The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 ›...
Transcript of The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 ›...
![Page 1: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/1.jpg)
The Future of British Odonata
Determining temporal range dynamics from distribution patterns and dispersal
Alyson Pavitt
MSc Conservation Science
Imperial College London, Silwood Park
A thesis submitted in partial fulfilment of the requirements for the degree of Master of
Science and the Diploma of Imperial College London.
PDF processed with CutePDF evaluation edition www.CutePDF.com
![Page 2: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/2.jpg)
© Alyson Pavitt, 2011 [email protected]
Contents
1. Introduction ..................................................................................................................................... 1
1.1. Statement of need ..................................................................................................................... 1
1.2. Range dynamics ........................................................................................................................ 2
1.3. Distribution and dispersal ....................................................................................................... 3
1.4. Aims and objectives .................................................................................................................. 4
1.5. Hypotheses ................................................................................................................................ 4
2. Background ...................................................................................................................................... 7
2.1. Analysis of range dynamics .................................................................................................... 7
2.2. Odonates as a model taxon ................................................................................................... 10
2.3. Spatial distribution patterns .................................................................................................. 12
2.4. Use of traits in understanding species distribution ........................................................... 15
3. Methods and materials ................................................................................................................. 20
3.1. Range dynamics ...................................................................................................................... 20
3.2. Patterns of distribution .......................................................................................................... 21
3.3. Trait data .................................................................................................................................. 22
3.4. Data analysis............................................................................................................................ 23
3.5. Analysis of measurement methods ...................................................................................... 24
4. Results ............................................................................................................................................. 25
4.1. Range dynamics ...................................................................................................................... 25
4.2. Patterns of distribution .......................................................................................................... 26
4.3. Morphometric analysis .......................................................................................................... 27
4.4. Dispersal capacity ................................................................................................................... 32
4.5. Representativeness of museum specimens ......................................................................... 33
5. Discussion ...................................................................................................................................... 34
5.1. Spatial dynamics ..................................................................................................................... 34
5.2. Patterns of Distribution .......................................................................................................... 36
5.3. Dispersal morphometrics ...................................................................................................... 37
5.4. Determining occupancy change ........................................................................................... 38
5.5. Suborder variation .................................................................................................................. 42
5.6. Value of historic collections ................................................................................................... 42
5.7. Recommendations for future research................................................................................. 43
6. Conclusions .................................................................................................................................... 45
References........................................................................................................................................... 46
Appendices ......................................................................................................................................... 62
![Page 3: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/3.jpg)
© Alyson Pavitt, 2011 [email protected]
Acronyms
AOO Area of Occupancy
BAP Biological Action Plan
BDS British Dragonfly Society
BRC Biological Records Centre
EOO Extent of Occupancy
IUCN International Union for Conservation of Nature
NBN National Biological Network
NHM Natural History Museum, London
T1 Time period 1 (1970-1996)
T2 Time period 2 (1997-2008)
VCA Variance Component Analysis
![Page 4: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/4.jpg)
© Alyson Pavitt, 2011 [email protected]
Figures
Page
4.1 The relationship between the fractal dimension of species distribution (D) and cell
occupancy. The change in shape between t1 and t2 is shown for each species,
displayed with the linear regression model for mean species residual D against mean
species cell occupancy (grey line). The dotted lines are species with a positive
residual D change, these species became more aggregated from t1 to t2. The solid
lines are species with a negative residual D change, and these are species which
became sparser across time.
27
4.2 Linear regression models for the relationships between (a) dry wing length and the
logs of abdominal length, (b) thoracic volume and the logs of abdominal length3, and
(c) wing aspect ratio and the logs of abdominal length2. Solid lines show the linear
regression fitted straight to the data, whilst dashed lines show the linear regression
once phylogenetic signal is controlled for.
28
4.3 Relationship between the log of wing aspect ratio and residuals of dry wing length
(controlling for phylogenetic signal) in anisopterans (filled points and solid
regression line) and zygopterans (empty points and dashed regression line).
30
4.4 Difference between the two primary methods of measuring odonate wing length.
This is based on length measurements taken from the right forewing of Coenagrion
puella (see Conrad et al. 2002; Hassall et al. 2008 for complete methodologies).
31
4.5 Distribution of occupied grid cells in the British ranges of (a) Orthopterum cancellatum
(a species increasing in occupancy) and (b) Calopteryx virgo (a species decreasing in
occupancy),for t1 (left) and t2 (right).
39
4.6 Distribution of occupied grid cells in the Britain range of Aeshna isosceles for t1 (left)
and t2 (right).
40
![Page 5: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/5.jpg)
© Alyson Pavitt, 2011 [email protected]
Tables
Page
2.1 Comparison of the two main methods used to measure species range shift.
10
2.2 Variables responsible for determining species distribution patterns, grouped
according to their influences as defined by Guisan & Thuiller (2005).
12
2.3 Comparison of methods used to measure spatial distribution pattern.
13
2.4 Selection of species traits which have been found to correlate significantly with
changes to invertebrate ranges.
16
2.5 Morphological traits used as proxies for dispersal capacity in a range of species.
18
4.1 The number of species which showed cell occupancy or range shift changes. The
binomial sign test shows whether a significantly greater proportion of species
showed a range expansion or northwards shift).
25
4.2 The number of species with more or less aggregated distributions (residual D), the
number of species changing over time (residual D change), and whether there was a
significant proportion with more aggregated or increasingly aggregated ranges
(binomial sign test).
26
4.3 The variance in dry wing length and aspect ratio accounted for by each of the nested
taxonomic scales (for n= 572 specimens).
29
4.4 The variance in dry wing length and aspect ratio accounted for by inter- (between
species and all higher taxonomic levels) and intraspecific (between and within sex)
variation. The binomial sign test shows whether a significantly greater proportion of
variation was accounted for interspecific variation.
30
4.5 The number of species with a positive change index (expansion in overall cell
occupancy from t1 to t2). The binomial proportion test shows whether there is a
significant difference in the proportion of each sub-order showing a cell occupancy
increase. Percentages given as a proportion of n.
32
4.6 Statistical outputs from Welch two sample t-tests investigating differences in wing
length between dry museum specimens and wet specimens (recently collected).
33
![Page 6: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/6.jpg)
© Alyson Pavitt, 2011 [email protected]
2
Abstract
Temporal range dynamics (range shift and change index), dispersal morphometrics
(wing length, wing aspect ratio, and thoracic volume), and distribution pattern
(residual D) were investigated in British Odonata. Initial analyses discounted range
shift from this study due to the absence of evidence that the data were showing
directional shift rather than non-directional expansion. A significant proportion of
change index (Telfer et al. 2002) was described by combining the three dispersal
traits and residual D. Species with increasing range size were those with long, broad
wings, large thoracic volumes and a more aggregated distribution. This morphology
is found in the anisopteran (dragonflies) suborder, which showed a substantially
greater increase in occupancy than the smaller and weaker zygopterans
(damselflies). In a preliminary investigation into the extant representativeness of
museum collections, there was found to be no differences in wing length with
recently caught, wet specimens.
Key words: Dragonfly, damselfly, anisoptera, zygoptera, change index, range shift, fractal
dimension
![Page 7: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/7.jpg)
© Alyson Pavitt, 2011 [email protected]
2
Acknowledgements
I thank my supervisors Dr. Nick Isaac (CEH) and Dr. Simon Leather (Imperial
College) for their comments, support and guidance; Steve Brooks (NHM) for all his
help and advice throughout the project; Gary Powney for his assistance with
statistical analyses; Dr. Christopher Hassall, Dr. Kelvin Conrad and Philip Bowles
for their kindness in making their data available to me; and to everyone else who has
supplied advice or comment on this work. I also extend thanks to the inexhaustible
support, hugs and cake-eating of SW3, and to tea, without which none of this would
have been possible.
Word count: 9,472
![Page 8: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/8.jpg)
© Alyson Pavitt, 2011 [email protected]
1
1. Introduction
1.1. Statement of need
Under the combined force of multiple environmental and anthropogenic drivers,
many species are undergoing extensive change to their global ranges. Past work has
focussed on the impact of rising global temperatures (Houghton et al. 2002), but this
is just one of the many pressures that species are under. The relative impacts of these
drivers are often species specific, limiting the potential for proactive conservation
without extensive, single-species research. Most species lack such detailed
information, and there is neither the time nor the available resources to fill these
knowledge gaps. In species with long historic records, however, it may be possible
to study long term changes in light of more general species traits, with the potential
to build a database of traits which are frequently associated with declining species
ranges. Achieving this requires the in depth analysis of multiple and disparate taxa,
with a heavy reliance on those for which substantial and long-running records
already exist.
The odonates (order Odonata) are one such taxon, with records from amateur
naturalists dating back centuries, although their value in this field has only recently
been recognised (e.g. Hickling et al. 2005; Hassall et al. 2010). Through the study of
their range dynamics and associated traits, this study will investigate the predictive
potential of dispersal morphology and distribution pattern, testing the belief that key
traits can indeed be used to predict species temporal changes. Findings of this study
will also be important for national odonate conservation by providing additional
means of determining species most at risk. Species with chronic declines, or
increasing range fragmentation, may otherwise be overlooked by conservationists
working with small snapshots of present distribution, particularly if they remain
fairly common.
![Page 9: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/9.jpg)
© Alyson Pavitt, 2011 [email protected]
2
1.2. Range dynamics
Focussing on changes to the odonates’ British distributions, this study will utilise
two different range dynamic metrics: range shift and occupancy change (see
Appendix I). Range shift is characterised by a series of low latitude extinctions and
high latitude colonisations (Malcolm et al. 2006; Thomas, 2010). Extensive research
almost exclusively attributes this to shifting climate envelopes under climate change
pressures (Elith & Leathwick, 2009). Past focus has primarily been on birds and
butterflies (Parmesan et al. 1999; Warren et al. 2001; Poyry et al. 2009), but similar
trends are increasingly being found across a wide variety of taxa (Hickling et al.
2006). Drivers of occupancy change, however, are far less certain. Although evidence
suggests that there is an indirect link with climate (Robinet & Roques, 2010), this is
only one of a suite of determinants. Other pressures include habitat loss and
fragmentation, invasive species, and over-exploitation (Gaston, 1996; Korkeamaki &
Suhonen, 2002; Hassall & Thompson, 2008). These are being counteracted in some
instances by improved habitat quality, remediation and conservation work (Willis et
al. 2009), resulting in complex effects and interactions.
Whilst regional scale studies are needed to fully understand range dynamics
(Parmesan, 1996), national or local level research can still be hugely important,
particularly in informing national conservation policy and management. Range shift
poses the most problems for localised monitoring, because it will not incorporate
both northern and southern range margins. Instead it can only be interpreted from
the uni-directional, polewards expansion of the range margin under study.
1.2.1. Study taxa
Macro-invertebrates are particularly useful for studying range dynamics because of
their short generation times (Parmesan, 1999; Thomas et al. 2004). Many also have
long running national data sets (e.g. in Britain) because their ease of collection and
charisma made them appealing to amateur naturalists. This, coupled with its
position at the dynamic margins of many of the species ranges (Wilson et al. 2004;
Hassall et al. 2010), makes Britain an ideal site for entomological range dynamics
![Page 10: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/10.jpg)
© Alyson Pavitt, 2011 [email protected]
3
study. Until recently, research in this field centred on butterfly responses (e.g. Hill et
al. 1999; Wilson et al. 2004; Willis et al. 2009), however odonates are now gaining
recognition as an order of equal, if not greater importance. Not only do odonates
encompass both aquatic and terrestrial habitats (Hickling et al. 2005), they are alos
sufficiently small in number (~6,000 species) to allow a large, representative
proportion to be studied in detail (Clausnitzer et al. 2009).
1.3. Distribution and dispersal
Species range dynamics are heavily influenced by internal distribution patterns (see
Appendix I), and the ability of individuals to disperse between patches of suitable
habitat (Hill et al. 1999). It is important to consider interactions between the two, as
landscape may be more permeable to a sparsely distributed but mobile species, than
to a more sedentary species with an aggregated distribution (Thomas et al. 1992;
McCauley, 2006). This has implications for both the colonisation of habitat beyond
current range boundaries, and the recolonisation of habitat following local
population extinction.
1.3.1 Patterns of distribution
Species are not evenly distributed across their range, but instead form unique
patterns of occupancy and absence through a complex assortment of biotic and
abiotic interactions (Wilcove et al. 1986; Fischer & Lindenmayer, 2007). Past work on
butterflies suggests that a fixed snapshot of this spatial pattern can predict the
species’ range dynamics across time, with more aggregated species showing range
expansions (Wilson et al. 2004). By applying methods used in this study to a new
taxon (Odonata), this project will investigate the value of distribution pattern as a
predictive metric of range dynamics.
1.3.2. Dispersal capacity
Because direct measures of dispersal distance are less accessible as predictors, and
are unknown for a number of species, this study will take three accepted
morphological proxies of dispersal instead: wing length, wing aspect ratio, and
![Page 11: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/11.jpg)
© Alyson Pavitt, 2011 [email protected]
4
thoracic volume (Malmqvist, 2000; Hughes et al. 2007; Meyer et al. 2008). The
relationship between wing length/aspect ratio and flight capacity are well
documented in the literature. Long wings and higher aspect ratios (narrow wings)
have long been associated with economical flight and dispersal (Conrad et al. 2002;
Rundle et al. 2007; Saino et al. 2010). Wing morphology, however, provides little
information on the energetic requirements for flight, and thus its relationship with
thoracic volume, heavily influenced by flight musculature (Pringle, 1949), is an
informative one.
1.4. Aims and objectives
The primary focus of this study is on identifying temporal changes in British
odonates, and whether they show any general trends with dispersal morphology
and distribution pattern. It also aims to investigate the value of museum collections
as sources of this morphological trait data.
With these in mind the objectives of the project are to:
1.4.1. Create an up-to-date picture of how species ranges are changing over time,
both in latitude (range shift) and overall size (occupancy change).
1.4.2. Investigate the potential for both static and dynamic measures of species
distribution pattern to predict future range changes.
1.4.3. Collate and generate dispersal trait data, and investigate their relationships
with range dynamics.
1.4.3. (a) Investigate variation in different methods of measuring morphology.
1.4.3. (b) Investigate whether traits show the greatest level of variation within or
between species.
1.4.4. Investigate how representative museum specimens are of extant populations.
1.5. Hypotheses
1.5.1. Overall, significantly more species will show a polewards shift at their
northern range margin. The range margins of all southern species will shift
![Page 12: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/12.jpg)
© Alyson Pavitt, 2011 [email protected]
5
northwards. Northern species will show no change at the polewards range margin
due to geographic barriers (e.g. British coastline).
1.5.2. Overall, significantly more species will increase their cell occupancy between
the two time periods as conditions become more suitable under shifting climate
envelopes. Northern species will be the only species to decrease in overall cell
occupancy as fewer areas become suitable for them.
1.5.3. There will be a significant positive correlation between northwards range shift
and change index as species ranges move further into Britain. There will not be a
corresponding movement at the eastern range margin, indicating a directional
polewards shift rather than non-directional range expansion.
1.5.4. Species with sparse distributions will decline in cell occupancy due to local
extinctions which were not recolonised. More aggregated species will remain stable,
or increase in cell occupancy. More aggregated species will also show a greater
northwards range shift.
1.5.5. Species becoming sparser will show a greater decline in cell occupancy as
localised extinctions are not replaced.
1.5.6. There will be a significant relationship between wing length, wing aspect ratio,
and thorax volume, and together these will describe a significant proportion of
variation in occupancy change and range shift.
1.5.7. The combination of species morphology and distribution pattern will describe
the most variance in occupancy change and range shift.
1.5.8. Wing length will not vary between different measurement methods.
1.5.9. The greatest variation in wing morphology will be seen above the level of the
species, rather than being intraspecific.
1.5.10. Wing morphology will not vary between museum specimens and wet
specimens.
1.6 Scope
This project is concerned with tracking temporal changes in the distribution of
odonates within Britain, and identifying whether these changes can be pre-empted
with an understanding of the species’ dispersal capacity and distribution pattern.
![Page 13: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/13.jpg)
© Alyson Pavitt, 2011 [email protected]
6
Two forms of temporal change will be measured; change in the number of grid cells
occupied (occupancy change) and latitudinal movement at the margins of the range
(range shift). Interspecific variation in three morphological dispersal traits (wing
length, wing aspect ratio, and thoracic volume), and in distribution pattern (Wilson
et al. 2004) will be analysed in relation to these range dynamics, both individually
and together. The morphological data, collected from Natural History Museum
(NHM) specimens, will also be compared to more recently collected wet
measurements, in order to investigate the assumption that morphometric data from
these old collections is representative of extant populations.
![Page 14: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/14.jpg)
© Alyson Pavitt, 2011 [email protected]
7
2. Background
2.1. Analysis of range dynamics
Whilst much can be learnt about a species’ ecology from the size and distribution of
its range (Gaston, 2003), this static measure is of limited use in understanding its
long term viability. Species with small ranges (geographically rare), for example,
may be stable and at optimal population size, whilst those with extensive ranges
may be rapidly declining and in need of conservation attention (Gaston, 1994). It is
therefore vital that not only the range size, but also temporal changes to that size, are
understood. This ensures that limited conservation resources are targeted at the
species with the most immediate risk.
2.1.1. Drivers
Whilst many studies investigating species range dynamics have focussed on the
escalating issue of climate driven changes (e.g. Parmesan, 1996; Parmesan & Yohe,
2003; Hickling et al. 2006; Termaat et al. 2010), it is important that this is not treated
as a default driving force of all change. Even where species ranges are shifting in
response to corresponding shifting climate envelopes (Elith & Leathwick, 2009), it is
unlikely that they are responding in isolation from other drivers. Both range shift
and occupancy change are ultimately determined by a unique combination of
internal pressures, such as habitat transformation and altered interspecific
interactions (Parmesan & Yohe, 2003; Hofmann & Mason, 2005). Some of these may
directly or indirectly result from climate change, but this is not always the case.
Some may be caused by population stochasticity generating “noisy biological data”
(Parmesan & Yohe, 2003), or result from other anthropogenic factors such as
pollution (Goffart, 2010) or land conversion. In such cases, range is limited by the
suitability of the habitat, not by temperature tolerances, and must be interpreted as
such. Indeed, Parmesan & Yohe (2003) claim the latter of these to have had more
influence on 20th century range shifts than any climate change effects.
![Page 15: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/15.jpg)
© Alyson Pavitt, 2011 [email protected]
8
2.1.2. Methods of analysis
Metrics of range change have traditionally been based on the difference in species
distribution between two separated time periods of fixed length (e.g. Hickling et al.
2006; Hassall et al. 2010). This convention has become less absolute in recent years,
however, with unequal time periods being used to account for record effort (e.g.
Termaat et al. 2010). Analyses tend to fall under one of two categories, either
measurements of changing range size (“occupancy change”) or shifts to the overall
species range.
Occupancy change
Change in a species’ habitat occupancy can be measured using a number of different
methods which are dictated by the form or quality of the input data: see Gaston &
Fuller (2009) for a summary of these. By putting a value on the actual area of a range,
area of occupancy (AOO) may provide the most information (e.g. Warren et al. 2001),
but only tends to work on the small scale (Termaat et al. 2010). For large scale
monitoring schemes such as species atlases and national databases, change needs to
be calculated from coarser measurements of grid cell occupancy (Tefler et al. 2002;
Hui, 2011). These rely on distribution maps built from binary records of species
presence or absence within a regular system of grid cells, and can therefore only
provide a measure of temporal change in the number of occupied cells at a given
scale. Other methods that have been utilised include extent of occupancy (EOO)
which maps the distance of the outer limits of the distribution (Gaston, 1996). Whilst
this can be useful (e.g. Jetz et al. 2008), it is likely to be of limited value when tracking
temporal changes, particularly in light of species conservation and management.
Whilst the value of range size and occupancy change cannot be overestimated
(Gardenfors et al. 1999), a huge variation in scale (from globally ubiquitous species to
single site endemics) means that no standard monitoring methodology has yet been
implemented (Segurado & Araujo, 2004).
![Page 16: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/16.jpg)
© Alyson Pavitt, 2011 [email protected]
9
Range shift
Because of its association with climate change, species range shift has been
extensively studied, although only a few groups (primarily birds and butterflies)
have been studied in great detail, making generalisations difficult. Recent analyses
by Hickling et al. (2006), however, found the responses of such groups to be
representative of trends seen across a range of less studied taxa. They analysed
British distribution data for 16 taxonomic groups with ranges in the south of Britain.
This included well studied species, such as birds (Aves) and butterflies
(Rhopalocera), and understudied species like millipedes (Diplopoda). Across the
taxa, 84% (of 329 species) were found to shift polewards at their range margin,
averaging 31-60 km between the two time periods studied.
There are two main methods which can be utilised to measure range shift (Table
2.1), the applicability of which vary depending upon the scale of the study and
quality of the available data. Ideally range shift analyses should consider the entire
range of a species, but in reality studies tend to focus on the more manageable local
level, and usually only at one latitudinal range margin. Whilst tracking the mean
range shift would be preferable, encompassing more individuals and taking into
consideration latitudinal clines in morphology and behaviour (Hassall et al. 2008,
2009: Table 2.1), this is unlikely to detect smaller scale changes (Shoo et al. 2006). In
light of this, and despite questions over the representativeness of the method, range
shift continues to be studied predominantly at the range margins (e.g. Hickling et al.
2006; Hassall et al. 2010).
![Page 17: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/17.jpg)
© Alyson Pavitt, 2011 [email protected]
10
Table 2.1: Comparison of the two main methods used to measure species range shift.
Marginal shift Central shift
Method
Shift is measured as change at (or
close to) the margin of the range
(Hickling et al. 2005, 2006; Hassall et al.
2010).
Shift is measured as a change to the
mean/ centre of the range (Mattila et
al. 2011).
Advantages
Detects small changes, making it
useful at the national and local level.
Provides a picture of overall range
change because it considers change to
the spread of a large number of
individuals under a range of different
conditions (Archaux, 2004).
Disadvantages
Cannot conclusively show range shift
is occurring from only a subsection of
the range (i.e. from only one margin).
Dependent on a small number of grid
cells, so is more affected by varying
recorder effort (Shoo et al. 2006).
Based on extreme sightings which
may fall outside the species’ normal
range.
Produces smaller range shift values,
so is unlikely to detect any changes at
the local level (Shoo et al. 2006).
2.2. Odonates as a model taxon
Globally there are around 6,000 known odonate species, 39 of which breed in Britain,
with a further 12 as regular migrants (Corbet & Brooks, 2008). In recent years several
migrants have established permanent colonies in Britain (Merritt et al. 1996; Parr,
2010), leading to disagreement over the exact number of “resident” species, though
the commonly accepted migrants and residents are listed in Appendix II.
![Page 18: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/18.jpg)
© Alyson Pavitt, 2011 [email protected]
11
2.2.1. Life history
Odonata is an ancient order of insects which has remained largely unchanged for
over 300 million years. The extant species are classified into two sub-orders based on
the morphology of their wings: the zygopterans (“similar wing”), small-bodied,
slender “damselflies”, and anisopterans (“dissimilar wing”), the larger bodied
“dragonflies”. Odonates are “morphologically generalised” (Corbet & Brooks, 2008)
insects, with a larval form resembling the adult in all but the presence of fully
developed wings. Almost all species have obligatory aquatic larvae before they
develop into airborne imagos.
2.2.2. A model taxon for range dynamics research
Odonates are one of the best studied insect taxa, with the British Dragonfly Society
(BDS) database holding records dating back as far as 1807 (Hassall et al. 2007).
Despite this, the impact of abiotic interactions on their spatial distribution is only just
beginning to receive attention. Within Britain, work is currently underway to
produce an up to date dragonfly atlas (due for publication by the BDS in 2013), and
the past few years have seen several publications investigating the causes and
impacts of range change in this taxa. Hassall et al. (2010) used range shift analyses to
show a northward shift in all British odonate families, predicting changes to species
distribution which could affect the reliability of odonate larvae as water quality
indicators. Hickling et al. (2005, 2006) found the same trend in data from the
Biological Records Centre (BRC) for both southerly and northerly species. The only
species to be adversely affected were two of the four northern species (Aeshna
caerulea and Leucorrhinia dubia), which showed overall range declines as their
margins move further into Scotland (Hickling et al. 2005).
Despite being less studied than lepidopterans, odonates may have the advantage as
a study group, in that there are sufficiently few species to allow most, if not all, to be
studied individually. Numbering in the few thousand means there is scope for a
more comprehensive species-level analysis, rather than relying on a small subset of
species. This has allowed for the first global assessment of an insect order to be
![Page 19: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/19.jpg)
© Alyson Pavitt, 2011 [email protected]
12
conducted. Clausnitzer et al. (2009) studied and mapped the distributions of over
25% of all known odonates (1,500 species), to identify the order’s threat level. Such
detailed analyses are not possible for groups like Lepidoptera where over 135,000
species have already been identified (Kristensen, 1999).
2.3. Spatial distribution patterns
Species ranges are not homogenous across a landscape, but are aggregated in
accordance with the suitability of the immediate environment and underlying
habitat. This depends on a wide range of factors including inter-specific interactions
such as predation, competition and parasitism (Williams et al. 2001; Begon et al. 2005;
Martinez et al. 2008), as well as interactions with the abiotic environment (Guisan &
Zimmermann, 2000; Huston, 2002; Guisan & Thuiller, 2005: Table 2.2). This
distribution pattern is a key consideration when attempting to understand the
determinants of species distribution and range changes.
Table 2.2: Variables responsible for determining species distribution patterns, grouped
according to their influences as defined by Guisan & Thuiller (2005).
The exact determinants of a species geographic distribution has long been
considered a key question in ecology (Levin, 1992), with little evidence that it will be
resolved any time soon (Hartley et al. 2004; Gelfand et al. 2006). One of the central
problems is that these variables work at different scales, and so are often not
compatible. In a recent botanical study comparing patterns at different spatial scales,
Influence Definition Examples
Regulators “factors controlling species
eco-physiology”
Temperature, soil
composition
Disturbances “perturbations affecting
environmental systems”
Natural or anthropogenic
habitat change
Resources “components that can be
assimilated by organisms”
Energy, water
![Page 20: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/20.jpg)
© Alyson Pavitt, 2011 [email protected]
13
Hartley et al. (2004) noted each scale to be driven by a different key variable. For
example, the “intermediate scale”, used in this analysis to describe national (British)
scale patterns, was determined predominantly by habitat distribution and
persistence. Meanwhile, “fine scale” patterns were found to be more dependent on
sampling consistency and intraspecific competition for space. As expected,
determinants of distribution pattern correlated at similar scales, but become
increasingly dissimilar as the difference between scales increased.
2.3.1. Methods of analysis
It has been proposed that by understanding how distribution patterns correlate with
range changes, predictions of future change could be made based on a present day
snapshot of their distribution (Wilson et al. 2004; Hui, 2011). This could have huge
implications for conservation, were specific patterns of distribution indicative of
species or populations most vulnerable to decline. Numerous methods have been
employed to provide a metric for spatial distribution pattern (Table 2.3), though
many focus on the patterns of individuals within a population, rather than at the
regional level (Davis et al. 2000).
Table 2.3: Comparison of methods used to measure spatial distribution pattern.
Nearest neighbour Ripley’s K (or L) function Scaling patterns of
occupancy (SPO)
Data
Individuals recorded as
point data within a
population, assuming
randomly selected
sample points and spatial
continuity.
Point data, assuming all
points have been selected
within a predefined study
area.
Binary point data on a grid
system based on
presence/absence data,
where distances between
sampling points are the
same (eg/ species atlas).
![Page 21: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/21.jpg)
© Alyson Pavitt, 2011 [email protected]
14
Table 2.3 cont: Comparison of methods used to measure spatial distribution pattern.
Nearest neighbour Ripley’s K (or L) function Scaling patterns of
occupancy (SPO)
Scale
Primarily distance
between individuals,
though recent
developments have
shown it to work at larger
scales (Davis et al. 2000).
Can describe patterns
between individuals at
multiple scales (e.g.
Wiegand & Moloney,
2004).
Can describe patterns at
multiple scales, but works
particularly well at larger
landscape scales (e.g.
Wilson et al. 2004)
Methodology
Compare the distance
between closest data
points with the expected
distance to (a) just the
nearest neighbour (Clark
& Evans, 1954) or (b)
every other point (e.g.
David et al. 2000).
Sums records within a
given radial distance
around each data point
(Ripley, 1977).
Compare summed
occupancy at two different,
nested scales (e.g. Wilson
et al. 2004).
Results
Identifies whether a
population is clumped,
uniform or randomly
distributed. Populations
will show uniform
distributions when all
individuals are a similar
distance apart (e.g. Smith
& Murphy, 1982).
Populations with
aggregated distributions
will have more occupied
cells in close proximity to
the central data point.
Sparse or randomly
distributed species will
have similar numbers of
records at all distances
from it (Dixon, 2002).
Relative to occupancy at
the fine (small) grid cell
scale, species with
aggregated distributions
show a lower occupancy at
the coarse (large) grid cell
scale, than those which are
sparsely distributed.
![Page 22: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/22.jpg)
© Alyson Pavitt, 2011 [email protected]
15
Whilst the nearest neighbour method and Ripley’s K function serve important roles
in understanding the distribution patterns of individuals within a population, this
cannot readily be scaled up for landscape level studies of range dynamics. In order
to generate a meaningful measure of species distribution pattern, extensive data sets
across large spatial scales are needed (e.g. species atlas and national surveys). Data is
usually presented in the form of binary values of grid cell occupancy (presence/
absence), rather than treating each individual sighting as its own data point. The
limited amount of data available at this scale means that these finer analyses are not
applicable, instead, coarser methods like SPO need to be utilised (e.g. Wilson et al.
2004).
2.4. Use of traits in understanding species distribution
It is possible to consider each species as a system comprising of a unique set of trait
variables (see Table 2.4). In understanding the relationships between these traits and
species distribution, it may be possible to predict how a species (or species with a
similar set of traits) will respond to change. In an ideal situation, each species would
be studied in comprehensive detail, however limited resources make this impossible.
If traits are found to significantly correlate with distribution change, as is suggested
by numerous studies (Dukes & Mooney, 1999; Hill et al. 1999; Holt, 2003; Rundle et
al. 2007; Poyry et al. 2009; Termaat et al. 2010), then they have the potential to predict
under-studied species most at risk. This will allow for better management and the
focussing of limited conservation resources. They may also be used to roughly
predict future responses to climate change events, enabling a proactive conservation
management approach.
![Page 23: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/23.jpg)
© Alyson Pavitt, 2011 [email protected]
16
Table 2.4: Selection of species traits which have been found to correlate significantly with
changes to invertebrate ranges.
Trait Reference
Morp
hological traits
Body size Body size positively correlated with
dispersal and range shift in odonates.
Angelibert &
Giani (2003)
Wing aspect
ratio
Aspect ratio was positively correlated
with dispersal capacity in odonates.
Hassall et al.
(2008)
Wing: body
length ratio
Wing: body length ratio was positively
correlated with range size in odonates.
Rundle et al.
(2007)
Wing length/
span
Wing length/ span were positively
correlated with dispersal, range expansion
and range size in butterflies and odonates.
Rundle et al.
(2007); Poyry et
al. (2009)
Wing loading Wing loading negatively correlated with
dispersal capacity in odonates.
Hassall et al.
(2008)
Eco
logical traits
Agricultural
tolerance
Butterflies found to tolerate agricultural
land showed greater range size and
expansion.
Poyry et al.
(2009)
Breeding habitat Butterflies breeding in the forest edge
showed a greater northwards expansion
than those in open grasslands.
Poyry et al.
(2009)
Diet breadth Food plant number was positively
correlated with range size in butterflies.
Slove & Janz
(2011)
![Page 24: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/24.jpg)
© Alyson Pavitt, 2011 [email protected]
17
Table 2.4 cont: Selection of species traits which have been found to correlate significantly with
changes to invertebrate ranges.
Trait Reference
Eco
logical traits
Distribution
size
Distribution size was positively correlated
with range expansion in butterflies.
Poyry et al.
(2009)
Growth rate Growth rate was positively correlated with
range expansion in butterflies.
Bale et al. (2002)
Habitat
preference
Beetle species in coniferous forests had
larger ranges than those in broadleaf.
Species of odonate in grasslands and lotic
waterways had much smaller ranges, and
range shifts than in other habitats.
Bense (1995):
Clausnitzer et al.
(2009)
Habitat
tolerance
There was a positive correlation between the
number of habitats occupied and range shift
in odonates.
Hill et al. (1999)
Larval host
plant
Species of butterfly found on woody plants
showed a greater northwards expansion
than those feeding on herbaceous plants or
grasses.
Poyry et al.
(2009)
Mobility Dispersal capacity was positively correlated
with range shift in odonates.
Hill et al. (1999)
Threat status
(red list)
Non-threatened species of butterfly showed
poleward shifting ranges, whilst threatened
(red list) species remained stationary.
Mattila et al.
(2011)
![Page 25: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/25.jpg)
© Alyson Pavitt, 2011 [email protected]
18
2.4.1. Dispersal ability
It is important to consider the capacity for individuals to disperse through the
landscape, particularly in light of growing threats such as climate change. Actual
measures of dispersal distance are rare however, so morphological traits tend to be
relied upon as proxies (e.g. Table 2.5).
Table 2.5: Morphological traits used as proxies for dispersal capacity in a range of species.
Morphology Dispersal capacity Examples
Aspect ratio (wing shape) Higher aspect ratio (narrow
wings) is associated with
migratory species with long
dispersal distances.
Odonates (Conrad et
al. 2002); Birds (Saino
et al. 2010)
Body size (proxies inc.
abdomen length and wing
length; Cordero, 1994)
Larger species have longer
(odonates & multiple species)
or shorter (butterflies)
dispersal distances.
Butterflies (Mattila et
al. 2011); Multiple
species (Jenkins et al.
2007); Odonates (Allen
et al. 2010)
Wing length Long wings are associated
with migratory species with
long dispersal distances.
Birds (Rayner, 1988);
Odonates (Conrad et
al. 2002)
2.4.2. Problems with the use of traits
With a few exceptions (e.g. Hassall & Thompson, 2008; Hassall et al. 2008, 2009) most
invertebrate studies have focussed on butterflies. This large bias means that many
conclusions drawn are largely based on the work from one or two taxa, and thus will
be hugely influenced by taxon-specific characteristics. This emphasises a need for
further research in a more representative range of taxa since many traits, particularly
those focussed on the morphology of the species, are likely to be fairly taxon specific.
For example, wing aspect ratio (the length/ width of the wing) has been found to
![Page 26: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/26.jpg)
© Alyson Pavitt, 2011 [email protected]
19
correlate positively with dispersal capacity in odonates (Hassall et al. 2008), but
negatively in some butterfly studies (Hill et al. 1999). Body size is another such
example. Recent research on one of the smaller British odonates, Ischnura pumilio,
suggests a positive relationship between species size, dispersal distance and range
shift (Allen et al. 2010). Work on butterflies suggests that the opposite is true, with
larger species being more prone to range contractions (Mattila et al. 2011).
Due to the heritable nature of species traits, their use can be statistically invalidated
if species relatedness isn’t considered and controlled for. One of the core
assumptions for statistical analysis is that data points are independent from one
another; comparing species which share a common ancestor breaches this
assumption. Furthermore, some species will be monophyletic to the exclusion of
others. This problem has been dealt with in a number of ways, the commonest of
which is through the use of independent contrast scores (Rundle et al. 2009),
generated from comparative analyses such as Comparative Analysis by Independent
Contrast (CAIC) (Purvis & Rambaut, 1995). This allows taxa to be compared
independent of their phylogenetic relatedness.
Although frequently used in the literature, direct body measurements should be
treated with caution because of the huge number of other variables which confound
them. Because the body size of individual insects is so heavily influenced by
variables such as temperature and diet (Ikeya et al. 2002; Nijhout et al. 2006; Parker &
Johnston, 2006), even minor variations in conditions may have a noticeable impact.
Individual physiological plasticity is not the only issue; whilst wings remain largely
unchanged, body size will vary substantially between “wet” and dry specimens
because of the water content of the body. Indeed, it has been estimated that insects
can lose up to 20% of their length during post-senescence drying (Schoener & Janzer,
1968). Subsequently, the conditions and duration of storage will have a significant
bearing on measurements obtained. Body mass measurements are also problematic,
particularly in species which eat as adults, as they will be highly dependent on the
food availability and time from emergence (Johansson, 2003).
![Page 27: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/27.jpg)
© Alyson Pavitt, 2011 [email protected]
20
3. Methods and materials
3.1. Range dynamics
3.1.1. Data sources
Data on species’ distributions within Britain were obtained from National Biological
Network (NBN) Ordnance survey maps of cell occupancy (for n=37 species),
primarily from records held with the BDS. Cell occupancy was grouped into two
time periods of approximately equal record size: 1 Jan. 1970- 31 Dec. 1996 (t1: n=
309,830 records), and 1 Jan. 1997- 31 Dec. 2008 (t2: n= 312,571 records). Cell
occupancy values were taken for 10 x 10 km grid cells at t1 (t1_10) and t2 (t2_10), and
for 100 x 100 km grid cells at t1 (t1_100) and t2 (t2_100) (see Appendix I).
3.1.2. Change index
Occupancy change was measured using the “change index” method (see Appendix
I) developed by Telfer et al. (2002). T1 and t2 (at 10 km scale) were converted to
proportions of the total survey area, logit transformed, and a linear regression model
was fitted. Species change index was taken from standardised residuals of this
model (see Telfer et al. 2002 and Telfer, 2003 for complete methodology). This
method assumes that recorders report all species within each grid cell equally, and
controls for the effects of variation in geographic coverage and recorder effort by
restricting analyses to grid cells recorded in both time periods.
3.1.3. Range shift
Using records of cell occupancy, range shift was calculated to investigate the
directionality of species range movements in response to climate change. Range shift
was quantified as the difference in the northern range margin between t1 and t2 on
the 10 km scale. For each time period, the mean northing was taken from the species’
top ten occupied grid cells within Britain, and the distance (in km) between them
was used to generate overall range shift (Warren et al. 2001; Hickling et al. 2005,
2006). This is the most frequently used range shift metric (Thomas & Lennon, 1999;
Warren et al. 2001; Hickling et al. 2005, 2006; Hassall et al. 2010) because it only
![Page 28: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/28.jpg)
© Alyson Pavitt, 2011 [email protected]
21
requires 20 grid cells to be occupied at each time period (i.e. is not hugely data
hungry). As with the change index method, it also accounts for variation in
geography and recorder effort by using only grid cells recorded in both time periods.
3.2. Patterns of distribution
Work on butterflies suggests that a species distribution pattern can be indicative of
its temporal range dynamics (Wilson et al. 2004). This could have important
implications for conservation if a stationary snapshot can be used to interpret species
trends without several decades’ worth of distribution data.
A metric for distribution pattern (fractal dimension) was calculated using methods
derived from Wilson et al. (2004). “Fractal dimension” (D) was defined as “a
descriptive measurement of spatial aggregation over a narrow range of scales”
(Wilson et al. 2004), rather than being a measurement of the true mathematical fractal
dimension at infinite scales. For each species the number of occupied cells at t1_10
and t1_100 were individually summed and converted to AOO. The log10 of cell side
length (a measure of scale) was plotted against the log of AOO, and D was calculated
from the resulting slope using the formula:
D= 2-(y2-y1)/(x2-x1)
where x is the log10 of the cell side and y is the log of AOO. D2 was calculated in the
same manner using the number of occupied cells at t2_10 and t2_100.
These were standardised for cell occupancy by taking residuals from D and D2
regressed against the logs of t1_10 and t2_10 respectively. Positive values were more
aggregated than expected for the relative occupancy, whilst negative values were
sparser. Temporal change in these residuals was calculated using the formula:
Residual D change= (residual D2- residual D)/log(t2_10/t1_10)
![Page 29: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/29.jpg)
© Alyson Pavitt, 2011 [email protected]
22
3.3. Trait data
3.3.1. Data sources
Trait data was obtained from several sources (listed in Appendix I). Abdominal
lengths were obtained from a meta-analysis of the field guide literature (Bowles,
2010). Measurements of wing length (using the computer package ImageJ; Rasband,
1997-2007) were obtained for Calopteryx splendens (n=129), Coenagrion puella (n=193),
Erythromma najas (n=123), and Pyrrhosoma nymphula (n=444) (Hassall, pers.comm.); for
complete collection methods refer to Hassall et al. (2008, 2009). Additional wing
lengths (using calipers) were obtained for C. puella (n=770) (Conrad, pers.comm) to
enable a comparison of measurement methods; see Conrad et al. (2002) for complete
methods.
3.3.2. Morphometrics
In addition to those collected from the literature, morphological measurements were
taken from dry specimens held in NHM collections. This enabled conclusions to be
drawn on the representativeness of museum specimens to extant populations.
The dry length and width (at the widest point) of each wing, and the three
dimensions of the thorax, were taken from specimens (for n=37 species: ♀=8, ♂=8 for
each species) using digital callipers (0.1 mm precision). Specimens which had
incomplete labels, damaged wings or thorax, or were recorded as “bred” were
omitted. One specimen for each species was re-measured to estimate measurement
error. Given wing symmetry and the significant correlation between fore- and
hindwings in NHM specimens (Appendix III), the left hindwing will be used across
this study as a proxy for all wings. Hindwings are preferable as their shape shows
greater intraspecific variation, and is thus more likely to inform species ecology.
3.3.3. Controlling for allometry
Direct body measurements are unlikely to be independent of overall body size. To
identify these allometric relationships, the three dispersal morphometrics were
regressed against a specific abdomen length measure: dry wing length against
![Page 30: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/30.jpg)
© Alyson Pavitt, 2011 [email protected]
23
abdomen length, aspect ratio against abdomen length2 and thoracic volume against
abdomen length3. Where strong correlations were identified, regression residuals
were used in the analyses.
3.3.4. Sample size
A power analysis was used to determine the minimum number of specimens which
needed to be measured in order to generate a representative sample size. The
analysis was based on four species obtained from previous research (see section 3.3.1
Hassall pers. comm.), and used the formula:
n= (z2 x σ2) / k2
where z is the error, taken at 1.96 at 95% confidence interval; σ is the standard
deviation; and k is the margin of error, defined as the difference between the actual
mean and the estimated mean.
k= µ x p
where µ is the mean of the sample data; and p is precision, the accepted variance of
the data around the mean (+/- 5% in this study) (Hayek & Buzas, 1997).
3.4. Data analysis
3.4.1. Controlling for phylogenetic signal
Many statistical analyses assume independence between data points, however this
assumption doesn’t hold true in across-species analyses because of varying degrees
of phylogenetic relatedness. Given that some species will be monophyletic to the
exclusion of others, this also means that shared trait properties may only exist
because of a close shared ancestry.
To ensure data point independence, this study controlled for phylogenetic signal
using the package “caper” in R 2.11 (Orme et al. 2011). Unless otherwise stated,
![Page 31: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/31.jpg)
© Alyson Pavitt, 2011 [email protected]
24
phylogenetic signal was controlled for in all analyses. A phylogenetic tree
(Appendix IV) was compiled for this analysis in FigTree v1.3.1 (Rambaut &
Drummond, 2009) from several sources (Artiss et al. 2001; Von Ellenrieder, 2002;
Saux et al. 2003; Fleck et al. 2008; Pilgrim & Von Dohlen, 2008; Dumont et al. 2010).
3.4.2. Variance component analysis (VCA)
Variance in residual dry wing length and aspect ratio were analysed across a
number of nested taxonomic scales using VCA. The levels used in increasing order
were: sex, species, genus, family and suborder. This was carried out in R 2.11 (R
Development Core Team, 2011) using the code:
NHM.varcomp<-varcomp(lme(value~1,random=~1|
suborder/family/genus/species/Sex),1)
where “value” is trait of interest; in this case, residual dry wing length and wing
aspect ratio. This required the installation of “nlme” and “ape” libraries (Paradis et
al. 2004; R Development Team, 2011). These analyses did not control for
phylogenetic signal.
3.5. Analysis of measurement methods
Comparisons were drawn between two different measurement methods for
Coenagrion puella right forewing length using a Welch two sample t-test. The two
data sets were collected using the computer software package ImageJ (Hassall et al.
2008) and calipers to 0.1 mm precision (Conrad et al. 2002).
A subset of NHM specimens (♂=3, ♀=3) was selected at random from the four
species noted in Section 3.1.1. Measurements were made to the first cross vein of the
left fore- and hindwings, allowing direct comparison with the wet specimen data
from Hassall et al. (2008, 2009). Differences in length between the two data sets were
investigated using Welch two sample t-tests, enabling conclusions to be drawn on
the representativeness of dried museum specimens.
![Page 32: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/32.jpg)
© Alyson Pavitt, 2011 [email protected]
25
4. Results
4.1. Range dynamics
This study found there to be an even split between the number of species showing a
decline in grid cell occupancy (n=19) and those showing an increase (n=18).
Following the removal of Aeshna isosceles and Coenagrion hastulatum which had
insufficient cell occupancies to comply with the minimum requirements of the range
shift metric (>20 grid cells: Hickling et al. 2006), there was also no significant
difference in the number of species moving north or south (Table 4.1). Even after
removing northern and widespread species which may be restricted by geographic
barriers, there was still not a significant proportion of species expanding their
ranges, or moving polewards at their northern extent (Table 4.1).
Table 4.1: The number of species which showed cell occupancy or range shift changes. The
binomial sign test shows whether a significantly greater proportion of species showed a range
expansion or northwards shift.
All Southern Northern Widespread
n species 37 22 4 11
Change
index
Increase 18 (49%) 14 (64%) 1 (25%) 3 (27%)
Decrease 19 (51%) 8 (36%) 3 (75%) 8 (73%)
Sign test: p-value 1 0.29 0.63 0.23
n species 35 21 3 11
Range
shift
Northwards 18 (52%) 9 (43%) 1 (33%) 8 (73%)
Southwards 15 (43%) 12 (57%) 2 (67%) 1 (9%)
Sign test: p-value 1 0.66 1 0.23
Although northwards and eastwards range shift did not strongly correlation (f=0.46,
df=33, p=0.64), a significant proportion of species showed corresponding outwards
(north and east shift) or inwards (south and west shift) movements at the north and
east range margins (binomial sign test: p=0.029). Furthermore, change index shows a
![Page 33: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/33.jpg)
© Alyson Pavitt, 2011 [email protected]
26
strong positive correlate with both northwards and eastwards range shift (f=7.90,
df=33, p=0.0016 and f=5.58, df=33, p=0.0082 respectively). This was sufficient to
question the degree to which the range “shift” metric was actually recording range
shift, and not just non-directional, marginal range expansion. In light of this doubt,
range shift was discounted from subsequent analyses and this study instead focused
on change index as the primary measurement of range dynamics.
4.2. Patterns of distribution
Controlling for cell occupancy at t1, a significant proportion of species were found to
be sparser than expected (Table 4.2). When considered across time, however, species
were equally likely to become more aggregated as they were to become sparser.
Table 4.2: The number of species with more or less aggregated distribution (residual D), the
number of species changing over time (residual D change), and whether there was a
significant proportion with more aggregated or increasingly aggregated ranges (binomial
sign test).
All
species
(More)
Aggregated
(More)
sparse
Sign test:
p value
Residual D 37 12 (32%) 25 (68%) 0.047
Residual D change 37 18 (49%) 19 (51%) 1
It was hypothesised that species with sparser distributions (negative residual D)
would decline in cell occupancy over time (negative change index), but this was not
found to be the case (f=0.77, df=35, p=0.47). Neither was there a correlation between
residual D change and change index (f=0.11, df=35, p=0.90), however a significant
proportion of the species did show a positive relationship between the two (Figure
4.1; binomial sign test: p<0.001).
![Page 34: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/34.jpg)
© Alyson Pavitt, 2011 [email protected]
27
4.3. Morphometric analysis
4.3.1. Data independence
Linear regressions in the presence and absence of a phylogenetic signal control are
shown for the three dispersal traits and their respect abdomen length metrics in
Figure 4.2. The significant correlations for dry wing length and thoracic volume
against their respective abdomen metrics justify the use of residuals (f=95.74, df=35,
p<0.001 and f=41.46, df=35, p<0.001 respectively; Appendix V). Wing aspect ratio
showed no correlation with abdomen length2 once phylogenetic signal was
2 3 4 5 6 7 8
0.4
0.6
0.8
1.0
1.2
1.4
1.6
log(cell occupancy)
D
Figure 4.1: The relationship between the fractal dimension of species distribution (D) and
cell occupancy. The residual D change between t1 and t2 is shown for each species,
displayed with the linear regression model for mean species D against mean species cell
occupancy (fine black line). The dotted lines are species with a positive residual D change,
these species became more aggregated from t1 to t2. The solid lines are species with a
negative residual D change; these became sparser across time.
![Page 35: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/35.jpg)
© Alyson Pavitt, 2011 [email protected]
28
controlled for (f=0.018, df=35, p=0.99; Appendix V). This indicates that aspect ratio
does not show allometric scaling, and thus does not need to be transformed further.
4.3.2. Body morphometrics
Residuals of dry wing length and thoracic volume were shown to be significantly
correlated (f=137.30, df=35, p<0.001); species with relatively longer wings also had
relatively larger thoraces. Considered together, these two traits also showed a
(a) (b)
3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4
2.8
3.0
3.2
3.4
3.6
3.8
4.0
log(abdomen length (mm))
log(d
ry w
ing length
(m
m))
9 10 11 12 13
3
4
5
6
7
8
log(abdomen length^3 (mm3))
log(t
hora
cic
volu
me (
mm
3))
6.0 6.5 7.0 7.5 8.0 8.5
0.8
1.0
1.2
1.4
1.6
1.8
log(abdomen length^2 (mm2))
log(w
ing a
spect
ratio(m
m2))
Figure 4.2: Linear regression models for the relationships between (a) dry wing length and
the logs of abdominal length, (b) thoracic volume and the logs of abdominal length3, and
(c) wing aspect ratio and the logs of abdominal length2. Solid lines show the linear
regression fitted straight to the data, whilst dashed lines show the linear regression once
phylogenetic signal is controlled for.
(c)
![Page 36: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/36.jpg)
© Alyson Pavitt, 2011 [email protected]
29
significant negative relationship with wing aspect ratio (f=6.28, df=33, p<0.001).
Controlling for both phylogenetic and allometric signals, this study found species
with longer wings and larger thoraces to have significantly broader (lower aspect
ratio) wings.
4.3.3. Level of morphometric variance
Whilst this study incorporates a control for phylogenetic signal, it is important to
ensure that the assumption of negligible intra-specific trait variation holds true in
order to justify across-species comparisons. VCAs for residual dry wing length and
wing aspect ratio both show most variation at the level of the suborder (Table 4.3).
Table 4.3: The variance in dry wing length and aspect ratio accounted for by each of the
nested taxonomic scales (for n= 572 specimens).
% variance accounted for by taxonomic scale
Taxonomic level Residual dry wing length Wing aspect ratio
Suborder 36.88 65.60
Family 33.61 24.94
Genus 13.53 3.28
Species 6.42 1.10
Sex 4.53 1.60
Within 5.03 3.46
Overall, significantly more variation is accounted for at, or above, the level of the
species (see Table 4.4: sign test), with only 9.56% (residual dry wing length) and
5.06% (aspect ratio) accounted for at an intraspecific scale.
![Page 37: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/37.jpg)
© Alyson Pavitt, 2011 [email protected]
30
Table 4.4: The variance in dry wing length and aspect ratio accounted for by inter- (between
species and all higher taxonomic levels) and intraspecific (between and within sex) variation.
The binomial sign test shows whether a significantly greater proportion of variation was
accounted for interspecific variation.
% variance accounted for by taxonomic scale
Level of variation Residual dry wing length Wing aspect ratio
Interspecific 90.44 94.92
Intraspecific 9.56 5.06
Sign test: p-value <0.001 <0.001
Differences between suborders account for the majority of the variation seen in wing
aspect ratio (65.6%), with anisopteran wings significantly broader (lower aspect
ratio) than those in zygopterans (t test: t=-9.15, df=16, p<0.001). For example,
regressions of aspect ratio against residual dry wing length show very different
trends between the suborders (Figure 4.3: zygoptera: f=15.22, df=13, p<0.001;
anisoptera: f=1.45, df=20, p=0.26).
-0.2 0.0 0.2 0.4 0.6
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
residual dry wing length
log
(win
g a
sp
ect
ratio
)
Figure 4.3: Relationship between the log of wing aspect ratio and residuals of dry wing
length (controlling for phylogenetic signal) in anisopterans (filled points and solid
regression line) and zygopterans (empty points and dashed regression line).
![Page 38: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/38.jpg)
© Alyson Pavitt, 2011 [email protected]
31
4.3.4. Measurement errors
NHM measurements were accurate to an average of 0.12 mm. Measurements taken
from the thorax were less accurate than those from the wings; with mean accuracies
of 0.20 mm and 0.086 mm respectively. Thoracic depth showed the greatest
variation, with a mean difference of 0.28 mm across all species.
When incorporating data from numerous sources, the method of collection can also
have implications for the accuracy and comparability of the data. Within this study,
absolute wing length was shown to vary significantly between the two predominant
methods of measuring odonate wings (Figure 4.4: t test: t= 46.38, df= 355, p<0.001).
Figure 4.4: Difference between the two primary methods of measuring odonate wing
length. This is based on length measurements taken from the right forewing of Coenagrion
puella (see Conrad et al. 2002; Hassall et al. 2008 for complete methodologies).
Calipers ImageJ
16
18
20
22
24
26
measurement method
rig
ht
fore
win
g le
ng
th (
mm
)
![Page 39: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/39.jpg)
© Alyson Pavitt, 2011 [email protected]
32
4.4. Dispersal capacity
Dispersal traits and their interactions were not sufficient to explain a significant
amount of the variation seen in change index (f=0.93, df=10, p=0.53). The inclusion
of residual D, however, significantly increased the explanatory power of the model
(f=3.244, df=21, p=0.0064). Following model simplification, the minimum adequate
model included all dispersal traits and residual D, as well as the interactions
between residual dry wing length, residual thoracic volume and residual D (f=6.30,
df=28, p<0.001). This described 54% of the variation seen in change index.
This study found species with longer, broader wings, larger thoraces, and relatively
aggregated distributions to be showing the greatest range expansions. This
morphology is most commonly seen in anisopterans, which have significantly longer
wings (t test: t=-4.048, df=26, p<0.001), larger thorax volumes (t test: t=-2.16, df=21,
p=0.043), and broader wings (t test: t=-9.15, df=16, p<0.001) than the zygopterans.
The positive link between anisopteran-like morphology and range expansion is
further reflected in the substantially higher proportion of anisopterans expanding
their ranges (Table 4.5: 64% compared to 27% of anisopterans).
Table 4.5: The number of species with a positive change index (expansion in overall cell
occupancy from t1 to t2). The binomial proportion test shows whether there is a significant
difference in the proportion of each sub-order showing a cell occupancy increase. Percentages
are given as a proportion of n.
Species n species Expanding
All 37 18 (49%)
Anisoptera 22 14 (64%)
Zygoptera 15 4 (27%)
Prop.test: p-value 0.061
![Page 40: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/40.jpg)
© Alyson Pavitt, 2011 [email protected]
33
4.5. Representativeness of museum specimens
There was no significant difference in wing length between museum specimens (n=6
for the species in Section 3.1.1) and measurements from wet field specimens (Table
4.6; Appendix VI).
Table 4.6: Statistical outputs from Welch two sample t-tests investigating differences in
wing length between dry museum specimens (NHM) and wet specimens (recently collected).
Species
Forewing Hindwing
T df P t df p
Calopteryx splendens -0.94 5 0.39 -0.89 5 0.41
Coenagrion puella -1.93 5 0.11 -0.11 5 0.92
Erythromma najas 0.86 5 0.43 -1.47 5 0.2
Pyrrhosoma nymphula 0.041 5 0.97 1.25 5 0.26
![Page 41: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/41.jpg)
© Alyson Pavitt, 2011 [email protected]
34
5. Discussion
5.1. Spatial dynamics
Past work has shown most British odonate species to be responding positively to
climate change, exhibiting widespread range expansions and poleward shifts (e.g.
Hickling et al. 2005, 2006; Hassall et al. 2010). This study was unable to substantiate
either of these trends (Table 4.1). Furthermore, findings cast doubt on whether the
range shift metrics can be considered exclusively representative of shifting species
range, independent of general occupancy change.
Disparity between these and previous findings may arise from differences in the
time periods covered. Whilst this study encompasses all years from 1970-2008
(622,401 records), past studies such as Hickling et al. (2005, 2006) have used two
short time periods of equal length set several years apart. The use of two unequally
sized, adjacent time periods is not extensive in the literature, despite having some
advantages over the conventional method. A longer first time period can account for
lower rates of recording in earlier years (Hill et al. 2002; Wilson et al. 2004; Matilla et
al. 2008; Termaat et al. 2010), whilst consecutive time periods ensures that all data is
utilised (Lenoir et al. 2008; Termaat et al. 2010).
5.1.1. Range shift
Range shift was measured in accordance with methods from previous odonate range
studies (Hickling et al. 2005). Whilst working with extreme records of presence (ten
north-most occupied cells) risks generating trends that do not represent the range as
a whole, this method was less data “hungry” than viable alternatives. Hassall &
Thompson (2010) proposed basing shift on changes at the 95th quantile to account for
extreme sightings, but this requires species to occupy over 45 grid cells at each time
period. Under this proposed method, around a quarter of species (9/37) would have
been excluded, as opposed to the two which this study removed.
![Page 42: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/42.jpg)
© Alyson Pavitt, 2011 [email protected]
35
When studies focus on a subsection of the range margin, shift can only be inferred
from the latitudinal dynamics of this area. Most studies, however, fail to consider
corresponding longitudinal dynamics, which may lead to premature conclusions
that latitudinal movements reflect directional range shift. Whilst there was no
correlation in the size of northwards and eastwards range shift, a significant
proportion of species were found to move outwards (north and east) or inwards
(south and west) at both north and east range margins (p=0.029), and both directions
showed strong positive correlations with change index. Whilst it may be that species
are just exhibiting corresponding latitudinal and longitudinal shifts, the significant
number of species showing this trend suggests that the two are related, even though
the distances moved are not proportional. This lack of correlation between north and
east range shift could be explained by factors such as unequal dispersal barriers (e.g.
British coastline). Apparent non-directionality in range margin movement was
sufficient to cast doubt on the representativeness of “range shift” as an unambiguous
measure of shift in a species range. It was for this reason that the study focussed
predominantly on changes in range occupancy.
5.1.2. Change index
Changes in occupancy depend on a species-specific combination of factors, of which
climate change is just one. This makes any conclusions about the relative impacts of
different drivers, including climate change, difficult to interpret. For example,
population contractions or declines may result from habitat destruction or
fragmentation, as is likely to be the case for the declining Coenagrion mercuriale which
has seen significant habitat loss in recent years (JNCC, 2007). Meanwhile, the
migration of species into previously occupied habitats could result from improving
water quality (Gibbins & Moxton, 1998; Environment Agency, 2008), an increase in
habitat (e.g. garden ponds and gravel pits: Mill et al. 2010), or the adoption of aquatic
pesticide buffer zones around waterways. The introduction of targeted habitat
restoration schemes, such as that for Aeshna isosceles, also corresponds with an
increase in cell occupancy.
![Page 43: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/43.jpg)
© Alyson Pavitt, 2011 [email protected]
36
Whilst the drivers themselves may be unknown, what can be drawn from range
change metrics is an indication of how species are responding to the combined
pressures they are under. Although this limits potential for proactive conservation,
understanding how species are changing is still important to ongoing management.
Indeed, changes to a species’ overall occupancy is considered so indicative of species
threat and stability that declining AOO is one of the criteria used by the
International Union for Conservation of Nature (IUCN) in their species Red List
(Gardenfors et al. 1999; Daguet et al. 2008).
5.2. Patterns of distribution
Significantly more species (68%) were sparser than expected for a given cell
occupancy. It can be speculated that species freshwater habitat preferences may
account for some of this, with substantially more species found in discrete lentic
(“still”) systems than lotic (“flowing”) systems (Aguilar et al. 1986; Brooks, 2005).
Rivers and streams (lotic) provide a network of permanent or semi-permanent
dispersal corridors linking neighbouring grid cells, whilst lentic waterways (e.g.
lakes and ponds) are often more geographically isolated and ephemeral (Angelibert
& Giani, 2003; Hof et al. 2006).
Lotic waterways function as the predominant dispersal pathway for lotic species
(Ward & Mill, 2007; Chaput-Bardy et al. 2008), mostly through imago dispersal, but
this is also likely to be enhanced by some nymphal drift (Beukema, 2002; Chaput-
Bardy et al. 2008). Lentic species, however, must rely wholly on terrestrial imago
dispersal, and are therefore more likely to establish local metapopulations around
key source sites. Whilst these lentic species are considered better dispersers on the
whole (Corbet, 1999; Clausnitzer et al. 2009), the isolation and unpredictability of
potentially suitable habitat may be such, that there is limited capacity for the
establishment of new sustainable populations far from a dispersal source. In support
of this, of the six obligatory lotic species recognised in this study (Calopteryx
splendens, Calopteryx virgo, Cordulegaster boltonii, Gomphus vulgatissimus, Platycnemis
![Page 44: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/44.jpg)
© Alyson Pavitt, 2011 [email protected]
37
pennipes and Coenagrion mercuriale) (Bowles, 2010), the only one to be more sparsely
distributed than expected was C. mercuriale. This is unsurprising given its rarity and
designation as a UK BAP (Biological Action Plan) species threatened by “isolation
and habitat scarcity” (JNCC, 2007; BDS, 2011). Whilst no conclusions can be drawn
without further investigation, this does suggest that when cell occupancy is
controlled for, lotic species are more likely to display aggregated distributions than
lentic species.
5.3. Dispersal morphometrics
With direct measures of dispersal distance lacking for most species, morphological
traits act as indirect indicators. Because past studies are often relied upon as source
of this data, it is vital that the exact methodologies and their implications are
understood. As shown here with wing length, different measurement methods can
result in significantly different values for seemingly identical traits. Wing lengths
obtained using the computer software ImageJ (Rasband, 1997-2007) were
significantly smaller than those obtained from hand measuring with calipers (Figure
4.4). Because ImageJ works with scanned images of the wing, the very base is
omitted from analysis and is therefore not comparable to other methods which
measure wing length to the base.
Residuals were taken for both dry wing length and thoracic volume to control for
body size, as both strongly correlated with metrics of abdomen length (body size
proxy). Whilst generating thoracic volume residuals from abdomen length3 may rely
on the use of an artificial metric, cubing the length values did ensure that allometric
scaling was corrected for in the right dimension. In general, measurements of
abdomen length are prone to inaccuracies, particularly in species with telescopic
abdomens like the odonates (Nijhout et al. 2006), however it was used in this study
because it was the most readily available proxy for body size (Serrano-Meneses et al.
2007). The value of these wing metrics in across-species comparisons was validated
using VCAs. In trait analyses, within-species variation is often assumed to be absent
but this is rarely the case (Ives et al. 2007). VCAs can serve an important function in
![Page 45: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/45.jpg)
© Alyson Pavitt, 2011 [email protected]
38
identifying the taxonomic or population level where most variation occurs. In this
study there was negligible intraspecific variation for both wing measurements, with
most variance found above the species level (Table 4.4). This increases the certainty
that it is interspecific differences, and not overlapping intraspecific variation, which
result in the trends observed in this study.
5.4. Determining occupancy change
5.4.1. Distribution pattern change
Although there was no significant correlation between the sizes of change index and
residual D change, a significant number of species (>86%) showed a positive
relationship between the two (Figure 4.1). This may arise from species declining due
to internal localised extinctions (Thomas, 2000; Tscharntke et al. 2002; e.g. Figure
4.5(b)), and increasing due to step-wise expansion from previously occupied cells
(e.g. Figure 4.5(a)). For example, improvements in river management and water
quality have seen populations of species such as Libellula fulva and Platycnemis
pennipes expanding their ranges along the lengths of riverine systems (Corbet &
Brooks, 2008).
![Page 46: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/46.jpg)
© Alyson Pavitt, 2011 [email protected]
39
(a)
(b)
Figure 4.5: Distribution of occupied grid cells in the British ranges of (a) Orthopterum
cancellatum (a species increasing in occupancy) and (b) Calopteryx virgo (a species
decreasing in occupancy),for t1 (left) and t2 (right).
![Page 47: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/47.jpg)
© Alyson Pavitt, 2011 [email protected]
40
Of the five species showing the reverse trend, becoming sparser with increasing cell
occupancy, the most notable was Aeshna isosceles. This rare species (Endangered on
the IUCN British Red List; Daguet et al. 2008) is highly sensitive to pollution, with a
British range confined to Norfolk and north-east Suffolk, where it breeds exclusively
in ditches supporting the aquatic plant Stratiotes aloides (water soldier) (Corbet &
Brooks, 2008). Because of the small number of occupied cells, single sightings far
beyond the species’ acknowledged range will disproportionately influence the
results. Three notably extreme records, each based on a single sighting, account for
over 11% of all occupied grid cells at t2 (Figure 4.6). Such unexpected sightings may
be explained by vagrant individuals transported by extreme weather, coupled with
an increase in recorder effort and species awareness since its appearance in the 1995
UK Biodiversity Strategy.
Figure 4.6: Distribution of occupied grid cells in the Britain range of Aeshna isosceles for
t1 (left) and t2 (right).
![Page 48: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/48.jpg)
© Alyson Pavitt, 2011 [email protected]
41
5.4.2. Distribution and dispersal
Neither distribution pattern (residual D), nor dispersal morphometrics (residual dry
wing length, wing aspect ratio, and residual thoracic volume) were sufficient in
describing change index when considered alone. The descriptive power of the
model, however, was significantly enhanced when the two were considered together
(accounting for 54% of the variation seen in change index). This reinforces the
importance of both dispersal capacity and underlying population distribution in
determining a species’ ability to colonise new areas.
Species with the greatest range expansion were dominated by those with long, broad
wings and large thoracic volumes, in keeping with previously published findings on
invertebrate dispersal (e.g. Malmqvist, 2000; Rundle et al. 2007). Long, broad wings
maximise surface area for efficient flight, whilst a large thoracic volume (indirect
measure of muscle mass; Schilder & Marden, 2004) is indicative of the greater power
needed to generate and sustain the wing beat frequency of these larger wings
The assumption that sparse distributions will always be declining towards extinction
(Wilson et al. 2004) does not hold true, because it fails to consider both variation in
the dispersal ability of individuals, and in the inherent sparsity of the habitat. As a
sole measure of species threat this metric makes the erroneous assumption that all
odonate species start with a relatively aggregated distribution. For example, lentic
habitats, as previously discussed in Section 4.2, are likely to be more distributed
across the landscape than lotic habitats.
When considered with weak dispersal capacity (small, narrow wings and small
thoracic volume) however, those species with sparser distributions were indeed
those with declining range occupancies. New areas are easier to colonise if there is a
concentrated source of potential colonists in close proximity, particularly in species
with the greatest capacity for dispersal. Conversely, the landscape is less permeable
to species with limited dispersal ranges and which are already sparsely distributed
(either through underlying habitat distribution, or through local extinctions which
![Page 49: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/49.jpg)
© Alyson Pavitt, 2011 [email protected]
42
have not been recolonised). Amongst these declining species are the four British
Coenagrion damselflies (C. hastulatum, C. mercuriale, C. puella and C. pulchellum). This
genus has been in decline here since the 1950’s, with two species becoming
nationally extinct between 1953-1983 (Freeland & Conrad, 2002). They have similar
morphologies, with small, delicate thoraces and relatively short, narrow wings; all
are sparsely distributed and all, with the exception of C. mercuriale, favour lentic
habitats. Their limited capacity for recolonisation, and the inherent isolation of their
primary habitats, makes this genus particularly vulnerable to habitat destruction and
fragmentation processes (Daguet et al. 2008; Lowe et al. 2008; BDS, 2011).
5.5. Suborder variation
In light of the morphological associations with occupancy change, the significant
difference in dispersal morphometrics between suborders could have crucial
implications for their respective abilities to respond to changing conditions. The
phenotype found in British anisopterans (significantly longer, broader wings and
larger thoracic volume) was that associated with range expansion, suggesting that
they are responding far better to changes in the past four decades than zygopterans.
This corresponds with extensive evidence that anisopterans perform longer
migrations (Ruppell, 1989), are better colonists (comprising over 75% of recent
British colonists; Appendix II), and exhibit faster flight across greater distances than
the more manoeuvrable but weaker zygopterans (Wakeling & Ellington, 1997).
5.6. Value of historic collections
In order for analyses to have any explanatory powers, a sufficient number of
specimens are needed to represent the population or species under study. When
species are rare, cryptic or difficult to find, collecting enough data points may be
difficult. Under these circumstances, natural history collections can provide an
invaluable source of data. These specimens will have been dried and are often
decades, if not centuries old, calling into question their value in representing extant
populations (Doucet & Hill, 2009). Although only preliminary, findings from this
study show the wing morphology of odonate NHM specimens to be
![Page 50: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/50.jpg)
© Alyson Pavitt, 2011 [email protected]
43
indistinguishable from those of recently caught, wet specimens. These results do, of
course, need to be treated with caution, and no claims can be made regarding their
representativeness of other morphological traits or taxonomic groups. Odonate
wings have little water content and so are unlikely to shrink or warp to any
significant degree when they dry. The same cannot be said for the body, particularly
the abdomen, where results are likely to be more variable.
5.7. Recommendations for future research
Whilst this study investigated general trends in odonate distribution, and looked at
the value of distribution pattern (residual D) and dispersal morphology in
describing these changes, no attempt was made to explore potential drivers of this
change. Future research needs to focus on pulling apart the relative impacts of
climate and habitat change in order to guide pre-emptive conservation action.
Climate change investigations need to move on from focussing on gradual
temperature change to encompass other variables. Maximum temperature
thresholds, the frequency of extreme weather (such as droughts and floods), and
changing humidity will all effect species distribution and survival potential.
Although this study does not dispute the validity of range shift metrics, it does
question the readiness to attribute all polewards range movement to climate change
induced range shift. There is the risk that polewards shift may be concluded
prematurely, particularly when there is no corresponding investigation into
longitudinal movements, or the possibility that trends may result from non-
directional occupancy change. Further investigation is needed in this area to ensure
that no rash conclusions are drawn which may overlook vital factors or driving
forces of change.
Despite being one of the best studied taxa in Britain, there remain many gaps in our
knowledge of simple odonate life history. Although it has fallen from conservation
favour in recent years, there is a pressing need for more targeted natural history
research, not only from data collected by amateur naturalists and enthusiasts, but
![Page 51: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/51.jpg)
© Alyson Pavitt, 2011 [email protected]
44
also from controlled scientific study and experimentation. For example, there is
limited information on many species’ dispersal capacity (this study is only able to
take morphology as a proxy). For a number of species, such information is only
speculative, or based on a small number of observations.
Further research is also needed into the value of museum specimens, an area only
briefly touched upon in the scope of this project. Whilst preliminary findings do
suggest that historic collections are representative of extant species morphometrics,
it is pertinent to note that this was only carried out on a subsample of specimens for
four species stored at NHM. It would be important to extend this to other species
and other collections, ensuring a spatially and temporally representative sample set.
It is known that some physical traits shrink when specimens are dried, however
there has been little research into different drying or storage methods, and whether
interspecific variation remains representative after drying.
![Page 52: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/52.jpg)
© Alyson Pavitt, 2011 [email protected]
45
6. Conclusions
Models are increasingly being used to predict large scale changes in species
distribution, particularly in light of climate change. Because there is insufficient data
on many taxa, the few that have been extensively studied, and which have long-
standing historic records, are being investigated for generally applicable trends.
Using a well studied group of British odonates, this study shows that the
combination of dispersal morphology (dry wing length and aspect ratio, and
thoracic volume) and distribution pattern can be used as an indicator of changes in
range occupancy over time.
Species showing the biggest range expansions were predominantly those with long,
broad wings and large thoracic volumes, particularly when showing aggregated
distributions. This was linked to suborder differences, with most anisopterans
(generally large winged and large bodied) exhibiting range expansions, whilst most
zygopterans (smaller, weaker dispersers) showed a decline. Given the importance of
obtaining reliable morphometrics, this study also investigated the value of museum
specimens as sources of morphometric data. Measurements obtained from NHM
were not found to differ from the variation seen in extant populations of the species,
drawing the, albeit preliminary, conclusion that these collections remain
representative of extant populations.
Whilst there were not a significant number of species with decreasing cell
occupancies, neither was there the overwhelming number of increasing species that
past studies have found (e.g. Hickling et al. 2005). This study therefore, urges caution
in prematurely writing off the British odonates as “safe” under predicted
environmental and climatic changes. It also highlights the risk of automatically
concluding a climate change induced range shift from the dynamics of a single range
margin.
![Page 53: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/53.jpg)
© Alyson Pavitt, 2011 [email protected]
46
References
Aguilar, J., Dommanget, J. & Prechac, R. (1986) A Field Guide to the Dragonflies of
Britain, Europe and North Africa. London, Collins.
Allen, K. & Thomas, D. (2010) Movement characteristics of the scarce blue- tailed
damselfly, Ischnura pumilio. Insect Conservation and Diversity, 3 (1), 5.
Altermatt, F. (2010) Climatic warming increases voltinism in European butterflies
and moths. Proceedings of the Royal Society B: Biological Sciences, 277 (1685), 1281-1287.
Angelibert, S. & Giani, N. (2003) Dispersal characteristics of three odonate species in
a patchy habitat. Ecography, 26 (1), 13.
Archaux, F. (2004) Breeding upwards when climate is becoming warmer: no bird
response in the French Alps. Ibis, 146, 138-144.
Artiss, T., Schultz, T., Polhemus, D. & Simon, C. (2001) Molecular phylogenetic
analysis of the dragonfly genera Libellula, Ladona, and Plathemis (Odonata:
Libellulidae) based on mitochondrial cytochrome oxidase I and 16S rRNA sequence
data. Molecular Phylogenetics and Evolution, 18 (3), 348-361.
Bale, J., Masters, G., Hodkinson, I., Awmack, C., Bezemer, T., Brown, V., Butterfield,
J., Buse, A., Coulson, J., Farrar, J., Good, J., Harrington, R., Hartley, S., Jones, T.,
Lindroth, R., Press, M., Symrnioudis, I., Watt, A. & Whittaker, J. (2002) Herbivory in
global climate change research: direct effects of rising temperature on insect
herbivores. Global Change Biology, 8 (1), 1.
BDS. (2010) Norfolk Biodiversity Action Plan: Norfolk Hawker. British Dragonfly Society.
Report number: Species Action Plan 27.
BDS. Southern Damselfly. [Online] Available from: http://www.british-
dragonflies.org.uk/species/southern-damselfly [Accessed 22/8/2011].
![Page 54: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/54.jpg)
© Alyson Pavitt, 2011 [email protected]
47
Begon, M., Townsend, C. & Harper, J. (2005) Ecology: From Individuals to Ecosystems.
4th edition. Oxford, Blackwell Science.
Bense, U. (1995) Longhorn beetles: illustrated key to the cerambycidae and vesperidae of
Europe. Germany, Margraf Verlag, Wiekersheim.
Beukema, J. (2002) Changing istribution patterns along a stream in adults of
Calopteryx haemorrhoidalis (Odonata: Calopterygidae): a case of larval-drift
complensation? International Journal of Odonatology., 5 (1), 1-14.
Bowles, P. (2010) Correlates of distribution and decline in UK species. Masters of Science.
Imperial College London.
Brittain, J. (2008) Mayflies, biodiversity and climate change. In: Hauer, F., Standford,
J. & Newell, R. (eds.) International Advances in the Ecology, Zoogeography, and
Systematics of Mayflies and Stoneflies. University of California Press, pp. 1-14.
Brooks, S. (ed.) (2005) Field Guide to the Dragonflies and Damselflies of Great Britain and
Ireland. Dorset., British Wildlife Publishing.
Chaput-Bardy, A., Lemaire, C., Picard, D. & Secondi, J. (2008) In-stream and
overland dispersal across a river network influences gene flow in a freshwater insect,
Calopteryx splendens. Molecular Ecology, 17, 3496-3505.
Clark, P. & Evans, F. (1954) Distance to nearest neighbor as a measure of spatial
relationships in populations. Ecology, 35 (4), 445-453.
Clausnitzer, V., Kalkman, V., Ram, M., Collen, B., Baillie, J., Bedjanic, M., Darwall,
W., Dijkstra, K., Dow, R., Hawking, J., Karube, H., Malikova, E., Paulson, D., Schutte,
K., Suhling, F., Villanueva, R., von Ellenrieder, N. & Wilson, K. (2009) Odonata enter
the biodiversity crisis debate: The first global assessment of an insect group.
Biological Conservation, 142, 1864-1869.
Conrad, K. (Personal communication, 2011).
![Page 55: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/55.jpg)
© Alyson Pavitt, 2011 [email protected]
48
Conrad, K., Wilson, K., Harvey, I., Thomas, C. & Sherratt, T. (1999) Dispersal
characteristics of seven odonate species in an agricultural landscape. Ecography, 22,
524-531.
Corbet, P. (1999) Dragonflies: Behaviour and Ecology of Odonata. NY, Cornell University
Press.
Corbet, P. & Brooks, S. (2008) Dragonflies. The New Naturalist Library, , Collins.
Corbet, P., Suhling, F. & Soendgerath, D. (2006) Voltinism of Odonata: a review.
International Journal of Odonatology., 9 (1), 1-44.
Cordero, A. (1994) Reproductive allocation in different-sized adults Ischnura
graellsii (Rambur) (Zygoptera: oenagrionidae). Odonatologica, 23, 271-276.
Daguet, C., French, G. & Taylor, P. (2008) The odonata red list data for Great Britain.
Species Status 11, Peterborough, Joint Nature Conservation Committee.
Davis, J., Howe, R. & Davis, G. (2000) A multi-scale spatial analysis method for point
data. Landscape Ecology, 15, 99-114.
DEFRA. (2011) Chemical river water quality. [Online] Available from:
http://www.defra.gov.uk/statistics/environment/inland-water/iwfg07-chrivqual/
[Accessed 26/4/2011].
Dixon, P. (2002) Ripley's K function. In: El-Shaarawi, A. & Piegorsch, W. (eds.)
Encyclopedia of Environmetrics. Volume 3 series. Chichester, UK, John Wiley & Sons
Ltd, pp. 1796-1803.
Doucet, S. & Hill, G. (2009) Do museum specimens accurately represent wild birds?
A case study of carotenoid, melanin, and structural colours in long-tailed manakins
Chiroxiphia linearis. Journal of Avian Biology, 40 (2), 146-156.
![Page 56: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/56.jpg)
© Alyson Pavitt, 2011 [email protected]
49
Dukes, J. & Mooney, H. (1999) Does global change increase the success of bioloical
invaders? Trends in Ecology and Evolution, 14 (4), 135-139.
Dumont, H., Vierstraete, A. & Vanfleteren, J. (2010) A molecular phylogeny of the
Odonata (Insecta). Systematic Entomology, 35, 6-18.
Elith, J. & Leathwick, J. (2009) Species distribution models: ecological explanation
and prediction across space and time. Annual Review of Ecology, Evolution, and
Systematics, 40, 677.
Environment Agency. (2008) Water Resources in England and Wales- Current State and
Future Pressures. Bristol, Environment Agency.
Ferreras-Romero, M. & Corbet, P. (1999) The life cycle of Cordulegaster boltonii
(Donovan, 1807) (Odonata: Cordulegastridae) in the Sierra Morena Mountains
(southern Spain). Hydrobiologia, 405, 39-48.
Fischer, J. & Lindenmayer, D. (2007) Landscape modification and habitat
fragmentation: a synthesis. Global Ecology and Biogeography, 16 (3), 265.
Fleck, G., Ullrich, B., Brenk, M., Wallnisch, C., Orland, M., Bleidissel, S. & Misof, B.
(2008) A phylogeny of anisopterous dragonflies (Insecta, Odonata) using mtRNA
genes and mixed nucleotide ⁄ doublet models. Journal of Zoological Systematics and
Evolutionary Research, 46 (4), 310-322.
Freeland, J. & Conrad, K. (2002) Genetic similarity within and among populations of
the Variable and Azure damselflies (Coenagrion pulchellum and C. puella).
Hydrobiologia, 479, 69-73.
Gardenfors, U., Rodriguez, J., Hilton-Taylor, C., Hyslop, C., Mace, G., Molur, S. &
Poss, S. (1999) Draft guidelines for the application of IUCN Red List Criteria at
national and regional levels. Species 31/32, 58-70.
Gaston, K. (1994) Rarity. Population and Community Biology Series. , Springer.
![Page 57: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/57.jpg)
© Alyson Pavitt, 2011 [email protected]
50
Gaston, K. (1996) Species-range-size distributions: patterns, mechanisms and
implications. Trends in Ecology and Evolution, 11 (5), 197-201.
Gaston, K. (2003) The Structure and Dynamics of Geographic Ranges. Oxford, Oxford
University Press.
Gaston, K. & Fuller, R. (2009) The sizes of species' geographic ranges. Journal of
Applied Ecology, 46, 1-9.
Gelfand, A., Silander, J., Wu, S., Latimer, A., Lewis, P., Rebelo, A. & Holder, M.
(2006) Explaining species distribution patterns through hierarchical modeling.
Bayesian Analysis, 1, 41-92.
Gibbins, C. & Moxton, J. (1998) Calopteryx splendens (Harris) at edge of range sites
in North East England. Journal of British Dragonfly Society, 14, 33-45.
Goffart, P. (2010) Southern dragonflies expanding in Wallonia (south Belgium): a
consequence of global warming? BioRisk, 5, 109-126.
Guisan, A. & Thuiller, W. (2005) Predicting species distribution: offering more than
simple habitat models. Ecology Letters, 8, 993-1009.
Guisan, A. & Zimmermann, N. (2000) Predictive habitat distribution models in
ecology. Ecological Modelling, 135, 147-186.
Hartley, S., Kunin, W., Lennon, J. & Pocock, M. (2004) Coherence and discontinuity
in the scaling of species' distribution patterns. Proceedings of the Royal Society B:
Biological Sciences, 271, 81-88.
Hassall, C. (Personal communication, 2011).
Hassall, C. & Thompson, D. (2010) Accounting for recorder effort in the detection of
range shifts from historical data. Methods in Ecology and Evolution, 1 (4), 343-350.
![Page 58: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/58.jpg)
© Alyson Pavitt, 2011 [email protected]
51
Hassall, C. & Thompson, D. (2008) The effects of environmental warming on
Odonata: a review. International Journal of Odonatology., 11 (2), 131-153.
Hassall, C., Thompson, D. & Harvey, I. (2008) Latitudinal variation in morphology in
two sympatric damselfly species
with contrasting range dynamics (Odonata: Coenagrionidae). European Journal of
Entomology, 105 (5), 939.
Hassall, C., Thompson, D. & Harvey, I. (2009) Variation in morphology between core
and marginal populations of three British damselflies. Aquatic Insects, 31 (3), 187-197.
Hassall, C. (2010) The impact of climate-induced distributional changes on the
validity of biological water quality metrics. Environmental Monitoring and Assessment,
160 (1), 451.
Hassall, C., Thompson, D., French, G. & Harvey, I. (2007) Historical changes in the
phenology of British Odonata are related to climate. Global Change Biology, 13 (5), 933.
Hayek, L. & Buzas, M. (1997) How many field samples? In: Surveying Natural
Populations. New York, Colombia University Press, pp. 66-83.
Hickling, R., Roy, D., Hill, J. & Thomas, C. (2005) A northward shift of range margins
in British Odonata. Global Change Biology, 11 (3), 502.
Hickling, R., Roy, D., Hill, J., Fox, R. & Thomas, C. (2006) The distributions of a wide
range of taxonomic groups are expanding polewards. Global Change Biology, 12 (3),
450-455.
Hill, J. (1999) Climate and habitat availability determine 20th century changes in a
butterfly's range margin. Proceedings of the Royal Society B: Biological Sciences, 266
(1425), 1197.
![Page 59: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/59.jpg)
© Alyson Pavitt, 2011 [email protected]
52
Hill, J., Thomas, C., Fox, R., Telfer, M., Willis, S., Asher, J. & Huntley, B. (2002)
Responses of butterflies to twentieth century climate warming: implications for
future ranges. Proceedings of the Royal Society B: Biological Sciences, 269, 2163-2171.
Hof, C., Brandle, M. & Brandl, R. (2006) Lentic odonates have larger and more
northern ranges than lotic species. Journal of Biogeography, 33 (1), 63-70.
Hofmann, T. & Mason, C. (2005) Habitat characteristics and the distribution of
Odonata in a lowland river catchment in eastern England. Hydrobiologia, 539, 137-
147.
Holt, R. (2003) On the evolutionary ecology of species' ranges. Evolutionary Ecology
Research, 5, 159-178.
Houghton, J., Ding, Y., Griggs, D., Noguer, M., van der Linden, P., Dai, X., Maskell,
K. & Johnson, C. (2002) Climate Change 2001: The Scientific Basis. Contribution of
working group I to the third assessment report of the Intergovernmental Panel on
Climate Change. Cambridge University Press., 57 (8), 267.
Hughes, C., Dytham, C. & Hill, J. (2007) Modelling and analysing evolution of
dispersal in populations at expanding range boundaries. Ecological Entomology, 32
(5), 437.
Hui, C. (2011) Forecasting population trend from the scaling pattern of occupancy.
Biological Modelling, 222, 442-446.
Huston, M. (2002) Introductory essay: critical issues for improving predictions. In:
Scott, J., Heglund, P., Morrison, M., Haufler, J., Raphael, M., Wall, W. & Samson, F.
(eds.) Predicting Species Occurrences: Issues of Accuracy and Scale. Covelo, California,
Island Press, pp. 7-21.
Ikeya, T., Galic, M., Belawat, P., Nairz, K. & Hafen, E. (2002) Nutrient-dependent
expression of insulin-like peptides from neuroendocrine cells in the CNS contributes
to growth regulation in Drosophila. Current Biology, 12 (15), 1293-1300.
![Page 60: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/60.jpg)
© Alyson Pavitt, 2011 [email protected]
53
Ives, A., Midford, P. & Garland, T. (2007) Within-species variation and measurement
error in phylogenetic comparative methods. Systematic Biology, 56 (2), 252-270.
Jenkins, D., Brescacin, C., Duxbury, C., Elliott, J., Evans, J., Grablow, K., Hillegass,
M., Lyon, B., Metzger, G., Olandese, M., Pepe, D., Silvers, G., Suresch, H.,
Thompson, T., Trexler, C., Williams, G., Williams, N. & Williams, S. (2007) Does size
matter for dispersal distance? Global Ecology and Biogeography, 16, 415-425.
Jetz, W., Sekerciogly, C. & Watson, J. (2008) Ecological correlates and conservation
implications of overestimating species geographic ranges. Conservation Biology, 22
(1), 110-119.
JNCC. (2007) Second Report by the United Kingdom under Article 17 on the
Implementation of the Directive from Janurary 2001 to December 2006. Peterborough,
JNCC. Report number: S1044- Coenagrion mercuriale- Southern damselfly.
Johansson, F. (2003) Latitudinal shifts in body size of Enallagma cyathigerum
(Odonata). Journal of Biogeography, 30, 29-34.
Korkeamaki, E. & Suhonen, J. (2002) Distribution and habitat specialization of
species affect local extinctions in dragonfly Odonata populations. Ecography, 25 (4),
459-465.
Kristensen, N. (ed.) (1999) Lepidoptera, moths and butterflies. Volume 1: Evolution,
systematics, and biogeography. Handbuch der Zoologie. Eine Naturgeschichte der Stämme
des Tierreiches / Handbook of Zoology. A Natural History of the phyla of the Animal
Kingdom, Berlin, Walter de Gruyte.
Legendre, P. (1993) Spatial autocorrelation: trouble or new paradigm? Ecology, 74 (6),
1659-1673.
Lenoir, J., Gegout, J., Marquet, P., Ruffray, P. & Brisse, H. (2008) A significant
upward shift in plant species optimum elevation during the 20th century. Science,
320, 1768-1771.
![Page 61: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/61.jpg)
© Alyson Pavitt, 2011 [email protected]
54
Levin, S. (1992) The problem of pattern and scale in ecology. Ecology, 73, 1943-1967.
Lowe, C., Harvey, I., Thompson, D. & Watts, P. (2008) Strong genetic divergence
indicates tha congeneric damselflies Coenagrion puella and C. pulchellum (Odonata:
Zygoptera: Coenagrionidae) do not hybridise. Hydrobiologia, 605, 55-63.
Malcolm, J., Liu, C., Neilson, R., Hansen, L. & Hannah, L. (2006) Global warming
and extinctions of endemic species from biodiversity hotspots. Conservation Biology,
20 (2), 538.
Malmqvist, B. (2000) How does wing length relate to distribution patterns of
stoneflies (Plecoptera) and mayflies (Ephemeroptera)? Biological Conservation, 93,
271-276.
Marchetti, K., Price, T. & Richman, A. (1995) Correlates of wing morphology with
foraging behaviour and migration distance in the genus Phylloscopus. Journal of
Avian Biology, 26, 177-181.
Marden, J. (1994) From damselflies to pterosaurs: how burst and sustainable flight
performance scale with size. American Journal of Physiology, 266, R1077-R1084.
Martinez, J., Martinez, J., Zuberogoitia, I., Zabala, J., Redpath, S. & Calvo, J. (2008)
The effects of intra- and interspecific interactions on the large-scale distribution of
cliff-nesting raptors. Ornis Fennica, 85, 13-21.
Mattila, N., Kaitala, V., Komonen, A., Paivinen, J. & Kotiaho, J. (2011) Ecological
correlates of distribution change and range shift in butterflies. Insect Conservation and
Diversity, (1), 8.
McCauley, J. (2007) The role of local and regional processes in structuring larval
dragonfly distributions across habitat gradients. Oikos, 116 (1), 121.
McCauley, S. (2006) The effects of dispersal and recruitment limitation on
community structure of odonates in artificial ponds. Ecography, 29 (4), 585.
![Page 62: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/62.jpg)
© Alyson Pavitt, 2011 [email protected]
55
McPeek, M. (1989) Differential dispersal tendencies among Enallagma damselflies
(Odonata) inhabiting different habitats. OIKOS, 56, 187-195.
Menendez, R. (2007) How are insects respondin to global warming? Tijdschrift Voor
Entomologie, 150, 355-365.
Meyer, C., Frund, J., Lizano, W. & Kalko, E. (2008) Ecological correlates of
vulnerability to fragmentation in Neotropical bats. Journal of Applied Ecology, 45 (1),
381.
Mill, P., Brooks, S. & Parr, A. (2010) Dragonflies (Odonata) in Britain and Ireland. In:
Maclean, N. (ed.) Silent Summer: The State of Wildlife in Britain and Ireland. Cambridge
University Press, pp. 471-494.
Morin, X., Viner, D. & Chulne, I. (2008) Tree species range shifts at a continental
scale: new predictive insights from a process- based model. Journal of Ecology, 96 (4),
784-794.
Nijhout, H., Davidowitz, G. & Roff, D. (2006) A quantitative analysis of the
mechanism that controls body size in Manduca sexta. Journal of Biology, 5, 16.
Norberg, U. (1995) How a long tail and changes in mass and wing shape affect the
cost for flight in animals. Functional Ecology, 9 (1), 48.
Opdam, P. & Wascher, D. (2004) Climate change meets habitat fragmentation:
linking landscape and biogeographical scale levels in research and conservation.
Biological Conservation, 117 (3), 285.
Orme, D., Freckleton, R., Thomas, G., Petzoldt, T., Fritz, S. & Isaac, N. (2011) Caper:
Comparative Analyses of Phylogenetics and Evolution in R. [R Package Verson 0.4/r73].
Ott, J. (2010) Dragonflies and climate change- recent trends in Germany and Europe.
BioRisk, 5, 253-286.
![Page 63: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/63.jpg)
© Alyson Pavitt, 2011 [email protected]
56
Par, A. (2010) Monitoring of Odonata in Britain and possible insights into climate
change. BioRisk, 5, 127-139.
Paradis, E., Claude, J. & Strimmer, K. (2004) APE: analyses of phylogenetics and
evolution in R language. Bioinformatics, 20, 289-290.
Parker, J. & Johnston, L. (2006) The proximate determinants of insect size. Journal of
Biology, 5 (5), 15.
Parmesan, C. (2006) Ecological and evolutionary responses to recent climate change.
Annual Review of Ecology, Evolution, and Systematics, 37, 637-669.
Parmesan, C. (1996) Climate and Species Range. Nature, 382 (6594), 765.
Parmesan, C. (1999) Poleward shifts in geographical ranges of butterfly species
associated with regional warming. Nature, 399 (6736), 579.
Parmesan, C. & Yohe, G. (2003) A globally coherent fingerprint of climate change
impacts across natural systems. Nature, 421 (6918), 37.
Paynter, Q. & Bellgard, S. (2011) Understanding dispersal rates of invading weed
biocontrol agents. Journal of Applied Ecology, 48 (2), 407-414.
Pilgrim, E. & Von Dohlen, C. (2008) Phylogeny of the Sympetrinae (Odonata:
Libellulidae): further evidence of the homoplasious nature of wing venation.
Systematic Entomology, 33, 159-174.
Poyry, J., Luoto, M., Heikkinen, R., Kuussaai, M. & Saarinen, K. (2009) Species traits
explain recent range shifts of Finnish butterflies. Global Change Biology, 15 (3), 732.
Pringle, J. (1949) The excitation and contraction of the flight muscles of insects.
Journal of Physiology, 108, 226-232.
![Page 64: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/64.jpg)
© Alyson Pavitt, 2011 [email protected]
57
Purse, B. & Thompson, D. (2003) Emergence of the damselfly Coenagrion mercuriale
and Ceriagrion tenellum (Odonata: Coenagrionidae), at their northern range
margins, in Britain. European Journal of Entomology, 100, 93-100.
Purvis, A. & Rambaut, A. (1995) Comparative analysis by independent contrasts
(CAIC): an Apple Macintosh application for analysing comparative data. CABIOS, 11
(3), 247-251.
R Development Core Team. (2011) R: A Language and Environment for Statistical
Computing. [2.13] Vienna, Austria, R Foundation for Statistical Computing.
Rambaut, A. & Drummond, A. (2009) FigTree. [v1.3.1] University of Edinburgh,
Edinburgh, Institute of Evolutionary Biology.
Rasband, W. (1997-2007) ImageJ. [Online] Available from: http://rsb.info.nih.gov/ij/
[Accessed 3/5/2011].
Rayner, J. (1988) Form and function in avian flight. Current Ornithology, 5, 1-66.
Ripley, B. (1977) Modelling spatial patterns. Journal of the Royal Statistical Society: B,
39, 172-212.
Robinet, C. & Roques, A. (2010) Direct impacts of recent climate warming on insect
populations. Integrative Zoology, 5, 132-142.
Rundle, S., Bilton, D., Abbot, J. & Foggo, A. (2007) Range size in North American
Enallagma damselflies correlates with wing size. Freshwater Biology, 52, 471-477.
Ruppell, G. (1989) Kinematic analysis of symmetrical flight manoeuvres of Odonata.
Journal of Experimental Biology., 144, 13-42.
Saino, N., Rubolini, D., Van Hardenberg, J., Ambrosini, R., Provenzale, A., Romano,
M. & Spina, F. (2010) Spring migration decisions in relation to weather are predicted
![Page 65: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/65.jpg)
© Alyson Pavitt, 2011 [email protected]
58
by wing morphology among trans- Mediterranean migratory birds. Functional
Ecology, 24 (3), 658.
Saux, C., Simon, C. & Spicer, G. (2003) Phylogeny of the dragonfly and damselfly
order Odonata as inferred by mitochondrial 12S ribosomal RNA sequences. Annals of
the Entomological Society of America, 96 (6), 693-699.
Schilder, R. & Marden, J. (2004) A hierarchical analysis of the scaling of force and
power production by dragonfly flight motors. Journal of Experimental Biology., 207,
767-776.
Schoener, T. & Janzer, D. (1968) Notes on environmental determinants of tropical
versus temperate insect size patterns. The American Naturalist, 102 (925), 207-224.
Segurado, P. & Araujo, M. (2004) An evaluation of methods for modelling species
distributions. Journal of Biogeography, 31, 1555-1568.
Serrano-Meneses, M., Azpilicueta-Amorin, M., Szekely, T. & Cordoba-Aguilar, A.
(2007) The development of sexual differences in body size in Odonata in relation to
mating systems. European Journal of Entomology, 104, 453-458.
Shoo, L., Williams, S. & Hero, J. (2006) Detecting climate change induced range
shifts: Where and how should we be looking? Australian Ecology, 31, 22-29.
Slove, J. & Janz, N. (2011) The Relationship between Diet Breadth and Geographic
Range Size in the Butterfly Subfamily Nymphalinae- Study of Global Scale. PLoS
ONE, 6 (1), e16057.
Smith, D. & Murphy, J. (1982) Spatial relationships of nesting golden eagles in
central Utah. Raptor Research, 16 (4), 127-132.
Stevens, V., Turlure, C. & Baguette, M. (2010) A meta-analysis of dispersal in
butterflies. Biological Reviews, 85 (3), 625.
![Page 66: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/66.jpg)
© Alyson Pavitt, 2011 [email protected]
59
Taylor, P. & Merriam, G. (1995) Wing morphology of a forest damselfly is related to
landscape structure. Oikos, 73 (1), 43.
Telfer, M. (2003) Change index: a measure of change in range size that is
independent of changes in survey effort. In: Proceedings 13th International Colloquium
European Invertebrate Survey, Leiden 2-5 September 2001. pp. 107-110.
Telfer, M., Preston, C. & Rothery, P. (2002) A general method for measuring relative
change in range size from biological atlas data. Biological Conservation, 107 (1), 99.
Termaat, T., Kalkman, V. & Bouwman, J. (2010) Changes in the range of dragonflies
in the Netherlands and the possible role of temperature change. BioRisk, 5, 155-173.
Thomas, C. (2000) Dispersal and extinction in fragmented landscapes. Proceedings of
the Royal Society B: Biological Sciences, 267, 139-145.
Thomas, C., Bodsworth, E., Wilson, R., Summons, A., Davies, Z., Musche, M. &
Conradt, L. (2001) Ecological and evolutionary processes at expanding range
margins. Nature, 411, 577-581.
Thomas, C. & Lennon, J. (1999) Birds extend their ranges northwards. Nature, 399,
213.
Thomas, D., Cameron, A., Green, R., Bakkenes, M., Beaumont, L., Collingham, Y.,
Erasmus, B., de Siqueira, M., Grainger, A., Hannah, L., Hughes, L., Huntley, B., van
Jaarsveld, A., Midgley, G., Miles, L., Ortega-Huerta, M., Peterson, A., Philips, O. &
Williams, S. (2004) Extinction risk from climate change. Nature, 427 (6970), 145.
Thomas, C. (2010) Climate, climate change and range boundaries. Diversity
Distributions, 16 (3), 488.
Thomas, C., Thomas, J. & Warren, M. (1992) Distributions of occupied and vacant
butterfly habitats in fragmented landscapes. Oecologia, 92 (4), 563.
![Page 67: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/67.jpg)
© Alyson Pavitt, 2011 [email protected]
60
Tobin, P., Nagarkatti, S., Loeb, G. & Saunder, M. (2008) Historical and projected
interactions between climate change and insect voltinism in a multivoltine species.
Global Change Biology, 14, 951-957.
Tscharntke, T., Steffan-Dewenter, I., Kruess, A. & Thies, C. (2002) Characteristics of
insect populations on habitat fragments: a mini review. Ecological Research, 17, 229-
239.
Von Ellenrieder, N. (2002) A phylogenetic analysis of the extant Aeshnidae
(Odonata: Anisoptera). Systematic Entomology, 27, 437-467.
Wakeling, J. & Ellington, C. (1997) Dragonfly flight: II. Velocities, accelerations and
kinematics of flapping flight. Journal of Experimental Biology, 200, 557-582.
Ward, L. & Mill, P. (2007) Long range movements by individuals as a vehicle for
range expansion in Calopteryx splendens (Odonata: Zygoptera). European Journal of
Entomology, 104, 195-198.
Warren, M., Hill, J., Thomas, J., Asher, J., Fox, R., Huntley, B., Roy, D., Telfer, M.,
Jeffcoate, S., Harding, P., Jeffcoate, G., Willis, S., Greatorex-Davies, J., Moss, D. &
Thomas, D. (2001) Rapid responses of British butterflies to opposing forces of climate
and habitat change. Nature, 414 (6859), 65.
Wiegand, T. & Moloney, K. (2004) Rings, circles, and null-models for point pattern
analysis in ecology. Oikos, 104, 209-229.
Wikelski, M., Moskowitz, D., Adelman, J., Cochran, J., Wilcove, D. & May, M. (2006)
Simple rules guide dragonfly migration. Biological Letters, 2 (3), 329-325.
Wilcove, D., McLellan, C. & Dobson, A. (1986) Habitat fragmentation in the
temperature zone. In: Soule, M. (ed.) Conservation Biology. MA, Sunderland, pp. 237-
256.
![Page 68: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/68.jpg)
© Alyson Pavitt, 2011 [email protected]
61
Williams, I., Jones, H. & Hartley, S. (2001) The role of resources and natural enemies
in determining the distribution of an insect herbivore population. Ecological
Entemology, 26, 204-211.
Willis, S., Hill, J., Thomas, C., Roy, D., Fox, R., Blakeley, D. & Huntley, B. (2009)
Assisted colonization in a changing climate: a test-study using two U.K. butterflies.
Conservation Letters, 2 (1), 46-52.
Wilson, R., Thomas, C., Fox, R., Roy, D. & Kunin, W. (2004) Spatial patterns in
species distributions reveal biodiversity change. Nature, 432 (7015), 393.
![Page 69: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/69.jpg)
© Alyson Pavitt, 2011 [email protected]
62
Appendices
Appendix I: Variables used in the analysis of species range dynamics.
Name Definition Source #of
species
Range variables
Cell occupancy The number of grid cells a species was
recorded in during the time periods t1
(1970-1996) or t2 (1997-2008).
Source data
from NBN
37
Change index
Metric for the relative change in the cell
occupancy between t1 and t2 (see
Methods; Telfer et al. 2002).
Source data
from NBN
37
D/ D2 Fractal dimension of species
distribution, calculated from 10 km and
100 km grid cell occupancy at t1 (D) and
t2 (D2) (see Methods; Wilson et al. 2004).
Source data
from NBN
37
Distribution
pattern
The level of population aggregation
within a species’ range, measured using
residual D.
37
Occupancy
change
Change in cell occupancy from t1 to t2,
measured using change index.
37
Range shift Directional movement at the northern or
eastern margins of a species’ British
range, based on the 10 north- or east-
most occupied grid cells (see Methods;
Hickling et al. 2006).
Source data
from NBN
37
![Page 70: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/70.jpg)
© Alyson Pavitt, 2011 [email protected]
63
Appendix I cont: Variables used in the analysis of species range dynamics.
Name Definition Source #of
species
Range variables
Residual D/
Residual D2
The residuals of D and D2 regressed
against cell occupancy at t1 and t2
respectively (10 km cells) (see Methods;
Wilson et al. 2004).
Source data
from NBN
37
Residual D
change
Change between residual D and
residual D2 (see Methods).
Source data
from NBN
37
Trait variables
Abdominal
length
Midpoint between the minimum and
maximum abdominal lengths, with
measurements taken from base to tip.
Bowles,
2010
37
Dry wing
length
Mean length of the left hindwing from
the tip to the base of the wing plug.
Measured using 0.1 mm digital calipers.
NHM 37
Residual
thoracic
volume
Residuals of thoracic volume (length x
width x depth) regressed against the
abdominal length3.
Source data
from NHM
37
Residual dry
wing length
Residuals of dry wing length regressed
against abdominal length.
Source data
from NHM
37
Wing aspect
ratio
Shape of the left hindwing, expressed as
the ratio between dry wing length and
width (taken at the widest point of the
wing).
Source data
from NHM
37
![Page 71: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/71.jpg)
© Alyson Pavitt, 2011 [email protected]
64
Appendix I cont: Variables used in the analysis of species range dynamics.
Name Definition Source #of
species
Trait variables
Wing length
(ImageJ)
Length of the left hindwing (unless
otherwise specified) from tip to the first
cross vein from the base. Measured
using the computer software package
ImageJ (Rasband, 1997-2007).
Hassall et al.
2008,2009
4
Wing length
(Calipers)
Length of the forewing from tip to the
base. Measured using 0.1 mm digital
calipers
Conrad et
al. 2002
1
![Page 72: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/72.jpg)
© Alyson Pavitt, 2011 [email protected]
65
Appendix II: List of all species commonly found in Britain (from Corbet & Brooks, 2008)
and freshwater habitat preferences.
Species name Described
Residents
Aeshna caerulea Azure hawker Strom, 1783
Aeshna cyanea Southern hawker Muller, 1764
Aeshna grandis Brown hawker Linnaeus 1758
Aeshna isosceles Norfolk hawker Muller, 1767
Aeshna juncea Common hawker Linnaeus, 1758
Aeshna mixta Migrant hawker Latreille, 1805
Anax imperator Emperor dragonfly Leach, 1815
Brachytron pratense Hairy hawker Muller, 1764
Calopteryx splendens Banded demoiselle Harris, 1782
Calopteryx virgo Beautiful demoiselle Linnaeus, 1758
Ceriagrion tenellum Small red damselfly De Villers, 1789
Coenagrion hastulatum Northern dragonfly Charpentier, 1825
Coenagrion mercuriale Southern damselfly Charpentier, 1840
Coenagrion puella Azure damselfly Linnaeus, 1758
Coenagrion pulchellum Variable damselfly Vander Linden, 1825
Cordulegaster boltonii Golden-ringed dragonfly Donovan, 1807
Cordulia aenea Downy emerald Linnaeus, 1758
Enallagma cyathigerum Common blue damselfly Charpentier, 1840
Erythromma najas Red-eyed damselfly Hansemann, 1823
Erythromma viridulum* Small red-eyed damselfly Charpentier, 1840
Gomphus vulgatissimus Common club-tail Linnaeus, 1758
Ischnura elegans Blue tailed damselfly Vander Linden, ,1823
Ishnura pumilio Scarce blue-tailed
damselfly
Charpentier, 1825
Lestes dryas Scarce emerald damselfly Kirby, 1890
Lestes sponsa Emerald damselfly Hansemann, 1823
Leucorrhinia dubia White-faced darter Vander Linden, 1825
![Page 73: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/73.jpg)
© Alyson Pavitt, 2011 [email protected]
66
Appendix II cont: List of all species commonly found in Britain (from Corbet &
Brooks, 2008) and freshwater habitat preferences.
Species name Described
Residents
Libellula depressa Broad-bodied chaser Linnaeus, 1758
Libellula fulva Scarce chaser Muller, 1764
Libellula quadrimaculata Four-spotted chaser Linnaeus, 1758
Orthetrum cancellatum Black tailed skimmer Linnaeus, 1758
Orthetrum coerulescens Keeled skimmer Fabricius, 1798
Platycnemis pennipes White- legged damselfly Pallas, 1771
Pyrrhosoma nymphula Large red damselfly Sulzer, 1776
Somatochlora arctica Northern emerald Zetterstedt, 1840
Somatochlora metallica Brilliant emerald Vander Linden, 1825
Sympetrum danae Black darter Sulzer, 1776
Sympatrum sanguineum Ruddy darter Muller, 1764
Sympetrum striolatum Common darter Charpentier, 1840
Migrants
Anax junius Green darner Drury, 1770
Anax parthenope Lesser emperor Selys, 1839
Anax ephippiger Vagrant emperor Burmeister, 1839
Coenagrion lunulatum Irish damselfly Charpentier, 1840
Crocothemis erythraea Scarlet darter Brulle, 1832
Lestes barbarous Southern emerald Fabricius, 1798
Lestes viridis Willow emerald damselfly Vander Linden, 1825
Pantala flavescens Wandering glider Fabricus, 1798
Sympatrum flaveolum Yellow-winged darter Linnaeus, 1758
Sympetrum fonscolombii Red-veined darter Selys, 1840
Sympetrum meridionale Southern darter Selys, 1841
Sympetrum vulgatum Vagrant darter Linnaeus, 1758
*although this species is now a confirmed British resident, it is too recent a colonist to be included in
these analyses.
![Page 74: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/74.jpg)
© Alyson Pavitt, 2011 [email protected]
67
Appendix III: Linear regression models showing highly significant positive correlations
between mean wing lengths and aspect ratios (length/width) for each of the four wings. Data
was taken from specimens held in NHM collections.
f d.f p
Wing
length
left vs right hindwing 4.364 35 <0.001
left vs right forewing 9.433 35 <0.001
left forewing vs hindwing 1.104 35 <0.001
right forewing vs hindwing 8.743 35 <0.001
Wing
aspect
ratio
left vs right hindwing 6.0573 35 <0.001
left vs right forewing 57 35 <0.001
left forewing vs hindwing 1.0092 35 <0.001
right forewing vs hindwing 41.8 35 <0.001
![Page 75: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/75.jpg)
© Alyson Pavitt, 2011 [email protected]
68
Appendix IV: Phylogenetic tree for the resident British odonates (n=37 species), synthesised
from published morphological (Von Ellenrieder, 2002) and molecular (Artiss et al. 2001;
Saux et al. 2003; Fleck et al. 2008; Pilgrim & Von Dohlen, 2008; Dumont et al. 2010)
phylogenetic analyses.
![Page 76: The Future of British Odonata - iccs.org.uk › wp-content › thesis › consci › 2011 › Pavitt.pdfchange index (Telfer et al. 2002) was described by combining the three dispersal](https://reader031.fdocuments.net/reader031/viewer/2022013021/5f12de8d984a28775f614464/html5/thumbnails/76.jpg)
© Alyson Pavitt, 2011 [email protected]
69
Appendix V: Linear models for dry wing length, thoracic volume and wing aspect ratio
regressed against abdomen lengths. The abdomen length metrics used were abdomen length
(mm), abdomen length3 (mm3) and abdomen length2 (mm2) respectively. Model outputs are
given for models with and without phylogenetic signal controlled for.
Appendix VI: Differences in left wing lengths between dry museum specimens (collected in
the time period 1905-1953) and “wet” (recently caught) specimens (collected in 1999; see
Hassall et al. (2008,2009)).
Body metric
With phylogenetic
control
Without phylogenetic
control
F Df p F df P
Dry wing length 95.74 35 <0.001 107.40 35 <0.001
Thoracic volume 41.46 35 <0.001 81.14 35 <0.001
Wing aspect ratio 0.0044 35 0.99 6.19 35 0.018
20
22
24
26
28
30
32
34
left
fore
win
g length
(mm
)
15
16
17
18
19
20
17
18
19
20
21
22
23
24
15
20
25
30
museum wet
18
20
22
24
26
28
30
32
Calopteryx splendens
left
hin
dw
ing length
(mm
)
museum wet
14
15
16
17
18
19
Coenagrion puella
museum wet
15
16
17
18
19
20
21
22
Erythromma najas
museum wet
14
16
18
20
22
24
26
28
Pyrrhosoma nymphula