Long-term changes of El Niño/Southern Oscillation (ENSO) are studied with the ensemble run of Simple Ocean Data Assimilation (SODA 2.2.6) and the Coupled Model Intercomparison Project Phase 5 (CMIP5). An eight member ocean reanalyses (SODA 2.2.6) from 1871 to 2008 is produced by using forcing from eight ensemble members of an atmospheric reanalysis. The ensemble reanalysis shows that El Niño has prominent decadal variability. Weak El Niños occur throughout the entire record whereas the occurrence of strong El Niños varies, with strong El Niño at the beginning and end of the record. The strength of La Niña is weaker than for El Niño, and has less variability. Although for any given El Niño year all ensemble members show the occurrence of El Niño, in some ensemble members the El Niño is strong while in others it is weak. When the timing of the onset of Westerly Wind Bursts (WWBs) occurs earlier in the year and the strength of WWBs is stronger, strong El Niño occurs. To study the importance of the background state in the tropical Pacific Ocean on ENSO, long-term trends of tropical Pacific SST, wind stress, subsurface temperature and the sub-tropical cells (STCs) are analyzed. The reanalysis shows that there is a slight cooling trend of SST in the central tropical Pacific due to an enhanced tropical Pacific circulation. Subsurface temperature also has a cooling trend. The STCs, which consist of equatorial upwelling, Ekman transport, extra-tropical subduction and pycnocline transport from the sub-tropical to the tropical region, strengthen from 1900 to 2008. When the STCs are accelerated, equatorial upwelling increases bringing cold water from the subsurface that cools the surface. ENSO variability is also analyzed in the CMIP5 historical experiments. Results show that most of the models have a realistic representation of the strength of ENSO; however, the location of warming generally extends too far to the west. Overall, properties of ENSO do not show a significant change in most of the CMIP5 models. One distinguishing difference between the CMIP5 models and SODA 2.2.6 is that ENSO in SODA 2.2.6 has prominent asymmetry between El Niño and La Niña, whereas ENSO in the CMIP5 models tends to have fairly symmetric El Niño and La Niña. In contrast with the reanalysis most of the CMIP5 models have warming trends at the surface and the transport of the STCs has a decreasing trend.

Human-induced alterations of the environment are causing biotic changes worldwide, including the extinction of species and a mixing of once disparate floras and faunas. One type of biological communities that is expected to be particularly affected by environmental alterations are herb layer plant communities of fragmented forests such as those in the west European lowlands. However, our knowledge about current changes in species diversity and composition in these communities is limited due to a lack of adequate long-term studies. In this thesis, I resurveyed the herb layer communities of ancient forest patches in the Weser-Elbe region (NW Germany) after two decades using 175 semi-permanent plots. The general objectives were (i) to quantify changes in plant species diversity considering also between-community (β) and functional diversity, (ii) to determine shifts in species composition in terms of species’ niche breadth and functional traits and (iii) to find indications on the most likely environmental drivers for the observed changes. These objectives were pursued with four independent research papers (Chapters 1-4) whose results were brought together in a General Discussion. Alpha diversity (species richness) increased by almost four species on average, whereas β diversity tended to decrease (Chapter 1). The latter is interpreted as a beginning floristic homogenization. The observed changes were primarily the result of a spread of native habitat generalists that are able to tolerate broad pH and moisture ranges. The changes in α and β diversity were only significant when species abundances were neglected (Chapters 1 and 2), demonstrating that the diversity changes resulted mainly from gains and losses of low-abundance species. This study is one of the first studies in temperate Europe that demonstrates floristic homogenization of forest plant communities at a larger than local scale. The diversity changes found at the taxonomic level did not result in similar changes at the functional level (Chapter 2). The likely reason is that these communities are functionally “buffered”. Single communities involve most of the functional diversity of the regional pool, i.e., they are already functionally rich, while they are functionally redundant among each other, i.e., they are already homogeneous. Independent of taxonomic homogenization, the abundance of 30 species decreased significantly (Chapter 4). These species included 12 ancient forest species (i.e., species closely tied to forest patches with a habitat continuity > 200 years) and seven species listed on the Red List of endangered plant species in NW Germany. If these decreases continue over the next decades, local extinctions may result. This biotic impoverishment would seriously conflict with regional conservation goals. Community assembly mechanisms changed at the local level particularly at sites that experienced disturbance by forest management activities between the sampling periods (Chapter 3). Disturbance altered community assembly mechanisms in two ways: (i) it relaxed environmental filters and allowed the coexistence of different reproduction strategies, as reflected by a higher diversity of reproductive traits at the time of the resurvey, and (ii) it enhanced light availability and tightened competitive filters. These limited the functional diversity with respect to canopy height and selected for taller species. Thirty-one winner and 30 loser species, which had significantly increased or decreased in abundance, respectively, were characterized by various functional traits and ecological performances to find indications on the most likely environmental drivers for the observed floristic changes (Chapter 4). Winner species had higher seed longevity, flowered later in the season and had more often an oceanic distribution compared to loser species. Loser species tended to have a higher specific leaf area, to be more susceptible to deer browsing and to have a performance optimum at higher soil pH values compared to winner species. Multiple logistic regression analyses indicated that disturbances due to forest management interventions were the primary cause of the species shifts. As one of the first European resurvey studies, this study provides indications that an enhanced browsing pressure due to increased deer densities and increasingly warmer winters are important drivers. The study failed to demonstrate that eutrophication and acidification due to atmospheric deposition substantially drive herb layer changes. The restriction of the sample to the most base-rich sites in the region is discussed as a likely reason. Furthermore, the decline of several ancient forest species is discussed as an indication that the forest patches are still paying off their “extinction debt”, i.e., exhibit a delayed response to forest fragmentation. / Umweltveränderungen beeinträchtigen weltweit die Artenvielfalt. Die Lebensgemeinschaften fragmentierter Lebensräume gelten als besonders anfällig für Veränderungen. In dieser Arbeit wurden Untersuchungen an Krautschichtgemeinschaften historisch alter Waldfragmente im Elbe-Weser-Dreieck nach zwei Jahrzehnten wiederholt. Ziel war es anhand von 175 semi-permanenten Aufnahmeflächen (i) die Veränderungen der Pflanzenartendiversität zu quantifizieren, (ii) Artenverschiebungen in Bezug auf Nischenbreite und funktionale Merkmale festzustellen und (iii) Hinweise auf die verantwortlichen Umweltveränderungen zu finden. Die α-Diversität (Artenzahl) stieg durchschnittlich um vier Arten an. Die β-Diversität (Artenturnover zwischen den Flächen) nahm tendenziell ab. Letzteres wird als Beginn einer floristischen Homogenisierung interpretiert. Diese Studie ist eine der ersten im gemäßigten Europa, die eine floristische Homogenisierung von Waldpflanzengemeinschaften auf einer größeren als der lokalen Ebene aufzeigt. Die Diversitätsveränderungen auf taxonomischer Ebene führten nicht zu ähnlichen Veränderungen auf funktionaler Ebene. Bereits einzelne Gemeinschaften wiesen den Großteil der funktionalen Vielfalt des regionalen Artenpools, also ein Maximum an funktionaler Diversität auf. Gleichzeitig waren sie untereinander funktional redundant, also bereits homogen. Die mit der beginnenden taxonomischen Homogenisierung verbundene floristische Verarmung wird als gering eingestuft, da die Homogenisierung primär das Ergebnis der Zuwanderung häufig vorkommender Standortgeneralisten war. Unabhängig von der Homogenisierung gingen 30 Arten signifikant in ihrer Abundanz zurück, darunter 12 an historisch alte Wälder gebundene Arten sowie sieben Rote-Liste-Arten. Ein weiterer Rückgang oder ein lokales Aussterben dieser Arten stünde im Widerspruch zu regionalen Naturschutzzielen. Nullmodelltests und der Vergleich funktionaler und taxonomischer Diversitätskomponenten lassen auf regionaler Ebene auf eine zeitliche Konstanz der grundlegenden Mechanismen der Artenvergesellschaftung schließen. Auf der lokalen Ebene veränderten sich die Vergesellschaftungsmechanismen erheblich, insbesondere auf forstwirtschaftlich gestörten Standorten. Einerseits ermöglichte dort eine Abschwächung der Umweltfilter die Koexistenz von Arten mit unterschiedlichen Reproduktionsstrategien. Andererseits führte die erhöhte Lichtverfügbarkeit zu einer Verstärkung der Konkurrenzfilter und einer Selektion hochwüchsiger Arten. Gewinner- und Verliererarten wurden anhand funktionaler Merkmale und ihres ökologischen Verhaltens charakterisiert, um Hinweise auf die verantwortlichen Umweltveränderungen zu finden. Gewinnerarten wiesen eine höhere Langlebigkeit der Samen auf, blühten später in der Vegetationsperiode und hatten öfter eine ozeanische Verbreitung. Verliererarten hatten eine höhere spezifische Blattfläche, einen höheren Attraktivitätswert als Wildäsung und ein ökologisches Optimum bei höheren pH-Werten. Logistische Regressionsanalysen zeigen, dass Störung durch forstwirtschaftliche Eingriffe hauptverantwortlich für die Artenverschiebungen war. Zusätzlich liefert diese Wiederholungsstudie als eine der ersten in Europa Hinweise darauf, dass ein erhöhter Äsungsdruck sowie zunehmend mildere Winter entscheidende Einflussfaktoren darstellen. Der Rückgang mehrerer an historisch alte Wälder gebundener Arten wird als Anzeichen für eine verspätete Reaktion auf die Waldfragmentierung diskutiert.

Forest understory vegetation in lowland woodlands is coming through systematical changes, which are documented from many regions. Abandonment of traditional coppicing management, nitrogen deposition and effect of high game densities in some regions are considered to be the main key drivers. Changes of seminatural woodland in protected area of Bohemian Karst were studied in this research. The abandonment of traditional management is probably the biggest environmental change in this area. Original data, semi-permanent plots with phytosociological relevés, were surveyed shortly after the last coppicing in the half of twentieth century. These plots were now re-visited. Partial task was to evaluate how important relocation accuracy is. Discovered changes have similar trends as in other lowland woodlands. After abandonment of coppicing, site conditions changed towards more shade and nutrient rich sites. Vegetation shifted to species assemblages of late succession dominated by shade-adapted and nutrient-demanding species. Slight decrease in gama-diversity, increase in alpha-diversity and decrease of beta-diversity indicated taxonomic homogenization was recorded. Re-visiting of semi-permanent plots was evaluated as appropriate tool for this type of research. Spatial heterogenity is partly affecting temporal change,...