Abstract. The EU Copernicus Climate Change Service (C3S) European Climatic Energy Mixes (ECEM) has produced, in close collaboration with prospective users, a proof-of-concept climate service, or Demonstrator, designed to enable the energy industry and policy makers assess how well different energy supply mixes in Europe will meet demand, over different time horizons (from seasonal to long-term decadal planning), focusing on the role climate has on the mixes. The concept of C3S ECEM, its methodology and some results are presented here. The first part focuses on the construction of reference data sets for climate variables based on the ERA-Interim reanalysis. Subsequently, energy variables were created by transforming the bias-adjusted climate variables using a combination of statistical and physically-based models. A comprehensive set of measured energy supply and demand data was also collected, in order to assess the robustness of the conversion to energy variables. Climate and energy data have been produced both for the historical period (1979–2016) and for future projections (from 1981 to 2100, to also include a past reference period, but focusing on the 30 year period 2035–2065). The skill of current seasonal forecast systems for climate and energy variables has also been assessed. The C3S ECEM project was designed to provide ample opportunities for stakeholders to convey their needs and expectations, and assist in the development of a suitable Demonstrator. This is the tool that collects the output produced by C3S ECEM and presents it in a user-friendly and interactive format, and it therefore constitutes the essence of the C3S ECEM proof-of-concept climate service.

No abstract

While ubiquitous, textual sources of information such as company reports, social media posts, etc. are hardly included in prediction algorithms for time series, despite the relevant information they may contain. In this work, openly accessible daily weather reports from France and the United-Kingdom are leveraged to predict time series of national electricity consumption, average temperature and wind-speed with a single pipeline. Two methods of numerical representation of text are considered, namely traditional Term Frequency -Inverse Document Frequency (TF-IDF) as well as our own neural word embedding. Using exclusively text, we are able to predict the aforementioned time series with sufficient accuracy to be used to replace missing data. Furthermore the proposed word embeddings display geometric properties relating to the behavior of the time series and context similarity between words.

In this paper, we propose an algorithmic framework to automatically generate efficient deep neural networks and optimize their associated hyperparameters. The framework is based on evolving directed acyclic graphs (DAGs), defining a more flexible search space than the existing ones in the literature. It allows mixtures of different classical operations: convolutions, recurrences and dense layers, but also more newfangled operations such as self-attention. Based on this search space we propose neighbourhood and evolution search operators to optimize both the architecture and hyper-parameters of our networks. These search operators can be used with any metaheuristic capable of handling mixed search spaces. We tested our algorithmic framework with an evolutionary algorithm on a time series prediction benchmark. The results demonstrate that our framework was able to find models outperforming the established baseline on numerous datasets.

<p align="justify">The energy demand will potentially be affected by climate change in the future. The heating needs are expected to decrease while the cooling needs are expected to increase under projected future global warming. The question addressed in this work in the impact and the quantification of this changes on the temporal fragmentation of energy demand during winter and summer. This fragmentation creates a need for flexibility in energy production which constitutes a challenge for energy systems. In this work, the question is addressed by exploiting a biais-corrected and downscaled climate projections from CMIP6 simulations using a statistical method at a spatial resolution of 0.25° x 0.25° over Europe. Ten variables were used to estimate the main change on energy demand related to heating degree days (HDD) and cooling degree days (CDD) under four scenarios (ssp1-2.6, ssp2-4.6, ssp3-7.0 and ssp5-8.5). The results showed a large decrease of HDD over Europe and an increase of CDD under all of scenarios ssp considered in this work. The analyses of the heating and air conditioning period duration and frequency showed a fragmentation of the periods of use of heating during winter in the future which can lead potentially to a fragmentation of energy demand related to heating. On the other hand, the periods of use of air conditioning in summer are expected to be more frequent and longer but still very fragmented in time compared to the present climate.</p>