In the next upgrade of ECMWF’s Integrated Forecasting System (IFS), humidity in the stratosphere will be analysed based on observations to improve the starting conditions of forecasts.
This procedure, which will use sonde humidity observations, has been found to improve forecasts of a range of parameters. An earlier attempt to analyse humidity in the stratosphere in 1999 resulted in systematic forecast errors.
The upgrade from IFS Cycle 49r1 to 50r1, which includes many other changes, is scheduled to take place in November 2025. The work described here is part of the Copernicus Atmosphere Monitoring Service (CAMS) EvOlution project (CAMEO) funded by the EU.
Analysing humidity in the stratosphere
The ‘analysis’ is a general name for the initial conditions on which weather forecasts are based. An analysis is obtained by combining the latest observations with a short-range forecast constrained by previous observations.
In 1999, a first attempt was made to use observations to analyse humidity in the stratosphere, but it was found to lead to systematic forecast errors and was abandoned in the following year.
Twenty-five years later, however, changes to the data assimilation system, which establishes the analysis, have made it possible to analyse humidity in the stratosphere using observations.
Those changes include the introduction of ensemble-based techniques for uncertainty estimation through the Ensemble of Data Assimilations (EDA) and progressive improvements in the fidelity of the IFS model.
In a stratospheric humidity analysis experiment, we assimilate an additional source of stratospheric humidity information by extending the vertical usage of humidity observations from RS41-type sondes up to 60 hPa. These sondes are recognised for accurately measuring humidity across a wide range of atmospheric conditions.
The chart below shows the change in mean relative humidity and mean temperature this brought about in forecasts of different lengths.
Change in the forecast mean relative humidity (left) and the forecast mean temperature (right) in a stratospheric humidity analysis experiment compared to the control over three months of experimentation (1 December 2022 to 28 February 2023) at forecast times of 96 and 192 hours. Areas marked with crosshatching are statistically significant at the 95% level.
The primary direct effect of assimilation is the significant drying of the lower stratosphere, as illustrated on the left-hand side of the figure, alongside a slight moistening below the tropopause. Both effects reduce known systematic deficiencies of the IFS.
A secondary effect is observed on the right-hand side of the figure: due to radiative effects, the temperature of the lower stratosphere becomes warmer, while the area under the tropopause cools, with this effect growing as the forecast time increases. Again, these changes significantly reduce known IFS analysis and forecast biases.
Better forecasts
The effects on the quality of forecasts can be seen in charts showing the normalised change in the root-mean-square error (RMSE) of forecasts.
The figure below shows this for relative humidity (left) and temperature (right) from the stratospheric humidity analysis experiment, compared to the control and verified against its own analysis.
Normalised change in root-mean-square error (RMSE), in the stratospheric humidity analysis experiment compared to the control, of relative humidity forecasts (left) and temperature forecasts (right), verified against own analysis, over three months of experimentation (1 December 2022 to 28 February 2023), at forecast times of 24 and 216 hours. Areas marked with cross-hatching are statistically significant at the 95% level. Blue cross-hatched areas indicate a significant reduction in forecast error and hence an improvement in forecast quality.
Reintroducing the analysis of stratospheric humidity improves forecasts across all lead times. The enhancements are more pronounced in the Upper Troposphere Lower Stratosphere (UTLS) layer, but there are indications that they propagate into the troposphere with increasing lead time.
Outlook
Humidity, like other trace gases, functions as a long-lived tracer in the stratosphere. Enhancing the accuracy of initial conditions for trace gases in the stratosphere can substantially improve forecast skill in the medium and extended range. Current efforts aim to unlock this potential in future IFS operational cycles.
More details on the analysis of humidity in the stratosphere can be found in an ECMWF Newsletter article published this spring.
The CAMEO project (grant agreement No 101082125) is funded by the European Union. Views and opinions expressed are, however, those of the authors only and do not necessarily reflect those of the European Union or the Commission. Neither the European Union nor the granting authority can be held responsible for them.