Forecasting convective precipitation over northern Italy: A comparison of lightning and GNSS-ZTD data assimilation

Improving the forecast of intense rain events is very important to save lives and properties. In this paper we show the impact of lightning and Global Navigation Satellite System – Zenith Total Delay (GNSS-ZTD) data assimilation on the prediction of intense convective rain events in Northern Italy at the short-range (up to 6 h). A total of 116 cases, from April to September 2019, are considered, based on radar observations over the Lombardy Region. Lightning is assimilated by nudging, while GNSS-ZTD by 3DVar. The Weather and Research Forecast (WRF) model is used for predictions. Results show that lightning data assimilation and GNSS-ZTD improve the rain forecast. A statistical test shows that the improvement is significant for several thresholds and time ranges. Assimilating both lightning and GNSS-ZTD has a larger positive impact than assimilating one data source alone and, in addition to statistics, few case studies are shown to focus on this point. The forecast performance slightly decreases with time, while the impact of data assimilation decreases faster. Finally, a sensitivity test, aimed at reducing false alarms, shows contradictory results.