Abstract
Sudden precipitations may bring troubles or even huge harm to people's daily lives. Hence a timely and accurate precipitation nowcasting is expected to be an indispensable part of our modern life. Traditionally, the rainfall intensity estimation from weather radar is based on the relationship between radar reflectivity factor (Z) and rainfall rate (R), which is typically estimated by location-dependent experiential formula and arguably uncertain. Therefore, in this paper, we propose a deep learning-based method to model the ZR relation. To evaluate, we conducted our experiment with the Shenzhen precipitation dataset. We proposed a combined method of deep learning and the ZR relationship, and compared it with a traditional ZR equation, a ZR equation with its parameters estimated by the least square method, and a pure deep learning model. The experimental results show that our combined model performs much better than the equation-based ZR formula and has the similar performance with a pure deep learning nowcasting model, both for all level precipitation and heavy ones only.
Highlights
Nowcasting has always played an important role in the field of weather forecast
Scientists have been beginning to deploy deep neural networks [10,11,13,14] to deal with the spatio-temporal inputs of precipitation nowcasting, such as Long Short-Term Memory (LSTM) [13,14], Gated Recurrent Unit (GRU) [15], 3-Dimensional Convolutional Neural Network (3D-CNN) [10], and the recent Transformer model which is based on the attention mechanism [11], were widely used in the applications
It can be found that the pure deep learning model performs the best, while the deep learning + ZR combined model gives an RMSE 18.01 which is very close to the best result (17.47), and is much better than those of the traditional ZR relationship Z = 200R1.6 and the ZR relation based on the least square estimation, showing that deep learning does extract more accurate features than restricting the ZR relationship into the simple equation
Summary
Nowcasting has always played an important role in the field of weather forecast. Whilst it works in predicting phenomenons such as lightning [1], Hailstorm [2], convective storm [3,4,5], straight-line convective wind [6] and tropical cyclone [7], most of the nowcasting efforts are applied to forecasting precipitation. There were quite a few efforts proposed to improve the ZR relation, for instance, [12,19] tried to introduce additional features to reduce the estimation error, such as type of precipitation, distance from the radar, etc., while [20,21,22] added seasonal, monthly, or multi-daily time scale feature information These features are not always available, and the results are not universally better than that of the simple ZR formula. To this end, in this paper, we propose a deep learning-based method to model the ZR relation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have