Dragonflies and climatic change-recent trends in Germany and Europe

In this paper the trends of dragonfl y expansions during the last decades in Germany and Europe are summarized. It is shown, that there is a general expansion of many species to the north: Mediterranean species expanded to Central and Northern Europe, whereas some African species expanded to Southern Europe, some are even new to the continent. In general this means an increase of biodiversity, but looking at the ecological eff ects, in the medium term a decrease can be expected for mooreland and alpine species. Dragonfl ies can be regarded as a good indicator group for climatic change. Already now in some areas or regions negative eff ects on waters bodies and their dragonfl y communities can be observed and more will occur if e.g. temperature rises or precipitation decreases. Th e consequences for nature conservation strategies – such as the NATURA 2000 network – are outlined and the general need for monitoring programmes is emphasised.


Introduction and some definitions
Since the 1990ies the discussions on the eff ects of climate change became increasingly intensive, in ecology as well as in nature conservation (e.g. Gates 1993). Climate change is now regarded as one of the most important factor threatening species, habitats, ecosystems and biodiversity in general (e.g., Lovejoy and Hannah 2005;IPCC 2007a,b;Settele et al. 2010aSettele et al. , 2010bEEA 2010).
Subsequently an overview is given on the reactions of dragonfl ies following those climatic and habitat changes which have so far been observed. Finally some conclusions for the protection of dragonfl ies are drawn.

Oscillations versus Trends
Th e changes of species` ranges as a result of climatic changes -here the main focus is on expansion -is a normal process, but it is important to diff erentiate between "oscillation" and "trend".
Oscillation can be described as a regular movement from one side to the next or a regular cycle and for a species in a new area it means "coming and going". E.g., if the weather in some years is favourable for the species, it expands and later on, in years with unfavourable weather, it retreats its range back to the former extension.
A trend on the other hand is a gradual development -maybe including some small oscillations -but in general resulting for a species in "coming and staying" in a new area and a constant expansion of the species` range.

Tipping points, scenarios and ecosystem services
Tipping points are described as the moment, when an object or situation is displaced from a state of stable equilibrium into a new, diff erent state; such a shift from one state to another is irreversible.
A scenario is a postulated sequence of possible events based on assumptions. In this context it is for example the increase of carbon dioxide emissions or temperature, as a consequence of economic growth, as it is worldwide coupled with the consumption and burning of fossil energy. It is projected on the basis of current circumstances and trends.
Natural as well as man-made ecosystems supply humans with a lot of diff erent resources and processes and all these benefi ts are defi ned as ecosystem services, such as pollination, drinking water or the decomposition of organic material. Natural systems in particular are increasingly infl uenced and altered by human activities.

Example: the expansion of Crocothemis erythraea in Germany and Europe
Th e best example of the expansion of a dragonfl y is without doubt the Scarlet Darter (Crocothemis erythraea (Brullé, 1832), see fi g 1). In the literature in Germany and other countries north of the Alps this species still was described as a "Mediterranean species", until about three decades ago, also being regarded as a typical vagrant species which only in rare occasions could be observed breeding in northern countries (e.g. Jurzitza 1978).
But then it became permanently indigenous in Germany, in the beginning only in southern federal states and in the lowlands along the Rhine River (Ott 1988(Ott , 1996. In the following years it was expanding to the north -fi nally reaching the border to Denmark in 2009 -and it was registered also in higher altitudes (Ott 2001a(Ott , 2007a(Ott , 2010a, see also fi g 2). Its expansion corresponded with the increase of temperature, even if it is still unclear which factor(s) is (are) the dominating one(s) (e.g. maximum temperature, mean temperature, duration of sunshine).
Th e positive temperature eff ect is also underlined by the observation, that in the extremely warm year of 2003 the species obviously had a second generation (Horn 2003).
Th is expansion was not only registered in Germany, but also in other European countries which is summarized for some countries in the following table (Table 1).
A comparable expansion was noted also for other European countries, like for the Ukraine (see Khrokalo, 2010) and Luxemburg. In the Ukraine -beside other Mediterranean species -Crocothemis erythraea expanded all over the country in the last three decades. In Luxemburg the species was fi rst discovered in 1986 while in 2005 it was present in 17 % of the grid cells (Proess 2006).
Th is expansion in Europe over the last decades fi nally can be regarded as a clear trend of expansion rather than only an oscillation of its range.

Some more examples: other species expanding their range in Germany
Th e example of the Scarlet Darter, a very striking species which is easy to detect at the water (see fi g. 1) is not the only one. Other "southern" species also expanded their   (Ott 2007a, updated) range in Germany northwards or invasions were noted much more frequently. Besides species like Erythromma viridulum (Charpentier 1840), Orthetrum brunneum (Fonscolombe 1837) and O. coerulescens (Fabricius, 1798) which crossed Germany already some time ago or became much more abundant (see Ott 2001aOtt , 2008a, some more recent examples are subsequently presented.
Anax parthenope Selys, 1839: Th e species presently expands its range in central and north-east Germany (e.g. Mauersberger et al. 2002), but also in regions where it has previously established new populations it becomes more abundant. E.g., in some gravel pits in the Rhine Valley  1960-1986Dommanget (1987 known from 28 French departements out of 96, more frequent in the south and very abundant in the Mediterranean < 2000 Grand and Boudot (2006) not known from 3 departements, in 18 rare or very rare, but in all the other departments common to very common, also in central and northern ones Th e Netherlands < 1983 Geijkens andvan Tol (1983) only one sure record from 1959 and 2 others from 1967 and 1968 have been confi rmed < NVVL (2002 steady expansion after the fi rst population was discovered in 1993, thereafter several populations present and increasing < 2007 Bouwman et al. (2008) between 1997 and 2007 registered in more than 250 localities (= 5 x 5 km grid cell) UK < 1995< Hammond (1977, Merrit et al. (1996) no observation ever    1950-1989Dolny et al. (2008 found in 5 grid cells (out of 659) 1990Dolny et al. (2008 found in 105 grid cells (out of 659) 2008-2009 Dolny pers. comm. found in another 10 grid cells near Ludwigshafen and north of Worms it is now more abundant than Anax imperator Leach, 1815, whereas in the mid 1980ies A. parthenope was very rare in this area and A. imperator was the dominating aeshnid in the summer (Ott unpubl. data).
Aeshna affi nis Vander Linden, 1820: A constantly increasing number of observations has been confi rmed in the last two decades. In the Rhine Valley and Lower Saxony the species became, for the fi rst time, indigenous in the mid 1990ies (Ott 1997;Drees et al.1996), later on in 2000 also in north-eastern Germany in the federal state of Brandenburg (Brauner 2005), where up to 2005 it was found breeding in 32 waters.

Boyeria irene (Fonscolombe, 1838):
Th is Mediterranean species inhabiting mainly running waters, but also big lakes, in Germany was found for the fi rst time in 2002 and then again in 2004 (Schmidt 2005). As also in France northward expansion is registered (up to the region Champagne-Ardenne in 2006: Ternois 2008), without any doubt the species sooner or later will populate more waters. Whether the new German population in Lower Saxony (river Ötze) -hundreds of kilometres north from the known sites -could already be regarded as a part of an expansion, needs further investigation (Clausnitzer et al. 2010).
Coenagrion scitulum (Rambur, 1842): Also this damselfl y is expanding its range and was found in Rhineland-Palatinate for the fi rst time in 2006 (Glitz 2008), where in some areas it expanded very much (Lingenfelder 2008). Also it was newly discovered for Bavaria (Karle-Fendt 2006). In North Rhine-Westfalia, where it also was rarely seen in the past decades, it is now much more abundant and shows increasing populations (Grebe et al. 2006).
Beside these species mentioned above, several more Mediterranean species recently became much more abundant and even common in Germany, like Sympetrum fonscolombii (Selys, 1840) or Orthetrum brunneum and O. coerulescens.
It shall be noted that nearly all "southern" species expanding in Germany showed the same pattern: expansions show a clear northward direction and often individuals or populations are also found in higher altitudes.
On the other hand, no "northern" species showed a comparable expansion to the south. Th ere are also some expansions, e.g. by Gomphus vulgatissimus (Linnaeus 1758) or Gomphus fl avipes (Charpenties 1825), but this might rather be a consequence of a better water quality in rivers, than the eff ect of a change in temperature or climate. Whether the new and remarkable expansion of Leucorrhinia caudalis (Charpenties 1825), in northern Germany (Mauersberger 2009;Deubelius and Jödicke 2009) is an eff ect of climatic change defi nitely needs further investigation.

The effects on waters: changing climate -changing communities?
A practical example: the "Kolbental" monitoring-project Th e changes in the fauna of an area can only be described in detail, if these changes can be followed permanently and over an extended period. Th is is the case for example in the "Kolbental" monitoring-project near Kaiserslautern (Ott 2001b). Th e nature reserve "Täler und Verlandungszone am Gelterswoog" (valleys and silted-up zone near lake Gelterswoog) is a ca. 55 ha wetland complex with a mosaic of very diverse biotopes (meadows, forests, abandoned land, lentic and lotic waters etc.). Th is reserve consists of 3 valleys with 11 standing waters (so called "Wooge") and some of these biotopes are protected according to national or international laws (EC habitat's directive: e.g. dystrophic waters and transition mires, Natura 2000-code: 3160 and 7140).
In this area a regional agency (ZWW/TWK) planned to extract about one million cubic meters of ground water for drinking water supply. Th is permission was only given by the regional authority under the prerequisite, that the agency is able to proof the sustainability for the environment by means of hydrological and ecological monitoring. Th ese two monitoring projects started in 1998, and the ecological monitoring consists of the collection and evaluation of abiotic data (climate, soil humidity, water analysis etc.), as well as investigations on the vegetation and the fauna. As indicator taxa for the monitoring project carabid beetles, butterfl ies, grasshoppers and dragonfl ies were chosen.
Whereas in the beginning fl ora and fauna remained relatively constant, dramatic changes occurred after the year 2003 with its extreme warm and dry summer. Th ese changes are still ongoing, even if recently -as a consequence of an increasing precipitation (see tab. 2) -some waters recovered (see also fi g. 3).
Still it is unclear, if and to which extent the extraction of ground water has additional impact on the wetlands. Hydrologists calculated a maximum additional lowering of the ground water table of 10-20 cm per year, which is much lower than the eff ects of a lack of precipitation. Th us they assume that extraction only has a minor eff ect (L.U.P.O. 2009). On the other hand, each additional lowering of the water table in the open waters as well as the ground water dependant ecosystems (GWDE) will prolong the periods of drought and consequently the stress on the species and ecosystems increases.

Th e "Kolbenwoog": an example for the eff ects of the extreme summer of 2003
During the summer of 2003 with its lack of precipitation the water table of the Kolbenwoog dropped and in the beginning only the shallow silted-up zone fell dry. Th is zone is very important for the larvae of many dragonfl y species: here they can hide and escape predation (in particular the sensitive Leucorhhinia-larvae -compare: Henrikson 1988). Also oviposition of many species takes place in these parts of the water with rich structured vegetation along the shoreline. But also in the consecutive 3 years precipitation was very low (see tab. 2) and consequently the water level continued to fall. In summer 2006 the whole lake was nearly dry: only about 20 sqm of shallow water (nearly 40°C water temperature, no oxygen -own measurments) were left (see fi g. 4). At this moment the water surface of the lake was reduced to ca. 0.25% (0.8 hectares under normal conditions) and the water body was reduced to ca. 0.07% (ca. 5400 cubic meters under normal conditions).
Th is nearly dry lake (see fi g. 4), now having wide open shores with only scarce vegetation was colonized by several species, previously not registered at this water before: Orthetrum cancellatum (Linnaeus 1758), Libellula depressa Linnaeus 1758, Gomphus pulchellus Selys 1840 and also a few individuals of Crocothemis erythraea appeared. Th ese species are typical for dynamic or secondary biotopes like gravel pits etc. and here -at dystrophic water bodies with mooreland biotopes (mires and bogs) -can be described as disturbance indicators. Also the immigration of Anax imperator was noted, which became much more abundant and also indigenous. It is well known that this species has aggressive larvae which without doubt can have a strong infl uence on the other dragonfl y species (see e.g. Beutler 1985), as in general dragonfl ies are important and prominent parts of the aquatic food chain (Turner and Chislock 2007).
In the same time period the typical mooreland species -Coenagrion hastulatum (Charpenties, 1825), Somatochlora arctica (Zetterstedt 1840), Aeshna juncea (Linnaeus 1758), Leucorrhinia dubia (Vander Linden 1825) -left the water and still did not return, even although the water table in the meantime recovered to its former level (see fi g. 4).
For these species the Kolbenwoog lost its value, as the water was like a "small bathtub" with wide open shores and all the important habitat structures for larval development were gone: there were no more roots or other dense vegetation left, where larvae could hide or live and where the adults of the endophytically ovipositing species could lay their eggs (see fi g. 3 and 5, also compare e.g. Henrikson 1993). It is well known and documented in many cases, that the complexity of the habitat structure is of general importance for the success of predators (see e.g. Warfe and Barmuta 2004).
Obviously this was a tipping point for the dragonfl y community and it changed fundamentally.

Changes in other waters bodies
A similar change of the dragonfl y fauna was observed in all the other waters of the three valleys of the monitoring area, only in two waters of the Erlental a small population of Coenagrion hastulatum and Leucorrhinia dubia survived (see fi g. 5). Somatochlora arctica was not registered anymore and became extinct in the monitoring area, whereas Aeshna juncea disappeared also as an indigenous species, only single dispersing individuals were registered, but no population was left. As these two species in general became very rare in the Palatinate and the adjacent regions with only very few isolated populations remaining (see Ott 2006aOtt , 2007bOtt , 2007cTrockur et al. in prep. -see fi g. 7) such extreme events may lead to a general extinction of the species on the regional level in the medium term.
A species which on the other side was profi ting from this situation was the damselfl y Ischnura pumilio (Charpentier 1825): all over the Palatinate it colonized these waters with open shores and extremely low water tables (Ott 2008b). Future investigation must show, whether this colonisation is successful for a longer period or whether the species becomes rare again, as it is normally found only in secondary waters.

Changes in the composition of regional faunas
Th e southern species did not only expand their range, they also increasingly dominated the regional faunas, as will be shown with the following examples.  Table 3 shows the change in the dragonfl y fauna (Anisoptera) in the last four and a half decades in two diff erent but neighbouring regions of the Palatinate (= part of the federal state of Rhineland-Palatinate). Whereas in the mid 1960ies the dragonfl y fauna in the lower situated and warmer "Vorderpfalz" was already a mixture of Mediterranean and Eurosiberian elements, in the cooler and higher Westpfalz it was still dominated by Eurosiberian species (see Itzerott 1965). For the author this was a normal and typical situation and the fauna was indicating very well the diff erent climatic frame conditions. About 30 years later the Westpfalz faced a big change: the species numbers increased and the Mediterranean species reached nearly the same percentage as in the "warmer" Vorderpfalz; at the same time temperatures increased about one to two degrees in the formerly "cooler" Westpfalz (Ott 1996(Ott , 2001a)! Again fourteen years later in 2009 the situation did not diff er that drastically anymore, but some changes still have been observed (Trockur et al. in prep.; Ott unpubl. data). Besides some turnovers a slight increase of species numbers in total could be registered. In the Vorderpfalz three new species were found in the meantime: Leucorrhinia rubicunda (Linnaeus 1758), L. caudalis (Charpentier 1840) and Stylurus fl avipes (Charpentier 1825). Th e latter two species are also autochthonous. In the Westpfalz Epitheca bimaculata (Charpentier 1825), Somatochlora fl avomaculata (Vander Linden 1825) and Libellula fulva O.F. Muller, 1764 were new.

First Example: Anisoptera of Mediterranean origin in the Palatinate
Interestingly all new species are Eurosiberian elements, but this does not mean the start of a return to a former situation: looking at the details it is more a stabilisation of the situation. As shown before -see e.g. fi g. 6 -especially the Eurosiberian elements became much rarer and were found in fewer sites. In the central Palatinate forest, a part of the Westpfalz, some Mediterranean species now are defi nitely or probably indigenous, like Crocothemis erythraea or Ashna affi nis (Ott 2010b). Erlentalweiher Walkmühltal In this context it must be considered, that this analysis is only done for the Anisoptera but not for the Zygoptera (as Itzerott 1965 published only on the Anisoptera and no data for Zygoptera are available in this detail). Within the Zygoptera quite some expansions of Mediterranean elements in the central Palatinate were registered in recent years, like for Lestes barbarus (Fabricius 1798), Ischnura pumilio and Coenagrion scitulum (Ott 2006b(Ott , 2008bLingenfelder 2008).

Second Example: the odonatofauna in the SLL+-region
Recently the dragonfl y fauna of the so called SLL+-region -consisting of the two German federal states Saarland and Rhineland-Palatinate, as well as Luxembourg, the French department Lorraine and the Belgian Wallonia -was investigated and analysed for an atlas project (Trockur et al. in prep). In this area, covering 65,401 sqkm, also the southern species increased in abundance and enlarged their ranges. When comparing the situation before and after 1990, many southern species increased in the numbers of grid cells where they were found (e.g. Crocothemis erythraea + 109, Erythromma viridulum (Charpentier, 1840) + 107, Aeshna mixta Latreille, 1805 + 73, Anax imperator + 72). Also species like Anax parthenope, Sympetrum meridionale and Sympetrum fonscolombii became more abundant and very recently another damselfl y -Coenagrion scitulum -showed a remarkable expansion (Lingenfelder 2008), being new for many parts of the area.
If this trend continues, without any doubt more species with a Mediterranean origin will appear in the near future, like Boyeria irene (already present near Lake Constance -see Schmidt 2005 or in the French Departement Haute-Marne -see Ternois 2008), Anax ephippiger (Burmeister, 1839) (already appearing several times as a guestsee Schorr 1989) and even the African Trithemis annulata (Palisot de Beauvois, 1805), which is expanding in southern France and now was also found in Lombardy (Boudot et al. 2009).
On the other hand Eurosiberean elements decreased (e.g. Lestes sponsa (Hansemann, 1823) -39, Coenagrion hastulatum -31). Especially Coenagrion hastulatum is facing a strong decrease in the Palatinate (see fi g. 7); obviously the species is very sensi-  (1965( : Itzerott 1965( , 1995Ott 1996Ott , 2009Trockur et al. 2010, Ott unpubl. data tive to lowered water tables. Some species even seem to be close to extinction, as their populations are very small, the quality of the biotopes is poor (moorelands already degraded) and the distances between the remaining biotopes are very long (high degree of fragmentation); this is in particular true for Somatochlora arctica (Ott 2006a), but also some other species face a similar -only slightly better -situation (Aeshna juncea, Leucorrhinia dubia, see Ott in press).

More examples: the odonatofauna in Bavaria and North Rhine-Westfalia
Th e same trends -increase of southern species, often accompanied by the decrease of mooreland species -were registered in several other federal German states or regions. In a region of Bavaria (Nordwest-Oberfranken) investigations on the dragonfl y fauna started in the 1970ies by the Bund Naturschutz (NGO in nature conservation). In the last years observations of a decrease in mooreland species increased. In 2006 all available old data were analysed and compared with data collected in this year (ÖBO 2007). If possible, the same waters as in former times were investigated to have a direct comparison. For this study a total of 41 water bodies were assessed. Figure 7. Distribution of the mooreland species Coenagrion hastulatum and Somatochlora arctica in the Sar-Lor-Lux-plus-region (Trockur et al. in prep.) For the fi rst time ever in this area Crocothemis erythraea was found now, which in two cases also appeared in typical mooreland waters, together with Leucorrhinia dubia. Th e latter species could not be registered anymore in half of the formerly populated waters, and Aeshna juncea disappeared even from 60 % of the waters colonised in the 1980ies and 1990ies. Also Coenagrion hastulatum dissapeared from 50 % of the formerly populated waters; in particular from the waters below 350 a.s.l. Leucorrhinia rubicunda (Charpentier 1825) was not found anymore, Leucorrhinia pectoralis also vanished from all its former waters and for Somatochlora arctica only one observation was made. So all mooreland species showed a strong decrease, whereas on the other hand Crocothemis erythraea now was found in three mooreland waters; the authors see the climatic changes as the reason for this change in the dragonfl y fauna, as until today the main eff ects occurred mainly in the climatically favourable lowlands.
For the federal state of North Rhine-Westfalia the increase and spread of thermophilous dragonfl ies in recent decades is shown by Conze et al. (2010) through an analysis of about 150,000 data sets which were collected by a working group. Also in this case Crocothemis erythraea was the "leading" species.

Biodiversity increase and consequences for the Red Lists
In the past decades biodiversity on all levels faced a more or less strong decrease, which is documented in the red lists of species and biotopes. If new species now arrive in an area, biodiversity -if we look only at the number of species -increases. To maintain a high biodiversity is one of the goals of nature conservation in general, and consequently the present situation should be regarded as positive and desired, also if we look at the fact that many southern species were on the last red list while now in the updated version many of them could be taken off (Ott and Piper 1998;Ott et al. in prep.).
But we still do not know, whether in the medium or long term, at least on a regional scale, biodiversity will rather decrease. As shown by the examples in Bavaria, the Palatinate or the SLL+region, in particular species of mooreland biotopes, which are more sensitive or stenoecious, seem to suff er from the present climatic situations (increased temperatures, falling water tables, drying out of waters, invasion of other species, e.g. with aggressive larvae). Beside the mooreland species also the alpine species are at risk: when looking at scenarios of future development for many regions, these species may not survive the next decades, as e.g. in Germany many regions will see some kind of "mediterranisation" of the climate (lack of water in the summer and higher temperatures). In particular the small water bodies in higher altitudes may easily dry out for a longer period, which most probably will lead to the extinction of many alpine dragonfl y species.
After some increase in dragonfl y biodiversity on the national level -as a consequence of the invasion of southern species -biodiversity will probably decrease, as we will lose the mooreland and alpine species.

Eff ects on the Natura 2000 web in the Palatinate forest
Th is decrease of biodiversity will probably also take place at the landscape level: due to the lack of water in extreme warm years many waters lost their habitat suitability (see fi g. 8) and thus the species indicating the typical communities of the Natura-2000-biotopes (according to Ssymank 1998). Th is is shown in tab. 4: whereas two decades ago the typical species were still present (see Niehuis 1984), in recent investigations they only could be found in signifi cant fewer numbers or even could not be found anymore at all (Ott 2007b, in press).
Th e species Aeshna juncea, Leucorrhina dubia, Coenagrion hastulatum only survived in very few and small populations, they could be classifi ed as loser; for Somatochlora arctica only one single population is left in the whole German part of the Biosphere reserve Pfälzerwald-Vosges du Nord (Ott 2006a(Ott , 2010b, see also fi g. 7).
Interestingly, several populations of these endangered species are not found in reserves or protected biotopes -dragonfl ies should be integrated more in reserve planning, what was recently also suggested by Heino et al. (2009), who showed that many protected areas were not delineated based on the requirements of freshwater organisms.
In principal the recolonisation of the waters from the French part of the reserve, the Vosges du Nord, is always possible, as there the situation seems to be still better Table 4. Dragonfl y fauna in some waters of the Natura 2000 web in the biosphere reserve Palatinate Forest (A = 1980-1995; B = 2005-2007, water for mooreland species (Duchamps and Morelle, pers. comm.). But on the other hand the waters in the Palatinate now have another abiotic quality and other communities are now established (see the example of the monitoring-project Kolbenwoog), consequently the dystrophic waters lost or will loose their characteristics. Without doubt this also will have consequences for the ecosystem services of these waters (water retention, landscape aesthetics etc.) and other -already existing impacts (fragmentation, lowering of the groundwater table) -will have additional eff ects.

Changes in the phenology
In general insect species react on the increase of temperature with a change in their phenology: increasingly observations are made very early or very late in the season. Th is process started in the 1990ies, since then with an increasing number of such observations.. Here some recent data for Germany are summarized. E.g., in mid December 1994 Sympetrum striolatum was still on the wing in Switzerland and in early November 1999 Lestes sponsa (Hansemann 1823) in Baden-Württemberg (Jödicke 2000). In 1994 and 1997 Somatochlora metallica was on the wing until October (Reder 1997) while in 2000 Gomphus fl avipes was registered in the Rhine Valley still in mid October (Reder 2001).
Extremely early and warm springs -like in 2007, one of the warmest years since climatic data are registered in Germany -did have an additional eff ect on the phenology: in Baden-Württemberg more than 30 species emerged earlier than ever registered before (Hunger 2007). Some for only a few days, but many species emerged even one (e.g. L. barbarus, E. najas (Hansemann 1823), L. caudalis (Carpentier 1840)), two (e.g. C. hastulatum and pulchellum (Vander Linden 1825), C. aenea (Linnaeus 1758)) or even three weeks (L. dubia) earlier then ever observed before.
Sometimes very late emergences are registered recently for Gomphus vulgatissimus. Niehuis and Heilig (2004)  Besides a prolonged phenology of the adults also an impact on the eclosion period was registered which is indicated by very late and "not normal" eclosion: at the end of September 1999 a teneral male of Gomphus vulgatissimus was seen in Lower Saxony (Fliedner and Fliedner 2000). In Baden-Württemberg 33 individuals of Lestes sponsa emerged during the fi rst week of September 2005 (Koch 2005).
In the mid or long term this might lead to a desynchronisation in the phenology. It is well known that thermal impacts on waters -e.g. by power plants -lead to earlier eclosion, even in winter. Consequently, also the general increase of the temperature will most probably have the same eff ect.

More generations: from semi-to univoltine, from uni-to bivoltine
A wide range of species now has a second generation: in Germany species which formerly were univoltine now became bivoltine, or migrating species had a second generation. In the mid 1990ies this was only registered for a few species and areas, like for S. fonscolombii in Bremen, Lower Saxony and northern Hesse (e.g. Pix 1994), or Ischnura elegans and I. pumilio in North Rhine-Westfalia (Inden-Lohmar 1997). In the consecutive years this phenomenon became much more widespread and was seen in an increasing number of species, as well as all over Germany. E.g. it was shown for E. cyathigerum, E. najas, I. elegans and S. fonscolombii in Baden-Württemberg (Schiel 2006;Koch 2002), and for I. elegans and pumilio, E. najas and S. fonscolombii in Bavaria (e.g. Burbach 2000). In Rhineland-Palatinate I.pumilio and E. cyathigerum in some years do have a second generation (Ott 2008b, unpubl. data).
Especially in hot summers, like the one in 2003, this eff ect is apparent and even the Mediterranean Crocothemis erythraea might have a second generation in Germany (Horn 2003). In some years (1999,2000,2006) this was also shown for Anax imperator in southern Germany in four localities, the species generally is known to be bivoltine only for the Mediterranean (Westermann and Weihrauch 2007).
Species formerly not known to have a univoltine cyclus in Central Europe, like Gomphus pulchellus and Leucorrhinia caudalis, show it now in southern Germany (Schirrmacher et al. 2007) or partly, like Leucorrhinia pectoralis and Brachytron pratense, in north-eastern Germany (Brauner 2006).

Moving to higher altitudes
Anax imperator is in general a lowland species, but recently also can be found in altitudes of more than 1000 m a.s.l. (Hunger et al. 2006) and even on 915 m a.s.l. an indigenous population was registered (Westermann 2003c). But also damselfl ies move to higher altitudes, Lestes viridis (Vander Linden 1825) in general found in the lowland is found breeding in 900 m a.s.l. (Westermann 2003a). Also Lestes barbarus was found in the Black Forest regularly up to 700 m a.s.l. where it is also breeding in some cases (Hunger et al. 2006;Salcher 2006), in the Palatinate it populated the higher centre expanding from the lowlands and is now found breeding in altitudes of ca. 400 m a.s.l. (Ott 2006b). Again in the Black Forest at an altitude of 1010 m a.s.l. another damselfl y -Erythromma najas -was found indigenous, which represents the highest reproduction site in Germany so far (Westermann and Westermann 2003) and in 2005 the highest elevation of an autochthonous population in central Europe was registered for Gomphus pulchellus (Selys 1840) (Westermann 2006). Th is western Mediterranean species started its east and northward expansion already in the last decades (Rudolph 1980) and now -besides the continuation of this expansion -also moves to higher elevations.

Changes on the European level -some recent trends
Even if the expansion of damsel-and dragonfl ies on the European level are not demonstrated for all countries in a totally comparable way -as in many countries data collection is done in a diff erent way and intensity -the general pattern however is very obvious.
To compare this trend the situation in the early/mid 1980ies is taken from Askew (1988) who for the fi rst time presented maps on the European distribution. For the situation in 2009 data and information were taken from diff erent publications and national atlas projects (e.g. Nielsen 1998;NVL 2002;Karjulainen 2002;Nelson and Th ompson 2004;Grand and Boudot 2006;Parr , 2009Bouwman et al. 2008;Dolny et al. 2008;Bernard et al. 2009), as well as websites of the national odonatological societies.  Askew (1988).

Species name Range expansion in
Ischnura pumilio New for Sweden in 1992, remarkable increase from 2007 onwards Coenagrion scitulum France, Germany, Belgium, Luxemburg, also in the East, e.g. in the Czech Republic

Erythromma lindenii
North-eastern France, parts of Belgium, northern and eastern Germany, new to UK

Erythromma viridulum
North-eastern France and Netherlands ( Especially when looking at the northern European countries, expansions from the south are obvious, or at the "other end of the continent": at the southern countries, where recently African species expand to the north as well (see below).
In Sweden for example beside the above mentioned new species otherswhich were already present in low numbers -currently expand in a remarkable way, like Calopteryx splendens (Harris 1782), Libellula depressa (Linnaeus 1758) and Sympetrum striolatum (Charpentier 1840). The same trend is seen in the UK, where species formerly mainly restricted to the south -e.g. Libellula depressa and Orthetrum cancellatum -expanded northwards. Libellula depressa reached Scotland where it was first recorded in 2003 and again twice in 2007 .
New for Lithuania in 2003 became Aeshna affi nis Vander Linden, 1820 (Bernard 2005), which is very likely to be also indigenous. For Latvia in 2008 Anax parthenope was registered for the fi rst time and in 2008 and 2009 it was found in fi ve localities (Kalnins 2009), where in one case it also eclosed. In Belarus several species are now new to the national fauna and/or expanded, like Sympecma fusca (Vander Linden 1820), Lestes viridis, Erythromma viridulum, Orthetrum brunneum and O. albistylum (Selys 1848) (Buczynski and Moroz 2008).
On the other hand in the Mediterranean countries we can observe a recently started process: the expansion of African species as shown in table 6 (again compared with Askew 1988; the present situation according to the websites mentioned above and Boudot et al. 2009).
Th e Violet Dropwing (Trithemis annulata, see fi g. 9), a typical species all over Africa and the Middle East, formerly occurred only up to southern Spain and central Italy (Askew 1988). But now it can be found even in southern France, having crossed entire Spain and also the Pyrenees and in Italy it now reached the area of Ferrara (Boudot et al. 2009). Another Trithemis species -the afro-tropical Orang-winged Dropwing (Trithemis kirbyi Selys 1891) -was not known for Europe at all, but then was discovered for the fi rst time on the isle of Sardinia in 2003 . After being discovered in 2007 also in southern Spain near Malaga by D. Chelmick, in 2008 for the fi rst time larvae of this species were found proofi ng its fi rst autochthony in Europe (Cano-Villegas and Conesa-Garcia 2009).
Th e third example of an African respectively Asian species expanding its range to the north is the Black Pennant (Selysiothemis nigra Vander Linen 1825), which in Italy is actually found up to the area of Trieste/Venice (Boudot et al. 2009) in the eastern part and up to Parma (M. Salvarani pers. comm.) in the western part.

Biological effects -Climatic change: a filter for different ecological strategies and species
Th e biological eff ects of the rise in temperature for Odonata could be summarized as follows (updated from Ott 2001a; see also Hickling et al. 2005;Corbet et al. 2006;Dingemanse and Kalkman 2008;Hassel and Th ompson 2008): -more prominent tendency for expansion -more northerly breeding, also breeding in higher altitudes -changes in the composition of the fauna -eclosion earlier in the season, overall alteration in the phenology -second generation, changes in voltism -more rapid larval development Th e tendency for expansion is in particular notable in warm years like 2003: see in this context e.g. Parr ( , 2009  Finally climatic changes can be regarded as a fi lter: they favour the species which are able to adapt to the new situations (higher temperatures, drying out of waters etc.) and eliminate the species which cannot cope with the new environmental conditions. Th ere are winners and losers of the situation.

The future is … hot?! What do scenarios tell us and what does this mean for dragonflies?
Even if in some countries (e.g. Germany) or even in the entire EC the emissions are stable or reduced (see EEA 2009a), worldwide emissions of carbon dioxide and other greenhouse gases will still be increasing (IPCC 2007a).
Assuming that these gases are the reasons for the climatic changes (on which there is a general agreement within the scientifi c community), the changes of the abiotic conditions (e.g. temperature increase, changes in the precipitation) will go on, as well as the eff ects on the biotopes and communities.
Th e diff erent scenarios -like the ones which are used by the IPCC (IPCC 2007a) or the ones in the ALARM-project (Settele et al. 2010a(Settele et al. , 2010b) -expect that there will be an additional increase of temperature of at least 2 degrees within the next decades, some scenarios expect an even higher increase.
Th is means that all the shown conditions for dragonfl ies in Europe, e.g. higher temperatures in the waters, lack of the precipitation in summer, falling water tables, Table 7. Winners and losers of climatic change (Ott 2001a, updated higher air temperatures, more sunshine etc. will continue and become even more intense in the future. As the range expansions to the north demonstrated here (see tab. 6) were the results of only an increase of about 1° Celsius in Central Europe in the last decades, in the future the changes of the waters and their communities -higher proportion of thermophilic and southern species etc. -could be expected to be even stronger and faster, fi nally these species will dominate the dragonfl y fauna.
However, also some Mediterranean species in the medium term may loose large parts of their distributional area in the south, as in particular in the Mediterranean many waters will dry out and so lose their biotope quality in general for aquatic species (see August and Geiger 2008;Ott 2010a).
Th e eff ects of climatic changes in the Mediterranean will however be much stronger on other taxa, like the Trichoptera and Plecotera, as these taxa have many endemics (Hering et al. 2009;Tierno de Figueroa et al. 2010).
On the European level there might be little concern for most of the southern generalist dragonfl y species level, while on the other hand the species of moorelands, higher altitudes and colder biotopes will continuously be eliminated, as in particular these biotopes will suff er in the next decades (Ott 2001a, in press). In higher areas biodiversity will increase as a result of the "invasion" of lowland species -see Oertli et al. (2008) and Oertli (2010) -but this will without any doubt have a negative eff ect on the more sensitive species of the mountains, which can not move any higher, as there are no more waters. In general mountains like the Alps or the Pyrenees will face strong impacts and changes of their water regimes (e.g. for the Alps: EEA 2009b) and in particular species with a small range and those which live in rare climates (e.g. interglacial relicts) will be reduced disproportionally (Ohlemüller et al. 2008).
Consequently these cold stenothermal species will be eliminated, which is also true for other sensitive species of the lower mountains or even the lowlands. In the Black Forest (see Hunger et al. 2006) Aeshna caerulea -a species restricted to peat bogs above 830 m a.s.l. -is more or less extinct, as the climatic circumstances became increasingly unfavorable for the species.
Th e same is true for Aeshna subarctica in northern Germany: their larvae are specialists of peat bogs, and monitoring studies in the federal states of Brandenburg and Mecklenburg-Vorpommern show their dramatic decrease (Peters 2008;Bönsel 2001). Th e reasons seem to be the eutrophication of the waters in general and the extreme summer heat in the waters, which has negative impacts on the larvae, as these have preferences for lower temperatures.
Furthermore there will also be many cumulative and synergistic eff ects, which hitherto have hardly been studied.
Th ese factors are for example: * acidifi cation of the waters ("acid rain", impacts on the aquatic communities) * eutrophication through immissions (leading to oxygen consumption, algal blooms and succession) * (ground-)water extraction (lowering of water table, impacts on ground water depending ecosystems) * higher concentrations of pollutants / toxic substances (mainly in the running waters) and also alien invasive species (alien crayfi sh, grass carps etc.) will play an increasingly important role.
For example the level -and thus also the lowering -of the groundwater determines the relative susceptibility of regions to changes in temperature and precipitation, ergo the extraction will be an additional and cumulative threat (see Maxell and Kollet 2008). Th is might be the case for the example "monitoring Kolbental" which was shown here: if at the end the extraction has a negative eff ect on the groundwater level, the whole area with its biotopes will be impacted even more.
Another example is the increase of the water temperature of rivers: the mean water temperature in the river Rhine increased by 3 degrees during the last 100 years, 2 degrees as a result of cooling water discharge and another one as an eff ect of climatic changes (BUND 2009). But not only the mean water temperature increased, also the days with water temperatures above 23° and 25° C, and the probability of extreme high water temperatures, passing a critical -or tipping -point of 28 degrees, where many species (e.g. fi sh, molluscs) die and then cause "toxic waves" of ammonium etc., leading to other impacts.
Th ese problems are reviewed and summarized for European rivers by WWF (2009), where it is shown, that the known and expected changes in the river temperatures do already have many ecological consequences -and will have even more in the future -as the communities of the waters are adapted to a certain temperature regime. Eff ects are posed on the abiotic conditions -e.g. lack or higher consumption rate of oxygen (Sand-Jensen and Pedersen 2005) -as well as on the composition of the communities (e.g. the mollusc fauna will be reduced: Mouthon and Daufresne 2006; the fi sh communities altered: Daufresne and Boet 2007). Finally complete food webs are changed (Emmerson et al. 2005).
It is also relevant that not only the general temperature increase will have an eff ect on the fl ora and fauna, but also the extreme situations (e.g. heat waves), which are expected to happen much more often .
Th is leads to another aspect: as shown before, temperature increase functions like a general fi lter (see tab. 7), but the increasing number of extreme events will function as a second fi lter.
Extreme climatic conditions -which used to be rare and localised -become more and more abundant and also new situations will occur, which until today did not happen at all (see Meehl et al. 2000 andWigley 2009). Th is will lead to many impacts, also on the communities' level (see the above mentioned examples of the rivers, or Th ibault and Brown 2008) and synergism of global warming and other stresses (e.g. habitat destruction) can disrupt the communities (Root et al. 2003).
Still it is unclear, whether the undoubtful increasing of the competition when new species arrive in a water body, also leads to an exclusion of the former fauna. Th is might be in particular the case, when new biotope types are colonised: e.g. Crocothemis erythraea enters now acid waters (Ott 2010b), whereas the species formerly preferred secondary waters like sand-and gravel pits (Ott 1996) and now may use another niche (see here: Broennimann et al. 2007). Emmerson et al. (2005) pointed out, that already small changes in the number of species in a food web can have consequences both for community structure and ecosystem processes, consequently community stability and ecosystem functioning is altered. In particular when a top predator arrives, what is shown by Byström et al. (2007): the extinction of the char was caused by the pike (trough predation and competition) which expanded to the north and eff ects on the whole food web were registered. Consequently, also eff ects on the dragonfl y fauna of northern countries can be expected, even if there are still many open questions in community ecology (see Booth et al. 2009: what will be the eff ects of the changes in phenology on the diff erent trophic levels?). Biotic interactions and feedback processes lead to highly complex, nonlinear and sometimes abrupt responses. To identify and quantify these processes remains a huge challenge (see Walther 2010).

Conclusions
In Europe, dragonfl ies have a moderate number of species, their ecology is mostly well known, and they are easy to identify, thus they are perfect indicators. Th is is in particular true -as shown here -for the eff ects of climatic changes on diff erent levels (single waters, landscape or national / European level).
Contrary to other taxa they depend only on waters, which are more or less omnipresent, and their expansion is only due to their dispersal and migration behaviour. Butterfl ies depend also on plants, and if these do not expand their range, also the butterfl ies are unable to do so; grasshoppers are transported sometimes by vehicles and so their expansion is some kind of artifi cial. Th is makes dragonfl ies some kind of unique as climate change indicators.
Here it is shown that in the recent decades there were massive range expansions of damsel-and dragonfl ies in Europe, leading to changes in the communities. Climatic changes are the reason for these expansions, leading to a higher biodiversity in many areas, but they are also the reason that now some species are threatened or will be threatened in the future. Th is is in particular true for mooreland species and species of higher altitudes, in the future maybe also for species of springs or species of smaller running waters (mainly in the Mediterranean).
Climate change can be seen as a threat for the dragonfl y fauna in addition to the impacts which already had been identifi ed. Many synergistic and cumulative eff ects do occur and will do so even more in the future (e.g. lack of precipitation and an increase of water demand and consumption). Alien invasive species (e.g. fi sh, crayfi sh) might also play an increasingly import role.
Th e eff ects -most of them negative -for the waters and (dragonfl y) communities have a consequence also for future strategies in nature conservation, as e.g. one of the most important concepts to protect biodiversity in Europe -the Natura-2000-network might not work anymore, as these biotopes are increasingly detiorated and lose their function.
To follow these processes and to identify the eff ects of global change phenomena, it is of crucial importance to establish and maintain European wide data collections and monitoring schemes.