Assessing the potential of assimilating wind turbine power output and smartphone data for future reanalysis

Improving the representation of lower boundary layer winds and temperatures in weather forecasting is still a major task, despite the increase in the spatial resolution. As applications strongly rely on high-quality data from weather and climate models, systematic errors and biases in relevant areas, e.g., boundary layer wind speed or near-surface temperatures, degrade the quality of such applications. In this regard, new observing systems can provide information to enhance the model state representation through data assimilation. Smartphones are widely used and are already equipped with build-in sensors, e.g., for air pressure, thus providing a potentially valuable source of meteorological information. We further see wind power data obtained as wind turbine power output as potential implicit measurements for boundary layer wind speeds. Through the FAIR project, examplary data from smartphopnes is provided by the University Duisburg-Essen and for wind power data from the industry partners BayWa r.e. In this study, we aim to assimilate these observations into the operational NWP model ICON-LAM to assess the potential benefit with respect to NWP and future regional reanalyses. A new forward operator was developed for the LETKF scheme to assimilate wind power output. To investigate the benefit of assimilating wind power and air pressure data from smartphones tuning experiments have been performed. The sensitivity of various parameter settings, such as horizontal localization scale, and appropriate observation error estimation using the Desroziers metric is tested. The simulation results are evaluated and compared against independent observations for different synoptical situations and they indicate a potential for improvement.