Useful local climate information for Germany (NUKLEUS) - Hybrid deterministic-stochastic methods and Downscaling for Climate Services
The joint project NUKLEUS represents the cross-sectional activity of RegIKlim (focus B) and deals with the provision of high-resolution climate information that is required for the Model Areas in Germany. In phase 1 of the project, the prototype of an unprecedented ensemble of regional climate model simulations will be developed, and complemented by statistical and statistical-dynamic downscaling methods towards a better estimation of the climate information range. NUKLEUS will provide the local climate information for the Model Areas via the FREVA portal. NUKLEUS is divided into the following work packages:
- Expansion of the local climate information database with high-resolution climate simulations for Germany;
- Evaluation and analysis of the reliability of local climate information;
- Establishment and optimization of interfaces to impact models together with the partners from the Model Areas.
The focus of the sub-project of the Justus Liebig University (JLU) lies within WP3 and consists in the creation and establishment of interfaces to both internal (focus B "Regional Climate Modelling") as well as external project partners (focus C "overall coordination" and focus A "Model Areas"). The interfaces between climate information and the applications in the Model Areas (urban climate models, impact models and integrated models) are defined and implemented with the model regions in WP3.2 (Definition and Implementation of Interfaces to the Model Areas). The need for diagnostics, but also other types of interfaces (e.g. further localization of the data, time scales, statistical key figures) should be developed in dialogue with the Model Areas from focus A. Thereby, sectors, events and the intended use of the climate data should be provided by the Model Areas. The transfer/links of this information to physical processes, meteorological quantities and the model-inherent spatio-temporal reference will be undertaken by the JLU in WP3.2.
The goal of WP3 is the development of novel postprocessing and bias adjustment methods to make climate model data consistent in the multivariate sense, and to correct systematic errors in the ensemble of regional climate model data. The sub-project “Hybrid-deterministic-stochastic methods and Downscaling for Climate Services” contributes mainly to the tasks WP3.2 of WP3 - Interfaces of the full proposal for the collaborative project NUKLEUS. The planned research work is divided into two central activity areas that form a clear line of action and are strongly guided by the needs and questions set in the Model Areas (see Fig. 1).
For applications of the “Model Areas”, which require a higher spatial and temporal scale than the model data offered by WP1, the needed simulations are statistically post-processed with a hybrid deterministic-stochastic method. First, a deterministic-bias correction method is applied, which is intensity dependent and therefore particularly well suited for weather and climate extremes. In the second step a stochastic weather generator (Rglimclim) - based on a generalized linear model - comes into play to scale down the data to a 1 km grid for Germany and individual weather stations. Postprocessing and bias adjustment procedures focused on specific tasks and climate processes will be integrated into FREVA and will also be used in WP1.4 and WP3 of the NUKLEUS project.
Thus, the research activities planned here form an important methodological and data-based bridge between the work packages WP1 - Modelling and WP2 - Evaluation and Assessment of the project. In accordance with these superordinate tasks, the research activities in this sub-project are embedded in the three NUKLEUS work packages (Fig. 1).
- Cooperation Partners
- Karlsruhe Institute of Technology (KIT)
- Helmholtz-Zentrum Geesthacht (HZG)
- Deutsches Klimarechenzentrum GmbH (DKRZ)
- Universität Würzburg (UW)
- Brandenburgische Technische Universität (BTU)
- Technische Universität Dresden (TUD)
- Exploitation plan
The research results and data products from NUKLEUS form the basis for services and products for the Model Areas from focus A and they will play also an important role for other research projects up to international synthesis reports like IPCC. To the national stakeholders are included research projects such as “ClimXtreme” and “Stadtklima im Wandel/Urban Climate under Change”, which research questions can benefit from the high-resolution climate data. At international level, the results will be used to support projects such as the EUCP and the WCRP CORDEX FPS on Convection Initiative with complementary data. In addition to the data products, the results and analytical methods will be published in scientific journals and presented at conferences.
The research results, data sets and methods developed in this sub-project will be used within and outside the RegIKlim project as well as beyond its duration and objectives. On the one hand, the bias-adjusted climate model data is the central interface between the foci A and B of RegIKlim, because in the Model Areas as much error-free and consistent climate information is required. On the other hand, the methodological approaches for compiling a spatio-temporal high-resolution multivariate validation data set and for multivariate bias adjustment of climate model data will serve as examples for research in the interplay between climate modelling on the convection-resolving scale and the various facets of climate impact research and modelling.
Methods and analyses, which are entered as plugins in FREVA in the first phase of NUKLEUS, are prepared for a potential second phase of RegIKlim. The creation of every software introduced to FREVA will use OpenSource licenses and will allow FREVA, including plugins to be available to the community for use and implementation after the end of the project.
This project is generously funded as part of the FONA framework program of the Federal Ministry of Education and Research (BMBF).