@phdthesis{Hahn2002,
  author    = {Hahn, Robert},
  title     = {Das Bl{\"u}te-Best{\"a}uber-Netz auf Brachfl{\"a}chen : bioz{\"o}nologische Untersuchung zur Bedeutung von Brachen in einer intensiv genutzten Agrarlandschaft},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-0000652},
  school      = {Universit{\"a}t Potsdam},
  year      = {2002},
  abstract  = {In der vorliegenden Dissertation wird die Bedeutung von Brachen f{\"u}r Artenvielfalt und Stabilit{\"a}t von Bl{\"u}te-Best{\"a}uber-Nahrungsnetzen in agrarisch genutzten Landschaften anhand ausgew{\"a}hlter bl{\"u}tenbesuchender Insektengruppen (Syrphidae, Lepidoptera) untersucht. Die Freilandarbeiten fanden von 1998-2000 im Raum der Feldberger Seenlandschaft, Mecklenburg-Vorpommern, statt. Es werden die beiden Hauptnahrungsquellen Nektar und Pollen betrachtet, dabei fanden Untersuchungen zur Intensit{\"a}t der Bl{\"u}te-Best{\"a}uber-Interaktion auf Stilllegungsfl{\"a}chen, zum fl{\"a}chenbezogenen quantitativen Nektarangebot im Jahresverlauf, zur individuellen Pollennutzung bei Syrphiden und zur Breite und {\"U}berlappung der Nahrungsnischen bei den dominanten Arten Episyrphus balteatus, Metasyrphus corollae, Syritta pipiens und Sphaerophoria scripta statt. Im Ergebnis zeigt sich eine hohe Bedeutung der Brachfl{\"a}chen f{\"u}r die Stabilit{\"a}t des Bl{\"u}te-Best{\"a}uber-Netzes, w{\"a}hrend die Diversit{\"a}t von anderen, eher landschaftsbezogenen Faktoren abh{\"a}ngig ist.},
  subject      = {Feldberger Seenlandschaft ; Agrarlandschaft ; Brache ; Samenpflanzen ; Best{\"a}uber ; Artenreichtum},
  language  = {de}
}
@phdthesis{TabaresJimenez2021,
  author    = {Tabares Jimenez, Ximena del Carmen},
  title     = {A palaeoecological approach to savanna dynamics and shrub encroachment in Namibia},
  doi       = {10.25932/publishup-49281},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-492815},
  school      = {Universit{\"a}t Potsdam},
  pages     = {121},
  year      = {2021},
  abstract  = {The spread of shrubs in Namibian savannas raises questions about the resilience of these ecosystems to global change. This makes it necessary to understand the past dynamics of the vegetation, since there is no consensus on whether shrub encroachment is a new phenomenon, nor on its main drivers. However, a lack of long-term vegetation datasets for the region and the scarcity of suitable palaeoecological archives, makes reconstructing past vegetation and land cover of the savannas a challenge. To help meet this challenge, this study addresses three main research questions: 1) is pollen analysis a suitable tool to reflect the vegetation change associated with shrub encroachment in savanna environments? 2) Does the current encroached landscape correspond to an alternative stable state of savanna vegetation? 3) To what extent do pollen-based quantitative vegetation reconstructions reflect changes in past land cover? The research focuses on north-central Namibia, where despite being the region most affected by shrub invasion, particularly since the 21st century, little is known about the dynamics of this phenomenon. Field-based vegetation data were compared with modern pollen data to assess their correspondence in terms of composition and diversity along precipitation and grazing intensity gradients. In addition, two sediment cores from Lake Otjikoto were analysed to reveal changes in vegetation composition that have occurred in the region over the past 170 years and their possible drivers. For this, a multiproxy approach (fossil pollen, sedimentary ancient DNA (sedaDNA), biomarkers, compound specific carbon (δ13C) and deuterium (δD) isotopes, bulk carbon isotopes (δ13Corg), grain size, geochemical properties) was applied at high taxonomic and temporal resolution. REVEALS modelling of the fossil pollen record from Lake Otjikoto was run to quantitatively reconstruct past vegetation cover. For this, we first made pollen productivity estimates (PPE) of the most relevant savanna taxa in the region using the extended R-value model and two pollen dispersal options (Gaussian plume model and Lagrangian stochastic model). The REVEALS-based vegetation reconstruction was then validated using remote sensing-based regional vegetation data. The results show that modern pollen reflects the composition of the vegetation well, but diversity less well. Interestingly, precipitation and grazing explain a significant amount of the compositional change in the pollen and vegetation spectra. The multiproxy record shows that a state change from open Combretum woodland to encroached Terminalia shrubland can occur over a century, and that the transition between states spans around 80 years and is characterized by a unique vegetation composition. This transition is supported by gradual environmental changes induced by management (i.e. broad-scale logging for the mining industry, selective grazing and reduced fire activity associated with intensified farming) and related land-use change. Derived environmental changes (i.e. reduced soil moisture, reduced grass cover, changes in species composition and competitiveness, reduced fire intensity) may have affected the resilience of Combretum open woodlands, making them more susceptible to change to an encroached state by stochastic events such as consecutive years of precipitation and drought, and by high concentrations of pCO2. We assume that the resulting encroached state was further stabilized by feedback mechanisms that favour the establishment and competitiveness of woody vegetation. The REVEALS-based quantitative estimates of plant taxa indicate the predominance of a semi-open landscape throughout the 20th century and a reduction in grass cover below 50\% since the 21st century associated with the spread of encroacher woody taxa. Cover estimates show a close match with regional vegetation data, providing support for the vegetation dynamics inferred from multiproxy analyses. Reasonable PPEs were made for all woody taxa, but not for Poaceae. In conclusion, pollen analysis is a suitable tool to reconstruct past vegetation dynamics in savannas. However, because pollen cannot identify grasses beyond family level, a multiproxy approach, particularly the use of sedaDNA, is required. I was able to separate stable encroached states from mere woodland phases, and could identify drivers and speculate about related feedbacks. In addition, the REVEALS-based quantitative vegetation reconstruction clearly reflects the magnitude of the changes in the vegetation cover that occurred during the last 130 years, despite the limitations of some PPEs. This research provides new insights into pollen-vegetation relationships in savannas and highlights the importance of multiproxy approaches when reconstructing past vegetation dynamics in semi-arid environments. It also provides the first time series with sufficient taxonomic resolution to show changes in vegetation composition during shrub encroachment, as well as the first quantitative reconstruction of past land cover in the region. These results help to identify the different stages in savanna dynamics and can be used to calibrate predictive models of vegetation change, which are highly relevant to land management.},
  language  = {en}
}