This layer shows a weather radar-based nowcasting product to highlight areas (shown as ellipses) where intense precipitation may cause flash flood impacts.
|
|
These products combine cell-based storm nowcasts (Rossi et al. 2015) with a machine learning (ML) model to predict storm hazard levels. The hazard levels are storm-estimated based on historical meteorological observations and weather impact reports following the concepts described by Rossi et al. (2013) and Tervo et al. (2019).
The storm impact layers combine meteorological data from various sources such as weather radar, Numerical Weather Prediction (NWP) models and ERA5 re-analyses. The weather impact reports are obtained from the European Severe Weather Database (ESWD). Probabilistic nowcasts for the future location of the classified thunderstorms for the coming 5-60 minutes are being produced by using a Kalman filter model (Rossi et al. 2015).
The product generation process is illustrated in Figure 1. Storm cells are identified and tracked from OPERA reflectivity composites (step 1). Hazard levels of the cells are then predicted with the ML model for the four different hazard types (heavy rainfall, wind gusts, hail and lightning). This is done by using meteorological information from additional data sources, such as ERA5 reanalyses (step 2). Ellipses are fitted to the storm cells, and probabilistic nowcasts of storm ellipse locations are produced by applying the Kalman filter-based model (step 2). The predicted hazard levels are combined with the storm locations, which gives the hazard nowcasts (step 3). Here we assume that the hazard levels remain constant during the forecast time range. The nowcast giving the total hazard level is generated by taking the maximum levels from each hazard type. Finally, the maximum hazard nowcast is combined with an exposure layer to produce an impact nowcast (step 4). This is done by using a single exposure data source combined with openly available datasets of population, transport and critical infrastructure (e.g. health, education and energy).
Figure 1. Different stages of generating the multi-hazard product for thunderstorms.
Figure 2 below shows examples of storm hazard and impact nowcasts. The hazard classes have colours white (no data), grey (no hazard), yellow (moderate hazard), orange (severe hazard) and red (extreme hazard). The three nonzero classes correspond to 10%, 50% and 75% percentiles of storm hazard levels in the pan-European training dataset consisting of a five-year period. The products shown in this example are based on the maximum level of each hazard type (heavy rainfall, wind gusts, hail and lightning). In addition, the uncertainty in the storm location is represented by the increasing transparency and spread of the ellipses as the forecast lead time is increased.
Figure 2. Storm hazard nowcasts (coloured ellipses) during the 27th August 2023 flood event in Helsinki. – to be replaced by EFAS screenshot
References
Rossi, P.J., Hasu, V., Halmevaara, K., Makela, A., Koistinen, J., Pohjola, H. 2013. Real-time hazard approximation of long-lasting convective storms using emergency data. Journal of Atmospheric and Oceanic Technology. 30, pp 538-555. https://doi.org/10.1175/JTECH-D-11-00106.1
Rossi, P. J., V. Chandrasekar, V. Hasu, and D. Moisseev, 2015: Kalman Filtering–Based Probabilistic Nowcasting of Object-Oriented Tracked Convective Storms. J. Atmos. Oceanic Technol., 32, 461–477, https://doi.org/10.1175/JTECH-D-14-00184.1.