Complex systems are often simulated by running computational models. As models are not ideal and measurements imperfect, simulation results are uncertain. To evaluate uncertainty, ensembles of forecasts started from slightly perturbed initial conditions are generated. This thesis proposes novel methods to model ensembles statistically and convey variability aspects effectively, by analyzing the stability of critical points for scalar ensembles, and performing a comparative visual analysis of local and global variability for vector ensembles.
«
Complex systems are often simulated by running computational models. As models are not ideal and measurements imperfect, simulation results are uncertain. To evaluate uncertainty, ensembles of forecasts started from slightly perturbed initial conditions are generated. This thesis proposes novel methods to model ensembles statistically and convey variability aspects effectively, by analyzing the stability of critical points for scalar ensembles, and performing a comparative visual analysis of loc...
»