Influence of radiosonde observations on the sharpness and altitude of the midlatitude tropopause in the ECMWF IFS

• Published in: Weather and Climate Dynamics
• Main Authors: Krüger, Konstantin, Schäfler, Andreas, Weissmann, Martin, Craig, George C.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2024
• Subjects: article; Verlagsveröffentlichung; North Atlantic
• Online Access: https://doi.org/10.5194/wcd-5-491-2024; https://noa.gwlb.de/receive/cop_mods_00072714; https://noa.gwlb.de/servlets/MCRFileNodeServlet/cop_derivate_00070910/wcd-5-491-2024.pdf; https://wcd.copernicus.org/articles/5/491/2024/wcd-5-491-2024.pdf

Initial conditions of current numerical weather prediction systems insufficiently represent the sharp vertical gradients across the midlatitude tropopause. Data assimilation may provide a means to improve tropopause structure by correcting the erroneous background forecast towards the observations. In this paper, the influence of assimilating radiosonde observations on tropopause structure, i.e., the sharpness and altitude, is investigated in the ECMWF's Integrated Forecasting System. We evaluate 9729 midlatitude radiosondes launched during 1 month in autumn 2016. About 500 of these radiosondes, launched on request during the North Atlantic Waveguide Downstream Impact Experiment (NAWDEX) field campaign, are used to set up an observing system experiment (OSE) that comprises two assimilation forecast experiments, one run with and one without the non-operational soundings. The influence on the tropopause is assessed in a statistical, tropopause-relative evaluation of observation departures of temperature, static stability (N2), wind speed, and wind shear from the background forecast and the analysis. Temperature is overestimated by the background at the tropopause (warm bias, ∼ 1 K) and underestimated in the lower stratosphere (cold bias, −0.3 K) leading to an underestimation of the abrupt increase in N2 at the tropopause. The increments (differences in analysis and background) reduce these background biases and improve tropopause sharpness. Profiles with sharper tropopause exhibit stronger background biases but also an increased positive influence of the observations on temperature and N2 in the analysis. Wind speed is underestimated in the background, especially in the upper troposphere (∼ 1 m s−1), but the assimilation improves the wind profile. For the strongest winds the background bias is roughly halved. The positive influence on the analysis wind profile is associated with an improved vertical distribution of wind shear, particularly in the lower stratosphere. We furthermore detect a shift in the analysis ...