bååth E., Frostegård Å., Pennanen T. & Fritze H. (1995) Microbial community structure and pH response in relation to soil organic matter quality in wood-ash fertilized, clear-cut or burned coniferous forest soils. Soil Biology & Biochemistry 27, 229–240.

Balvanera P., Pfisterer A.B., Buchmann N., He J.-S., Nakashizuka T., Raffaelli D., et al. (2006) Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecology Letters 9, 1146–1156.

Barko J.W., Gunnison D. & Carpenter S.R. (1991) Sediment interactions with submersed macrophyte growth and community dynamics. Aquatic Botany 41, 41–65.

Batzer D.P. & Wissinger S.A. (1996) Ecology of insect communities in nontidal wetlands. Annual Review of Entomology 41, 75–100.

Beche L.A., Mcelravy E.P. & Resh V.H. (2006) Long-term seasonal variation in the biological traits of benthic-macroinvertebrates in two Mediterranean-climate streams in California, U.S.A. Freshwater Biology 51, 56–75.

Beltman B. (1983) Van de wal in de sloot: een typologisch onderzoek van macrofaunacoenosen. PhD Thesis, Wageningen University, Wageningen, The Netherlands.

Bijkerk R. (2010) Handboek Hydrobiologie III: Biologisch onderzoek voor de ecologische beoordeling van Nederlandse zoete en brakke oppervlaktewateren. STOWA, Amersfoort, The Netherlands.

Bilton D.T., Freeland J.R. & Okamura B. (2001) Dispersal in freshwater invertebrates. Annual Review of Ecology and Systematics 32, 159–181.

Bligh E.G. & Dyer W.J. (1959) A rapid method of total lipid extraction and purification. Canadian Journal of Biochemistry and Physilogy 37, 911–917.

Bonada N., Prat N., Resh V.H. & Statzner B. (2006) Developments in aquatic insect biomonitoring: a comparative analysis of recent approaches. Annual Review of Entomology 51, 495–523.

Boon P.I., Virtue P. & Nichols P.D. (1996) Microbial consortia in wetland sediments: a biomarker analysis of the effects of hydrological regime, vegetation and season on benthic microbes. Marine and Freshwater Research 47, 27–41.

References

138

Bootsma M.C., Barendregt A. & van Alphen J.C.A. (1999) Effectiveness of reducing external nutrient load entering a eutrophicated shallow lake ecosystem in the Naardermeer nature reserve, The Netherlands. Biological Conservation 90, 193–201.

Bunn S.E. & Arthington A.H. (2002) Basic Principles and Ecological Consequences of Altered Flow Regimes for Aquatic Biodiversity. Environmental Management 30, 492–507.

Burnham K.P. & Anderson D.R. (2002) Model selection and multimodel inference: a practical information-theoretic approach, 2nd edn. Springer-Verlag, New York, USA.

Busch G. (2006) Future European agricultural landscapes—What can we learn from existing quantitative land use scenario studies? Agriculture, Ecosystems & Environment 114, 121–140.

Camargo J.A., Alonso A. & Salamanca A. (2005) Nitrate toxicity to aquatic animals: a review with new data for freshwater invertebrates. Chemosphere 58, 1255–1267.

Caraco N.F., Cole J.J. & Likens G.E. (1989) Evidence for sulphate-controlled phosphorus release from sediments of aquatic systems. Nature 341, 316–318.

Cardinale B.J., Ives A.R. & Inchausti P. (2004) Effects of species diversity on the primary productivity of ecosystems: extending our spatial and temporal scales of inference. Oikos 104, 437–450.

Cardinale B.J., Palmer M.A. & Collins S.L. (2002) Species diversity enhances ecosystem functioning through interspecific facilitation. Nature 415, 426–429.

Carpenter S.R. & Lodge D.M. (1986) Effects of submersed macrophytes on ecosystem processes. Aquatic Botany 26, 341–370.

Chase J.M. (2003) Community assembly: when should history matter? Oecologia 136, 489–498.

Clay J. (2004) World Agriculture and the Environment: A Commodity by Commodity Guide to Impacts and Practices. Island Press, Washington, DC.

Colmer T.D. (2003) Long-distance transport of gases in plants: a perspective on internal aeration and radial oxygen loss from roots. Plant, Cell and Environment 26, 17–36.

Connell J.H. (1978) Diversity in tropical rain forests and coral reefs. Science 199, 1302–1310.

Conrad K.F., Willson K.H., Harvey I.F., Thomas C.J. & Sherratt T.N. (1999) Dispersal characteristics of seven odonate species in an agricultural landscape. Ecography 22, 524–531.

Crichton M.I., Fisher D. & Woiwod I.P. (1978) Life histories and distribution of British Trichoptera, excluding Limnephilidae and Hydroptilidae, based on the Rothamsted Insect Survey. Holarctic Ecology 1, 31–45.

De Lange L. (1972) An ecological study of ditch vegetation in The Netherlands. University of Amsterdam, Amsterdam, The Netherlands.

De Szalay F.A. & Resh V.H. (1997) Responses of wetland invertebrates and plants important in waterfowl diets to burning and mowing of emergent vegetation. Wetlands 17, 149–156.

De Szalay F.A. & Resh V.H. (2000) Factors influencing macroinvertebrate colonization of seasonal wetlands. Freshwater Biology 45, 295–308.

Dickson L., Bull I.D., Gates P.J. & Evershed R.P. (2009) A simple modification of a silicic acid lipid fractionation protocol to eliminate free fatty acids from

References

139

glycolipid and phospholipid fractions. Journal of Microbiological Methods 78, 249–254.

Dijkman N.A. & Kromkamp J.C. (2006) Phospholipid-derived fatty acids as chemotaxonomic markers for phytoplankton: application for inferring phytoplankton composition. Marine Ecology Progress Series 324, 113–125.

Dolédec S., Phillips N., Scarsbrook M., Riley R.H. & Townsend C.R. (2006) Comparison of structural and functional approaches to determining landuse effects on grassland stream invertebrate communities. Journal of the North American Benthological Society 25, 44–60.

Donald P.F. & Evans A.D. (2006) Habitat connectivity and matrix restoration: the wider implications of agri-environment schemes. Journal of Applied Ecology 43, 209–218.

Donald P.F., Sanderson F.J., Burfield I.J. & van Bommel F.P.J. (2006) Further evidence of continent-wide impacts of agricultural intensification on European farmland birds, 1990–2000. Agriculture, Ecosystems & Environment 116, 189–196.

Edington J.M. & Hildrew A.G. (1995) A revised key to the caseless caddis larvae of the British Isles, with notes on their ecology. The Freshwater Biological Association, Cumbria, United Kingdom.

Edlund A., Nichols P.D., Roffey R. & White D.C. (1985) Extractable and lipopolysaccharide fatty acid and hydroxy acid profiles from Desulfovibrio species. Journal of Lipid Research 26, 982–988.

Elliott J.M. & Humpesch U.H. (2010) Mayfly larvae (Ephemeroptera) of Britain and Ireland: keys and a review of their ecology. The Freshwater Biological Association, Cumbria, United Kingdom.

Elliott J.M. & Mann K.H. (1979) A key to the British freshwater leeches, with notes on their cycles and ecology. The Freshwater Biological Association, Cumbria, United Kingdom.

Fahrig L. (2003) Eeffects of habtitat fragmentation on biodiversity. Annual Review of Ecology, Evolution, and Systematics 34, 487–515.

Feio M.J., Almeida S.F.P., Craveiro S.C. & Calado A.J. (2007) Diatoms and macroinvertebrates provide consistent and complementary information on environmental quality. Fundamental Applied Limnology 168, 247–258.

Findlay R.H., King G.M. & Watling L. (1989) Efficacy of phospholipid analysis in determining microbial biomass in sediments. Applied and Environmental Microbiology 55, 2888–2893.

Findlay R.H., Yeates C., Hullar M.A.J., Stahl D.A. & Kaplan L.A. (2008) Biome-level biogeography of streambed microbiota. Applied and Environmental Microbiology 74, 3014–3021.

Folke C., Carpenter S., Walker B., Scheffer M., Elmqvist T., Gunderson L., et al. (2004) Regime shifts, resilience, and biodiversity in ecosystem management. Annual Review of Ecology, Evolution, and Systematics 35, 557–581.

Folt C.L., Chen C.Y., Moore M.V. & Burnaford J. (1999) Synergism and antagonism among multiple stressors. Limnology and Oceanography 44, 864–877.

Foote A.L. & Hornung C.L.R. (2005) Odonates as biological indicators of grazing effects on Canadian prairie wetlands. Ecological Entomology 30, 273–283.

Frostegård Å., Tunlid A. & Bååth E. (1991) Microbial biomass measured as total lipid phosphate in soils of different organic content. Journal of Microbiological Methods 14, 151–163.

References

140

Frostegård Å., Tunlid A. & Bååth E. (2011) Use and misuse of PLFA measurements in soils. Soil Biology & Biochemistry 43, 1621–1625.

Gao X., Olapade O.A. & Leff L.G. (2005) Comparison of benthic bacterial community composition in nine streams. Aquatic Microbial Ecology 40, 51–60.

Gauch Jr. H.G., Whittaker R.H. & Wentworth T.R. (1977) A comparative study of reciprocal averaging and other ordination techniques. Journal of Ecology 65, 157–174.

Gessner M.O., Inchausti P., Persson L., Raffaelli D.G. & Giller P.S. (2004) Biodiversity effects on ecosystem functioning: insights from aquatic systems. Oikos 104, 419–422.

Geurts J.J.M., Sarneel J.M., Willers B.J.C., Roelofs J.G.M., Verhoeven J.T.A. & Lamers L.P.M. (2009) Interacting effects of sulphate pollution, sulphide toxicity and eutrophication on vegetation development in fens: a mesocosm experiment. Envrionmental Pollution 157, 2072–2081.

Giller P.S., Hillebrand H., Berninger U.-G., Gessner M.O., Hawkins S., Inchausti P., et al. (2004) Biodiversity effects on ecosystem functioning: emerging issues and their experimental test in aquatic environments. Oikos 104, 423–436.

Giller P.S. & McNeill S. (1981) Predation strategies, resource partitioning and habitat selection in Notonecta (Hemiptera/Heteroptera). Journal of Animal Ecology 50, 789–808.

Goedkoop W., Gullberg K.R., Johnson R.K. & Ahlgren I. (1997) Microbial response of a freshwater benthic community to a simulated diatom sedimentation event: interactive effects of benthic fauna. Microbial Ecology 34, 131–143.

Goedkoop W., Sonesten L., Markensten H. & Ahlgren G. (1998) Fatty acid biomarkers show dietary differences between dominant chironomid taxa in Lake Erken. Freshwater Biology 40, 135–143.

Goldenberg-Vilar A., van Dam H., van Loon E.E., Vonk J.A., van der Geest H., Admiraal W. (2014) Eutrophication decreases distance decay of similarity in diatom communities. Freshwater Biology.

Gonzalez J.M., Basaguren A. & Pozo J. (2001) Life history and production of Caenis luctuosa (Burmeister) (Ephemeroptera, Caenidae) in two nearby reaches along a small stream. Hydrobiologia 452, 209–215.

Gotelli N.J. & Colwell R.K. (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecology letters 4, 379–391.

Gotelli N.J. & Entsminger G.L. (2011) EcoSim: Null models software for ecology. Acquired Intelligence Inc. & Kesey-Bear, Jericho, VT, USA.

Goyke A.P. & Hershey A.E. (1992) Effects of fish predation on larval chironomid (Diptera: Chironomidae) communities in an arctic ecosystem. Hydrobiologia 240, 203–211.

Griffen B.D. & Drake J.M. (2008) Effects of habitat quality and size on extinction in experimental populations. Proceedings of the Royal Society of Biological Sciences 275, 2251–2256.

Hansen J.P., Wikström S.A., Axemar H. & Kautsky L. (2011) Distribution differences and active habitat choices of invertebrates between macrophytes of different morphological complexity. Aquatic Ecology 45, 11–22.

References

141

Harrison E.T., Norris R.H. & Wilkinson S.N. (2007) The impact of fine sediment accumulation on benthic macroinvertebrates: implications for river management. In: Proceedings of the 5th Australian Stream Management Conference. Australian Rivers: Making a Difference. pp. 139–144.

Harvey H.R., Fallon R.D. & Patton J.S. (1986) The effect of organic matter and oxygen on the degradation of bacterial membrane lipids in marine sediments. Geochimica et Cosmochimica Acta 50, 795–804.

Hendrickx F., Maelfait J.-P., Van Wingerden W., Schweiger O., Speelmans M., Aviron S., et al. (2007) How landscape structure, land-use intensity and habitat diversity affect components of total arthropod diversity in agricultural landscapes: Agricultural factors and arthropod biodiversity. Journal of Applied Ecology 44, 340–351.

Herzon I. & Helenius J. (2008) Agricultural drainage ditches, their biological importance and functioning. Biological Conservation 141, 1171–1183.

Higler L.W.G. & Verdonschot P.F.M. (1989) Macroinvertebrates in the Demmerik ditches (the Netherlands): the role of environmental structure. Hydrobiological Bulletin 23, 143–150.

Hinojosa-Garro D., Mason C.F. & Underwood G.J.C. (2010) Influence of macrophyte spatial architecture on periphyton and macroinvertebrate community structure in shallow water bodies under contrasting land management. Fundamental Applied Limnology 177, 19–37.

Hofman C.C. (2007) Visstandonderzoek Wormer- en Jisperveld: KRW-monitoring Hollands Noorderkwartier. Hoogheemraadschap Hollands Noorderkwartier, Edam, The Netherlands.

Holmer M. & Storkholm P. (2001) Sulphate reduction and sulphur cycling in lake sediments: a review. Freshwater Biology 46, 431–451.

Hooper D.U., Chapin Iii F.S., Ewel J.J., Hector A., Inchausti P., Lavorel S., et al. (2005) Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecological monographs 75, 3–35.

Horppila J. & Nurminen L. (2001) The effect of an emergent macrophyte (Typha angustifolia) on sediment resuspension in a shallow north temperate lake. Freshwater Biology 46, 1447–1455.

Hughes A.R. (2012) Disturbance and diversity: an ecological chicken and egg problem. Nature 3, 48–56.

Hunting E.R., Whatley M.H., van der Geest H.G., Mulder C., Kraak M.H.S., Breure A.M., et al. (2012) Invertebrate footprints on detritus processing, bacterial community structure, and spatiotemporal redox profiles. Freshwater Science 31, 724–732.

Huston M.A. (1979) A general hypothesis of species diversity. The American Naturalist 113, 81–101.

Huston M.A. (1994) Biological diversity: the coexistence of species on changing landscapes. Cambridge University Press, Cambridge, UK.

Janse J.H. & Van Puijenbroek P.J. (1998) Effects of eutrophication in drainage ditches. Environmental Pollution 102, 547–552.

Janssen R., Goosen H., Verhoeven M.L., Verhoeven J.T.A., Omtzigt A.Q.A. & Maltby E. (2005) Decision support for integrated wetland management. Environmental Modelling & Software 20, 215–229.

References

142

Jeffries M. Invertebrate colonization of artificial pondweeds of differing fractal dimension. Oikos 67, 142–148.

Johnson S.L. & Ringler N.H. (2014) The response of fish and macroinvertebrate assemblages to multiple stressors: A comparative analysis of aquatic communities in a perturbed watershed (Onondaga Lake, NY). Ecological Indicators 41, 198–208.

Jones J.G. (1979) Microbial nitrate reduction in freshwater sediments. Journal of General Microbiology 115, 27–35.

Junier S.J. & Mostert E. (2012) The implementation of the Water Framework Directive in The Netherlands: does it promote integrated management? Physics and Chemistry of the Earth 47-48, 2–10.

Kaur A., Chaudhary A., Kaur A., Choudhary R. & Kaushik R. (2005) Phospholipid fatty acid - a bioindicator of environment monitoring and assessment in soil ecosystem. Current Science 89, 1103–1112.

Keast A. (1984) The introduced aquatic macrophyte, Myriophyllum spicatum, as habitat for fish and their invertebrate prey. Canadian Journal of Zoology 62, 1289–1303.

Kleijn D., Berendse F., Smit R., Gilissen N., Smit J., Brak B., et al. (2004) Ecological effectiveness of agri-envrionment schemes in different agricultural landscapes in The Netherlands. Conservation Biology 18, 775–786.

Kovalenko K.E., Dibble E.D. & Slade J.G. (2010) Community effects of invasive macrophyte control: role of invasive plant abundance and habitat complexity. Journal of Applied Ecology 47, 318–328.

Lamers L.P.M., Smolders A.J.P. & Roelofs J.G.M. (2002) The restoration of fens in the Netherlands. Hydrobiologia 478, 107–130.

Lande R. (1996) Statistics and partitioning of species diversity, and similarity among multiple communities. Oikos 76, 5–13.

Laprise R. & Dodson J.J. (1993) Nature of environmental variability experienced by benthic and pelagic animals in the St. Lawrence Estuary, Canada. Marine Ecology-Progress Series 94, 129–129.

Laurance W. (2008) Theory meets reality: how habitat fragmentation research has transcended island biogeographic theory. Biological Conservation 141, 1731–1744.

Laurance W.F. (2002) Hyperdynamism in fragmented habitats. Journal of Vegetation Science 13, 595–602.

Lenat D.R. (1993) A biotic index for the southeastern United States: derivation and list of tolerance values with criteria for assigning water-quality ratings. Journal of the North American Benthological Society 12, 279–290.

Léveillé J.-C., Amblard C. & Bourdier G. (1997) Fatty acids as specific algal markers in a natural lacustrian phytoplankton. Journal of Plankton Research 19, 469–490.

Liang K.Y. & Zeger S.L. (1986) Longitudinal data analysis using generalized linear models. Biometrics 73, 13–22.

Ligtvoet W., Beugelink G.P. & Franken R. (2008) Evaluation of the Water Framework Directive in the Netherlands; costs and benefits. Netherlands Environmental Assessment Agency (PBL), Bilthoven, The Netherlands.

Loreau M. (1998) Biodiversity and ecosystem functioning: a mechanistic model. Proceedings of the National Academy of Sciences 95, 5632–5636.

Lucena-Moya P. & Duggan I.C. (2011) Macrophyte architecture affects abundance and diversity of littoral microfauna. Aquatic Ecology 45, 279–287.

References

143

Lunde K.B. & Resh V.H. (2012) Development and validation of a macroinvertebrate index of biotic integrity (IBI) for assessing urban impacts to Northern California freshwater wetlands. Environmental Monitoring and Assessment 184, 3653–3674.

Lytle D.A. & Poff N.L. (2004) Adaptation to natural flow regimes. Trends in Ecology & Evolution 19, 94–100.

Macan T.T. (1977) A key to the British fresh and brackish water Gastropods, with notes on their ecology. The Freshwater Biological Association, Cumbria, United Kingdom.

MacArthur R.H. & Wilson E.O. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton, USA.

Malmqvist B. (2002) Aquatic invertebrates in riverine landscapes. Freshwater Biology 47, 679–694.

McCabe D.J. & Gotelli N.J. (2000) Effects of disturbance frequency, intensity, and area on assemblages of stream macroinvertebrates. Oecologia 124, 270–279.

McLaughlin D.B. & Harris H.J. (1990) Aquatic insect emergence in two Great Lakes marshes. Wetlands Ecology and Management 1, 111–121.

Menezes S., Baird D.J. & Soares M.V.M. (2010) Beyond taxonomy: a review of macroinvertebrate trait-based community descriptors as tools for freshwater biomonitoring. Journal of Applied Ecology 47, 711–719.

Molden D. (2007) Water for food, water for life: a comprehensive assessment of water management in agriculture. Earthscan, London; Sterling, VA.

Moller Pillot H.K.M. (2009) Chironomidae larvae, biology and ecology of the Chironomini. KNNV Publishing, Zeist, The Netherlands.

Moser M., Prentice C. & Frazier S. (1996) A Global Overview of Wetland Loss and Degradation. In: Proceedings of the 6th Meeting of the Conference of the Contracting Parties. Ramsar, Brisbane, Australia.

Moss B. & Timms M. (1989) Sediment stability and food availability as determinants of the benthic invertebrate fauna in two shallow lakes. Hydrobiologia 185, 249–257.

Mousseau T.A. & Dingle H. (1991) Maternal effects in insect life histories. Annual review of entomology 36, 511–534.

Murkin E.J., Murkin H.R. & Titman R.D. (1992) Nektonic invertebrate abundance and distribution at the emergent vegetation-open water interface in the delta marsh Manitoba, Canada. Wetlands 12, 45–52.

Newman R.M. (1991) Herbivory and detritivory on freshwater macrophytes by invertebrates. Journal of the North American Benthological Society 10, 89–114.

Nijboer R.C. & Verdonschot P.F.M. (2004) Rare and common macroinvertebrates: definition of distribution classes and their boundaries. Arch. Hydrobiol. 161, 45–64.

Nilsson A. (2005) Aquatic insects of north Europe: a taxonomic handbook, CD-ROM. Apollo Books, Stenstrup, Denmark.

O’Toole C., Donohue I., Moe S.J. & Irvine K. (2008) Nutrient optima and tolerances of benthic invertebrates, the effects of taxonomic resolution and testing of selected metrics in lakes using an extensive European data base. Aquatic Ecology 42, 277–291.

Oksanen J. (2013) Multivariate analysis of ecological communities in R: vegan tutorial. University of Oulu, Oulu, Finland.

References

144

Orendt C. & Spies M. (2010) Bestimmungsschlüssel Chironomini (Diptera: Chironomidae: Chironominae). Unter Verwendung vorwiegend makroskopischer Merkmale. Leipzig, Germany.

Orr B.K. & Resh V.H. (1992) Influence of Myriophyllum aquaticum cover on Anopheles mosquito abundance, oviposition, and larval microhabitat. Oecologia 90, 474–482.

Painter D. (1999) Macroinvertebrate distributions and the conservation value of aquatic Coleoptera, Mollusca and Odonata in the ditches of traditionally managed and grazing fen at Wicken Fen, UK. Journal of Applied Ecology 36, 33–48.

Palmer M.A., Covich A.P., Lake S., Biro P., Brooks J.J., Cole J., et al. (2000) Linkages between aquatic sediment biota and life above sediments as potential drivers of biodiversity and ecological processes. Bioscience 50, 1062–1075.

Peeters E.T.H.M., Gylstra R. & Vos J.H. (2004) Benthic macroinvertebrates community structure in relation to food and environmental variables. Hydrobiologia 519, 103–115.

Petchey O.L., Downing A.L., Mittelbach G.G., Persson L., Steiner C.F., Warren P.H., et al. (2004) Species Loss and the Structure and Functioning of Multitrophic Aquatic Systems. Oikos 104, 467–478.

Pimm S.L. (1984) The complexity and stability of ecosystems. Nature 307, 321–326. Piotrowska-Seget Z. & Mrozik A. (2003) Signature lipid biomarker (SLB) analysis in

determining changes in community structure of soil microorganisms. Polish Journal of Environmental Studies 12, 669–676.

Poff N.L. (1997) Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. Journal of the North American Benthological Society 16, 391–409.

Pons L.J. (1992) Holocene peat formation in the lower parts of the Netherlands. In: Fens and bogs in the Netherlands: vegetation, history, nutrient dynamics and conservation. (Ed. J.T.A. Verhoeven), pp. 7–79. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Potts M., Olie J.J., Nickels J.S., Parsons J. & White D.C. (1987) Variation in phospholipid ester-linked fatty acids and carotenoids of desiccated Nostoc commune (Cyanobacteria) from different geographic locations. Applied and Environmental Microbiology 53, 4–9.

Quinn G.P. & Keough M.J. (2002) Experimental design and data analysis for biologists. Cambridge University Press, Cambridge, UK.

R Development Core Team (2012) R: A language and environment for statistical computing. Foundation for Statistical Computing, Vienna, Austria.

Radomski P. & Goeman T.J. (2001) Consequences of human lakeshore development on emergent and floating-leaf vegetation abundance. North American Journal of Fisheries Management 21, 46–61.

Rantala M.J., Ilmonen J., Koskimäki J., Suhonen J. & Tynkkynen K. (2004) The macrophyte, Stratiotes aloides, protects larvae of dragonfly Aeshna viridis against fish predation. 38, 77–82.

Rip W.J., Rawee N. & de Jong A. (2006) Alternation between clear, high-vegetation and turbid, low-vegetation states in a shallow lake: the role of birds. Aquatic Botany 85, 184–190.

References

145

Roelofs J.G.M. (1991) Inlet of alkaline river water into peaty lowlands: effects on water quality and Stratiotes aloides L. stands. Aquatic Botany 39, 267–293.

Sala O.E., Chapin III F.S., Armesto J.J., Berlow E., Bloomfield J., Dirzo R., et al. (2000) Global biodiversity scenarios for the year 2100. Science 287, 1770–1774.

Sand-Jensen K., Prahl C. & Stokholm H. (1982) Oxygen release from roots of submerged aquatic macrophytes. Oikos 8, 349–354.

Savage A.A. (1989) Adults of the British aquatic Hemiptera Heteroptera, a key with ecological notes. The Freshwater Biological Association, Cumbria, United Kingdom.

Scheffer M. (1990) Multiplicity of stable states in freshwater systems. Hydrobiologia 200/201, 475–486.

Scheffer M., Achterberg A.A. & Beltman B. (1984) Distribution of macro-invertebrates in a ditch in relation to vegetation. Freshwater Biology 14, 367–370.

Scheffer M., Hosper S.H., Meijer M.-L., Moss B. & Jeppesen E. (1993) Alternative equilibria in shallow lakes. Trends in Ecology and Evolution 8, 275–279.

Scheffer M. & van Geest G.J. (2006) Small habitat size and isolation can promote species richness: second-order effects on biodiversity in shallow lakes and ponds. Oikos 112, 227–231.

Schutten J., Dainty J. & Davy A.J. (2005) Root anchorage and its significance for submerged plants in shallow lakes. Journal of Ecology 93, 556–571.

Shurin J.B., Winder M., Adrian R., Keller W. (Bill), Matthews B., Paterson A.M., et al. (2010) Environmental stability and lake zooplankton diversity - contrasting effects of chemical and thermal variability. Ecology Letters 13, 453–463.

Siegenthaler P.-A. & Murata N. eds (1998) Lipids in photosynthesis: structure, function and genetics. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Sinke A.J.C., Cornelese A.A., Keizer P., van Tongeren O.F.R. & Cappenberg T.E. (1990) Mineralization, pore water chemistry and phosphorus release from peaty sediments in the eutrophic Loosdrecht lakes, The Netherlands. Freshwater Biology 23, 587–599.

Smith A.J., Bode R.W. & Kleppel G.S. (2007) A nutrient biotic index (NBI) for use with benthic macroinvertebrate communities. Ecological Indicators 7, 371–386.

Smolders A.J.P., Lamers L.P.M., Lucassen E.C.H.E.T., van der Velde G. & Roelofs J.G.M. (2006) Internal eutrophication: how it works and what to do about it - a review. Chemistry and Ecology 22, 93–111.

Smoot J.C. & Findlay R.H. (2001) Spatial and seasonal variation in a reservoir sedimentary microibal community as determined by phospholipd analysis. Microbial Ecology 42, 350–358.

Søndergaard M., Jensen J.P. & Jeppesen E. (2003) Role of sediment and internal loading of phosphorus in shallow lakes. Hydrobiologia 506, 135–145.

Southwood T.R.E. (1977) Habitat, the templet for ecological strategies? Journal of Animal Ecology 46, 337–365.

Stearns S.C. (1976) Life-history tactics: a review of the ideas. The Quarterly Review of Biology 51, 3–47.

Storey R. (2007) Aquatic invertebrate diversity and distribution at Aammiq Marsh. A Rocha, Lebanon.

Suhonen J., Suutari E., Kaunisto K.M. & Krams I. (2013) Patch area of macrophyate Stratioites aloides as a critical resource for declining dragonfly Aeshna viridis. Journal of Insect Conservation 17, 393–398.

References

146

Sundh I., Nilsson M. & Borga P. (1997) Variation in microbial community structure in two boreal peatlands as determined by analysis of phospholipid fatty acid profiles. Applied and Environmental Microbiology 63, 1476–1482.

Suren A.M., Lambert P., Image K. & Sorrell B.K. (2008) Variation in wetland invertebrate communities in lowland acidic fens and swamps. Freshwater Biology 53, 727–744

Tachet H., Bournaud M., Richoux P. & Usseglio-Polatera P. (2002) Invertébrés d’eau douce : systématique, biologie, écologie. CNRS Editions, Paris.

Taylor J. & Parkes R.J. (1983) The cellular fatty acids of the sulphate-reducing bacteria, Desulfobacter sp., Desulfobulbus sp. and Desulfovibrio desulfuricans. Journal of General Microbiology 129, 3303–3309.

Ter Braak C.J.F. & Smilauer P. (2002) CANOCO Reference Manual and CanoDraw for Windows User’s Guide: Software for Canonical Community Ordination (version 4.5). Microcomputer Power, Ithaca, NY, USA.

Townsend C.R. & Hildrew A.G. (1994) Species traits in relation to a habitat templet for river systems. Freshwater Biology 31, 265–275.

Traunspurger W., Bergtold M. & Goedkoop W. (1997) The effects of nematodes on bacterial activity and abundance in a freshwater sediment. Oecologia 112, 118–122.

Twisk W., Noordervliet M.A.W. & ter Keurs W.J. (2000) Effects of ditch management on caddisfly, dragonfly and amphibian larvae in intensively farmed peat areas. Aquatic Ecology 34, 397–411.

Twisk W., Noordervliet M.A.W. & ter Keurs W.J. (2003) The nature value of the ditch vegetation in peat areas in relation to farm management. Aquatic Ecology 37, 191–209.

Underwood G.J.C. & Thomas J.D. (1990) Grazing interactions between pulmonate snails and epiphytic algae and bacteria. Freshwater Biology 23, 505–522.

University of Wisconsin (2010) Transparency - a water clarity measurement. Wisconsin, Wisconsin, USA.

Usseglio-Polatera P., Bournaud M., Richoux P. & Tachet H. (2000) Biological and ecological traits of benthic freshwater macroinvertebrates: relationships and definition of groups with similar traits. Freshwater Biology 43, 175–205.

Van Dam H. (2009) Evaluatie basismeetnet waterkwaliteit Hollands Noorderkwartier: Trendanalyse hydrobiology, temperatuur en waterchemie 1982-2007. Hoogheemraadschap Hollands Noorderkwartier, Edam, The Netherlands.

Van Dam P.J.E.M. (2001) Sinking peat bogs: environmental change holland 1350-1550. Environmental History 6, 32–45.

Van der Bolt F., van den Bosch R., Brock T., Hellegers P., Kwakernaak C., Leenders D., et al. (2003) Aquarein: gevolgen van de Europese Kaderrichtlijn Water voor landbouw, natuur, recreatie en visserij. Alterra, Research Instituut voor de Groene Ruimte, Wageningen, The Netherlands.

Van der Hammen H. (1992) De macrofauna van Noord-Holland. PhD Thesis, Universiteit Nijmegen, Nijmegen, The Netherlands.

Van Eerden M.R., Lenselink G. & Zijstra M. (2010) Long-term changes in wetland area and composition in The Netherlands affecting the carrying capacity for wintering waterbirds. ARDEA 98, 265–282.

References

147

Van Nes E.H., Scheffer M., van den Berg M.S. & Coops H. (2002) Dominance of charophytes in eutrophic shallow lakes - when should we expect it to be an alternative stable state? Aquatic Botany 72, 275–296.

Van Zuidam J.P. (2013) Macrophytes in drainage ditches, functioning and perspectives for recovery. PhD Thesis, Wageningen University, Wageningen, The Netherlands.

Van Zuidam J.P. & Peeters E.T.H.M. (2012) Cutting affects growth of Potamogeton lucens L. and Potamogeton compressus L. Aquatic Botany 100, 51–55.

Van Zuidam J.P., Raaphorst E.P. & Peeters E.T.H.M. (2012) The role of propagule banks from drainage ditches dominated by free-floating or submerged plants in vegetation restoration. Restoration Ecology 20, 416–425.

Varga I. (2003) Structure and changes of macroinvertebrate community colonizing decomposing rhizome litter of common reed at Lake Fertö/Neusiedler See (Hungary). Hydrobiologia 506-509, 413–420.

Verberk W.C.E.P., Kuper J.T., Lamers L.P.M., Christianen M.J.A. & Esselink H. (2007) Restoring fen water bodies by removing accumulated organic sludge: what are the effects for aquatic macroinvertebrates. Proceedings of the Netherlands Entomological Society Meeting 18, 115–124.

Verberk W.C.E.P., Siepel H. & Esselink H. (2008a) Life-history strategies in freshwater macroinvertebrates. Freshwater Biology 53, 1722–1738.

Verberk W.C.E.P., Siepel H. & Esselink H. (2008b) Applying life-history strategies for freshwater macroinvertebrates to lentic waters. Freshwater Biology 53, 1739–1753.

Verberk W.C.E.P., van Duinen G.A., Brock A.M.T., Leuven R.S.E.W., Siepel H., Verdonschot P.F.M., et al. (2006) Importance of landscape heterogeneity for the conservation of aquatic macroinvertebrate diversity in bog lanscapes. Journal for Nature Conservation 14, 78–90.

Verberk W.C.E.P., van der Velde G. & Esselink H. (2010) Explaining abundance-occupancy relationships in specialists and generalists: a case study on aquatic macroinvertebrates in standing waters. Journal of Animal Ecology 79, 589–601.

Verberk W.C.E.P., van Noordwijk C.G.E. & Hildrew A.G. (2013) Delivering on a promise: integrating species traits to transform descriptive community ecology into a predictive science. Freshwater Science 32, 531–547.

Verbruggen F., Heiri O., MeriläInen J.J. & Lotter A.F. (2011) Subfossil chironomid assemblages in deep, stratified European lakes: relationships with temperature, trophic state and oxygen: Chironomid assemblages in deep, stratified lakes. Freshwater Biology 56, 407–423.

Verdonschot R.C.M. (2010) Optimizing the use of activity traps for aquatic biodiversity studies. Journal of North American Benthological Society 29, 1228–1240.

Verdonschot R.C.M. (2012) Drainage ditches, biodiversity hotspots for aquatic invertebrates, defining and assessing the ecological status of a man-made ecosystem based on macroinvertebrates. PhD Thesis, Wageningen University, Wageningen, The Netherlands.

Verdonschot R.C.M., Didderen K. & Verdonschot P.F.M. (2012a) Importance of habitat structure as a determinant of the taxonomic and functional composition of lentic macroinvertebrate assemblages. Limnologica 42, 31–42.

Verdonschot R.C.M., Keizer-Vlek H.E. & Verdonschot P.F.M. (2011) Biodiversity value of agricultural drainage ditches; a comparative analysis of the aquatic

References

148

invertebrate fauna of ditches and small lakes. Aquatic Conservation: Marine and Freshwater Ecosystems 21, 715–727.

Verdonschot R.C.M., Keizer-Vlek H.E. & Verdonschot P.F.M. (2012b) Development of a multimetric index based on macroinvertebrates for drainage ditch networks in agriculutral areas. Ecological Indicators 13, 232–242.

Verdonschot R.C.M. & Verdonschot P.F.M. (2013) Shading effects of free-floating plants on drainage-ditch invertebrates. Limnology.

Vestal J.R. & White D.C. (1989) Lipid analysis in microbial ecology. Bioscience 39, 535–541.

Viljoen A., Cyrus D. & Wepener V. (2001) Comparison of the density and species composition of aquatic invertebrates found between the roots of Eichhornia crassipes plants from two coastal lakes in northern KwaZulu-Natal. African Journal of Aquatic Science 26, 57–66.

Vinebrooke R.D., Cottingham K.L., Norberg J., Scheffer M., Dodson S.I., Maberly S.C., et al. (2004) Impacts of multiple stressors on biodiversity and ecosystem functioning: The role of species co-tolerance. Oikos 104, 451–457.

Vlek H.E., Verdonschot P.F.M. & Nijboer R.C. (2004) Towards a multimetric index for the assessment of Dutch streams using benthic macroinvertebates. Hydrobiologia 516, 173–189.

Vos J.H., van den Brink P.J., van den Ende F.P., Ooijevaar M.A.G., Oosthoek J.F., Postma J.F., et al. (2002) Growth response of a benthic detritivore to organic matter composition in sediments. Journal of the North American Benthological Society 21, 443–456.

Wagenhoff A., Townsend C.R., Phillips N. & Matthaei C.D. (2011) Subsidy-stress and multiple-stressor effects along gradients of deposited fine sediment and dissolved nutrients in a regional set of streams and rivers: Sediment and nutrients in streams. Freshwater Biology 56, 1916–1936.

Wallace I.D., Wallace B. & Philipson G.N. (2003) Keys to the case-bearing caddis larvae of Britain and Ireland. The Freshwater Biological Association, Cumbria, United Kingdom.

Warfe D.M. & Barmuta L.A. (2004) Habitat structural complexity mediates the foraging success of multiple predator species. Oecologia 141, 171–178.

Wassen M.J., Barendregt A., Bootsma M.C. & Schot P.P. (1989) Groundwater chemistry and vegetation of gradients from rich fen to poor fen in the Naardermeer (the Netherlands). Vegetatio 79, 117–132.

Whatley M.H., van Loon E.E., van Dam H., Vonk J.A., van der Geest H.G. & Admiraal W. (2014) Macrophyte loss drives decadal change in benthic invertebrates in peatland drainage ditches. Freshwater Biology 59, 114–126.

White D.C., Bobbie R.J., Herron J.S., King J.D. & Morrison S.J. (1979) Biochemical measurements of microbial mass and activity from environmental samples. In: Native aquatic bacteria: enumeration activity and ecology. (Eds J.W. Costerton & R.R. Colwell), pp. 69–81. American Society for Testing and Methods, Philadelphia, USA.

Williams P., Whitfield M., Biggs J., Bray S., Fox G., Nicolet P., et al. (2003) Comparative biodiversity of rivers, streams, ditches and ponds in an agricultural landscape in Southern England. Biology and Fertility of Soils 115, 329–341.

References

149

Winemiller K.O., Flecker A.S. & Hoeinghaus D.J. (2010) Patch dynamics and environmental heterogeneity in lotic ecosystems. Journal of North American Benthological Society 29, 84–99.

Wise D.H. & Molles M.C.J. (1979) Colonization of artificial substrates by stream insects: influence of substrate size and diversity. Hydrobiologia, 69–74.

Wood P.J. & Armitage P.D. (1997) Biological effects of fine sediment in the lotic environment. Environmental Management 21, 203–217.

Yuan L.L. (2004) Assigning macroinvertebrate tolerance classifications using generalised additive models. Freshwater Biology 49, 662–677.

Summary

150

Summary Agricultural environments cover about a third of the earth’s land area and contribute significantly to biodiversity, yet agricultural intensification is leading to a global decline in biodiversity with aquatic systems being the most at risk. The low-lying landscape of North Holland, The Netherlands contains an interlacing network of drainage ditches and its peatlands are remnants of a once vast system of floodplains and raised bogs, which can still support a range of aquatic invertebrates. However, these drainage ditches are subject to regular disturbances from the inlet of mineral rich river derived water, eutrophication, regular vegetation removal and sediment dredging, leading to a decline in the diversity of invertebrates and plants over recent decades. To prevent further biodiversity loss it is essential to identify the dominant environmental drivers and mechanisms underlying the distribution of aquatic invertebrates in this heavily impacted ecosystem. This thesis seeks to elucidate how aquatic invertebrate communities are structured by environmental drivers in North Holland’s agricultural drainage ditches. To this end the following questions were defined:

What are the key environmental drivers determining aquatic invertebrate community composition in agricultural ditches?

What are the mechanisms underlying the response of aquatic invertebrates to these environmental drivers?

How can management practices be adjusted to improve ecological conditions in the agricultural ditches?

Peat degradation causes the accumulation of fine particles in the aquatic environment and can lead to the decline of submerged macrophytes, which provide necessary habitat for benthic invertebrates. Therefore in Chapter 2 the decadal (1985 – 2007) trends in benthic species richness were investigated in 29 peatland ditches by reviewing long-term monitoring data from the province of North Holland. Monitoring data were analysed in conjunction with a complementary field experiment, in which submerged artificial macrophytes, natural sediments and emergent bank vegetation were sampled in degraded peatland ditches, to determine patterns of macroinvertebrate habitat occupancy. The long-term monitoring data shows that chemical conditions in agricultural peat ditches have improved (slightly) over recent decades, yet the diversity of benthic invertebrates have declined, concomitant to a decline in submerged macrophyte diversity. The dependence of macroinvertebrates on macrophytes was reinforced by our field experiment which revealed that invertebrate density was highest in submerged artificial plants and invertebrate species richness was highest in natural emergent vegetation. Conversely, degraded peat sediments supported extremely few invertebrates. Thus the degradation of peat

Summary

151

predominantly had an indirect effect on benthic invertebrate diversity via the loss of essential macrophyte habitat. As described above the availability of submerged macrophyte habitat is an important driver of aquatic invertebrate communities, yet eutrophication and peat degradation are leading to the decline of submerged vegetation. Emergent vegetation is able to persist in eutrophicated ditches, however vegetation removal is carried out annually and can reduce the availability of this habitat. Thus, in Chapter 3 we applied the landscape filtering approach to determine how the absence of emergent vegetation structured aquatic insect communities in peatland drainage ditches under different trophic conditions. To this end, aquatic insects were sampled in emergent vegetation stands in one mesotrophic and one eutrophic peatland. In addition to taxonomic responses, functional community composition was studied by assigning life-history strategies to insect species to determine the influence of vegetation structure and peatland type on taxonomic and functional community structure. The findings showed that ditches in the eutrophic peatland were dominated by insects adapted to abiotic extremes, while species with good dispersal abilities were strongly related to emergent vegetation cover. These results indicate that while peatland type appeared to primarily determining the pool of species within each wetland, emergent vegetation acted as a secondary filter by structuring functional community composition within ditches. The inlet of mineral rich, river derived water can alter abiotic conditions and is expected to influence drainage ditch communities. Yet, it is often impossible to disentangling, interconnected abiotic parameters in the field. The potential of benthic microbial composition to reveal environmental drivers of insect communities has not been explored. Thus, in Chapter 4 we investigated benthic microbial community composition in 25 peatland ditches by analysing phospholipid fatty acid (PLFA) profiles and investigate correlations with aquatic insects. Furthermore, we examined relationships between microbial and insect data alongside abiotic parameters, emergent and submerged vegetation. PLFA composition indicated the dominance of eukaryotic algae, cyanobacteria, sulphate reducing bacteria, (gram positive) anaerobic bacteria and gram negative bacteria in the microbial community. Moreover, ditches that were distinguished by their microbial communities differed significantly in insect composition, in particular Odonata, Trichoptera and Chironomus larvae. The main environmental factors underlying this pattern were the presence of submerged and emergent vegetation and concentrations of bicarbonate, sulphate and nutrients. These particular abiotic parameters are known to be associated with the inlet of mineral rich, river derived water. These findings demonstrated that the inlet of external mineral rich waters was negatively impacting ditch community composition. The inlet of mineral rich waters is also expected to cause variability in abiotic conditions. Yet the importance of abiotic variability has not been

Summary

152

investigated for invertebrate communities in Dutch drainage ditches. In Chapter 5 the role of temporal abiotic variability as a driver of invertebrate community composition in agricultural ditches is investigated by analysing monitoring data covering 84 ditches and three soil types (sand, peat and clay). We examined correlations between abiotic conditions (variability and average values of parameters) and macroinvertebrate diversity, determined as local (α diversity), species-turnover (β diversity) and regional diversity (γ diversity). In addition, functional community responses were examined by analysing the expression of insect life-history strategies in relation to abiotic conditions. This study shows that abiotic variability of nutrients and macro-ions was significant in structuring aquatic invertebrate diversity in the drainage ditches and the effect appears to be scale dependent. In addition, the response of different species could be related to their life-history strategy. This thesis provides an overview of the interrelationships between different fractions of the ecosystem in North Holland’s intensively managed agricultural drainage ditches. It has been demonstrated that both submerged and emergent vegetation strongly influence invertebrate assemblages. Moreover, specific abiotic factors driving invertebrate diversity in these waters are associated with the inlet of mineral rich, river derived water and include concentrations of nutrients, bicarbonates and sulphate. Nutrients and macro-ions cause degradation of peat soils, leading to the accumulation of amorphous layers of mud in remnant peatland environments. In addition, temporal abiotic variability is structuring macroinvertebrate taxonomic and functional composition in the landscape. Agricultural intensification is placing increased pressures on aquatic ecosystems, via inputs of nutrients, suspended sediments and water abstraction. In the province of North Holland the demand for water is leading to greater influence of mineral rich waters which is ultimately degrading the aquatic environment. Moreover, the annual removal of vegetation is weakening the plant community which is already stressed by eutrophication and turbidity and this has a knock-on effect in causing a decline in invertebrate diversity. Despite the implementation of habitat improvement schemes, such as nature friendly banks in the Netherlands, there appears to be a lack of evidence supporting their effectiveness in promoting biodiversity. The isolation and size of these habitat creation measures are likely to be partially responsible. Small patches of suitable habitat, surrounded by a matrix of intensively managed agricultural land limit the ability of species to colonize these isolated habitats. For management strategies to successfully increase biodiversity in agricultural landscapes the requirements of habitat size and quality for biota must be met. Fortunately, with the large number of ditches and canals in North Holland’s landscape there is good potential to provide the habitat necessary to support a diverse range of aquatic biota, granted management seeks to do so.

Samenvatting

153

Samenvatting Ongeveer één derde van het landoppervlak van de aarde is in gebruik voor de landbouw. Ondanks het ruimtebeslag kunnen deze landbouwgebieden in potentie wel een belangrijke bijdrage leveren aan de biodiversiteit. Echter, in de afgelopen decennia heeft een sterke intensivering van de landbouw geleid tot een wereldwijde afname van de biodiversiteit, waarbij vooral aquatische systemen onder druk kwamen te staan. Het laaggelegen agrarische landschap in de provincie Noord-Holland (Nederland) herbergt een omvangrijk vlechtwerk van sloten en kanalen, in klei- en zandgronden, en in overgebleven veengebieden. Ook deze gebieden vormen in potentie een belangrijk resevoir van biodiversiteit van met name aquatische organismen. Echter, als onderdeel van het lopende beheer worden de sloten in Noord-Holland regelmatig verstoord door de inlaat van gebiedsvreemd en mineraalrijk water, afspoeling van nutriënten, verwijdering van de waterplanten en uitbaggeren. Al deze verstoringen zijn waarschijnlijk verantwoordelijk voor een sterke daling van de diversiteit van ongewervelde zoetwaterdieren en waterplanten in de afgelopen decennia. Om een verdere afname van de biodiversiteit te voorkomen, en om kansrijke herstelmaatregelen te kunnen nemen, is het noodzakelijk om de belangrijkste stuurfactoren van deze achteruitgang te achterhalen. Daarom heb ik in dit proefschrift de volgende onderzoeksvragen geformuleerd:

Wat zijn de belangrijkste milieufactoren die de samenstelling van gemeenschappen van ongewervelde zoetwaterdieren in de sloten in Noord-Holland bepalen?

Wat zijn de mechanismen die ten grondslag liggen aan de reactie van ongewervelde waterdieren op deze milieufactoren?

Hoe kan het beheer van deze gebieden worden aangepast om de ecologische toestand te verbeteren?

Eén van de belangrijke veranderingen in de agrarische veengebieden in Noord-Holland is de toename in troebelheid van het water doordat het veen afbreekt en de daarbij gevormde slibdeeltjes zich ophopen in de sloten. In hoofdstuk 2 van dit proefschrift hebben we onderzocht hoe de veenafbraak uiteindelijk kan leiden tot een afname in de diversiteit van bodembewonende ongewervelde zoetwaterdieren. Hiertoe hebben we lange-termijn gegevens over de verspreiding van dieren en planten (verzameld door het Hoogheemraadschap Hollands Noorderkwartier in de periode 1985-2007) geanalyseerd, veldexperimenten uitgevoerd met plastic waterplanten (als kunstmatig substraat) en zelf extra bemonsteringen uitgevoerd in verschillende habitats in de sloten. Uit deze studie is gebleken dat in de afgelopen decennia de waterkwaliteit in de veensloten welliswaar (in geringe mate) is verbeterd, maar

Samenvatting

154

dat tegelijkertijd de diversiteit van zowel ondergedoken waterplanten als van de ongewervelde dieren sterk is afgenomen. Uit het experiment met de plastic planten bleek dat juist de waterplanten een belangrijk habitat vormen voor de dieren, die zich in grote dichtheden vestigenden, terwijl in de slibrijke waterbodems zonder planten nauwelijks dieren voorkwamen. Op basis van deze studie kunnen we concluderen dat de afbraak van veen en de ophoping van slib heeft geleid tot een afname van geschikte habitats (waterplanten) en een toename van ongeschikt oppervlak (slib) en daarmee dus indirect de oorzaak vormt van de waargenomen sterke afname van het aantal bodembewonende waterdieren. Hoewel de meeste ondergedoken waterplanten in de veensloten zijn verdwenen, hebben zich langs de oevers van de voedselrijke sloten nog wel emergente (uit het water omhoog groeiende) waterplanten zoals riet en lisodde kunnen handhaven. Deze oeverplanten vormen ook een potentieel belangrijke leefgebied voor waterdieren, echter in de meeste sloten worden deze planten jaarlijks weggemaaid. Om te onderzoeken wat het effect is van deze beheersmaatregelen hebben we in hoofdstuk 3 onderzocht hoe de aan- of afwezigheid van emergente waterplanten (in sloten met verschillende waterkwaliteit) van invloed is op het voorkomen van vooral aquatische insecten. Om te achterhalen wat de mechanismen zijn achter de effecten hebben we tevens gekeken naar de karakteristieke eigenschappen van de insecten (life-history strategieën) die de omgang van de specifieke soorten met hun omgeving karakteriseren. De resultaten van deze studie lieten in de eerste plaats zien dat in de sloten in voedselrijke veengebieden vooral insecten soorten voorkomen die goed aangepast zijn aan extreme milieuomstandigheden. Daarnaast resulteert de aanwezigheid van de emergente waterplanten in een toename van het aantal insecten soorten die gekenmerkt worden door hun vermogen tot snelle verspreiding. We kunnen hieruit concluderen dat de waterkwaliteit bepaalt welke soorten er binnen een bepaald gebied voorkomen, en dat de aanwezigheid van de emergente waterplanten in de individuele sloten binnen zo’n gebied bepaalt welke soorten met hun karakteristieke eigenschappen uiteindelijk wel of niet aanwezig zijn. Een dergelijk proces is ook wel bekend als ‘environmental filtering’. Naast de invloed van de aan- of afwezigheid van waterplanten is uiteraard ook de waterkwaliteit van grote invloed op de soortensamenstelling van de beesten in een sloot. Door de inlaat van gebiedsvreemd water om het waterpeil te reguleren kan de waterkwaliteit sterk veranderen in de tijd. Doordat echter veel van de parameters (indirect) aan elkaar gekoppeld zijn is het vaak lastig om te achterhalen wat nu precies de stuurfactoren zijn die de soortsamenstelling van leefgemeenschappen in een sloot bepalen. In hoofdstuk 4 hebben we onderzocht of we aan de hand van de microbiële samenstelling van de waterbodem meer inzicht kunnen verkrijgen in deze stuurfactoren. Hiertoe hebben we in 25 verschillende veensloten de waterkwaliteitsparameters

Samenvatting

155

bepaald, een inventarisatie gemaakt van de waterplanten, gekeken naar zowel de soortssamenstelling van de insecten als een analyse gemaakt van de samenstelling van de microbiële gemeenschap (door het analyseren van fosfolipide vetzuren, PLFA). Deze PLFA’s zeggen iets over de aanwezigheid van algen, cyanobacteriën, sulfaat-reducerende bacteriën (gram positieve), anaërobe bacteriën en gram-negatieve bacteriën in de waterbodem. Uit de combinatie van deze waarnemingen is gebleken dat er een relatie bestaat tussen de microbiële samenstelling in de waterbodem en het voorkomen van insectensoorten, zoals waterjuffers, kokerjuffers en dansmuggen, en dat deze relatie zowel wordt bepaald door verschillen in waterplanten als door verschillen in waterkwaliteit (bicarbonaat, sulfaat en nutrienten). Van deze waterkwaliteitsparameters is bekend dat ze gedomineerd worden door de inlaat van mineraalrijk boezemwater, en daarom kunnen we concluderen dat de inlaat van gebiedsvreemd water een indirect negatief effect heeft op de biodiversiteit van aquatische insecten in de veensloten. Omdat de inlaat van gebiedsvreemd water alleen gebeurt wanneer het waterpeil in een gebied te laag dreigt te worden, leidt deze inlaat niet alleen tot een verandering in de waterkwaliteit, maar ook tot variaties in de samenstelling van het water over de tijd. In hoofdstuk 5 hebben we gekeken hoe deze schommelingen van invloed zijn op de soortsamenstelling van ongewervelde zoetwaterdieren in verschillende typen sloten in agrarische gebieden in Noord-Holland. We hebben waterkwaliteitsgegevens van 84 sloten (zand , veen en klei gebieden) onderzocht en gekeken of we correlaties konden vinden tussen abiotische randvoorwaarden (variabiliteit en de gemiddelde waarden van de waterkwaliteitsparameters) en de diversiteit van de invertebrate fauna. Hierbij hebben we wederom niet alleen gekeken naar de taxonomische samenstelling, maar ook naar de eerder genoemde life-history strategieën van de dieren en hun vermogen zich over het gehele gebied te verspreiden. Uit deze studie is gebleken dat deze schommelingen (van met name nutrienten en de macro-ionen) inderdaad net zo belangrijk zijn voor het voorkomen van diersoorten zijn als de gemiddelde concentraties van dezelfde stoffen. Dat was voorheen alleen bekend uit onderzoek aan rivieren, die inherent grote variabiliteit kennen. De hoofdstukken 4 en 5 laten dus allebei zien hoe bepalend de inlaat van gebiedsvreemd water is voor de diversiteit van de fauna. In dit proefschrift geef ik een overzicht van een groot aantal interacties tussen biologische en abiotische componenten van de sloot-ecosystemen in agrarische gebieden in Noord-Holland. Zowel bodemsamenstelling, aan- of afwezigheid van waterplanten en de waterkwaliteit zijn direct en indirect van invloed op het voorkomen van de ongewervelde zoetwaterdieren. Een toenemende intensivering van het agrarisch landgebruik leidt tot een toenemde druk op de aquatische ecosystemen door de toevoer van nutriënten, ophoping van slib en een grotere watervraag. In de provincie Noord-Holland leidt deze

Samenvatting

156

vraag naar water tot meer inlaat van gebiedsvreemd water met de schadelijke effecten zoals die in dit proefschrift zijn beschreven. Bovendien wordt door het wegmaaien van de waterplanten (waarvan een belangrijk deel door de eutrofiering en slibophoping al zijn verdwenen) een onevenredig grote schade toegebracht aan de resterende gemeenschappen van waterdieren. Op dit moment worden veel maatregelen genomen, zoals het aanleggen van natuurvriendelijke oevers, die er toe moeten leiden dat de biodiversiteit in de sloten toeneemt. Er is echter nog weinig wetenschappelijk bewijs dat deze maatregelen ondersteunt. Het verdient aanbeveling dat bij dergelijke maatregelen gericht op het creeren van nieuwe leefgebieden voor de waterdieren, ook gelet wordt op de lokale waterkwaliteit (isoleren van de inlaat van gebiedsvreemd water) en op de ecologische verbindingszones van deze nieuwe habitats (i.v.m. kolonisatie). Gelukkig, zijn er in Noord-Holland met zijn grote aantal sloten en kanalen goede mogelijkheden om geschikte leefgebied met een grote diversiteit aan organismen te handhaven; de keuze is aan de bestuurders en beheerders om dit effectief te laten zijn.

Acknowledgements

157

Acknowledgments

This thesis would not have come to fruition without the advice and support I’ve received during my studies. Firstly I thank the Water Authority (Hoogheemraadschap Hollands Noorderkwartier) and Stichting Waterproef for funding my PhD. I thank Gert van Ee, Emile Nat, Ron van Leuken and Herman van Dam. You were the sounding boards for many of my ideas and offered good insights, local knowledge and helped me to grow by challenging me. I extend my gratitude to my committee members, Patrick Armitage, Karsten Kalbitz, Jos Verhoeven, Wilco Verberk and Piet Verdonschot – thank you for being my committee and taking the time to critique my work, I look forward to the defence. To my promoter, Wim – your patience and good nature were always appreciated. You frequently offered me pearls of wisdom. There were times you gave me clarity and at other times you kept me guessing. To my co-promoter, Harm – your spontaneous style and easy-going attitude challenged my sensibilities, I really enjoyed problem solving with you and developing ideas together. You taught me a lot and I couldn’t have wished for a better supervisor. Arie – thanks for inspiring me to go further with my work, you frequently came up with great ideas and your astute attention to detail didn’t go amiss either. To Michiel – you were my first contact at AEE and your energy and boisterous laughter made a great impression on me! Thank you for introducing me to the group and offering kind words and encouragement when I needed to hear them. Helen – I remember celebrating together the day I was appointed to the PhD position. Your gentle way and warm smile brightens my day. One of the wonderful things about the whole PhD project was the great friends I’ve made along the way. To my colleagues, Sascha, Susanne, Ciska, Bram, Marino, Vesna, Ellard, Raúl and Britt – I’m thankful to have such a fantastic group of friends to see me through the tough times and to celebrate the good ones, you made this PhD a colourful experience. We didn’t get the reputation of party department for nothing! Ale – it’s been a joy to share this experience with you. Discussing our work together was always insightful and I learnt a lot form you. Your charismatic nature and good humour helped me to find my way through the valley of (….)! To Judith – you are just starting your PhD journey, I wish you all the best for the future, I have the feeling that you will do a great job. To Coen – my good friend and ever helpful assistant, it was great fun working in the field together. You also stopped me from going crazy when everything seemed to go wrong, like that day at Naardermeer when the so called “waterproof” laptop drowned in the rain (yet again)!? To Jasper – near the end of my project you gave me some very good advice, “take your time on the discussion and enjoy it” I followed 50% of that advice and enjoyed it. To Bas you reminded me that time spent drinking beer in the sun is time well

Acknowledgements

158

spent. To Emiel – always busy yet always ready to help. Thank you for getting me on the road to using R. All the tuition you gave me and discussions we had about the data was not only helpful but enjoyable. To Chiara – you offered me and my students a lot of helped. You took me under your wing when I was wading through mud samples. You taught me so much, not just about lipid extraction but also about patience and generosity. I wish you all the best for the road ahead and hope that soon you will have the time you deserve for yourself. To Joke – thank you for helping me out and your patience when things got broken. To Casper – the boat trips in the Weerribben and Naardemeer were a lot fun, despite the biting insects. Thanks for the boat trips and sharing your knowledge of wetland vegetation with me. To my students Linda, Boris and Soraya – you were hard workers and really applied yourselves to your projects. At times you were left to your own devices, especially Soraya who started her internship shortly before I left to get married in New Zealand. Thank you all for your enthusiasm, hard work and what you taught me along the way, I wish you all the best for the road ahead. I thank Martin Meirink for his help with the extensive and convoluted HHNK datasets. To the team at Waterproef – there are so many of you who helped me during this project. I extend a special thanks to Mark, Casper, Annie, Willie, Frank, Angelique, Gerrit and (of course) Pim. Pim - when I began this project I had never identified invertebrates before, you familiarized me with the basics and gave me the push I needed to get started and soon I was addicted. At this point I would like to acknowledge all the small creatures that were sacrificed during my PhD. This work would not have been possible without the thousands who (unintentionally) gave their bodies to science. I also extend my gratitude to the Natuurmonumenten, especially Ed Zijp, Annemieke Ouwehand and Andre Timmer. Some of my fondest memories during this PhD were the times spent in the Wormer en Jisperveld and Naardermeer. I save the last words of my acknowledgements for my family, to my Mum (Judy), Dad (Michael), Anisa, Jason, Liam, Jessica, Nan and June. One of the toughest things about doing this PhD was the time spent away from you and being so very far from home. I thank you all for giving me your blessing and believing in me. And to my husband Nigel – simply said I could never have done this without you, you gave me strength when I needed it most.

159

Curriculum vitae

Merrin Whatley was born on the 23 September 1980 in Auckland New Zealand. She spent her childhood on the family farm in Waiuku, situated on the west coast of the North Island (New Zealand). The property supports a number of small streams and from a very young age Merrin was fascinated by water and the creatures she found living there. She also become aware that these aquatic environments were heavily impacted by human activity, this was evident because she kept losing her gumboots in the accumulating mud which

came from upstream market gardens. This prompted her to commence a BSc in Resource and Environmental Planning at the University of Waikato, New Zealand. After graduating Merrin spent a year managing the family farm, but she was eager to get involved in the management of inland aquatic environments. In 2005 Merrin started working at the Department of Conservation (DoC) as the Biodiversity Ranger for Freshwater in the Auckland region. This position gave her the opportunity to attend national meetings and engage in monitoring programmes for native fish and pest fish species. She also worked on public awareness campaigns to highlight key issues in the Auckland region. After two years working at DoC Merrin wanted to continue her tertiary education to further her scientific understanding of freshwater environments. At this time she also had a strong desire to travel and decided to combine travel and education. In 2007 Merrin commenced her MSc in Ecology at the VU University, Amsterdam. During her MSc she conducted research on the influence of invertebrate species diversity on microbial activity and detrital processing. This internship was conducted at the University of Amsterdam in the department of Aquatic Ecology and Ecotoxicology and it paved the way towards her PhD position, resulting in the work detailed in this dissertation. After her graduation Merrin plans on returning to New Zealand where she would like to work as an environmental consultant and specialise in the management of inland waters in agricultural landscapes. She also hopes to start a business and educational centre on the family farm where she first discovered her passion for water.

201518-os-Whatley.indd 1 02-04-14 09:22