REtuning Climate Model Outputs (RECMO method) at regional and local level in the Carpathian Region, Phase II: climate indices

Climate model simulations often differ from observational data, resulting in different projections of future temperature characteristics for a given geographical region. While the magnitude of relative climate changes tends to be consistent across different models, the absolute temperature characteristics can show substantial variation when evaluated against observations. Global Climate Models (GCMs) provide valuable insights into climate change on a global scale at predefined Warming Levels (WLs). A WL is defined as a specific higher temperature threshold relative to a designated reference period or observational baseline. One of the main sources of uncertainty in WL assessment is the timing of when the threshold is reached in relation to the reference period (e.g. 1976–2005). Moving from a global to a regional scale requires downscaling from the coarser resolution typically found in GCMs (approximately 100-150 km) to the finer resolution found in RCMs (approximately 10 km). Consequently, the timing of reaching a specific WL can be accurately assessed at regional and local scales by using high-resolution RCM simulations and the corresponding high-resolution observational data for the region of interest. The REtuning Climate Model Outputs (RECMO) method is introduced as a strategy to mitigate discrepancies among various RCM simulations. This method targets the reduction of uncertainties arising from the divergent climatic baselines described by different models across various WLs. The reference for the WLs is based on observations, and not on the raw model outputs. Our study focuses on the expected changes under different WLs, namely: 1.5 °C, 2 °C and 3 °C. The RECMO methodology is applied here to seven high-resolution raw and bias-corrected EURO-CORDEX and Med-CORDEX outputs for the Carpathian Region. The major towns and cities of the region (including four capitals) are involved in the research as follows: Budapest and Debrecen (Hungary), Bratislava and Kosice (Slovakia), Uzhhorod (Ukraine), Bucharest and Cluj (Romania), Beograd (Serbia). Present research consists of climate indices (e.g. tropical night, summer day, consecutive dry days) based on the following daily meteorological variables: precipitation, minimum and maximum temperature, mean temperature. Our results show that the timing of certain expected changes in these climate indices can differ by up to a decade, depending on whether the computation is based on raw or bias-corrected data. It is clear that the temporal aspect is a crucial factor in preparing for expected changes and developing adaptation strategies. Our findings also highlight the importance of bias-corrected RCM data and reliable high-resolution observational data in the field of climate science.