@phdthesis{Brosinsky2015,
  author    = {Brosinsky, Arlena},
  title     = {Spectral fingerprinting},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-83369},
  school      = {Universit{\"a}t Potsdam},
  pages     = {VI, 117},
  year      = {2015},
  abstract  = {Current research on runoff and erosion processes, as well as an increasing demand for sustainable watershed management emphasize the need for an improved understanding of sediment dynamics. This involves the accurate assessment of erosion rates and sediment transfer, yield and origin. A variety of methods exist to capture these processes at the catchment scale. Among these, sediment fingerprinting, a technique to trace back the origin of sediment, has attracted increasing attention by the scientific community in recent years. It is a two-step procedure, based on the fundamental assumptions that potential sources of sediment can be reliably discriminated based on a set of characteristic 'fingerprint' properties, and that a comparison of source and sediment fingerprints allows to quantify the relative contribution of each source. This thesis aims at further assessing the potential of spectroscopy to assist and improve the sediment fingerprinting technique. Specifically, this work focuses on (1) whether potential sediment sources can be reliably identified based on spectral features ('fingerprints'), whether (2) these spectral fingerprints permit the quantification of relative source contribution, and whether (3) in situ derived source information is sufficient for this purpose. Furthermore, sediment fingerprinting using spectral information is applied in a study catchment to (4) identify major sources and observe how relative source contributions change between and within individual flood events. And finally, (5) spectral fingerprinting results are compared and combined with simultaneous sediment flux measurements to study sediment origin, transport and storage behaviour. For the sediment fingerprinting approach, soil samples were collected from potential sediment sources within the Is{\´a}bena catchment, a meso-scale basin in the central Spanish Pyrenees. Undisturbed samples of the upper soil layer were measured in situ using an ASD spectroradiometer and subsequently sampled for measurements in the laboratory. Suspended sediment was sampled automatically by means of ISCO samplers at the catchment as well as at the five major subcatchment outlets during flood events, and stored fine sediment from the channel bed was collected from 14 cross-sections along the main river. Artificial mixtures of known contributions were produced from source soil samples. Then, all source, sediment and mixture samples were dried and spectrally measured in the laboratory. Subsequently, colour coefficients and physically based features with relation to organic carbon, iron oxide, clay content and carbonate, were calculated from all in situ and laboratory spectra. Spectral parameters passing a number of prerequisite tests were submitted to principal component analyses to study natural clustering of samples, discriminant function analyses to observe source differentiation accuracy, and a mixing model for source contribution assessment. In addition, annual as well as flood event based suspended sediment fluxes from the catchment and its subcatchments were calculated from rainfall, water discharge and suspended sediment concentration measurements using rating curves and Quantile Regression Forests. Results of sediment flux monitoring were interpreted individually with respect to storage behaviour, compared to fingerprinting source ascriptions and combined with fingerprinting to assess their joint explanatory potential. In response to the key questions of this work, (1) three source types (land use) and five spatial sources (subcatchments) could be reliably discriminated based on spectral fingerprints. The artificial mixture experiment revealed that while (2) laboratory parameters permitted source contribution assessment, (3) the use of in situ derived information was insufficient. Apparently, high discrimination accuracy does not necessarily imply good quantification results. When applied to suspended sediment samples of the catchment outlet, the spectral fingerprinting approach was able to (4) quantify the major sediment sources: badlands and the Villacarli subcatchment, respectively, were identified as main contributors, which is consistent with field observations and previous studies. Thereby, source contribution was found to vary both, within and between individual flood events. Also sediment flux was found to vary considerably, annually as well as seasonally and on flood event base. Storage was confirmed to play an important role in the sediment dynamics of the studied catchment, whereas floods with lower total sediment yield tend to deposit and floods with higher yield rather remove material from the channel bed. Finally, a comparison of flux measurements with fingerprinting results highlighted the fact that (5) immediate transport from sources to the catchment outlet cannot be assumed. A combination of the two methods revealed different aspects of sediment dynamics that none of the techniques could have uncovered individually. In summary, spectral properties provide a fast, non-destructive, and cost-efficient means to discriminate and quantify sediment sources, whereas, unfortunately, straight-forward in situ collected source information is insufficient for the approach. Mixture modelling using artificial mixtures permits valuable insights into the capabilities and limitations of the method and similar experiments are strongly recommended to be performed in the future. Furthermore, a combination of techniques such as e.g. (spectral) sediment fingerprinting and sediment flux monitoring can provide comprehensive understanding of sediment dynamics.},
  language  = {en}
}
@phdthesis{Luft2015,
  author    = {Luft, Laura Charlotte},
  title     = {Bridging the gap between science and nature conservation practice},
  school      = {Universit{\"a}t Potsdam},
  pages     = {173},
  year      = {2015},
  language  = {en}
}
@phdthesis{Allroggen2015,
  author    = {Allroggen, Niklas},
  title     = {Observation of subsurface flow from the surface : applications of ground-penetrating radar},
  school      = {Universit{\"a}t Potsdam},
  pages     = {67},
  year      = {2015},
  language  = {en}
}
@phdthesis{Kalbe2016,
  author    = {Kalbe, Johannes},
  title     = {Stepping stones hominin dispersal out of Africa},
  school      = {Universit{\"a}t Potsdam},
  pages     = {122},
  year      = {2016},
  language  = {en}
}
@phdthesis{Lenz2016,
  author    = {Lenz, Josefine},
  title     = {Thermokarst dynamics in central-eastern Beringia},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-101364},
  school      = {Universit{\"a}t Potsdam},
  pages     = {XII, 128, A-47},
  year      = {2016},
  abstract  = {Widespread landscape changes are presently observed in the Arctic and are most likely to accelerate in the future, in particular in permafrost regions which are sensitive to climate warming. To assess current and future developments, it is crucial to understand past environmental dynamics in these landscapes. Causes and interactions of environmental variability can hardly be resolved by instrumental records covering modern time scales. However, long-term environmental variability is recorded in paleoenvironmental archives. Lake sediments are important archives that allow reconstruction of local limnogeological processes as well as past environmental changes driven directly or indirectly by climate dynamics. This study aims at reconstructing Late Quaternary permafrost and thermokarst dynamics in central-eastern Beringia, the terrestrial land mass connecting Eurasia and North America during glacial sea-level low stands. In order to investigate development, processes and influence of thermokarst dynamics, several sediment cores from extant lakes and drained lake basins were analyzed to answer the following research questions: 1. When did permafrost degradation and thermokarst lake development take place and what were enhancing and inhibiting environmental factors? 2. What are the dominant processes during thermokarst lake development and how are they reflected in proxy records? 3. How did, and still do, thermokarst dynamics contribute to the inventory and properties of organic matter in sediments and the carbon cycle? Methods applied in this study are based upon a multi-proxy approach combining sedimentological, geochemical, geochronological, and micropaleontological analyses, as well as analyses of stable isotopes and hydrochemistry of pore-water and ice. Modern field observations of water quality and basin morphometrics complete the environmental investigations. The investigated sediment cores reveal permafrost degradation and thermokarst dynamics on different time scales. The analysis of a sediment core from GG basin on the northern Seward Peninsula (Alaska) shows prevalent terrestrial accumulation of yedoma throughout the Early to Mid Wisconsin with intermediate wet conditions at around 44.5 to 41.5 ka BP. This first wetland development was terminated by the accumulation of a 1-meter-thick airfall tephra most likely originating from the South Killeak Maar eruption at 42 ka BP. A depositional hiatus between 22.5 and 0.23 ka BP may indicate thermokarst lake formation in the surrounding of the site which forms a yedoma upland till today. The thermokarst lake forming GG basin initiated 230 ± 30 cal a BP and drained in Spring 2005 AD. Four years after drainage the lake talik was still unfrozen below 268 cm depth. A permafrost core from Mama Rhonda basin on the northern Seward Peninsula preserved a full lacustrine record including several lake phases. The first lake generation developed at 11.8 cal ka BP during the Lateglacial-Early Holocene transition; its old basin (Grandma Rhonda) is still partially preserved at the southern margin of the study basin. Around 9.0 cal ka BP a shallow and more dynamic thermokarst lake developed with actively eroding shorelines and potentially intermediate shallow water or wetland phases (Mama Rhonda). Mama Rhonda lake drainage at 1.1 cal ka BP was followed by gradual accumulation of terrestrial peat and top-down refreezing of the lake talik. A significant lower organic carbon content was measured in Grandma Rhonda deposits (mean TOC of 2.5 wt\%) than in Mama Rhonda deposits (mean TOC of 7.9 wt\%) highlighting the impact of thermokarst dynamics on biogeochemical cycling in different lake generations by thawing and mobilization of organic carbon into the lake system. Proximal and distal sediment cores from Peatball Lake on the Arctic Coastal Plain of Alaska revealed young thermokarst dynamics since about 1,400 years along a depositional gradient based on reconstructions from shoreline expansion rates and absolute dating results. After its initiation as a remnant pond of a previous drained lake basin, a rapidly deepening lake with increasing oxygenation of the water column is evident from laminated sediments, and higher Fe/Ti and Fe/S ratios in the sediment. The sediment record archived characterizing shifts in depositional regimes and sediment sources from upland deposits and re-deposited sediments from drained thaw lake basins depending on the gradually changing shoreline configuration. These changes are evident from alternating organic inputs into the lake system which highlights the potential for thermokarst lakes to recycle old carbon from degrading permafrost deposits of its catchment. The lake sediment record from Herschel Island in the Yukon (Canada) covers the full Holocene period. After its initiation as a thermokarst lake at 11.7 cal ka BP and intense thermokarst activity until 10.0 cal ka BP, the steady sedimentation was interrupted by a depositional hiatus at 1.6 cal ka BP which likely resulted from lake drainage or allochthonous slumping due to collapsing shore lines. The specific setting of the lake on a push moraine composed of marine deposits is reflected in the sedimentary record. Freshening of the maturing lake is indicated by decreasing electrical conductivity in pore-water. Alternation of marine to freshwater ostracods and foraminifera confirms decreasing salinity as well but also reflects episodical re-deposition of allochthonous marine sediments. Based on permafrost and lacustrine sediment records, this thesis shows examples of the Late Quaternary evolution of typical Arctic permafrost landscapes in central-eastern Beringia and the complex interaction of local disturbance processes, regional environmental dynamics and global climate patterns. This study confirms that thermokarst lakes are important agents of organic matter recycling in complex and continuously changing landscapes.},
  language  = {en}
}
@phdthesis{Baese2016,
  author    = {B{\"a}se, Frank},
  title     = {Interception loss of changing land covers in the humid tropical lowland of Latin America},
  school      = {Universit{\"a}t Potsdam},
  pages     = {ix, 85 Seiten},
  year      = {2016},
  abstract  = {Das Gebiet der feuchten Tropen ist die am st{\"a}rksten durch den Landnutzungswandel betroffene Region der Erde. Vor allem die Rodung tropischer W{\"a}lder, um Platz f{\"u}r Rinderweiden oder den Anbau von Soja zu schaffen, aber auch seit j{\"u}ngster Zeit die Bem{\"u}hungen um Wiederaufforstungen pr{\"a}gen diesen Landnutzungswandel. Dabei beeinflusst die {\"A}nderung der Vegetationsbedeckung den regionalen Wasserhaushalt auf vielf{\"a}ltige Weise. Betroffen ist unter anderem die Verdunstung von feuchten Oberfl{\"a}chen. Die so genannte Interzeptionsverdunstung bzw. der Interzeptionsverlust tr{\"a}gt erheblich zum Wasserdampfgehalt in der unteren Atmosph{\"a}re und schließlich zur Niederschlagsbildung bei. Ziele dieser Dissertation waren (1) die experimentelle Untersuchung der Interzeptionsverlustunterschiede zwischen einem nat{\"u}rlichen, tropischen Wald und einer Sojaplantage im s{\"u}dlichen Amazonasgebiet, (2) die Modellierung des Interzeptionsverlustes dieser beiden Vegetationsformen im Vergleich zu einem jungen Sekund{\"a}rwald unter dem Aspekt der Unsicherheiten bei der Ableitung notwendiger Modellparameter sowohl im S{\"u}damazonas als auch im Einzugsgebietes des Panamakanals sowie (3) die Wasserhaushaltsanalyse eines vom Landnutzungswandel gepr{\"a}gten Teileinzugsgebietes des Panamakanals in Hinblick auf die Ver{\"a}nderung der Interzeptionsverdunstung durch sich ver{\"a}ndernde Landnutzung und der {\"A}nderung der klimatischen Bedingungen. Die Messung des Interzeptionsverlustes zeigte, dass in der Hauptwachstumsphase vom Soja von dessen Oberfl{\"a}che mehr Wasserverdunstet als von der Oberfl{\"a}che des Waldes. Allerdings ist in der Jahresbilanz der Interzeptionsverlust vom Wald h{\"o}her, da diese Studie nur eine Momentaufnahme zur Zeit der vollen Vegetationsentwicklung des Sojas mit einem Zeitfenster von zwei Monaten widerspiegelt. Durch die geringere ganzj{\"a}hrige Verdunstung von den mit Soja bestandenen Fl{\"a}chen, wird hier der Niederschlag schneller dem Abfluss zugef{\"u}hrt und schell aus der Region ausgetragen. Somit tr{\"a}gt der Landnutzungswandel von Wald zu Soja zu einer mittelfristigen Reduktion des in der Region verf{\"u}gbaren Wassers bei. Die anschließende Modellierung des Interzeptionsverlustes zeigte Einerseits einen starken Einfluss der Datenqualit{\"a}t auf die Plausibilit{\"a}t der Ergebnisse und Andererseits, dass die Sensitivit{\"a}t der einzelnen Parameter zwischen den Untersuchungsgebieten variiert. Eine Schl{\"u}sselrolle nimmt die Wasserspeicherkapazit{\"a}t der Vegetationskrone ein. Dennoch ist die Evaporationsrate die treibende Gr{\"o}ße im Interzeptionsprozess, so dass von ihr die gr{\"o}ßte Unsicherheit ausgeht. Je nach verwendeter Methode zur Ableitung dieses Parameters unterscheiden sich die gewonnenen Parameterwerte erheblich. Die Wirkungsanalyse der Interzeptionsverdunstung auf den Wasserhaushalt im Wirkungsgeflecht der {\"A}nderungen von Temperatur, Niederschlag und Landnutzung im Landschaftsmosaik eines Flusseinzugsgebiets mit Hilfe eines Wasserhaushaltsmodels zeigte den Einfluss der Landnutzungs{\"a}nderung auf die Abflussbildung mittels verschiedener Landnutzungsszenarien. Die Ergebnisse belegen, dass die Landnutzungs{\"a}nderung im Gebiet nur einen geringen Einfluss auf den Jahresabfluss hat. St{\"a}rker scheint sich der gemessene Temperaturanstieg auf die Verdunstung auszuwirken. Der mit einer h{\"o}heren Temperatur einhergehende Anstieg der Transpiration und Interzeptionsverdunstung gleicht die gemessene Zunahme des Gebietsniederschlages aus, sodass keine signifikanten {\"A}nderungen im Jahresabfluss nachgewiesen werden konnten. Die Ergebnisse der drei Studien verdeutlichen den Einfluss der Landnutzung auf die Interzeptionsverdunstung. Allerdings veranschaulichten die Resultate der Wasserhaushalts-modellierung, wie sehr dieser Einfluss durch die Ver{\"a}nderung der {\"a}ußeren Rahmenbedingungen, vor allem durch den Anstieg der Temperatur, {\"u}berpr{\"a}gt werden kann. Dies belegt, dass eine einfache {\"U}bertragung der Ergebnisse zwischen den Untersuchungsgebiet nicht m{\"o}glich ist. Somit bleibt die experimentelle Erhebung von Vegetationsparametern sowie des Interzeptionsverlustes an den jeweils zu untersuchenden Standort f{\"u}r die Anwendung von Modellen unerl{\"a}sslich.},
  language  = {en}
}
@phdthesis{Draeger2016,
  author    = {Dr{\"a}ger, Nadine},
  title     = {Holocene climate and environmental variability in NE Germany inferred from annually laminated lake sediments},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-103037},
  school      = {Universit{\"a}t Potsdam},
  pages     = {xv, 144 Seiten},
  year      = {2016},
  abstract  = {Understanding the role of natural climate variability under the pressure of human induced changes of climate and landscapes, is crucial to improve future projections and adaption strategies. This doctoral thesis aims to reconstruct Holocene climate and environmental changes in NE Germany based on annually laminated lake sediments. The work contributes to the ICLEA project (Integrated CLimate and Landscape Evolution Analyses). ICLEA intends to compare multiple high-resolution proxy records with independent chronologies from the N central European lowlands, in order to disentangle the impact of climate change and human land use on landscape development during the Lateglacial and Holocene. In this respect, two study sites in NE Germany are investigated in this doctoral project, Lake Tiefer See and palaeolake Wukenfurche. While both sediment records are studied with a combination of high-resolution sediment microfacies and geochemical analyses (e.g. µ-XRF, carbon geochemistry and stable isotopes), detailed proxy understanding mainly focused on the continuous 7.7 m long sediment core from Lake Tiefer See covering the last ~6000 years. Three main objectives are pursued at Lake Tiefer See: (1) to perform a reliable and independent chronology, (2) to establish microfacies and geochemical proxies as indicators for climate and environmental changes, and (3) to trace the effects of climate variability and human activity on sediment deposition. Addressing the first aim, a reliable chronology of Lake Tiefer See is compiled by using a multiple-dating concept. Varve counting and tephra findings form the chronological framework for the last ~6000 years. The good agreement with independent radiocarbon dates of terrestrial plant remains verifies the robustness of the age model. The resulting reliable and independent chronology of Lake Tiefer See and, additionally, the identification of nine tephras provide a valuable base for detailed comparison and synchronization of the Lake Tiefer See data set with other climate records. The sediment profile of Lake Tiefer See exhibits striking alternations between well-varved and non-varved sediment intervals. The combination of microfacies, geochemical and microfossil (i.e. Cladocera and diatom) analyses indicates that these changes of varve preservation are caused by variations of lake circulation in Lake Tiefer See. An exception is the well-varved sediment deposited since AD 1924, which is mainly influenced by human-induced lake eutrophication. Well-varved intervals before the 20th century are considered to reflect phases of reduced lake circulation and, consequently, stronger anoxic conditions. Instead, non-varved intervals indicate increased lake circulation in Lake Tiefer See, leading to more oxygenated conditions at the lake ground. Furthermore, lake circulation is not only influencing sediment deposition, but also geochemical processes in the lake. As, for example, the proxy meaning of δ13COM varies in time in response to changes of the oxygen regime in the lake hypolinion. During reduced lake circulation and stronger anoxic conditions δ13COM is influenced by microbial carbon cycling. In contrast, organic matter degradation controls δ13COM during phases of intensified lake circulation and more oxygenated conditions. The varve preservation indicates an increasing trend of lake circulation at Lake Tiefer See after ~4000 cal a BP. This trend is superimposed by decadal to centennial scale variability of lake circulation intensity. Comparison to other records in Central Europe suggests that the long-term trend is probably related to gradual changes in Northern Hemisphere orbital forcing, which induced colder and windier conditions in Central Europe and, therefore, reinforced lake circulation. Decadal to centennial scale periods of increased lake circulation coincide with settlement phases at Lake Tiefer See, as inferred from pollen data of the same sediment record. Deforestation reduced the wind shelter of the lake, which probably increased the sensitivity of lake circulation to wind stress. However, results of this thesis also suggest that several of these phases of increased lake circulation are additionally reinforced by climate changes. A first indication is provided by the comparison to the Baltic Sea record, which shows striking correspondence between major non-varved intervals at Lake Tiefer See and bioturbated sediments in the Baltic Sea. Furthermore, a preliminary comparison to the ICLEA study site Lake Czechowskie (N central Poland) shows a coincidence of at least three phases of increased lake circulation in both lakes, which concur with periods of known climate changes (2.8 ka event, 'Migration Period' and 'Little Ice Age'). These results suggest an additional over-regional climate forcing also on short term increased of lake circulation in Lake Tiefer See. In summary, the results of this thesis suggest that lake circulation at Lake Tiefer See is driven by a combination of long-term and short-term climate changes as well as of anthropogenic deforestation phases. Furthermore, the lake circulation drives geochemical cycles in the lake affecting the meaning of proxy data. Therefore, the work presented here expands the knowledge of climate and environmental variability in NE Germany. Furthermore, the integration of the Lake Tiefer See multi-proxy record in a regional comparison with another ICLEA side, Lake Czechowskie, enabled to better decipher climate changes and human impact on the lake system. These first results suggest a huge potential for further detailed regional comparisons to better understand palaeoclimate dynamics in N central Europe.},
  language  = {en}
}
@phdthesis{Chen2016,
  author    = {Chen, Kejie},
  title     = {Real-time GNSS for fast seismic source inversion and tsunami early warning},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-93174},
  school      = {Universit{\"a}t Potsdam},
  pages     = {xii, 81},
  year      = {2016},
  abstract  = {Over the past decades, rapid and constant advances have motivated GNSS technology to approach the ability to monitor transient ground motions with mm to cm accuracy in real-time. As a result, the potential of using real-time GNSS for natural hazards prediction and early warning has been exploited intensively in recent years, e.g., landslides and volcanic eruptions monitoring. Of particular note, compared with traditional seismic instruments, GNSS does not saturate or tilt in terms of co-seismic displacement retrieving, which makes it especially valuable for earthquake and earthquake induced tsunami early warning. In this thesis, we focus on the application of real-time GNSS to fast seismic source inversion and tsunami early warning. Firstly, we present a new approach to get precise co-seismic displacements using cost effective single-frequency receivers. As is well known, with regard to high precision positioning, the main obstacle for single-frequency GPS receiver is ionospheric delay. Considering that over a few minutes, the change of ionospheric delay is almost linear, we constructed a linear model for each satellite to predict ionospheric delay. The effectiveness of this method has been validated by an out-door experiment and 2011 Tohoku event, which confirms feasibility of using dense GPS networks for geo-hazard early warning at an affordable cost. Secondly, we extended temporal point positioning from GPS-only to GPS/GLONASS and assessed the potential benefits of multi-GNSS for co-seismic displacement determination. Out-door experiments reveal that when observations are conducted in an adversary environment, adding a couple of GLONASS satellites could provide more reliable results. The case study of 2015 Illapel Mw 8.3 earthquake shows that the biases between co-seismic displacements derived from GPS-only and GPS/GLONASS vary from station to station, and could be up to 2 cm in horizontal direction and almost 3 cm in vertical direction. Furthermore, slips inverted from GPS/GLONASS co-seismic displacements using a layered crust structure on a curved plane are shallower and larger for the Illapel event. Thirdly, we tested different inversion tools and discussed the uncertainties of using real-time GNSS for tsunami early warning. To be exact, centroid moment tensor inversion, uniform slip inversion using a single Okada fault and distributed slip inversion in layered crust on a curved plane were conducted using co-seismic displacements recorded during 2014 Pisagua earthquake. While the inversion results give similar magnitude and the rupture center, there are significant differences in depth, strike, dip and rake angles, which lead to different tsunami propagation scenarios. Even though, resulting tsunami forecasting along the Chilean coast is close to each other for all three models. Finally, based on the fact that the positioning performance of BDS is now equivalent to GPS in Asia-Pacific area and Manila subduction zone has been identified as a zone of potential tsunami hazard, we suggested a conceptual BDS/GPS network for tsunami early warning in South China Sea. Numerical simulations with two earthquakes (Mw 8.0 and Mw 7.5) and induced tsunamis demonstrate the viability of this network. In addition, the advantage of BDS/GPS over a single GNSS system by source inversion grows with decreasing earthquake magnitudes.},
  language  = {en}
}
@phdthesis{Wolter2016,
  author    = {Wolter, Juliane},
  title     = {Mid- to Late Holocene environmental dynamics on the Yukon Coastal Plain and Herschel Island (Canada) - envidence from polygonal peatlands and lake sediment},
  school      = {Universit{\"a}t Potsdam},
  pages     = {176},
  year      = {2016},
  language  = {en}
}
@phdthesis{Rieckh2016,
  author    = {Rieckh, Helene},
  title     = {Hydropedological analysis of erosion-affected soils in a hummocky ground-moraine landscape - interactions of water flow, dissolved carbon and particle transport, grop growth, and pedogenesis},
  school      = {Universit{\"a}t Potsdam},
  pages     = {154},
  year      = {2016},
  language  = {en}
}