Abstract In the Democratic People's Republic of Korea (DPRK), hot and humid air is influx from the sea by the monsoon and the temperature is high in summer, so that the phenomenon of mesoscale free convection due to thermal instability is frequently observed. For these weather scenarios, the Weather Research and Forecasting (WRF) model produces more rainfall distributions than the observations. In this paper, the sensitivity of the autoconversion threshold in the microphysics scheme was analyzed through the study of convective rainfall events that occurs frequently in summer in the DPRK. WRF–ARW version 4.5.1 was used and Thompson microphysics scheme was selected to change the autoconversion threshold. WRF results show that it gives many false alarms in the study area and the higher the terrain height, the morefalse alarms. Based on the relationship between terrain height and the ratio of false alarms, this study has proposed five approaches that reflect the effects of terrain height differently to the autoconversion process which cloud water converted into raindrop: a linear change of autoconversion threshold with terrain, three parabolic changes similar to the trend of increasing false alarm rates with terrain, and an exponential change with opposite trend. Through the analysis of 24 haccumulated rainfall and statistical indices for four convective rainfall events, an approach was selected that not only reduce the false alarms but also simulate the rainfall well. Methods that allow autoconversion thresholds vary with terrain significantly reduce false alarms in the study area; the False Rate (FR) was decreased from 50% to 30%. However, Probability of Detection (POD) does not decrease, but rather increases slightly. Dividing the autoconversion threshold change methods proposed in this paper into two groups methods with much false alarms and methods with less false alarms, two groups simulate clearly different profiles for cloud water mixing ratio and rain droplet mixing ratio. And the best performing method simulates large amounts of solid-phase hydrometeors compared to the original WRF simulation. This study shows that adjusting the autoconversion threshold is one way to control the rainfall mechanism. It is also shown that effective adjustment of microphysical schemes can reduce the false alarms and simulate the mesoscale convective events well. Research Highlights <UnorderedList Mark="Bullet"> <ItemContent> The WRF model predicts much of the false precipitation for convective rainfall events, especially the predicted rainfall is much larger than the observed one in the regions with the high altitude. </ItemContent> <ItemContent> Adjusting the autoconversion process of the microphysical schemes is one way to improve the accuracy of rainfall simulation. </ItemContent> <ItemContent> Effective representation of the topographic effects into the microphysical scheme can significantly reduce the false alarm of the numerical prediction models. </ItemContent> <ItemContent> Adjusting the autoconversion threshold causes a large change in the hydrometeor concentrations, which has a large impact on the predicted rainfall field. </ItemContent> </UnorderedList>

Sensitivity of autoconversion thresholds on cloud microphysics schemes: a case study on convective precipitations during summer in the Democratic People’s Republic of Korea

期刊论文
发表日期：

2026年5月19日

查看全文

Ju-Song Kim,
Hyok-Chol Kim,
Sang-Il Pak,
Kyong-Ung Ri,
Song-Il Kim

摘要

Abstract

In the Democratic People's Republic of Korea (DPRK), hot and humid air is influx from the sea by the monsoon and the temperature is high in summer, so that the phenomenon of mesoscale free convection due to thermal instability is frequently observed. For these weather scenarios, the Weather Research and Forecasting (WRF) model produces more rainfall distributions than the observations. In this paper, the sensitivity of the autoconversion threshold in the microphysics scheme was analyzed through the study of convective rainfall events that occurs frequently in summer in the DPRK. WRF–ARW version 4.5.1 was used and Thompson microphysics scheme was selected to change the autoconversion threshold. WRF results show that it gives many false alarms in the study area and the higher the terrain height, the morefalse alarms. Based on the relationship between terrain height and the ratio of false alarms, this study has proposed five approaches that reflect the effects of terrain height differently to the autoconversion process which cloud water converted into raindrop: a linear change of autoconversion threshold with terrain, three parabolic changes similar to the trend of increasing false alarm rates with terrain, and an exponential change with opposite trend. Through the analysis of 24 haccumulated rainfall and statistical indices for four convective rainfall events, an approach was selected that not only reduce the false alarms but also simulate the rainfall well. Methods that allow autoconversion thresholds vary with terrain significantly reduce false alarms in the study area; the False Rate (FR) was decreased from 50% to 30%. However, Probability of Detection (POD) does not decrease, but rather increases slightly. Dividing the autoconversion threshold change methods proposed in this paper into two groups methods with much false alarms and methods with less false alarms, two groups simulate clearly different profiles for cloud water mixing ratio and rain droplet mixing ratio. And the best performing method simulates large amounts of solid-phase hydrometeors compared to the original WRF simulation. This study shows that adjusting the autoconversion threshold is one way to control the rainfall mechanism. It is also shown that effective adjustment of microphysical schemes can reduce the false alarms and simulate the mesoscale convective events well.

Research Highlights

The WRF model predicts much of the false precipitation for convective rainfall events, especially the predicted rainfall is much larger than the observed one in the regions with the high altitude.

Adjusting the autoconversion process of the microphysical schemes is one way to improve the accuracy of rainfall simulation.

Effective representation of the topographic effects into the microphysical scheme can significantly reduce the false alarm of the numerical prediction models.

Adjusting the autoconversion threshold causes a large change in the hydrometeor concentrations, which has a large impact on the predicted rainfall field.