Posters

Abstract

The design of offshore wind platforms requires detailed knowledge of thewind and waveclimateat different spatial (regional, local, high resolution) and temporal (daily, intraannual and interannual variability, climate change projections) scales. Generally, this is based on dynamic downscaling using a series of nested atmospheric and oceanographic numerical models that require large amounts of computing resources. To reduce these resources, weather-based climate emulators have emerged. These emulators are statistical models capable of emulating meteo-ocean parameters by identifying their relationship to a set of synoptic patterns representing atmospheric pressure variability. The use of weather types allows the analysis of the historical climate variability of the relationship between the atmosphere and the different variables that influence the design of wind platforms (e.g. waves, wind, currents) and provides information on the spatial and temporal variability in the region. This provides information on the potential risks associated with extreme weather events, such as storms and hurricanes, which are relevant to design and can damage offshore infrastructures during their lifetime. To this end, TESLA (Anderson et al 2021), the emulator used in this work, allows the generation of sets of future projections and the quantification of the associated uncertainty with climate model simulations, using the outputs of climate models under different greenhouse gas emissions scenarios. These projections offer valuable information on the potential risks associated with climate change that affect the deployment and operation of offshore wind farms, which is critical for policymakers and energy analysts seeking to maximize the potential of offshore wind energy in the face of climate change over the next few decades. TESLA is a hybrid modelling framework that merges statistical and dynamic models and is offered to the scientific community through the GeoOean group's GitLab repository as open source software. Developed in Python, it has been designed to run on Jupyter Notebooks environments. This environment not only ensures reproducibility of experimental results, but also facilitates collaboration between researchers, as these Notebooks can be easily shared. TESLA optimises the use of computational resources in a variety of infrastructures, ranging from a PC to a compute cluster, including cloud infrastructures. Easy installation and deployment on different types of infrastructure makes it easy to conduct experiments without the need for advanced computer skills. We will present two different applications of TESLA: a simple example aimed at those approaching the framework for the first time and wishing to learn how to use it, and a more complex experiment focused on the North Sea area. This experiment will be carried out using the GeoOcean group's computational cluster and will demonstrate how TESLA facilitates the transparent execution demanding experiments, both from a computational and data access perspective, through a web interface. REFERENCES Anderson, D.L., Ruggiero, P., Mendez, F.J., Barnard, P.L., Erikson, E.H., O'Neill, A.C., Merrifield, M., Rueda, A., Cagigal, L., Marra, J. (2021) Projecting Climate Dependent Coastal Flood Risk with a Hybrid Statistical Dynamic Model, Earth's Future, doi: 10.1029/2021EF002285