Biodiversity Differences between Young and Historically Old Steppe Grasslands in Thuringia

Table of Contents

Abstract

1. Introduction
1.1. Background

2. Material and Methods
2.1. The Study Area
2.2. Area Analysis and Vegetation Sampling
2.3. Laboratory Soil Analysis
2.4. Species Biological Traits, Indicator Values and other Parameters
2.5. Statistical Analysis

3. Results
3.1. Diversity Measures
3.2. Indicator Species

4. Ordinations
4.1.1. DCA
4.1.2. NMDS

5. Discussion
5.1. Field Work
5.2. Ordinations
5.2.1. Environmental Variables & Life Strategy Types
5.2.2. Species Dispersal & Seed Longevity
5.3. Diversity and Conservational Values
5.4. Management
5.5. Conclusion

6. Appendix

7. Acknowledgements

8. References

Abstract

Generally there are only few studies that focus on the interrelation of species composition, habitat properties and indicator species of young and old grasslands (Karlík & Poschlod, 2009), and the few existing oftentimes have contradictory findings (e.g. Karlík & Poschlod, 2009, Ruprecht et al. 2009, Waesch & Becker 2009). This is surprising, as decision-makers are oftentimes forced to concentrate efforts on areas of highest conservation value due to financial restrictions. In this study, we assumed that habitat age and land-use history have verifiable effects on the present vegetation community. We analyzed a total of 14 study sites at the southern slopes of the Kyffhäuser Mountains, at the northern border of the federal state of Thuringia, Germany. The area is renowned for its richness in fauna and especially flora, constituting the western-most outpost of Eurasian steppic grasslands. The study area was analyzed with the help of historical maps, post-war aerial survey photographs and present-state orthophotographs. For each study site 12 relevés, 6 on historically old and 6 on relatively young grasslands on former arable land were conducted. Using DCA and NMDS ordinations, a total of 88 environmental variables ranging from Ellenberg Indicator Values, orographic and edaphic conditions, soil contents and properties, to species inherent traits, were tested for their power to explain the observed occurrence and abundance patterns. Results clearly show that land use history and habitat age has a significant effect on species composition. Concordantly, an Indicator species analysis (ISA) identified 28 indicator species for old and 21 for young calcareous grasslands. Most significant differences between the two groups were the Ellenberg Indicator Values for moisture and nutrients for the occurring species, which were much higher for young sample sites. Accordingly, there was a higher proportion of ruderalists on grasslands of younger age, while old patches were dominated by stress-strategists. In total, strong relationships could be found for 17 of the tested variables (Pearson´s r ≥ 0.5, p ≤ 0.01). While α-diversity was only slightly higher for old sites, occurrence of threatened species was more than doubled (young sites: 1.98, old: 4.71 mean species/sample site p < 0.001). Thus, we conclude that the identification of historic calcareous grassland patches is desirable and, where absolutely necessary, old sites should take precedence in management efforts over sites of younger origin. However, considering recent findings in literature, ideally a mosaic of different successional stages should be aspired for, as it can harbor the widest range of plant species and associated invertebrates.

1. Introduction

1.1. Background

The majority of the world´s highly diverse ecosystems can be found in areas that, for a variety of reasons, were essentially left untouched by humans. The biggest exemption to this rule of thumb is Europe, where a long cultural history gave birth to landscapes that were greatly influenced by humans, yet developed a pattern of diverse ecosystems of high qualities (Naveh, 1998; Vos & Meekes, 1999). Calcareous grasslands are among these semi-natural habitats (e.g. WallisDeVries, Poschlod & Willems 2002). Previous to human influences this habitat type is thought to have been restricted to small isolated patches such as `hilly domes or outcrops in the Jurassic mountains´ and to remnants of steppic vegetation (Poschlod & WallisDeVries, 2002). The role of herbivores remains disputed (WallisDeVries, Poschlod, & Willems, 2002), but most authors agree that, to a lesser extent, grazing wild animals contributed to grassland species survival by keeping smaller patches clear (Karlík & Poschlod, 2009; Pärtel, Bruun & Sammul, 2005; Poschlod & WallisDeVries, 2002; Willems, 2001). These microhabitats than functioned as species source pools when, with the Neolithic age, husbandry and manmade deforestation thinned out the primeval and dominant forest cover, and gave way to the first expansions of secondary grasslands (Bredenkamp, Spada, & Kazmierczak, 2002; WallisDeVries, Poschlod, & Willems, 2002). Pykälä (2000) hypothesized, that ancient cultural practices, for instance livestock grazing and mowing, compensates for human-suppressed natural processes such as fires and floods, and thus enabled the continued existence of many species. With growing population numbers, agricultural progress and upheavals, calcareous grasslands expanded further, with highest growth rates during the times of the Roman Empire and the Medieval Ages (Poschlod & WallisDeVries, 2002; Van Swaay, 2002 as cited in Pott, 1998). The peak point was reached between 17th and 19th century, when large flocks of domestic livestock, mostly sheep, where driven hundreds of kilometers between summer and winter pastures in central Europe, creating vast connected areas of calcareous grasslands (Poschlod & WallisDeVries, 2002). From the beginning of the early 19th century grasslands and grassland connectivity started to deplete (Pärtel, Bruun, & Sammul, 2005; Poschlod & WallisDeVries, 2002), a trend that accelerated tremendously after the Second World War (Fischer & Stöcklin, 1997). The reasons for this development vary locally, but have some common denominators. Long distance livestock drives were simply hindered by urbanization and growing infrastructure networks (Pärtel, Bruun, & Sammul, 2005; Thompson & Jones, 1999). Furthermore traditional shepherding became increasingly uneconomical with increasing meat and wool imports from Australia and New Zealand (Poschlod & WallisDeVries, 2002). Most importantly, manpower became a determining factor in commodity-chains, and rising opportunity costs of labor-intense work (Strijker, 2005) lead to a rapid mechanization and intensification of agricultural processes (e.g. Cremene, et al. 2005; Fischer & Stöcklin, 1997; Vos & Meekes, 1999). This trend continued over the last few decades (Kahmen, Poschlod, & Schreiber, 2002). Today calcareous grasslands are rare habitats that are severely endangered (Balmer & Erhardt, 2000), mostly by abandonment, subsequent bush encroachment and afforestation (WallisDeVries, Poschlod, & Willems, 2002). No figure describing the overall loss in size can be found in literature, but regional estimates range between 60 and 90 % reduction compared to former maximum known expansions (Adriaens, Honnay, & Hermy, 2006; Helm, Hanski, & Pärtel, 2006). With their shrinking extend came the perception of the immense conservation value at stake. Dry, nutrient-poor and especially calcareous grasslands are renowned to be among the habitats richest in species in Europe's ancient cultural landscape (Cremene, et al., 2005; Karlík & Poschlod, 2009). They harbor a high diversity of vascular plant species, many of which are specialists that highly depend on the continued existence of these microhabitats (Steffan-Dewenter & Tscharntke, 2002; WallisDeVries, Poschlod, & Willems, 2002). While this habitat type is also accepted as an important habitat for many other invertebrates (Balmer & Erhardt, 2000; Rico, Boehmer, & Wagner, in press; WallisDeVries, Poschlod, & Willems, 2002) such as grasshoppers and carabidae (Mayr, Wolters, & Dauber, 2007), a positive correlation in species richness could only be validated between butterflies and vascular plants (Cremene, et al., 2005). Van Swaay (2002) found that for butterflies, `calcareous grasslands rank as the most species-rich habitat in Europe with 274 species reported.´, of which a high proportion is threatened. A study of butterfly communities in calcareous grassland remnants in south-western Germany, showed a strong decline of more than 50% in ubiquitous species between 1972 and 2001 (Wenzel et al., 2006), a trend reflecting the worrisome state of many grassland remnants. Remaining stands constitute invaluable refuges for many threatened open-land species, but are often of such infinitesimal size and heavily fragmented, that they lose much of their value for associated species. Fragmentation can interfere with plant-pollinator and predator-prey interactions (Goverde et al., 2002; Steffan-Dewenter & Tscharntke, 2002), and reduces the overall likelihood of genetical transfers via diaspore dispersal. Grazing by domestic livestock is the key factor in the dispersal of grassland species (Fischer & Poschlod, 1996; Karlík & Poschlod, 2009) as many adapted over the centuries of traditional agricultural practices, and developed various types of zoochory as a dispersal mechanism. Increased fragmentation plus the omission of sheep flocks as `moving corridors´ (WallisDeVries, Poschlod, & Willems, 2002), reduced the connectivity of calcareous grassland essentially, and degraded subpopulations to isolated stands. This ultimately resulted in inbreeding of populations and finally in extinction events (Fischer & Stöcklin, 1997; WallisDeVries, Poschlod, & Willems, 2002). First studies also suggest that the changed environmental conditions are already reflected in community compositions (Kahmen, Poschlod, & Schreiber, 2002; Maurer, Durka, & Stöcklin, 2003). Maurer, Durka & Stöcklin (2003), found that `…traits enhancing persistence are more important for the frequency of occurrence of a species in calcareous grassland than traits affecting dispersal´. They conclude that this implies a reduced likelihood of recolonization events of locally extinct species by far-distance dispersal to a near zero. Observed species composition in many meta-populations might not only be traced back to seed dispersal and recruitment success between sites. Seed persistence is believed to decelerate extinction processes to an extent that current remnants of calcareous grasslands might actually draw a good picture of original plant species compositions (Helm, Hanski, & Pärtel, 2006). Thus, in addition to already recorded losses in biodiversity (Fischer & Poschlod 1996; Poschlod & WallisDeVries 2002; Stöcklin & Fischer, 1999), it is feared that for some areas, the consequences of habitat loss, fragmentation and reduced connectivity will eventually become apparent when the slow response to environmental change sets in, and `extinction debts´ will further reduce species richnesses (Eriksson, Cousins & Bruun 2002; Helm, Hanski & Pärtel 2006; Lindborg & Eriksson 2004).

Because of their unique evolutionary history and main area of distribution, calcareous grasslands are a European cultural and natural heritage, with all implied responsibilities. Their importance for the conservation of biodiversity in Europe and the urgency of their protection has been acknowledged by the European Union by listing this particular habitat type in the Annex I of the Natura 2000 Habitat Directive (92/43/EEC) (Ssymank et al., 1998). As many of the sites listed for Europe, calcareous grasslands need a constant management regime to maintain a favorable conservation status (Hampicke & Roth, 2000; Ostermann, 1998; WallisDeVries, Poschlod, & Willems, 2002), as an abandonment of low-intensity agricultural practices would inevitably led to shrub encroachment (Barbaro, Dutoit, & Cozic, 2001; Rico, Boehmer, & Wagner, in press; WallisDeVries, Poschlod, & Willems, 2002).

The accurate form of managing calcareous grasslands however is as intricate as debated. Depending on the set objective, management schemes often focus either on a set of particular species or follow a more integrated approach on the process and system level, both of which have disadvantages (WallisDeVries, Poschlod, & Willems, 2002). Also it has been criticized, that both in scientific research and practical conservation, particular attention is paid to the diversity of vascular plants, while associated and partially dependent, fauna is often disregarded (WallisDeVries, Poschlod, & Willems, 2002; Woodcock et al., 2005). First publications shed light on the effects of different management types on invertebrates. For instance showed Woodcock et al. (2005), that different management forms (ungrazed, long-term cattle and long-term sheep grazing) did not alter beetle abundance or species richness, but respective areas varied significantly in community structures. Balmer & Erhardt (2000) reported, that in sheer contrast to vascular plants (Poschlod & WallisDeVries, 2002), butterfly diversity and the number of threatened species does increase for fallow calcareous grassland where grazing has been intermitted for 2-3 years. This example shows how different management plans can turn out with varying main objectives. Most calcareous grasslands today are under protection, and as traditional land-use practices are highly uneconomical, their continuity almost exclusively depends on environmental contracting (Barbaro, Dutoit, & Cozic, 2001). These however are very cost intensive. In the timespan from 1994 to 2003, the European Union has spend approximately € 24.3 billion on agri-environment schemes (AES) with biodiversity conservation aims (Kleijn & Sutherland, 2003). Thus, on a regional level, financial feasibility is often among the decisive criteria for further management objectives. This is mirrored by the relatively high number of studies, that focus on alternative remuneration of land users for their ecological services (Hampicke & Roth, 2000), and on alternatives to low-intensity grazing (e.g. for mowing and mulching; Kahmen, Poschlod & Schreiber, 2002; or ploughing; Poschlod & WallisDeVries, 2002, as cited in Kleyer, 2000). Summarizing, calls for new focal points in management regimes have emerged in recent literature.

1) Maintenance of a mosaic of different successional stages (Balmer & Erhardt, 2000; Poschlod & WallisDeVries, 2002) as well as variety of management practices (Köhler et al., 2005; Marton, Ruprecht & Deák, 2008) in order to provide for the full span width of habitat types.
2) A more balanced focus of floral and faunal biodiversity in research and management schemes (Steffan-Dewenter & Tscharntke, 2002; WallisDeVries, Poschlod & Willems, 2002).
3) Where grazing was an integral component of the formation history, a continuation of this practice as well as a thought through cross-linking with flocks in-between stands (Rico, Boehmer & Wagner, in press), is essential in order to maintain vectors for dispersal and genetical flux (Karlík & Poschlod 2009; Poschlod & WallisDeVries 2002; for general matrix restoration: Donald & Evans, 2006).

On a regional level, financial restrictions oftentimes force decision makers, to pinpoint core areas of highest conservational value, as former expansions of calcareous grasslands are simply to large in order to restore and/or maintain them. Besides biodiversity levels, occurrence of threatened species and habitat connectivity, habitat age can function as a further eligibility criterion. Ancient grasslands are of special cultural importance, and constitute valuable reference systems e.g. for restoration efforts (Alard et al., 2005; Piqueray et al. 2001). Also, once overgrown calcareous grasslands quickly forfeit their restoration potential and value, as seed banks are mainly transient (Bistea & Mahy, 2005) and ´many characteristic species cannot rely on this mechanism as a buffer against local extinction´ (Stöcklin & Fischer, 1999). In respect to Climate Change, old grasslands are also likely to be more resilient to varying environmental conditions than those of younger age (Grime et al., 2000), and might thus gain even more importance as a refuge for many endangered species.

The main objective of this study, was to investigate the interdependencies of habitat age and properties with species compositions and associated conservational values.

2. Material and Methods

2.1. The Study Area

Abbildung in dieser Leseprobe nicht enthaltenFig. 1: The study area at the northern border of Thuringia and the 14 study sites, each comprising 12 plots, 6 on old grassland and 6 on grassland of young origin (See text for exact disambiguation).

The setting for this study was the Mountains located at the northern border of the federal state of Thuringia in Central Germany. With a length of 19km and an approximate width of 7km, the Kyffhäuser is known as the smallest of Germany´s Central German Uplands. The region is a characterized by a distinctive geology that lead to a pattern of quickly alternating sediment types from the Upper Permian Zechstein Period (Barthel & Pusch, 1996) such as deposits of relatively pure gypsum, clay shale (so-called `Rote Letten´) and bituminous lime stone (so-called `Stinkschiefer´) at its southern edges.

The rich structuring of the landscape, consisting of vast plains intermitted by the Kyffhäuser Mountains and its foothills with plateaus, steep slopes and even cliffs, is accompanied by a broad spectrum of microclimates. Three mountain ranges, from north the Harz, from south the Thuringian Forest and from the south-west the Hainlaite range, cast a rain shadow over the region, resulting in a bland and dry climate with precipitation amounts that can fall below 500 mm per annum (Barthel & Pusch, 1999; Various Authors, 1976). The low precipitation, as well as high differences between summer and winter temperatures result in a continental climate. This climatic effect is amplified edaphically by the high solubility and hydraulic conductivity of the present poriferous rocks. Quick weathering leaves an oftentimes heavily fissured karst structures that function as a subsurface seepage and further reduces soil moisture contents, thus abetting xerophilous plant communities even more (Barthel & Pusch, 1999; Nickel, et al. 2001). The interplay of geological, climatic, orographic and edaphic conditions as well as centuries of agricultural use resulted in a diversity in flora and fauna that is distinguished by an outstanding number of threatened species and a habitat composition that is unmatched in the European landscape (Barthel & Pusch, 1996; Stolze 2011, pers. com.). More than 100 species found here are mentioned in the Thuringian red list, 6 of those are endemic for the federal state (Ludwig & Schnittler 1996; Stolze 2011, pers. com.). Due to the climatic idiosyncrasy, 47 of the occurring phanerogams reach their most north-westerly geographic distribution in the Kyffhäuser Mountains (Stolze 2011, pers. com.). Because of its conservation value, a nature park of 30.500 ha covers large parts of the region (BfN, 2011), encompassing all fourteen study sites. Furthermore all but two study sites (`Herzchen´ and `Kippenhügel´) also from part of the designated Natura2000 site `Kyffhäuser-Badraer Schweiz-Solwiesen´ (site nr.: 4632-302), and are incorporated into four different conservation areas (all IUCN category IV – habitat species management area) (BfN, 2011). Primordially the Kyffhäuser Mountains were covered with forest with few exceptions at steeps, shallow soils and patches of extreme aridity (Various Authors, 1976). First human appearances in the form of hunters and gatherers could be proven for the Paleolithic. The landscape started to change in the early Neolithic, when the Bandkeramik culture started to claim land for the cultivation of grains and livestock breeding (Various Authors, 1976). In medieval times much of the forest was cleared for the immense requirements by the saline in Bad Frankenhausen and the flourishing copper shale mining industry (Stolze 2011, pers. com.). In the following centuries much off the cleared land was made arable and used for a variety of agricultural purposes. Most of the slopes were used as pastures for large sheep flocks. Gradually food shortages forced locals to extend their acreage into meager areas, a development that peaked with the beginning of World War II. In this process, were ever declivity and soil conditions promised minimal yields, many of the old sheep pastures were converted into crop fields. With the amelioration of living conditions in post-war times, these inhospitable lands were abandoned. In the 1960´s and 70´s industrialization and a consequential rural exodus caused a strong decline of traditional land use activities. The large-scale xerotherm calcareous grasslands on the southern and south-western slopes stood on the verge of abandonment and became subject to agri-environmental schemes. Today in the area old calcareous grasslands that have endured over centuries can be found alongside with young, recently re-established areas that were gradually recolonized by characteristic species after 1945.

2.2. Area Analysis and Vegetation Sampling

Area analysis was done on the basis of historical aerial photographs that were taken for a post-war aerial survey by the Allies in 1945, and another aerial survey in 1955. In consultation with Dr. Jürgen Pusch from the `Kyffhäuser administrative office for environment, nature and water´, fourteen study sites with old and young grasslands were chosen along the southern slopes of the Kyffhäuser Mountains (Fig.1). Old patches of the study sites were defined as areas, where aerial photographs showed no traces of arable farming for 1945 or thereafter. Areas where forms of land use were not clearly distinguishable were avoided during vegetation samplings. Conclusive signs of recent or existent usage were for example ploughing marks that consigned light parallels were gypsums were hoisted to the surface. Other important orientation marks that could be partially traced both, in the photographic material as well as in the field were piles of field stones and field edges where soil accumulated to small mounds. Chosen study sites were reviewed on the basis of historical maps (Prussian general census 1853-1855) part of which charted land use forms including grasslands. Thus for the sites `Stöckei´, `Herzchen´, `Mittelberg´, `Solberg´ and `Badra Lehde-Sachsen Wäldchen´ the history of the old sampling plots could be ascertained to be over a 150 years. The old patches of the remaining 9 study sites were assumed to be of similar age, which could not substantiated with historical maps, but are at least 66 years old indicated by the post-war aerial pictures. For each site at least one, but up to four different aerial photographs (varying map scales, 9, 10, 22 or 40 thousand : 1) from 1945 and 1953, as well as modern digital orthophotos, topographical maps (TLVerGeo 2011) and the above described historical maps, were superimposed and interpreted using ArcGIS©10 (Fig.2). For each of the fourteen study sites, twelve samples were conducted, six on old and six on young calcareous grassland patches. After pre-coordination with the constructed maps, the exact location for these sample sites was decided upon in the field.

Abbildung in dieser Leseprobe nicht enthalten

Where possible, one old and one young sample site were treated as a sampling pair, in order to grant for similar abiotic conditions such as exposition and soil type. Other circumstances that could affect species composition and occurrence, such as proximity to possibly fertilized fields, shrubs or tracks were avoided by a buffer zone of at least 10 meters from the edge of the respective sample site. Sampling of the in total 184 sites, was done from June till July 2011. Each plot had a size of 3x3m that was delimited with measuring tape during sampling. Because the early summer month had been exceptionally dry and many species showed a delay in growth, each site was visited again within the first week of August, and vegetation cover and species occurrences were revised.

For each sampling site the following data was collected:

- Date
- GPS coordination & elevation (Garmin GPSmap 76CSX)
- Substrate
- Exposition & Inclination
- Soil depth (A 1cm diameter iron rod was trusted 5 times at the corners and middle of the plot into the soil, mean value)
- Species occurrence and % coverage of the sampling site
- % of litter, herb, moss, lichen, open soil and rock layer coverage
- Estimated height of herb layer
- Grazing scheme

2.3. Laboratory Soil Analysis

Soil was taken at all sample sites in the beginning of August 2011. For each site, a core cutter of 5 cm in diameter was driven 20 cm into the soil with a rubber mallet. The hole depth and length of the undisturbed core specimen was noted. This process was repeated 5 times at the four edged and of the middle of each plot, to grant for representative soil sample. Were the soil was very shallow, more cores of shorter length were taken to gather sufficient soil material for further analyses. Samples were cooled directly with entrained thermal packs, and further stored in refrigerators in the laboratory. The heat and oxygen sensitive dissolved nutrients NH4, NO3, and phosphorus were measured first, by continuous-flow UV-spectrophotometry. ICP mass spectrometry was conducted in order to determine a range of compounds (Na, Mn, Mg, K, Fe, Ca). Furthermore, for each of the 84 samples, soil moisture content, conductivity, pH-value (soil diluted in distilled water), and the cation-exchange capacity (CEC) as an approximation value for nutrient retention capacity, were measured.

2.4. Species Biological Traits, Indicator Values and other Parameters

For each of the 193 species, an array of biological traits and indicator values were collected from various sources (Tab.1). Intentionally parts of this information are overlapping or iterant, e.g. the Indicator Values classifying light, temperature and moisture for the respective species.

Abbildung in dieser Leseprobe nicht enthalten

Tab. 1: List of biological traits, indicator values and further parameters deployed for further statistical analysis, arranged according to their sources.

For numerical variables the arithmetic means, and for categorical values the represented proportions were built for each of the 84 sample plots. A total of 87 variables were thus incorporated in all direct ordinations and post ordination correlation-searches for explaining factors and patches. The Heat Load, a formula accounting for the inclination, exposition and geographical latitude was calculated after the Formula from (McCune & Keon, 2002), with a predefinition of south-south-west as the warmest aspect following Becker, Andres & Dierschke (2011).

2.5. Statistical Analysis

An Indicator Species Analysis was conducted with PC-ORD 5.10 (McCune & Grace, 2002), following the method of Dufrene & Legendre (1997) for the calculation of Indicator Values (IV). For ordinations, the percentage coverage data was standardized by logarithmic transformation after a small number had been added to all data points, in order to bypass the problem of log(0) being undefined (McCune & Grace, 2002). Initially a DCA (Detrended Correspondence Analysis; Hill & Gauch, 1980) was conducted to access the length of gradients (LG) (Leyer & Wesche, 2008), as a measure for the length of underlying ecological gradients. Species occurring in less than three sample sites were deleted prior to calculations (n=136), as they do not allow for realistic estimations of their environmental requirements and optima (Leyer & Wesche, 2008), and are furthermore known to have distorting effects on LG calculations in the DCA (Eilertsen, et al. 1990) and final ordination results (McCune & Grace, 2002). In addition, remainig rare species were downrated. According to the results, DCA was employed for further data analysis, as this ordination technique assumes a unimodal distribution, in which the abundance of species describes an optimum somewhere along the environmental gradient. The rescaling algorithm inherited to the DCA, sections the LG into ecologically interpretable SD units (average standard deviation of species turnover) (Eilertsen et al., 1990; Leyer & Wesche, 2008), which were used to get an estimate of the beta diversity of the two sample site types, old and young. Not only because of the interpretable SD units, is the DCA among the most frequently used multivariate ordination techniques in ecology (Leyer & Wesche, 2008). However it is also among the most criticized, as method inherited mathematical corrections `…impose assumptions about the distribution of sample units and species in environmental space´ (McCune & Grace, 2002), that can ultimately corrupt results. Therefore calculations were recapitulated with presence-absence data, which can yield in more robust results (McCune & Grace, 2002). As the reliability of DCA results are heavily disputed (Kent, 2011), and duplicate tests are recommended (Leyer & Wesche, 2008; McCune & Grace, 2002), a NMDS (Non-metric Multi Dimensional Scaling; Kruskal 1964; Shepard 1962) was also performed, to compare and verify obtained results. NMDS followes an entirely different mathematical approach, in which a dissimilarities among samples are ranked and used for the analysis instead of absolute distances (Jannicke, Ptacnik, & Ptacnik, 2007). McCune & Grace (2002), regard this increasingly used method as the most progressive one, with less flaws and better detection of underlying structures in most community data sets. The same standardized data set as for the DCAs was used. Following the recommended procedure (McCune & Grace, 2002), a preliminary NMDS ordination for a 6 up to to a 1 dimensional solution was run. A Monte Carlo test, assessing the probability that stress values have been derived by chance, showed a significant distinction (p = 0.0096) of randomized and real data results (# runs with real data 50, # runs randomized data 50, stability criterion 0.00001, max. # of iterations 300). Depicting the results of the Monte Carlo test, a Scree Plot (Fig. 9) was utilized to decide on the ideal number of dimensions for the final solution. Stability values for all dimensional calculations were satisfactory (=0.00001). As for the DCA, for all NMDS ordinations the Bray-Curtis similarity algorithm was applied. For the actual NMDS run a three dimensional solution was computed, reapplying the best starting configuration of the preliminary NMDS (# of runs with real data = 1, no randomization or step-down in dimensionality, stability criterion 0.00001, max. # of iterations 300). Results were rotated with the simultaneous Varimax rotation, to maximize the loadings of the variables for each of the axes (McCune & Mefford, 2006, adapted from Mather 1976). DCA and NMDS results were interpreted with the help of overlaid vector biplots, representing the individual environmental variables. Ordination axes were correlated with all environmental variables implementing Pearson´s r². All above mentioned ordinations were performed with the multivariate statistics program PC-ORD 5.0 (McCune & Mefford, 2006). Graphs depicting ordination results were remodeled with SigmaPlot 12.0.

3. Results

3.1. Diversity Measures

Abbildung in dieser Leseprobe nicht enthaltenAbbildung in dieser Leseprobe nicht enthalten

Fig. 3: Species Area Curves – SACs (heavy line) displayed separately for `old´ and `young´ sample sites (42 each). Dotted lines represent ± 1 standard deviation. The distance curve (light line, right axis) describes the average Sørensen distance between the single samples and the total of 42 samples, calculations where done with PC-ORD 5.10 (McCune & Grace, 2002)

Species Area Curves were conducted separately for the two groups of old and young sample sites, using PC-ORD 5.10 (McCune & Grace, 2002). Both curves (Fig.3) clearly show that the chosen number of samples for each group (42) sufficed, as the number of recorded species levels off and an increased sampling effort would result in only small increases of the number of species encountered. Information about the species abundances is allowed for in the distance curve, which depicts the `average distance between the centroid of a subsample and the centroid of the whole sample´ (McCune & Grace, 2002). Remarkable here, is the great resemblance of both curves between old and young sample groups, which also becomes apparent in the low disparities in diversity measures (Tab.1). Species richness was only slightly but significantly higher for old sample plots (t (82) = 0.0132).

Tab. 2: Comparative result chart of old and young sample sites. (Ø here means average over the 42 respective samples)

Abbildung in dieser Leseprobe nicht enthalten

[...]