Exploring physical processes related to past climate proxies: lake sediments and stable water isotopes

Proxy data are essential for the investigation of climate variability on time scales larger than the historical meteorological observation period. The potential value of a proxy depends on our ability to understand and quantify the physical processes that relate the corresponding climate parameter and the signal in the proxy archive. These processes can be explored under present-day conditions. In this thesis, both statistical and physical models are applied for their analysis, focusing on two specific types of proxies, lake sediment data and stable water isotopes.rnIn the first part of this work, the basis is established for statistically calibrating new proxies from lake sediments in western Germany. A comprehensive meteorological and hydrological data set is compiled and statistically analyzed. In this way, meteorological times series are identified that can be applied for the calibration of various climate proxies. A particular focus is laid on the investigation of extreme weather events, which have rarely been the objective of paleoclimate reconstructions so far. Subsequently, a concrete example of a proxy calibration is presented. Maxima in the quartz grain concentration from a lake sediment core are compared to recent windstorms. The latter are identified from the meteorological data with the help of a newly developed windstorm index, combining local measurements and reanalysis data. The statistical significance of the correlation between extreme windstorms and signals in the sediment is verified with the help of a Monte Carlo method. This correlation is fundamental for employing lake sediment data as a new proxy to reconstruct windstorm records of the geological past.rnThe second part of this thesis deals with the analysis and simulation of stable water isotopes in atmospheric vapor on daily time scales. In this way, a better understanding of the physical processes determining these isotope ratios can be obtained, which is an important prerequisite for the interpretation of isotope data from ice cores and the reconstruction of past temperature. In particular, the focus here is on the deuterium excess and its relation to the environmental conditions during evaporation of water from the ocean. As a basis for the diagnostic analysis and for evaluating the simulations, isotope measurements from Rehovot (Israel) are used, provided by the Weizmann Institute of Science. First, a Lagrangian moisture source diagnostic is employed in order to establish quantitative linkages between the measurements and the evaporation conditions of the vapor (and thus to calibrate the isotope signal). A strong negative correlation between relative humidity in the source regions and measured deuterium excess is found. On the contrary, sea surface temperature in the evaporation regions does not correlate well with deuterium excess. Although requiring confirmation by isotope data from different regions and longer time scales, this weak correlation might be of major importance for the reconstruction of moisture source temperatures from ice core data. Second, the Lagrangian source diagnostic is combined with a Craig-Gordon fractionation parameterization for the identified evaporation events in order to simulate the isotope ratios at Rehovot. In this way, the Craig-Gordon model can be directly evaluated with atmospheric isotope data, and better constraints for uncertain model parameters can be obtained. A comparison of the simulated deuterium excess with the measurements reveals that a much better agreement can be achieved using a wind speed independent formulation of the non-equilibrium fractionation factor instead of the classical parameterization introduced by Merlivat and Jouzel, which is widely applied in isotope GCMs. Finally, the first steps of the implementation of water isotope physics in the limited-area COSMO model are described, and an approach is outlined that allows to compare simulated isotope ratios to measurements in an event-based manner by using a water tagging technique. The good agreement between model results from several case studies and measurements at Rehovot demonstrates the applicability of the approach. Because the model can be run with high, potentially cloud-resolving spatial resolution, and because it contains sophisticated parameterizations of many atmospheric processes, a complete implementation of isotope physics will allow detailed, process-oriented studies of the complex variability of stable isotopes in atmospheric waters in future research.rn

Advanced data deduplication techniques and their application

Data deduplication describes a class of approaches that reduce the storage capacity needed to store data or the amount of data that has to be transferred over a network. These approaches detect coarse-grained redundancies within a data set, e.g. a file system, and remove them.rnrnOne of the most important applications of data deduplication are backup storage systems where these approaches are able to reduce the storage requirements to a small fraction of the logical backup data size.rnThis thesis introduces multiple new extensions of so-called fingerprinting-based data deduplication. It starts with the presentation of a novel system design, which allows using a cluster of servers to perform exact data deduplication with small chunks in a scalable way.rnrnAfterwards, a combination of compression approaches for an important, but often over- looked, data structure in data deduplication systems, so called block and file recipes, is introduced. Using these compression approaches that exploit unique properties of data deduplication systems, the size of these recipes can be reduced by more than 92% in all investigated data sets. As file recipes can occupy a significant fraction of the overall storage capacity of data deduplication systems, the compression enables significant savings.rnrnA technique to increase the write throughput of data deduplication systems, based on the aforementioned block and file recipes, is introduced next. The novel Block Locality Caching (BLC) uses properties of block and file recipes to overcome the chunk lookup disk bottleneck of data deduplication systems. This chunk lookup disk bottleneck either limits the scalability or the throughput of data deduplication systems. The presented BLC overcomes the disk bottleneck more efficiently than existing approaches. Furthermore, it is shown that it is less prone to aging effects.rnrnFinally, it is investigated if large HPC storage systems inhibit redundancies that can be found by fingerprinting-based data deduplication. Over 3 PB of HPC storage data from different data sets have been analyzed. In most data sets, between 20 and 30% of the data can be classified as redundant. According to these results, future work in HPC storage systems should further investigate how data deduplication can be integrated into future HPC storage systems.rnrnThis thesis presents important novel work in different area of data deduplication re- search.