A Novel Machine Learning Based Bias Correction Method and Its Application to Sea Level in an Ensemble of Downscaled Climate Projections

A new machine learning based bias correction method is presented and applied to sea level in a regional climate model. The corrections derived using this depend on the state of model it corrects. This contrasts with conventional methods that operate distributions output variables. dependence states allows for better performance classical skill scores, but also limits applicability models can perform hindcasts. very large dataset corrected hourly levels from many different emission scenarios created. In total contains over 2600 years exists seven tide-gauge stations Swedish Baltic Sea coast. prevalence significant trends yearly maximum found be independent scenario, suggesting anthropogenic change no driver storm surge variability area. Lastly, used estimate return long periods, block length computation affect result at some stations. suggests commonly annual approach not always applicable determining level.