-------Received 14 January 2008; revised and accepted 21 May 2008------- Habitat requirements of Orthetrum coerulescens and management of a secondary habitat in a highly man-modified landscape (Odonata: Libellulidae) Hansruedi Wildermuth Haltbergstrasse 43, CH-8630 Ruti, Switzerland. <[email protected]> Key words: Odonata, dragonfly, Orthetrum coerulescens, habitat use, habitat recognition, habitat management, conservation, Switzerland. ABSTRACT Due to the destruction of its primary habitats, the West Palaearctic libellulid Orthe trum coerulescens has suffered much decline in central Europe. However, at the re gional scale it has survived in a variety of secondary habitat, such as draining ditches. In order to find adequate measures for its conservation and promotion, habitat use and habitat recognition of 0. coerulescens were investigated by description and ex perimentation at fenland ditches in a small nature reserve in the Swiss Plateau. This breeding habitat, which harbours a viable population, had been restored and main tained for 25 years. The most densely populated sites comprised small ditches be tween 40-70 em wide, with rather sparse vegetation of narrow-leaved plants and that had parts of the water surface uncovered; the peaty, mud ground was partly overgrown with submerged pads of stonewort (Chara spp.). Water was mainly sup plied by seepage springs with a mixture of local slow flow that were hardly recog nizable and shallow sites, which were used for oviposition. In hot summer spells the water temperature could exceed 30°C. Some freezing occurred in winter, but the mud was permanently ice-free. The development of the breeding population, which comprised more than 200 individuals in 2006, was followed over two subsequent years. My data indicate that conservation and promotion of 0. coerulescens popu lations in small ditches can be achieved by relatively simple habitat maintenance, such as a rotational strategy of clearing ditches, using of small weirs to prevent or protract desiccation and annual cutting of the surrounding litter meadows. INTRODUCTION In industrialized and densely populated countries the primary habitats of many Odonata species have largely been destructed or degraded (Corbet 1999: 564 ff.). As result of the landscape changes a number of ecologically specialized species have declined or become extinct on a regional or national scale, as demonstrated for Neha lennia speciosa (Charpentier) in Europe (Bernard & Wildermuth 2005). Yet, some species have persisted in the face of anthropogenic habitat modification, often by International journal of Odonatology 11 (2) 2008: 261-276 261 ~-------------Wildermuth-------------- colonizing secondary biotopes that are similar to the ecological conditions of their primary habitats. In contrast to primary biotopes small secondary water bodies need to be maintained in order to function as viable larval habitats of specialists. In some cases this requires a sophisticated system of management such as the rotational model aimed at Leucorrhinia pectoralis (Charpentier), another species that is en dangered in Europe (Wildermuth & Schiess 1983; Wildermuth 2001). As a prere quisite for effective practical conservation, an appropriate knowledge on the specific habitat requirements of a focus species is needed (e.g. Wildermuth 1992). One species that has adopted certain secondary habitats is Orthetrum coerulescens, a libellulid that is widespread in Europe (Kalkman 2006). 0. coerulescens inhabits natural streams, flushes, seepages and runnels in areas of moorland, but also secon dary waters like draining ditches, streamlets in gravel pits and rivulets in partly ex ploited peat-bogs. While many authors have described the variety habitats used by this species (e.g. Buchwald 1983, 1989; Wildermuth & Krebs 1983; Clausnitzer 1988; Muller 1992; Weihrauch 1998; Buchwald & Schmidt 1990; Sternberg & Buchwald 2000, White 2006), as well as its reproductive behaviour (e.g. Heymer 1969; Parr 1983; Huber 1984; Miller & Miller 1989; Lee 1994; White 2008), com paratively few studies have quantified aspects of its population biology (see Juillerat 2002; Kery & Juillerat 2004). As a consequence of human impact many central Euro pean populations of 0. coerulescens have declined or become extinct (e.g. for Switzerlerland cf. Ris 1886; Juillerat 2005). Therefore, this species is considered, on a national scale, to vary from near threatened to endangered in Austria (Raab 2006), Belgium (De Knijf 2006; Goffart 2006), Germany (Ott & Piper 1998), The Nether lands (Ketelaar & Kalkman) 2002 and Switzerland (Gonseth & Monnerat 2002). In this study I report on a viable population in the Swiss Plateau that established itself in recently cleared fenland that possesses restored and maintained drainage ditches. The study was aimed at conservation and promotion of 0. coerulescens in a secon dary habitat and comprised various ecological aspects such as size and phenology of the annual populations. I demonstrate for 0. coerulescens that successful population conservation may be achieved by relatively simple modes of habitat management. STUDY SITE AND METHODS The study was carried out in a 1.5 ha fen near Hinwil Switzerland (47°17.87'N, 08°48.82'E; 550 m a.s.l.). The small area named 'Chuderriet' was part of the nature reserve 'Drumlinlandschaft Zurcher Oberland'. This nature reserve consisted of a 1. 7 km2 core zone, which comprised a number of mostly forested drumlins that were interspersed with fragmented, largely exploited, bogs and litter meadows and was surrounded by intensely used agricultural and urban landscape. Between ca 1970 and 1980 the 'Chuderriet' was not used as litter meadow. It became densely over grown by large shrubs and therefore unhabitable for Odonata. In winter 1981/82 the area was cleared and the meadows were restored, including the original system of draining ditches with an overall length of ca 330m. The ditches were mainly fed by water originating from scattered seepages at the foot of the adjacent drumlin. In 1989 it was first recognized that the ditches had been colonized by Orthetrum coerulescens, some 4 km away from the nearest known site with a permanent popu- 262 International journal oi Odonatology 11 (2) 2008: 261-276 --------Promotion of Orthetrum coerulescens in a secondary habitat-------- lation of this species. For 25 years subsequently, the area has been regularly main tained by annual cutting of the vegetation, removal of neophytes (Solidago gigantea) and by a sporadic manual or mechanical clearing of the ditches. In order to delay des iccation of the ditches during hot summer periods with no or little precipitation, six simple and controllable weirs have been installed successively (Figs 1, 2c). Extensive monitoring of 0. coerulescens began in 1989 and was intensified in 2006 and 2007. In order to assess the annual population size and to collect pheno logical data, the ditches were divided into six sections (Fig. 1) and were inspected every one to three days depending on the weather conditions between 2.5 h before and 2.5 h after solar noon. Counts were made of tenerals, exuviae (though these were not systematically searched for), territorial males, females, copulations and ovipositions. Random samples were made of all Odonata larvae that inhabited the ditches. The following habitat parameters were recorded: width of the ditches at underground pipework A ditch sections (flow) -II- weirs 51-56 q birch (Betula pendu/a) 'j' pine (Pinus sy!vestris) ~ wood, bushes ®-® numbering of ditch sections .. •o .. 0 20 40 60 80 100 m • 0 0 Figure 1: The study site 'Chuderriet' with the system of drainage ditches. The monitoring tran sects followed the ditches that were divided into six sections 1-6. 51-56 show the sites of con trollable weirs that largely protract desiccation of the ditches during dry and hot periods. International journal of Odonatology 11 (2) 2008: 261-276 263 --------Wildermuth-------- water level, depth of the ditches between shoulder and firm ground, density of emer gent vegetation, visibility of open water, temperature, water flow, depth of water and depth of the muddy sediment. The sections of the various ditches differed in some abiotic features, therefore en abling a comparison of factors that presumably contributed to the attractiveness for reproductively active adults. In order to test the hypothesis that adult 0. coerulescens recognize the oviposition sites by small patchy or linear, freely visible water surfaces, three experiments were conducted: ( 1) Ditch section 2 featured lush emergent vegetation, mainly consisting of Car ex acutiformis and Phragmites australis, almost covering the entire water surface. Here all the plant stems and leaves (20-25/dm2) were cut just above the water surface within an area of 0.6 x 0.8 m. After removal of the cut material the water was clearly visible, only interspersed by cut stems. The presence and behaviour of adults were recorded before and after the habitat manipulation, and a comparison was made between the cut area and the adjacent vegetated parts of the ditch. (2) A similar experiment was conducted on the moderately overgrown ditch section 4, where the vegetation was cleared from a 4 m-stretch (Fig. 2d). Next, the water surface was covered with aluminium foil (4.0 x 0.6 m), which reflects sunlight but does not polarize light horizontally as the surface of small water bodies does (Schwind 1991; Bernath et al. 2002). The presence and behaviour of the adults was compared between situations with and without aluminium foil. (3) A sheet of black plastic foil (0.2 x 3m), which reflects horizontally polarized sunlight similar to flat water surfaces and therefore is attractive to Odonata (cf. Wildermuth 1998), was placed on completely dry vegetation where the water course of the narrow ditch 4 ran underground in pipework. All experiments were conducted under optimal weather conditions (air temperature > 20°C, sunshine, calm) that corresponded to high population density and frequent reproduction activity of 0. coerulescens. For statistical analyses, r-test, Kolmo gorov-Smirnov-test and Mann-Whitney-test were used (Zar 1984). RESULTS Habitat characteristics The ditches ran within litter meadows containing mixed plant communities, mainly consisting of gramineous and rather low vegetation -mostly Caricetum davallianae and, to a lesser extent, Molinietum caeruleae - interspersed by Phragmites australis in a park-like landscape with scattered trees (Pinus sylvestris, Betula pendula) and bushes (Salix spp.). The rendezvous-sites and the larval habitats of Orthetrum coeru lescens were characterized by the following features: (1 ) ditches mostly 40-70 em wide (range 25-200 em) and 10-40 em deep between ditch shoulder and water sur face, depth of water at most sites 1-20 em (Figs 2a, b) but increasing to up to 50 em behind weirs (Fig. 2c), longitudinal slope of ditches low (~ 8%a), no or just slightly visible water flow that rarely exceeded 5-15 cm/s especially near the seepages; 264 International Journal of Odonatology 11 (2) 2008: 261-276 ---------Promotion of Orthetrum coerulescens in a secondary habitat--------- Figure 2: Rendezvous-sites and larval habitats at draining ditches in the 'Chuderriet' - (a) small ditch with shallow water and sparse emergent vegetation ideal for reproduction of Orthe-trum coerulescens; (b) the same ditch as in (a) in winter- as this ditch is fed by seepages ice is only formed superficially and the mud is left unfrozen; (c) simple wooden weir with plastic plug for emptying the ditch during the harvest period of the adjacent litter meadows; (d) small ditch after untimely cutting of the vegetation - experimental situation at ditch section 4. International journal of Odonatology 11 (2) 2008: 261-276 265 --------------Wildermuth--------- (2) submerged vegetation (mainly consisting of Chara spp.) absent or patchy; emer gent vegetation sparse to lush with narrow-leaved plants - predominantly Carex davalliana and C. (lava near spring mire seepages, C. acutiformis and C. rostrata at lower reaches of ditches, and Phragmites australis; (3) ground covered with loose mud consisting of peat on clayey ground mostly 15-35 em, reaching a maximum depth of 50 em, and with many sites rust-coloured by bacteria-produced iron oxides (Figs 4a, c); (4) water surface patchily covered with thin floating layer of biogene manganese oxides (Fig. 4b). The water temperature near the surface largely followed the air temperature over long distances except from the stretches close to the seep ages along section 4, which exhibited a temperature gradient. During hot summer spells the water temperature sometimes exceeded 30°C. In winter only superficial icing of the water was observed, leaving the mud permanently ice-free. Phenology and population size Developmental timing differed considerably between the two focus years. In 2006, after a cool and wet May, the emergence period started in early June and the repro ductive period ceased at the end of July, reaching a peak in mid-June; June and July were hot and dry, but August had heavy precipitation. In 2007 (Fig. 3), after a mild winter and an unusually warm and dry April, the emergence period started during the second half of May. Between June and August, however, the weather was vari able and often wet. Accordingly, the reproductive activity of 0. coerulescens was in terrupted, attaining a peak in early July and ending in the second half of August. Adult censuses were much larger in 2006, with the maximal number of territorial males per transect reaching 115 (mean= 72.6 ± 30.3, range= 2-115, n = 16) in con trast with 2007 when a maximum of 64 (mean= 28.9 ± 15.6, range= 1-64, n = 30) males were observed. The corresponding values for females equal18 (mean= 6.4 ± 5.8, range= 0-18, n = 16) and 6 (mean= 1.4 ± 1.9, range= 0-6, n = 30). Territorial males occupied all six ditch sections throughout the reproduction pe riod. The highest male density per transect, calculated per 10 m stretches for the total length of all ditches, equalled 3.5 in 2006 (4 July) and 1.9 in 2007 (20 June). However, the abundance differed between sections 1-6, for example between 0.9 and 8.0 on 4 July 2006 and between 0.5 and 3.4 on 20 June 2007; these differences were not significant (Kolmogorov-Smirnov-test, 2006: D = 0.281, n = 19.2, p < 0.1; 2007: D = 0.133, n = 12.0, p > 0.2). Tenerals, copulation and ovipositing behaviour were observed, without exception, at all sections during 2006-2007, implying that all six sections were suitable for reproduction. Density of vegetation Stretches with dense vegetation, where the cover was more than 60%, sedges were up to 0.8 m high and surmounted by single reeds that attained 1.5-2 min height, and hardly visible open water were generally avoided by both sexes. In such areas, there were between 200-250 leaves/stems of Carex acutiformis and two to three reed stems m-2 Sites with less dense vegetation, e.g. localities with 80-100 blades/stalks and five • reed stems or no Carex-vegetation with 15 reed plants m-2 were almost permanently , occupied by territorial males. At a very densely overgrown stretch of section 2, a 0.5 m2 section of the emergent vegetation was completely cut near the water surface. As a consequence the formerly unoccupied site was occupied subsequently by one male at every (n = 14) monitoring 266 International journal of Odonatology 11 (2) 2008: 261-276 ---------Promotion of Orthetrum coeru/escens in a secondary habitat-------- May june july August 30 "' Iii ~20 .l!l "'10 0 J':j Ill I I I 60 50 "-o 40 i ~30 "' "' 20 II il 10 II I I I I 0 ••1 I !':j •• f"c:' ':j I .. I • . ~ I, • 1.1 Jl I "' "' i'oj ·•I ·;;; 0 I • • • • 0 "' May june july August Figure 3: Seasonal development of the population of Orthetrum coerulescens in the 'Chuder riet' in 2007. International journal of Odonatology 11 (2) 2008: 261-276 267 ---------------Wildermuth---- -------~-~ --- period during the next eight days. No males were present at the adjacent stretches with dense vegetation. Copulation was recorded at two occasions and the site was vigorously defended against conspecific intruders as well as Somatochlora flavo maculata (Vander Linden) males. For the second experimental site a 4 m stretch at section 4 was selected. Due to less dense vegetation the locality was mostly occupied by two males prior to the experimental change. After clearing the site along the shal low ditch (including the border on both sides; see Fig. 2d), the site was permanently occupied by two to three males during 20 monitoring periods over the subsequent 17 days. In addition fights, copulations and ovipositions were repeatedly observed. A lack of upright vegetation did not prevent reproductive behaviour as males and copulating pairs were observed perching on, or near to, the ground. Solar radiation Most parts of the six sections were generally sunlit during clear days from ca 3.5 h before and 3.5 h after solar noon. Only the lower part of section 2 was shaded by trees in the morning up to about solar noon, presenting a contrast to sunlit stretches, not occupied by territorial males in the forenoon. During this time of the day the den sity of males (per 10m) varied between 0 (n = 4) and 2.5 (n = 4) at the shaded areas and the adjacent sunlit stretch respectively, with similarly dense vegetation (cover - 40-60%) along the whole section. Towards solar noon, when the ditch section was about to experience full sunlight, two or three males were observed in the sunlit meadow nearby possibly waiting for the available habitat to warm. On the other hand, when the air temperature exceeded 28-30°C two to three males also perched in a tree shaded place of the habitat at the end of section 6 that was generally avoided at lower temperatures. Visibility of open water Due to hot and dry weather in June and July 2006 some stretches of all ditch sections desiccated gradually thereby presenting a natural comparison between the attrac tiveness of wet (Figs 4a-c) and dry (Fig. 4d) stretches of ditch for 0. coerulescens. Males significantly (Mann-Whitney-test U = 0, n = 14, n = 14, p < 0.001) preferred 1 2 sections with open water visible to the human eye (average of 6.2 males per 10 m; range 3.5-11.7) compared with dried-out stretches (mean= 0.8; range= 0-3.3) males. Copulations and ovipositions were seen exclusively at stretches with open water. In respect to attractiveness, no significant difference was found between sites with clear and semitranslucent water surfaces that had a thin layer of manganese oxides (Mann Whitney-test, U = 26.5, n = 7, n = 8, p > 0.05). Some males established their terri 1 2 tories also at sites with wet mud and a rough, irregularly shiny surface (Fig. 4c). However, females did not use wet mud for oviposition. Since the overall population of 0. coerulescens was large in 2006 the gradual desiccation of the ditches caused greater crowding and male densities reached 10.0- 11.7 per 10m. At section 4, where the vegetation had been cleared (Fig. 3d, see above) and the water surface was covered with aluminium foil, all territorial males (n = 2-3) moved away from the site and occupied only the adjacent ditch stretches where there was visible water. An absence of males was noted on each of the four monitoring periods. 268 International journal of Odonatology 11 (2) 2008:261-276 --------Promotion of Orthetrum coerulescens in a secondary habitat-------- Shortly after the dry stretch between the open ditch sections was lined with black plastic foil it became occupied by a territorial male. No individuals were present at this site before and after the experiment. Hence, the presence of the black plastic proved attractive to 0. coerulescens, with 1-2 males (mean 1.4) present during 19 out of 20 checks, either perched on a reed stem, patrolling or fighting conspecifics. Moreover, one female performed dipping oviposition flight over the plastic foil. Depth of water Territorial males were mostly present at sites with shallow water (Fig. 5). This was recognizable at section 6 (20 m long) that was bisected by a weir into two equal parts, with relatively deep (up to 50 em), clear water upstream and shallow (1-5 em deep) water downstream. In the upper and lower sections between zero and one male (average= 0.27) and one to five males (average= 2.47) were present (Mann-Whit ney-test U = 1.5, n = 8, n = 8, p < 0.01). Copulations (n = 135) and ovipositions (n 1 2 = 26) were almost exclusively observed at sites with shallow (1-5 em deep) water and also those with deeper water but occupied by pads of Chara spp. that nearly reached or outgrew the water surface, thus feigning shallow water (Fig. 4a). Females oviposited sometimes in isolated areas as small as 7 em diameter and 1-2 em deep (Fig. 6). Figure 4: Oviposition substrates of Orthetrum coerulescens in fen land-ditches-(a) almost stagnant water with pad of Chara sp. growing up to near the surface, covered by precipitated Fe-oxides; (b) water surface covered by semitranslucent layer of Mn-oxides; (c) ditch site dry ing out with emergent Chara-pad; (d) Chara-pad completely dried out. Situations (a-c), are attractive for territorial males. For oviposition only (a, b) are suitable. (d) neither attracts males nor females. International Journal of Odonatology 11 (21 2008: 261-276 269 ---------------Wildermuth---~--~---------- Habitat use by adults Territorial males typically stayed near or above the water, using vertical to horizon tal plant stems or leaves as perches that allowed the oviposition site to be surveyed. Occasionally they made short feeding flights, patrolled the territory or chased odonate intruders. Copulation was initiated in mid-air and ended on a perch near to the water. After copulation the separated partners remained perched close to each other. On experimental stretches with cleared vegetation both sexes perched on re mainders of the vegetation close to the ground. Oviposition was normally guarded and took place within the male territory. Tenerals, immatures and matures were reg ularly encountered in the adjacent litter meadows. Microhabitat of larvae Freely moving larvae were not observed except immediately prior to emergence. Ran dom samples of larvae from the mud at sections 2 and 4 with a small hand net (05 vi 2006 and 16 x 2007 respectively) provided 66 0. coerulescens, six Aeshna cyanea (Muller), five Pyrrhosoma nymphula (Sulzer), one Coenagrion puella (Linnaeus) and one Cordulegaster boltonii (Donovan). Hence, 0. coerulescens clearly dominated this Odonata assemblage. Exuviae were found of all these species as well as a few ex uviae of Somatochlora flavomaculata and Libellula quadrimaculata Linnaeus. Most exuviae were covered with a thin layer of rust-colour originating from iron oxides in the mud. The teleost Phoxinus phoxinus and the amphibians Triturus alpestris and T. helveticus, which are potential vertebrate predators, were present in most ditch sections but in low densities. Figure 5: Territorial male Orthetrum coerules Figure 6: Female Orthetrum coerulescens cens perched on a cut reed stem over shallow ovipositing in shallow water at the edge of water in a small ditch. a ditch. 270 International Journal of Odonatology 11 (2) 2008: 261-276