Table of contents
1 Background overview
Our country has a vast territory and complex natural conditions, so there are many types of disastrous weather with large regional differences. Among them, strong convective weather such as thunderstorms, hail, tornadoes, and short-term heavy precipitation are the most serious types of disastrous weather that cause economic losses and endanger life safety [1]. Taking 2022 as an example, the number of deaths and missing persons and direct economic losses caused by wind and hail disasters caused by strong convective weather in my country accounted for 73% and 69% respectively. Since severe convective weather has the characteristics of suddenness, strong locality, short life history, and severe disasters, its short-term (0 to 12 hours) and nowcast (0 to 2 hours) forecasts are usually difficult in weather forecasting.
Traditional severe convective weather nowcasting mainly relies on radar and other observation data, combined with storm identification and tracking technology to perform radar extrapolation forecasting. That is, the radar reflectivity factor at the future moment is obtained through extrapolation, and further uses the radar reflectivity factor and precipitation. The empirical relationship between them (i.e. ZR relationship) estimates the precipitation in the future [2]. In recent years, with the accumulation of big data and the development of computer computing power, artificial intelligence and deep learning technology have developed rapidly. The deep learning method is a type of data-driven method. In theory, its performance improves as the amount of training data increases, so it is very suitable for the field of short-term forecasting where a large amount of radar observation data is accumulated. At present, there are two main types of short-term forecast models based on deep learning in the world. One type is based on Convolutional Neural Networks (CNNs), such as U-Net and other models [3]; the other type is based on Recurrent Neural Networks (Recurrent Neural Networks, RNNs), such as ConvLSTM, DGMR and other models [4, 5].
Raindrops are affected by air resistance during the falling process, and their shape can be oblate spherical or steamed bun-shaped. Generally speaking, the larger the raindrops, the flatter they are. Therefore, the reflection characteristics of raindrops for horizontally polarized electromagnetic waves (the direction of electric field vibration is in the horizontal plane) and vertically polarized (the direction of electric field vibration is in the vertical plane) are different. Traditional radar can only transmit and receive electromagnetic waves in one polarization direction, while the new dual-polarization radar can simultaneously transmit and receive electromagnetic waves in both horizontal and vertical polarization directions. According to the intensity difference of the echoes in the two polarization directions, Information such as phase relationship is used to obtain information such as the size, phase state, and water content of precipitation particles [6]. This information is collectively referred to as microphysical information. Research in recent years has shown that the microphysical information reflected by dual-polarization radar variables contains key information such as the evolution state of the convective system and the spatial dynamic structure [7, 8]. Therefore, the application of dual-polarization radar variables is theoretically important for severe convection forecasting.
2 Problem details
In order to better apply dual-polarization radar to improve short-term forecast of severe convective precipitation, please answer the following questions:
- How to effectively apply dual polarization variables to improve severe convection forecasting is still a key and difficult issue in current meteorological forecasting. Please use the data provided in the question to establish a mathematical model that can extract microphysical feature information from dual-polarization radar data for severe convection nowcasting. The input of nowcasting is the radar observations (ZH, ZDR, KDP) of the previous hour (10 frames) , and the output is the ZH forecast of the following hour (10 frames ) .
- Some current data-driven algorithms tend to generate forecasts close to the mean when making strong convection forecasts, that is, there is a "regression to the mean" problem, so the forecasts always tend to be blurry. Based on question 1, please design a mathematical model to alleviate the blurring effect of the forecast and make the predicted radar echo details more complete and realistic.
- Please use the Z H , Z DR and precipitation data provided in the question to design an appropriate mathematical model and use Z H and Z DR to conduct quantitative precipitation estimation. The model inputs are Z H and Z DR and the output is precipitation. (Note: The algorithm cannot use K DP variables.)
- Please design a mathematical model to evaluate the contribution of dual-polarization radar data in nowcasting of severe convective precipitation, and optimize the data fusion strategy to better deal with sudden and local severe convective weather.
3 noun explanation
- Dual polarization radar: A new type of weather detection radar that can provide richer physical information than traditional radar. It obtains information such as the size, phase, and moisture content of precipitation particles by measuring the reflection of electromagnetic waves in both horizontal and vertical directions by precipitation particles. This information is collectively called microphysical information and can help us better predict severe convective weather. The three most commonly used variables of dual-polarization radar are: 1) Z H, horizontal reflectivity factor, which is the echo intensity in the horizontal direction, usually in dBZ, mainly reflecting the intensity of precipitation; 2) Z DR, differential reflectivity, That is, the difference in echo intensity between horizontal and vertical directions mainly reflects the size of precipitation particles in the observation area; 3) K DP, specific differential phase shift, which is the phase difference between horizontal and vertical echoes caused by precipitation particles at unit distance, mainly Reflects liquid moisture content.
- ZR relationship: The empirical relationship between radar reflectivity and precipitation, usually expressed as
, where R is the precipitation, Z is the radar reflectivity, a and b are empirical parameters, which usually differ in different regions and different precipitation types. .
4 Questions and data
Link: https://pan.baidu.com/s/1cLfyT3KNOzgOeb7YkgAYPA
extraction code: kl75
-- Sharing from Baidu Netdisk Super Member V4
5 Python&Matlab code download
Reply: [2023 Huawei Cup Question F] Strong convective precipitation now forecast