Plant species diversity in fringe communities: Investigations from Wendland / Lower Saxony Maike Eisenberg, Jürgen Dengler* and Carsten Hobohm Institute for Ecology and Environmental Chemistry, University of Lüneburg, Scharnhorststraße 1, D-21335 Lüneburg E-mail of corresponding author (*): dengler@uni-lueneburg.de Introduction Transition zones like skirts of woods run as linear structures through the cultural landscape of Central Europe. They are said to contain a high diversity of plant species because they are influenced by both the adjacent forest and the open landscape next to them. A study dealing with syntaxonomy, site conditions and phytodiversity of fringe vegetation was carried out in NE Lower Saxony (Eisenberg 2003). Based on the collected data we tried to elucidate several questions concerning the phytodiversity of fringe vegetation: • Is the species density in the fringe vegetation higher than in adjacent vegetation types? • Where do the species in the fringe vegetation come from? • Which parameters affect the species density of fringe vegetation? • How much do fringes contribute to the regional phytodiversity? Methods In summer 2002 the vegetation of forest edges was investigated on the sheet no. 2832 (Dannenberg [Elbe] Nord) of the German topographic map series 1 : 25.000 (Wendland, NE Lower Saxony). We tried to represent all syntaxa of fringe (‘Saum’) vegetation of anhydromorphic sites in our data that occur in the investigation area. 2 different sampling methods were applied, always taking into consideration vascular plants, bryophytes and lichens: Firstly, we compared the richness of plant species with that of neighbouring vegetation types. For that 21 data sets were collected. Each of these transects contains a sequence of 5 parallel plots. The plots are of the same size and shape (1 m × 5 m). The transects were orientated in parallel extension to the border of the forest. One plot was positioned in the fringe vegetation itself. To the right and the left the other two plots were arranged in a distance of 5 m apart (Figure 1A). For each plot the total number of plant species was recorded. Furthermore, we quantified the number of joint species between each pair of the plots within the transects. Secondly, the different vegetation types were documented in a phytosociological survey by 196 vegetation records. All relevés had the same plot size of 5 m 2 . Several additional parameters caracterising the vegetation structure and the site conditions were collected. We used this data for a correlation analysis after Spearman. Results (1) Species density and species compo- sition along the transects The species densities show a considerable and nearly symmetrical increase from both the open landscape (acres, fallow fields, meadows and pastures) and the forest vegetation towards the forest fringe (Figure 1B). They reach a local maximum in the fringe vegetation itself with a mean of 15.9 species per 5 m². When the number of species identities between pairs of plots within the individual transects is calculated and summed up for all 21 transects, interesting results can be found (Figure 1C). In terms of floristic composition the two forest plots W10 and W5 show the closest resemblance between each other. The next greatest similarity exists between the plots W5 and ‘Saum’ (fringe vegetation). Even W10 shows more species identities with the ‘Saum’ plot than the much closer plot F5 in the open landscape. The resemblance between wood (W5/10) and open landscape (F5/10) is very low. For instance, only 15 species identities occur between W10 and F10, i. e. they share less than 1 species in average. Figure 1D shows this aspect in another form: It can be seen that from the 15.9 species of the fringe plot 5.5 and 4.0 occur in the forest plots W5 and W10 as well. The respective numbers for the open landscape plots F5 and F10 are only about half as high. If the 2 wood and the 2 open landscape plots are united each they contain 40 % and 27 % of the species occuring in the fringe plot respectively. When all plots other than the fringe are joined they only contain 57 % of the respective fringe species, i. e. 43 % of these are unique to the fringes on average. Tab. 1: Spearman rank correlation of environmental and structural parameters with the species density of fringe communities. Figure 1: Transects through wood fringes (n = 21): A - Schematic overview of the sampling design. B - Number of plant species found in 5 m² plots of different position (mean ± SD). Significant differences to the fringe plots are indicated by asterisks. C - Number of joint occurences of species within pairs of plots (sum for 21 transects). The arrows are arranged according to decreasing numbers. D - Number of plant species shared by different plots with the fringe plot (mean ± SD). Discussion/Conclusion Our survey partly confirmed our hypotheses, partly revealed unexpected results. The most important aspects will be mentioned in the following and possible interpretations given: • The species density in the fringes is considerably higher than in the adjacent vegetation types. This is due to the fact that both species from the forest and such from the open landscape join here with a remarkable number of exclusive ‘Saum’ species. • In the area of investigation the floristic relationship of the fringe communities in general is much higher to the adjacent forest than to the herbaceous vegetation next to it. This contrasts to the results from Dierschke (1974), who in fringe communities in S Lower Saxony / N Hesse found higher affinities towards the communities of the open landscape. This differences may be due to the fact that in Dierschke’s research area most often extensively used semi-dry grasslands were bordering to the investigated forest fringes, whereas in NE Lower Saxony the landscape is much more intensively used by agriculture. The fact that ‘inner’ fringes contain more species than outer ones here, points in the same direction. • Forest and hedge fringes contain an enormous part of the regional floristic stock. Though our investigation was restricted to communities from anhydromorphic sites and our 196 relevés altogether only covered 0.008 % of the total area of the map sheet, they contained 31 % of its known vascular plant flora. Zacharias (1990) and Oppermann (1998) found similar results. This emphasises the great importance of well developed fringe communities for species conservation. References Dierschke, H. (1974): Saumgesellschaften im Vegetations- und Standortgefälle an Waldrändern. – Scr. Geobot. 6: 246 p., 5 tab., Goltze, Göttingen. Eisenberg, M. (2003): Saumgesellschaften NO-Niedersachsens – Soziologie und Pflanzenartenvielfalt. – 112 + X p., 5 tab., Diploma thesis, Institut für Ökologie und Umweltchemie, Univ. Lüneburg. Ellenberg, H., Weber, H. E., Düll, R., Wirth, V., Werner, W., Paulißen, D. (1991): Zeigerwerte von Pflanzen in Mitteleuropa. – Scr. Geobot. 18: 248 p., Goltze, Göttingen. Haeupler, H., Schönfelder, P. (1988): Atlas der Farn- und Blütenpflanzen der Bundesrepublik Deutschland. – 768 p., Ulmer, Stuttgart. Oppermann, F. W. (1998): Die Bedeutung von linearen Strukturen und Landschaftskorridoren für Flora und Vegetation der Agrarlandschaft. – Diss. Bot. 289: 214 p., Cramer, Berlin. Zacharias, D. (1990): Flora und Vegetation von Waldrändern in Abhängigkeit von der angrendzenden Nutzung – unter Berücksichtigung auch der floristisch schwer charakterisierbaren Bestände. – Riewenherm, S., Lieth, H. [eds.]: Ökologie und Naturschutz im Agrarraum. – Verh. Ges. Ökol. 19(2): 336–345, Osnabrück. (2) Correlation of species density with environ- mental parameters The correlation analysis between species densities in different types of fringe communities with environmental and structural parameters revealed several significant relationships (Tab. 1): There’s a highly significant positive correlation between species densities and the pH of the Ah horizont of the soil, i. e. the lower the acidity of the soil in the root space, the higher the phytodiversity. On the other hand species density is negatively correlated with the coverage of litter. The structural parameters show opposite influences: Whereas the species density is negatively correlated both with the cover of the phanerophyte layers and that of the herb layer, there’s a strong positive relationship with the cover of the moss layer. Neither the broadth of the ‘Saum’ vegetation, nor the position of the relevé within the fringe have any significant influence on the number of species. The potential duration of daily sunshine hasn’t either. When calculating the mean indicator values after Ellenberg & al. (1991) per plot the most of them showed no significant correlation with the species densities at all. Only the mean temperature value was negatively correlated with the number of species. When comparing ‘Saum’ communities at outer forest edges with those within forests, i. e. along forest paths, these 2 groups show a highly significant difference. The ‘inner’ fringes with 19.4 species per 5 m 2 on average contained 3.5 taxa more than the ‘outer’ ones with 15.9 species (n tot = 161). (3) Contribution of fringe communities to the regional phytodiversity The 196 plots of the phytosociological survey altogether contained 322 different plant taxa (275 vascular plants, 39 bryophytes, 8 lichens). The number of vascular plant species equals 31 % of the total species stock of the investigation area (cf. Haeupler & Schönfelder 1988: 43). Figure 2: ‘Outer‘ fringes of forests and hedges run as linear structures through the cultural landscape of Central Europe. The photograph shows a mesophytic fringe community with Agrimonia eupatoria in the investigation area. Figure 3: ‘Inner’ fringes along forest paths considerably contribute to the total species diversity of forests. The photograph shows a nitrophytic fringe community with Arctium nemorosum in the investigation area. Figure 4: The Lathyro linifolii-Melampyretum pratensis Passarge 1967 - here with Lathyrus linifolius and Hieracium laevigatum - is an acidophytic fringe community. On average it contained 19.7 species per 5 m 2 in the investigation area. pH of the Ah soil horizont potential daily sunshine duration (h) distance to the trunks of the first trees (m) distance to the open landscape (m) broadth of the fringe (m) total cover phanerphytes (%) cover herb layer (%) cover moss layer (%) cover of litter (%) number of plant species Correlation Coefficient .319** -.037 -.040 -.064 -.040 -.160* -.173* .474** -.291** Significance (2-tailed) .000 .613 .951 .377 .519 .012 .016 .000 .000 n 187 193 190 192 194 196 195 196 196 *Correlation is significant at the 0.05 level (2-tailed). **Correlation is significant at the 0.01 level (2-tailed). 5 m 1mx5m 5 m 1mx5m 5 m 1mx5m 5 m 1mx5m 1mx5m 0 5 10 15 20 25 plot 1 plot 2 plot 3 plot 4 plot 5 number of plant species 2.2 2.4 - 2.0 2.9 SD 4.0 5.5 15.9 2.7 2.4 n 9.3*** 12.8 15.9 12.8* 9.6*** C 106 58 27 25 15 19 79 130 118 50 D B A