Near-Surface Temperature Biases in Regional Climate Models over Complex Orography: A Physically Grounded Diagnosis

Regional climate models (RCMs) remain affected by limited spatial resolution and systematic biases relative to observations, posing a major challenge for the provision of actionable climate information at the local scale. Bridging the gap between coarse, biased model outputs and site-specific climate needs is therefore both urgent and an active area of research. Here, we present a diagnosis of near-surface air temperature biases, based on the vertical structure of the atmospheric column, in an ensemble of EURO-CORDEX simulations over a complex Alpine region. We show how the conventional and commonly applied temperature bias correction, aimed at removing the discrepancy between the model representation of orography and the actual terrain elevation, still leaves biases of a magnitude comparable to or even greater than the applied correction. After the orographic correction, residual biases originate from both free-atmosphere temperature biases and low-level biases. Each contribution to the TAS bias is quantified through an inter-model seasonal analysis. This study highlights the limitations of RCMs in representing complex orography and boundary-layer conditions in Alpine regions. It also provides interpretative keys for a better understanding of the underlying site-specific physical causes of TAS biases and cautions against the potential pitfalls of a straightforward application of bias-correction methods.