Ocean climate and biogeochemistry scenarios for the Baltic and North Seas
Abstract
The project Havsmiljöscenarier, that we here apply for additional computational resources for, started in 2019. It gathers much of the work SMHI's oceanographic research group is performing in the field of climate and marine environment scenarios that is not directly coupled to and financed by specific external projects. In 2021, a large portion of the work was directed towards the validation of our new biogeochemical model NEMO-Nordic-SCOBI covering all Swedish seas (adding Kattegat and Skagerrak compared to the older RCO-SCOBI model) and the North Sea. Forcing (e.g. winds, riverflows), was also produced for our upcoming climate scenario runs and the first stages of these scenario runs, the historical climate scenarios were performed during 2021. The scenario runs that will be produced by the project are so called dynamic downscalings of climate scenarios that have been run with coarse resolution global climate models. This downscaling step is necessary when looking at impacts of climate change in coastal waters. In general, coastal waters and even coastal oceans such as the North Sea are not well resolved by global models. In practise, dynamic downscaling means that a complicated chain of numerical models is set-up that uses output from low resolution global models as forcing for high resolution regional models. Some of our forcing is also produced using machine learning methods. One example is wind driven sea level variability at our regional model's open boundary in the North Sea.
For 2022 the main aim of the project is to continue performing climate scenario projections. A total of 27 projections will be done within the project: encompassing three global climate models, three emission scenarios for greenhouse gases and three scenarios for nutrient deposition into the sea. This 27 member ensamble is small in the sense that it downscales only about 1/10 of the availiable global models, however, computationally it is very costly. Our plan, which requires using nearly half of the groups computational resources at the cluster Bi for these simulations lets us finish roughly half of these 27 simulations during 2022. The aim of this application is to secure computational resources so that a part of the remaining simulations can be finished on Tetralith. The application is also a way for us to test the waters for a potential future application for a bigger allocation. Downscalings of climate scenarios is an ever growing enterprise, and it is already clear to us that our research group at SMHI cannot in reasonable time create these datasets using only our own hardware.
These runs will form the basis for a number of scientific publications describing climatic changes in the Baltic and the North Seas. They will also be integrated into SMHI's climate scenario service (https://www.smhi.se/klimat/framtidens-klimat/fordjupade-klimatscenarier/oce/-/temperature/rcp45/2071-2100/year/reference), where climate projections are made available to the general public. Many results from these integrations are also used by stakeholder to inform decision making.
