Snow Microphysical Characteristics And Snowfall Estimation In East China Based On A 2D Video Disdrometer And Dual Polarization Radar

Snowfall is a common type of winter precipitation.Heavy snowfall can cause a negative impact on society.Fully understanding the microphysical characteristics of snow and the vertical structure of snow cloud is of great significance for comprehending the mechanism of snowfall and improving the accurate estimation of snow.In recent years,with the development of observation techniques and numerical models,the understanding of microphysical characteristics of winter snowfall have been gradually improved,and there are more in-depth studies on quantitative snowfall estimation.However,due to the limitation of observation in the past,there were few studies on microphysical characteristics of snow,vertical structure characteristics of snowfall system,and quantitative snowfall estimates in China.In this study,the data of a two-dimensional video disdrometer(2DVD),weighing precipitation gauges,an S-band Doppler weather radar,and a C-band dual-polarization Doppler radar in the winter snowfall events in Nanjing from 2015 to 2019 were collected.We analyzed the microphysical characteristics of ground snowfall,and explored vertical structure characteristics of snowfall system,and established a quantitative snowfall estimation formula suitable for this area.The main conclusions are as follows:First,by combining datasets from the 2DVD and weighing precipitation gauges,we analyzed the statistical characteristics of terminal velocity,density,and particle size distribution(PSD)of snowfall in Nanjing.Compared with the previous research results,the snow terminal velocity observed in this area is larger than that estimated using an empirical velocity-diameter relation.Two types of snowfall events are classified by terminal velocity,i.e.,a type of high terminal velocity and another with low terminal velocity.The differences in terminal velocity of snowflakes between snowfall events may be explained by differences in snowflakes type.Based on the observation data,we developed a snow density-size relationship.Compared to the result reported by Brandes et al.(2007),the mean snow density was greater in the current study when D₀<3.2 mm,which may be responsible for higher snowflakes terminal velocity.The statistical analysis of the snow PSD shows that the particle size of the peak number concentration of the average snow PSD is small,about 0.3 mm.Due to the low collision-coalescence rate of small particles during the growth of snow particles,the number concentration decreases rapidly with the increase of size.When the particle size reaches a certain size(>8 mm),the concentration of snow particles is between 1?4 mm^-1m^-3.The interrelationships among three parameters of Gamma model in terms of snow rate(SR)were analyzed.The concentration of snow particles in all size ranges increased as SR increased.We then evaluated the contribution of particle diameter and total number concentration to snowfall intensity.The increase in SR was influenced by both the diameter and concentration of snow particles.Secondly,the equivalent radar reflectivity factor(Ze),SR,and ice water content(IWC)are derived based on ground-observed data of snowfall microphysical characteristics.The simulated Ze using the density relation fitted is more accurate than using the empirical relation,by the mean absolute error(MAE)dropped from 21.9%to17.1%.The-SR and-power-law relationships can be then derived.Then we quantitatively validate the radar-derived snowfall estimation.The average MAE of snowfall estimation by the general-SR relation among the snowfall events in NJ was 16%,of which the MAE of low terminal velocity snowfall events is23.3%,and the MAE of high terminal snowfall events is 8.9%.Furthermore,according to the observation of dual-polarization radar,we analyzed the ice-phase microphysical characteristics in vertical structure and the growth mechanism of ice crystals of two types of snowfall events(i.e.low and high terminal velocity snowfall events).The results show that,for the low(high)velocity snowfall events,the echo height is low(high),the wet layer is not deep(deep),and the ice water content is low(high);the pristine ice crystal layer is shallow(deep),and the ice crystal growth layer is near the top(middle)of the echo,differential reflectivity(Z_DR)and special differential propagation phase shift(K_DP)in the ice crystal growth layer are enhanced significantly(less significantly);the average value of horizontal reflectivity factor(Z_H)for the entire layer is small(large),while the average Z_DR is large(small).The characteristics of the polarimetric radar variables shows that the growth processes of the crystals in the ice crystal growth layer of the low vertical velocity snowfall event are dominated by deposition and the characteristics of branching growth are significant.The ice crystals have a single composition type,and weak aggregation.As a result,the snow particle size is small,and the terminal velocities of the snowflakes are lower.For the snowfall event classified as the high-vertical-velocity type,the pristine ice crystal layer is deep and ice crystals inside have various habits,which has a strong seeding effect on the growth of the underlying ice crystals.The characteristics of branching growth are not significant in the ice crystal growth layer,and ice crystals are mainly grow by aggregation.Therefore,there are various types of snow particles with larger sizes and a wider range of terminal velocity values.Overall,the microphysical characteristics of snow observed on the ground are related to the height of the snowfall clouds,the vertical structural characteristics of the snowfall system,and the ice crystal growth mechanism.Finally,we explore the correlation between the radar observations at high altitudes and the snowfall rate on the ground of the two types of snowfall events.For low terminal velocity snowfall events,there are high correlation between the high-level K_DPand the low-level Z_H.The high-level K_DP have potential prediction on SR.Besides,the SR peak values are affected by the high-level K_DP and Z_DR.When the SR peaks,the K_DPand Z_DR of the upper layer are inversely correlated,and the SR decreases as the Z_DR of the upper layer increases.For high terminal velocity snow events,the correlation between the high-level K_DP and low-level Z_H and SR is not high.The predictability of high-level K_DP to SR is lower than that of low terminal velocity snowfall events.The correlation of high-level Z_H and SR is significantly high,indicating that high-level Z_Hhave higher potential prediction on SR than high-level K_DP.This paper reveals the microphysical characteristics of snowfall events in eastern China,effectively improves the accuracy of quantitative snowfall estimation,and analyzes the ice-phase microphysical characteristics of vertical structure of snowfall process and ice crystal growth mechanism by the observation of dual polarization radar.The research results have important guiding significance for the prevention and mitigation of winter blizzard.