Characterization of particle dark matter via multiple probes

The dark matter problem is one of the most striking puzzles in physics today. Cosmological and astrophysical observations have provided strong evidence that over 80% of the matter in the Universe is dark. However, direct proof for the existence of dark matter particles from laboratory experiments is still lacking, so that the physical nature of dark matter remains unknown. Possible solutions are found in theoretical models of new physics, which propose new particles that are excellent dark matter candidates, thus presenting a fundamental connection between elementary particle physics and the astrophysical dark matter. In this thesis, I adopt a multi-messenger approach towards the identification and characterisation of the dark matter particle. I apply advanced statistical and numerical techniques to probe theoretical models and derive robust constraints on the nature and properties of dark matter in light of the full range of existing experimental results. I present global fits analyses of three models of supersymmetry (the cMSSM, the NUHM and the MSSM-15), including data from collider searches for new physics, cosmology experiments, astro-particle dark matter searches, and the Higgs boson discovery. A strong complementarity between the LHC and astro-particle experiments is observed, highlighting the benefits of a combined analysis. I find that constrained models, such as the cMSSM and the NUHM, that were appropriate targets for global fits prior to the start of LHC operations, have been placed under strong pressure by recent data sets. I present the first statistically convergent profile likelihood maps of a 15-dimensional MSSM, which is only weakly constrained by the existing data, and is a much more suitable framework for phenomenological studies of supersymmetry. I derive robust and statistically meaningful constraints on the supersymmetric parameters and dark matter properties in this model. Detection prospects for the cMSSM and the NUHM are positive, while fully probing the rich phenomenology of the MSSM-15 is more difficult. I present the regions of the parameter spaces that are most promising to explore with future searches and pinpoint the signatures characteristic of supersymmetric dark matter in these models. A very effective experimental strategy is the direct detection of dark matter. I explore the statistical limitations of next-generation direct detection experiments in the case of a significant detection. I find that the uncertainty and bias in the reconstructed WIMP properties is particularly severe for heavy WIMPs, but can also be significant for intermediate-mass WIMPs leading to several hundreds of events. I demonstrate that the precision and accuracy of the WIMP characterisation can be considerably improved by exploiting the complementarity between different target materials, and by increasing the experimental exposure.Open Acces