Do Saharan dust outbreaks affect weather forecast skill in Europe?

Mineral dust is a prominent aerosol in the atmosphere. Dust aerosol directly interacts with radiation, altering the radiative budget and temperature structure in the atmosphere. Dust aerosol also affects clouds by acting as cloud condensation nuclei and ice nucleating particles. Models currently deployed for operational weather forecasting simplify dust by using climatologies for representing dust concentrations, hence misrepresenting dust effects during conditions which differ greatly from the climatology. Outbreaks of Saharan dust towards Europe occur several times a year and lead to elevated mineral dust concentrations over Europe. Such dust events often coincide with errors in operational weather forecasts.
In this thesis we systematically investigate errors in model analysis and short term forecasts during
events of Saharan dust over Central Europe. We perform a case study for a dust event in spring
2021 and investigate the cause of model errors. For this we assess models which are currently
used for operational weather forecasting as well as models with prognostic calculation of dust,
and compare model analysis and forecast data with the observations from satellite and in-situ
... mehr
measurements. Furthermore, we construct an event catalog with 38 dust days between 2018 and
mid 2021 and analyze the typical synoptic situation leading to dust events over Central Europe. We subsequently quantify the median model errors in cloudiness and surface radiation from a model used for operational weather forecasting by comparison against satellite products. For this we use brightness temperature as proxy for cloudiness and surface radiation as a measure for error propagation into further meteorological quantities.
We find misrepresentation of mineral dust the source of model errors during the event in spring 2021.
The inclusion of prognostic aerosol and direct radiative effects from dust improves the forecast in
surface radiation during clear-sky conditions, but is not sufficient for reproducing cloudiness during the dust events. We find two typical synoptic situations leading to Saharan dust over central Europe.
Most frequently, an omega situation with the high pressure center over Central Europe, and a
situation with a trough over the Iberian peninsula, northwest Africa, and a weakly pronounced ridge over Central Europe. We find weak indications for a longer transport path of dust towards Central Europe for the former, and a more direct transport for the latter synoptic situation. For dust days in the event catalog we find a significant overestimation of brightness temperature (underestimation of dust direct radiative effects, cloud top height, cloud optical depth) for cases with dust compared to cases without dust. For surface radiation we find a median overestimation of about 15% during cloudy conditions with dust compared to cloudy conditions without dust. For clear-sky conditions we do not find reliable differences, as the satellite retrieval is biased during such conditions.
Our findings show that the models deployed for operational weather forecasting do not adequately reproduce cloudiness during events with Saharan dust over Central Europe. The missing implementation of prognostic dust, in particular indirect effects on cloud formation, still leads to significant underestimations of cloudiness and overestimations of surface radiation during dust events. Our study therefore highlights the need of including such effects into future forecast systems.