| Summary: | 碩士 === 國立臺灣大學 === 大氣科學研究所 === 105 === In this study, a simulation was conducted by WRF model coupled with CLR microphysics scheme to investigate the effect on cloud properties and precipitation when ice nuclei concentration increases, and the difference of this effect when different ice nucleation parameterizations are applied. 3 ice nucleation parameterizations are tested, the first one is only dependent on ambient temperature, the second one is only dependent on ambient supersaturation, and the other one is dependent on both ambient supersaturation and ice nuclei concentration. For each parameterization, 3 values of ice nuclei concentration are tested, with the other two having a difference of 10 times larger or smaller compared with control runs, and there are 9 simulations in total.
The results show that increasing ice nuclei enhances the heterogeneous nucleation of ice crystals between 500 and 700hPa, thus there are more ice crystals and snow particles, and depositional growth process is enhanced. These ice-phase particles grow by deposition via Bergeron process at the expense of liquid water, reducing cloud drops significantly. The decrease of cloud drops leads to the decline of homogeneous nucleation process upon 500hPa, resulting in less ice crystals in the upper troposphere, as well as the decline of riming process of snow particles, cancelling out the enhancement of depositional growth. Therefore, no significant changes are shown in total precipitation and its spatial distribution.
When different ice nucleation parameterizations are applied, the changes of hydrometeors and cloud microphysics processes due to the increase of ice nuclei concentration have different amplitude. Simulating the difference of heterogeneous ice nucleation between the three sensitivity tests with low and high ice nuclei concentration, there is only significant difference in the upper troposphere where the temperature is low when the parameterization dependent on temperature is used, while significant difference also exist in the middle troposphere when the other two parameterization are used. Meanwhile, the amplitude of the changes simulated by the parameterization dependent on both supersaturation and ice nuclei concentration is less than the parameterization which is only dependent on supersaturation because the former one will diagnose weaker heterogeneous ice nucleation when ambient ice nuclei are consumed. With a significant change of heterogeneous ice nucleation simulated between 500 and 700hPa, the parameterization dependent only on supersaturation can also simulate a significant change of the depositional and riming growth of snow consequently.
|
|---|
