Trends in occurrence of thermophilous dragonfly species in North Rhine-Westphalia ( NRW )

Since 1996 the “Workgroup Odonata in North Rhine-Westphalia” (“AK Libellen NRW”) has built up a data base including about 150.000 data sets concerning the occurrence of dragonfl ies in North RhineWestphalia (NRW). Th is data confi rms an increase and spread of some thermophilous dragonfl y species in NRW, and the eff ects of climate change evidenced by an increasing average temperature, are considered to be important reasons for this process.


The area
Covering 34.100 km 2 North Rhine-Westphalia (NRW) is one of the biggest federal states of Germany with the largest number of inhabitants (about 17 million people).Th e Rhine-Ruhr-area in the center of NRW is the largest and most densely populated area in Europe.
Th e country is mainly fl at (65 % of its area is situated between 10 and 150 m a.s.l.).Th e southern parts are occupied by mountainous areas with altitudes up to 840 m a.s.l.Th e lowland is a mainly open landscape dominated by highly developed agriculture with several urban and industrial centers.Th e mountains are dominated by woodlands, nowadays the majority consisting of spruce forests (LÖBF 2005).

The climate
North Rhine-Westphalia is situated in a transition area of the atlantic and continental climate.Th e predominating western winds carry mainly wet air from the atlantic low pressure areas.Th erefore NRW has a warm-moderate rain climate with mild winters and relatively cool summers (DWD 1989).Corresponding to the orography NRW is also climatically divided into lowlands and mountainous areas.Th e annual average temperature equates to more than 9°C in the lowlands and 5°C in the mountains respectively.As a region of even higher temperatures the conurbation in the Rhine-Ruhrarea is in an exceptional position.
Th e temporal distribution of precipitation has a summerly maximum in the lowlands and an additional winterly one in the mountainous countries.Eff ects of luff and lee in the mountains produce diff erences in the amounts of precipitation and intensify the diff erences of regional climates.So the precipitation remains below 750 mm in major parts of the Northwestern region "Niederrheinische Bucht" whereas it reaches more than 1.100 mm in the mountainous regions.Th erefore NRW has a high diversity of regional climatic diff erences which for example also cause large amplitudes in the vegetation periods.

The inventory of waterbodies and its historical development
North Rhine-Westphalia is a country with abounding water.It is strongly marked by a dense net of river systems including the main rivers Rhine, Ems and Weser which contribute to the North Sea.Nevertheless especially in the lowlands only a few seminatural rivers and brooks are left, the most running waters being regulated and paved.
Th e quality of water instead is much better today than it used to be some years ago, especially in the age of industrialization about a hundred years ago when this had been much worse.Also a growing number of projects restoring parts of rivers and brooks to its former natural conditions indicate a positive development of the running water bodies.
Natural lakes do not exist in NRW.Th e only particular exception is the "Heiliges Meer" in the North of the country.Th is area consists of a group of small lakes, resulting from erosion and the caving in of salty layers in the underground.Th e lakes are only some hundred years old, several are even younger.But since a hundred years ago a lot of artifi cial lakes have been constructed as reservoirs or gravel pits all over the country.Th ey are mainly concentrated in the southern mountainous country called "Sauerland" and within the alluvial fl oodplains of the rivers Rhine and Weser.
Th e mountainous areas are rich of springs and brooks.Here small water bodies exist e.g.man-made mill ponds or barrages.Until the 18 th century a lot of swamps and bogs had existed in the lowlands, but only small rests have remained until today.Th e ground water level has decreased nearly everywhere and the diversity and density of small water bodies of the historical landscape does not exist any more.Although construction of new ponds and wetland areas is often practiced, these water bodies are subject to an accelerated succession due to eutrophication (in a high amount caused by air pollution) and also infl uenced by higher temperatures.
The "Workgroup Odonata North Rhine-Westphalia" (AK Libellen NRW) In 1996 this honorary working group was founded to work on the protection and investigation of Odonata in NRW.Th e AK organizes annual meetings for all collaborators giving new information and exchanging experiences with the participants.Circular letters and mailings also off er recent information to every member.In every summer a weekend-meeting is organized to investigate dragonfl ies and collect data in areas which have not been well-investigated yet.
Active collaborators can choose between two ways of ascertaining data: fi rst by observing only a few selected water bodies intensively and continuously to get a close image of the entire inventory of dragonfl ies autochthonous in these water bodies.Th ey have to undertake several excursions (at least 5) in a year for 2-4 years in a row.Th e second way is trying to cover a larger area and observing numerous water bodies but undertaking only a small number of excursions (1-3) to each of them.Despite of this all data is welcome.
Information is also available from our homepage (www.ak-libellen-nrw.de)containing dates of meetings, recent news of interesting dragonfl y-records, a download of the recording manual, the current red list, a bibliography as well as distribution maps of all known dragonfl y species in NRW.
Th e AK organizes training courses on exuviae and adult dragonfl ies.It supports dissertations and other works on dragonfl ies.Th ere is also an exhibition available to be rent for presentations in order to give information on dragonfl ies to a great public.
Th e AK is part of an effi cient network together with the LANUV (the major governmental institution for ecology and nature conservation in NRW), the "Biologische Stationen" (institutions for nature conservation on the level of districts), the "GdO" ("Gesellschaft deutschsprachiger Odonatologen") and other fi eldworking groups (e.g. the "Work Group on amphibians and reptiles", "AK Herpetofauna NRW").
Aim of this work is a publication about the dragonfl ies in NRW and the cooperation with other dragonfl y-protection-groups all over the world.
Due to this intensive work on dragonfl ies in the last years interesting new knowledge could be attained but also new questions arose.One obvious result is the observa-tion of positive trends in occurrence of several thermophilous species and other observations (such as a change in the phenology of some species) which indicate a warmer climate in NRW.

The climate change
Th e known global eff ects of the climate change are also visible in NRW.A report of regional climate scenarios for NRW (Gerstengarbe et al. 2004) shows the following results: "Between 1951 and 2000 signifi cant climatic changes could be observed in NRW.Th e most important changes are: -an increase of temperature in the annual average up to 1,5 K -a decrease of days with frost in average up to 20 days per year -an increase of summer days (meteorologically: days with more than 25°C) in average up to 20 days per year -an increase of precipitation in the annual average of more than 100 mm in some areas -an increase of days with strong precipitation up to 8 days a year and a highly signifi cant decrease of days (up to 40) with no or only small precipitation".
In respect to the climatic change in the near future  the report points to the following trends based on statistically solid proved probabilities: "Th e temperature will increase for at least 2 K and therefore the number of meteorologically "cold days" (with frost) will decrease whereas the number of "summer days" will increase.
Th e precipitation will increase on a lower level.Th e increase of water steam pressure will be overcompensated for the reason of increasing temperature and therefore the relative atmospheric humidity will decrease.
In connection with a clear increase of the air pressure the time of sunshine and global radiation will raise as well as the clouding will decrease slightly."

The database
Due to the almost complete interpretation of publications on dragonfl ies in NRW and other sources like unpublished reports, collections and the current data from the AKobservations a solid database with now about 150.000 data sets has been established.Th e spatial and temporal distribution of the acquired data is shown in the fi gures 1a and 1b.It refl ects a nearly complete coverage of NRW and a very strong increase of data based on the intensive and systematical work of the AK in the last decade.Th e peaks before 1996 represent data from publications on dragonfl ies which summarize the knowledge about the species for certain periods of time (for example Le Roi 1915 or Kikillus and Weitzel 1981).
Figure 1a.Th e distribution of data on dragonfl ies registered in the data base of the AK Libellen NRW on the base of large scale map (1:25.000)(m) -quadrants, diff erentiated for three time periods (before 1950, 1950-1995, since 1996); one grid cell represents an area of about 5,5 × 5,5 km.
Nevertheless the fact of an increasing amount of data collected in the last decade has to be considered for the interpretation of any results.It is not always possible to determine whether the "increasing" of a dragonfl y species results from climatic eff ects, other factors (f.e. off er of habitats or changing of habitats for example due to eutrophication) or only because of intensifi ed observations.

Methods
Th e trends in occurrence of Odonata species in North Rhine-Westphalia were analyzed on the basis of the number of observations for each year and species.A data record corresponds to the observation of a species in an investigated area for each day, independent from number or autochthonousness of the species.To take into account the diff erent intensity of this observation (the development of the number of observations is shown in fi g. 1b) this number was set in relationship with the total number of observations per year.So for each species the respective proportion was computed and its relative frequency was determined.With these annual portions an inventory trend analysis on the basis of the Spearman rank-order correlation coeffi cient (Spearman's rho) with the statistical program SPSS 11.5 was computed.Th e classifi cation of areal types ("Faunenelemente") corresponds with St. Quentin (1960), Sternberg (1998) and Sternberg andBuchwald (1999, 2000).

Results
Th e results of trend analyses are given for some thermophilous species in table 1.In addition the below table (Table 1) refers to publications for their fi rst records ("year") in NRW and current new works on their occurrence in our country.
Table 1 shows the currently known fi rst records of some thermophilous species in NRW and its origins.It elucidates that apart from a few species (Coenagrion scitulum, Crocothemis erythraea, Anax parthenope) single observations of most of them have already been made in NRW since the 19 th century.Th is indicates the dynamic distribution patterns of a very mobile dragonfl y species and the infl uence that already a few years with favourable weather conditions could have.1. Year and referred publication of fi rst records of thermophilous species in NRW and important new publications for the occurrence of the species, the Spearman rank-order correlation coeffi cient (Spearman's rho) and its level of signifi cance for the trends of occurring for two regarded periods.

Species
But it also proves signifi cantly the increase of thermophilous dragonfl y species in North Rhine-Westphalia.Especially in the last 25 years the positive trend is at least signifi cant for eight out of ten species.Th is development obviously is connected with the increase of average temperature in the same period of time.Furthermore the rapidity and intensity of the outspread of some species has been advanced by the change of habitat availability.Th ose developments have been observed f. e. for Crocothemis erythraea and Erythromma viridulum in the alluvial fl oodplains of the rivers Rhine and Weser, where many suitable waterbodies had arisen from the exploitation of gravel pits.
Th e occurrence of species in areas that were formerly not populated is also eff ected by climatic changes.For the fi rst time populations of Erythromma lindenii or individuals of Orthetrum brunneum have been recorded in the mountainous parts of NRW in the last fi ve years.In addition phenological changes have been observed for many species occurring nowadays much earlier and longer than in the 1980s.But additional factors may also infl uence the outspread and remain to be analysed in detail.

Comparison of groups of dragonfly species with different areal types due to their inventory development for the time span 1980-2005
Fig. 2 is a depiction of the diff erentiation between two groups of areal type (Mediterranean species / Eurasian, Westsibirian and Pontocaspian species) and displays the por-tion of the species with distinct trends of occurrence (signifi cant positive, signifi cant negative or not signifi cant).In detail, 43% of species with signifi cant increasing trends in their group "Mediterranean species" is clearly higher than in the group of "Eurasian, Westsibirian and Pontocaspian" which is 17%.However with "Eurasian, Westsibirian and Pontocaspian" 36% show a signifi cant negative development.Th ese trends underline a change in species' frequency within favour of the thermophilous species in North Rhine-Westphalia.

Diff erent examples
A well documented example for the increase and spread of thermophilous dragonfl y species is Crocothemis erythraea (Ott 1996(Ott , 2006(Ott , 2007)).Th is species is now widespread in the lowlands of NRW, too.Autochthonous populations have been recorded from several places especially in the area of the Rhineland, where it occurs together with Coenagrion scitulum (Rodenkirchen 2004).Particularly this very conspicious and easily determinable species is a good indicator species showing increasing trends affected by climate change (see also fi gure 3).
Th e infl uence of systematic examinations and monitoring programs is shown in fi gure 4 for Coenagrion mercuriale.Th e erratically increasing numbers of observations in the last years are to be explained by special investigations in connection with the announcement of special protected areas for the "Natura 2000" network of the EU.On the basis of these investigations a thorough knowledge of this endangered species could be gained.But no "real" increase in the populations could be observed.In the Rhineland, an area with a higher average temperature, only one small and very isolated population could be observed.For a long time it has been found in a calcerous springbog.In the "Westfälische Bucht" several partially very large populations are to be found in systems of brooks and ditches.By searching for this species in its best fl ight period and at typical habitats known for the occurency of this species, some new populations have been observed here in the last years.But an increase of Conagrion mercuriale for the reason of climate change cannot be proved.

Regional aspects
For some species the trends of occurrence on the regional level diff er from those in the entire country.In Figure 5 diff erent developments for Gomphus pulchellus in North Rhine-Westphalia and in the region of the upper River Weser valley ("Oberweser") situated in the east of NRW are presented.For the entire country this species does not show a signifi cant trend in the regarded periods of time.But in the valley of the Weser the species has strongly increased by number of occurrences as well as number of individuals since 1995.

Conclusions
Th e analysis of trends statistacally tested and based on a large area and long period of time can only be carried out on the basis of extensive data collection.But often additional information and circumstances must be considered to explain trends in occurrence.Th erefore the data base of the "AK Libellen NRW" is an important fund which has to be continued and further on interpretated closely in the next years.
Th e increase of thermophilous species in North Rhine-Westphalia can be proved on the basis of the collected data.Th e climate change is obviously the most probable reason for it.But there remain other aspects to be regarded, for example the eff ects of climate change on the whole species inventory (also on the base of regional landscapes), the inter-and intraspecifi c concurrence, the decrease of supposed "northern" species (for example S. arctica), the phenology and the local distribution of species.Th erefore not only the data collection has to be continued but also special monitoring programs have to be developed and carried out including currently common species.

Appendix I
Year of fi rst records of Odonata species in NRW, Spearman rank-order correlation coeffi cient (Spearman's rho) and its level of signifi cance for the trends of occurring for three regarded periods.In addition the classifi cation of areal types according to St. Quentin (1960), Sternberg (1998) and Sternberg andBuchwald (1999, 2000) is given.

Figure 1b .
Figure 1b.Development of dragonfl y observations for NRW in the last 155 years, diff erentiated for observations (blue), observations with proved autochthony (red) and coverage of quadrants of the large scale map "1:25.000"(m) (grey).

Figure 2 .
Figure 2. Portion of species with diff erent signifi cant trends in the "areal type groups" of the Mediterranean and the Eurasian, Westsiberian and Pontocaspian species (for the time span 1980-2005).

Figure 3 .
Figure 3. Th e inventory development of C. erytrhaea in NRW

Figure 4 .
Figure 4. Inventory development for C. mercuriale in NRW (additional species-specifi c investigations in special protected areas started in 2003)

Figure 5 .
Figure 5. Diff erent developements for Gomphus pulchellus in NRW and the upper valley of the River Weser ("Oberweser")