Succession versus grazing

Inland sand ecosystems are among the most threatened habitats in Central Europe. In recent decades, a decrease of site-specific plant species accompanied by a ruderalisation tendency was observed. This has often been explained by the influences of an enhanced nutrient supply and more widely distributed sites of spontaneous succession due to the abandonment of extensive land-use. The aim of this thesis was to develop a successional model of inland sand ecosystems as a basis for elucidating effects of ecosystem functioning, such as the phytodiversity-productivity relationship, and to simulate effects of an increased nutrient supply. As these habitats depend on a grazing impact, the effects of three grazing regimes (sheep grazing, donkey grazing and sheep followed by donkey grazing) were investigated. The main study site was the nature protection area “Ehemaliger August-Euler-Flugplatz von Darmstadt” located in the northern Upper Rhine Valley of Germany. This area served as a model for primarily calcareous inland sand ecosystems representing vegetation complexes of the Koelerion glaucae and the Armerion elongatae. Since 1999 the area has been grazed by sheep. Beginning in 2002 parts of the area are donkey-grazed and in 2003, at distinct areas, successive sheep and donkey grazing was established. For an analysis of the successional pathways, sand ecosystems with spontaneous succession in the vicinity of Darmstadt were investigated. For the creation of a successional model of sand ecosystems, two tall grass species were chosen as representatives of mid-successional stages. Stipa capillata served as a representative of non-ruderalised mid-successional stages with high conservation value and high diversity of site-specific plant species. In contrast, Calamagrostis epigejos was chosen as a representative of low-diversity sites with ruderal grass-encroachment. Vegetation relevés of 23 permanent plots studied for 5-7 years were analysed in combination with the vegetation relevés of 71 plots where in addition soil analyses were carried out. Mineral and total nitrogen content as well as phosphate and potassium contents and pH values of the soil were determined. Additionally, moisture values calculated on the basis of Ellenberg indicator values of the plant species were analysed for each plot. It was shown that two successional pathways leading from Koelerion glaucae plots to mid-successional stages exist. In the case of dry conditions (optimal moisture value 3.2) and with low phosphate contents (<1 15 mg * kg-1) the non-ruderalisation pathway leading to Allio-Stipetum stands is most likely to occur. The cover of Calamagrostis epigejos instead increases with increasing soil phosphate contents. This competitive grass species is not able to establish dominance stands if one of the following limits is reached: moisture value < 3.0, upper soil (0-10 cm soil depth): Ntotal < 0.04 %, K < 20 mg * kg-1, lower soil (10-30 cm): Ntotal < 0.01 %, Nmin < 0.5 mg * kg-1, K < 10 mg * kg-1. The conclusion can be drawn that site-specific plant communities are restricted to dry and nutrient-poor soils, and grass-encroachment is likely to occur in the case of nutrient-enriched plots. As an experimental approach, the effect of nutrient addition on early successional stages of sand ecosystems was studied. A 5-times replicated randomised block design was carried out to examine the effects of organic carbon (C), phosphorus (P), nitrogen (low [n] and high dosages [N]), NP, NPK and NPK+ micro and macronutrients (NPKM) in relation to a control on 10 m² plots. Data on species composition, vegetation layer, above- and below-ground phytomass, vegetation height, seed production and soil nutrient content were analysed. The results show a two-fold increase of phanerogam phytomass production after N, NP, NPK and NPKM addition, indicating the existence of nutrient limitation. Species composition and phytodiversity were not significantly affected but a significant increase of the height of 6 species, a tendency for increase of 11 species, a decrease of 1 species and a significant increase of seed production for 3 species was shown. Therefore, nearly all species benefit from a nutrient intake and after 4 treatment years, competition is not yet strong enough for species to outcompete one another. With regard to a grazing impact, it can be concluded that accumulation of faeces or urine can lead to an increase of phytomass and ruderalisation, whereas a homogeneous distribution of dung pellets does not influence the vegetation. For analysing the relationship between productivity, plant species richness and livestock diet in case of a sheep-grazed sand ecosystem two different approaches at two scales were used. At fine-grain scale (2 m²), productivity was measured by means of weighed aboveground phytomass, and data on livestock diet (grazed phytomass, crude protein content) were obtained. As a second means of analysing productivity, colour-infrared (CIR) aerial orthophotos were chosen and examined at fine-grain and at broad scale (79 m²). For measuring phytodiversity, at both scales, total numbers of vascular plant species and numbers of endangered vascular plant species were extracted from current vegetation relevés. A vegetational gradient from pioneer Koelerion glaucae stages and Koelerion glaucae grassland stages to consolidated Armerio-Festucetum trachyphyllae stands was investigated. Mixed linear model analyses reveal an influence of the year on all variables and as shown by correlation and regression analyses, relationships between variables are not significant in every year. At fine-grain scale, species richness and number of endangered species are negatively related to productivity while crude protein content and grazed phytomass are positively related to productivity. At broad scale the diversity-productivity relationship is unimodal, with highest species numbers in Koelerion glaucae grassland stages. The numbers of endangered species are higher in all pioneer stages than in the consolidated Armerio-Festucetum trachyphyllae stages. It can be concluded that scale and year have a strong impact on the diversity-productivity relationship. A well-balanced management method has to be developed to combine the aims of nature conservation value (high diversity) and livestock nutrition. In 2003 and 2004, the phytomass extraction and the diet preferences in the cases of sheep grazing, donkey grazing and sheep followed by donkey grazing were investigated with two closely related vegetation types (Armerio-Festucetum trachyphyllae and Festuca trachyphylla/ duvalii-community). In both vegetation types, each grazing approach was threefold replicated and paddocks of around 700 m² were grazed once a year for a short period. By means of 5 (respectively 4) 2-m² mini-exclosures per paddock, phytomass was sampled on the ungrazed plots and adjacent grazed plots and each sample was differentiated into the three plant functional types graminoids, Fabaceae and other herbs. It turned out that total phytomass extraction is highest with the two-species approach, while the quantitative phytomass extraction by sheep and donkeys is not significantly different. Concerning preferences for particular plant functional types, sheep extract more percentual phytomass of herbs than of graminoids, whereas donkeys select for Fabaceae. For successive two-species grazing no difference between the grazing extraction of all plant functional types is observed. As a consequence, successive two-species grazing is an ideal method for both reducing grass-encroachment and achieving maximum phytomass extraction, and thus to restore and to preserve high phytodiversity. Within the same threefold-replicated sampling design at the 700 m² paddocks, the effects of the three grazing regimes (sheep grazing [s], donkey grazing [d], successive two-species grazing [s+d]) on the vegetation were investigated. During 2002 and 2005 the vegetation was recorded by 1629 relevés of 25 m² plots. Additionally, data on species composition, nitrogen input and phytomass were obtained at donkey-generated microsites (faeces- and urine accumulation sites, wallows, scratch sites). Mixed linear model analyses revealed a strong influence of “year” and “vegetation type” on vegetation parameters. Species number, especially therophytes, and gap creation were significantly higher for donkey grazing than for sheep grazing. The vegetation layers were influenced by all treatments but cover of bryophytes and graminoids was influenced to a stronger extent for donkey grazing than for sheep grazing. Successive two-species grazing was not significantly different from the other treatments. Concerning donkeys’ faeces accumulation sites, a nitrogen input of 223 kg N * ha-1 at 4 % of the paddock could be observed. Although the phytomass at faeces sites was lower after 15 weeks and identical after 45 weeks in relation to controls, the species composition suggests a ruderalisation tendency. In conclusion, sheep grazing is adequate as a conservation method for rather intact sand ecosystems while donkey grazing is ideal for an intensive gap creation in case of consolidated stages. Due to a partial ruderalisation of the paddock, donkey grazing should not be repeated. Successive two-species grazing is also an adequate management method, but is probably of greater importance in ruderalised areas. This thesis first explained the occurrence of a ruderalisation and a non-ruderalisation pathway in inland sand ecosystems. While for early successional stages mainly nitrogen was proved to be the limiting factor for phytomass production, species composition was not significantly affected by an enhanced nutrient supply. In mid-successional stages low phosphate availability and dry conditions are decisive for the non-ruderalisation pathway to be taken. Productivity was found to be unimodally related to phytodiversity, with highest phytodiversity at Koelerion glaucae grassland stages. To preserve the open structure of these rather intact habitats, sheep grazing turned out to be an ideal management method. Donkey grazing is a very effective method for gap creation at more consolidated stages but due to the occurrence of ungrazed faeces accumulation sites, parts of the paddock can become ruderalised after repeated donkey grazing. The ruderalisation pathway of succession will develop in case of an enhanced nutrient availability. Successive two-species grazing is inferred to be the best management method for restoring these areas and directing the ruderal successional pathway into the non-ruderal direction