Double-Tropopause Events over the Mediterranean Region: Climatology, Trends and Dynamical Context from ERA5

Abstract

Double tropopauses are recurrent features of the extratropical upper-troposphere–lower-stratosphere (UTLS), but their climatology and long-term evolution over the Mediterranean region remain poorly characterized. Here, ERA5 pressure-level data for 1979–2025, complemented by ERA5.1 for 2000–2006, are used to investigate the seasonal distribution, vertical structure, trends, and geographical variability of double-tropopause events over the Mediterranean sector. Lapse-rate tropopauses are identified on vertically interpolated pressure-level profiles using a calibrated recovery threshold of 2.5 K km−1, evaluated against ERA5 model-level profiles retaining the standard 3.0 K km−1 WMO recovery threshold. The principal analysis focuses on double-tropopause profiles with the first lapse-rate tropopause (LRT1) at or above 8 km (DT–UTLS). DT–UTLS occurrence shows a marked seasonal structure, with winter maxima over the southern part of the domain and a pronounced northward displacement of the climatological distribution in summer. The detected events exhibit a layered vertical structure, with a lower and more variable LRT1 and a second lapse-rate tropopause (LRT2) several kilometres higher. Long-term changes are spatially heterogeneous, but the clearest seasonal frequency signal is a positive trend in March–May (MAM), embedded in an overall positive annual evolution of DT–UTLS occurrence. Positive MAM trends are found across all predefined Mediterranean subregions, and their sign remains stable across the tested thermal-recovery thresholds. The positive springtime tendency persists across the tested recovery thresholds, while statistical significance depends on the diagnostic and multiple-testing procedure. Detrended regional associations show that enhanced MAM DT–UTLS occurrence is most consistently associated with changes in the relative vertical configuration of the thermal and dynamical tropopauses, particularly a reduced LRT1–dynamical-tropopause separation and an increased LRT2–LRT1 separation, whereas associations with seasonal strong-jet occurrence are comparatively weak. Individual DT–UTLS events preferentially occur in dynamically active and more cyclonic upper-level environments, but the corresponding seasonal-mean diagnostics do not provide a simple explanation for the long-term MAM increase. No robust evidence is found for a systematic long-term geographical displacement of the DT–UTLS regime. These results identify Mediterranean DT–UTLS occurrence as a useful indicator of the seasonally evolving UTLS transition zone and highlight spring as the season exhibiting the clearest long-term change.