Assessment of global upper-air wind speed: a comparison between radiosonde observations and reanalysis, 1979-2023

Upper-air wind speed (UWS) is a crucial variable in climate change assessments and is also necessary in various socioeconomic areas such as evaluating wind energy production at the turbine height and optimizing commercial aviation routes. While reanalysis datasets have been widely utilized as essential tools for climate change analysis due to their comprehensive spatial coverage and temporal continuity, their performance in representing global UWS remains uncertain. This study evaluates four major reanalysis datasets (i.e., ERA5, ERA-Interim, MERRA-2, and JRA-55) by assessing their performance in capturing the spatio-temporal characteristics of global UWS, including climatological mean, variability, and linear trends seasonally and annually. The assessment is conducted through comparisons with radiosonde observations from two datasets: the Integrated Global Radiosounding Archive version 2 (IGRA-v2) and the homogenized Radiosounding HARMonization (RHARM) dataset distributed by the Copernicus Climate Change Service (C3S) for 1979-2023. The radiosonde observations reveal that UWS exhibits distinct vertical and zonal patterns. In the lower and middle troposphere, UWS generally remains below 10 m s⁻¹ with relatively weak non-significant trends in most stations. In contrast, the upper troposphere and lower stratosphere show pronounced zonal patterns with values up to 45 m s⁻¹, accompanied by significant increasing trends reaching 3.412 m s⁻¹ per decade. These zonal patterns demonstrate clear seasonal variations, appearing more linear in winter and wave-like in summer. While all four reanalysis datasets successfully capture the climatological patterns and seasonal variations of UWS, they show varying degrees of biases in trend estimation. These comparisons provide valuable insights for understanding UWS characteristics.