| Summary: | A transition to renewable energy is needed to mitigate climate change. This transition has been led by wind energy, and it is expected to continue to be the largest source of renewable energy through to 2030 (Sawyer et al., 2017). Both energy demand and production are sensitive to meteorological conditions and atmospheric variability at multiple time scales. To accomplish the required balance between these two variables, critical conditions of high demand and low wind energy supply must be considered in the design of energy systems. The aim of this thesis is twofold. Firstly, investigate the impacts of large-scale weather regimes on cold and weak-wind events during the extended boreal winter season (NDJFM). Secondly, to establish a methodology for modeling the joint distributions without making any assumptions about the marginal distributions. The analysis of 38 years of hourly high-resolution ERA5 reanalysis data proves that weather regimes are important predictors for both low temperature and low wind speed events over Europe. Blocking conditions, such as those observed during the negative phase of the North Atlantic Oscillation and the Scandinavian Blocking, are associated with cold and weak wind events. Compound events are observed more than 10% of the days overlarge geographical areas during blocking conditions. Nevertheless, high probabilities are also observed during AR, and to some extent, during the positive phase of the North Atlantic Oscillation. Dependency between cold events and weak wind events is proved to be statistically significant. The correlations between the events are higher when computed for each month separately compared to the entire winter season,revealing a strong seasonality. The highest correlations values are associated with the negative phase of the North Atlantic Oscillation, ρ=0.84, but values as high as 0.7 are registered for all the regimes. A methodology for modeling the bivariate joint distributions of low temperature and low-wind speed events is described. In this context, the concept of Gaussian copulas is used to mathematically model the correlated nature among them. The marginal distributions are modeled with logistic regressions defining two sets of binary variables for the weather regimes and months predictors.
|
|---|