Predicting the direct and indirect effects of atmospheric aerosol on photovoltaic power generation

<p>For an efficient integration of photovoltaic (PV) energy into the power grids, more accurate forecasts of the expected PV-power production are needed. However, most operational numerical weather prediction models rely on an aerosol climatology and ignore the spatio-temporal variability of the atmospheric aerosol. For specific weather conditions like during mineral dust outbreaks or major wildfire events, the negligence of prognostic aerosol often leads to significant deficiencies in the operational forecasts, however.</p><p>At Deutscher Wetter&#173;dienst (DWD) and Karlsruhe Insti&#173;tute of Technology (KIT) the project &#8220;PermaStrom&#8221; aims at the operational prediction of various natural aerosol species to improve radiation forecasts. Emission, transport and deposition of mineral dust, black carbon from vegetation fires, and sea salt are thus explicitly simulated in the ICON-ART model system. In the model, direct aerosol effects on radiation are considered using state-of-the-art optical properties. Microphysical effects of aerosol acting as cloud condensation nuclei (CCN) and ice nucleating particles (INPs) are investigated in a high-resolution regional model with the long-term goal to improve the parameterization of aerosol-cloud effects in global models.</p><p>Aerosol-cloud-radiation effects are studied in a regional ICON-ART model with 2 km grid spacing with an aerosol-aware two-moment bulk microphysics scheme. In addition, first steps are made towards a global ensemble system for aerosol forecasts using ICON-ART. This will allow to quantify the uncertainty of the forecasts. A multi-fidelity ensemble, which combines ICON-ART and ICON simulations to optimally sample the aerosol- and flow-dependent variability, is used to keep the computational processing manageable. The ICON-ART simulations are validated with aerosol and radiation measurements at surface stations as well as cloud, aerosol and radiation products from satellites and ceilometers.</p><p>We will give an overview of the ICON-ART configuration of the pre-operational real-time global aerosol prediction system at DWD. This includes aspects like mineral dust, sea salt, and wildfire emissions. For the latter, a machine learning emulator of the plume rise model is currently being developed.</p>