Classified Early Warning and Forecast of Severe Convective Weather Based on LightGBM Algorithm

Severe convective weather can lead to a variety of disasters, but they are still difficult to be pre-warned and forecasted in the meteorological operation. This study generates a model based on the light gradient boosting machine (LightGBM) algorithm using C-band radar echo products and ground observations, to identify and classify three major types of severe convective weather (i.e., hail, short-term heavy rain (STHR), convective gust (CG)). The model evaluations show the LightGBM model performs well in the training set (2011-2017) and the testing set (2018) with the overall false identification ratio (FIR) of only 4.9% and 7.0%, respectively. Furthermore, the average probability of detection (POD), critical success index (CSI) and false alarm ratio (FAR) for the three types of severe convective weather in two sample sets are over 85%, 65% and lower than 30%, respectively. The LightGBM model and the storm cell identification and tracking (SCIT) product are then used to forecast the severe convective weather 15 - 60 minutes in advance. The average POD, CSI and FAR for the forecasts of the three types of severe convective weather are 57.4%, 54.7% and 38.4%, respectively, which are significantly higher than those of the manual work. Among the three types of severe convective weather, the STHR has the highest POD and CSI and the lowest FAR, while the skill scores for the hail and CG are similar. Therefore, the LightGBM model constructed in this paper is able to identify, classify and forecast the three major types of severe convective weather automatically with relatively high accuracy, and has a broad application prospect in the future automatic meteorological operation.

A Case Study on the Rapid Rain-to-Snow Transition in Late Spring 2018 over Northern China:Effects of Return Flows and Topography

Phase changes in the precipitation processes of early winter and late spring in midlatitude regions represent challenges when forecasting the timing and magnitude of snowfall.On 4 April 2018,a heavy snow process occurred in Beijing and northwestern Hebei Province,becoming the most delayed occurrence of heavy spring snow ever recorded over Beijing in the last 30 years.This paper uses observational and numerical simulation data to investigate the causes for the rapid rain-to-snow(RRTS)phase transition during this process.The following results are obtained.(1)Return flows(RFs),an interesting type of easterly wind,including those at 1000,925,and 800 hPa,played an important role in this heavy snow process and presented a characteristic"sandwich"structure.The RFs,complex topography,and snow particles that dominated the clouds,were the three key factors for the RRTS transition.(2)The RRTS transition in the plains was directly related to the RF at 925 hPa,which brought about advective cooling initiated approximately 4-6 h before the onset of precipitation.Then,the RF played a role of diabatic cooling when snow particles began to fall at the onset of precipitation.(3)The RRTS transition in the northern part of the Taihang Mountains was closely related to the relatively high altitude that led to a lower surface temperature owing to the vertical temperature lapse rate.Both immediately before and after the onset of precipitation,the snow particles in clouds entrained the middle-level cold air downward,causing the melting layer(from surface to the 0℃-isotherm level)to become very thin;and thus the snow particles did not have adequate time to melt before falling to the ground.(4)The rapid RRTS over the Yanqing mountainous area in the northwest of Beijing could have involved all the three concurrent mechanisms:the advective cooling of RF,the melting cooling of cloud snow particles,and the high-altitude effect.Compared with that in the plain area with less urbanization the duration of the RRTS in the plain area with significant urbanization was extended by approximately 2 h.