Impacts of aerosol-cloud-radiation interactions on photovoltaic generation: case of Saharan dust outbreaks in March 2021

Aerosols interact with radiation and clouds and thereby disturb radiative budget and temperature structure in the atmosphere. To account for these effects, numerical weather prediction models rely on climatological mean concentrations. This simplification may lead to large errors in the forecasted cloud cover and radiative fluxes especially during major aerosol events. For example, Saharan dust events often coincide with significant errors in shortwave radiation and thus, day-ahead photovoltaic forecasts in Europe. In this study we investigate errors in the short-range forecasts during Saharan dust outbreaks in March 2021, analyze possible causes and explore the solutions. We use the data from pre-operational forecasts performed with the ICOsahedral Nonhydrostatic model with Aerosols and Reactive Trace gases (ICON-ART) based on two experiments: without dust effects and with direct dust effect only. We compare model data with the measurements from satellite and in-situ instruments. The results reveal that the inclusion of direct radiative effects from prognostic dust improves the forecast in surface radiation during clear-sky conditions. However, dusty Cirrus clouds are strongly underestimated, pointing to the importance of representing indirect effects. To fill this gap, we develop and test corresponding sub-grid parameterization for dusty Cirrus in the ICON-ART model. Only with help of this parameterization ICON-ART is able to simulate the formation of the dusty cirrus, which leads to substantial improvements in cloud cover and radiative fluxes compared to simulations without this parameterization. This study confirms that a reliable photovoltaic forecast requires explicit treatment of aerosol-cloud-radiation in numerical weather forecast systems.