A New Method for Estimating Atmospheric Turbulence from Global High-Resolution Radiosonde Data and Comparison with the Thorpe Method

Abstract. This study proposes a new method for estimating atmospheric turbulence from high vertical-resolution radiosonde data (HVRRD) using the minimum Richardson number (Ri_min). While previous studies using HVRRD have primarily been based on the Thorpe method, which detects turbulence only in regions of local potential temperature overturning (Ri < 0) and does not explicitly account for wind shear, the proposed approach overcomes these limitations. By incorporating the effects of gravity waves on the static stability and vertical wind shear, this method enables the detection of turbulence not only in regions of Ri < 0 but also within statically stable layers characterized by strong shear (0 < Ri < 0.25), where Kelvin–Helmholtz instability is likely to occur. Additionally, comparison with turbulence observations from commercial flights demonstrates that the time series of turbulence derived from the Ri_min method exhibits a significantly higher positive correlation with flight observations than that derived from the Thorpe method. Utilizing 10 years of global operational HVRRD, this study further analyzed the climatological distributions of turbulence derived from the Ri_min method. Results show that turbulence under positive Ri conditions occurs most frequently in winter and less frequently in summer, reflecting the seasonal variability of the jet stream. In contrast, negative Ri cases exhibit a summertime maximum and wintertime minimum in the troposphere, and the opposite seasonal variation in the stratosphere. Regionally, turbulence is most pronounced over Asia, South America, and Antarctica for both positive and negative Ri cases. We upload the datasets produced from the current work at publicly available sites: https://doi.org/10.5281/zenodo.16899801 (Ko and Chun, 2025a), https://doi.org/10.5281/zenodo.16899803 (Ko and Chun, 2025b), https://doi.org/10.5281/zenodo.16899805 (Ko and Chun, 2025c), https://doi.org/10.5281/zenodo.16810246 (Ko and Chun, 2025d), and https://doi.org/10.5281/zenodo.16899789 (Ko and Chun, 2025e).