Nowcasting of Yes/No rain situations at a station using soft computing technique to the radar imagery

Abstract

A soft computing model for nowcasting of Yes/No rain situations, with a lead time of 2 h, is developed over DWR station at Satish Dhawan Space Centre (SDSD), Shriharikota (13.66°N, 80.23°E) using Doppler weather radar (DWR) reflectivity imageries. Primarily, precipitating systems of mesoscale, i.e. meso-gamma (2-20 km), meso-beta (20-200 km) and meso-alpha (200-2000 km) are considered for the present study. The main components of soft computing approach are: analysis of two-dimensional reflectivity imageries from DWR and utilization of artificial neural network (ANN) for training of input/output data. About 15 input and one output parameters are extracted from radar imageries. The image analysis and training of ANN are carried out on MATLAB platform. After training of ANN, 91 and 77% results are matched with the observed values for No rain and Yes rain situations, respectively. The probability of detection (POD) for the nowcasting of Yes/No rain situations is found to be 0.84. The significant improvement in the nowcasting of Yes/No rain situations is observed by the developed methodology as compared to linear multivariable regression method. The POD of Yes/No rain for other two locations, namely Chennai (12.99°N, 80.18°E) and Tiruvallur (13.09°N, 79.57°E) are found to be 0.81 and 0.77, respectively. Overall, reasonably good results are obtained by the newly developed soft computing model.