Numerical Simulation of Dust-Cloud Interactions Using Regional Models

沙塵是大氣中最主要自然氣膠來源之一；由於其具有做為冰核與雲凝結核的能力，可分別影響冷雲與暖雲生成的微物理過程，進而影響降水或是輻射收支。此外，由於沙塵中帶有許多微量元素，像是其中的鐵，經傳送沉降至海洋後，可以做為浮游植物的營養鹽；而浮游植物是海洋生態循環的基底，沙塵對浮游植物的影響可能會更進一步的影響整個海洋生態系統。本研究的重點即是建立一套詳盡的沙塵與雲微物理模組，用以討論沙塵對於大氣中雲、降水、輻射的影響以及其對海洋營養鹽的貢獻量。 為瞭解大氣沙塵對海洋的貢獻量，本研究中將起沙模組放入CMAQ模式之中，以保留原始物理意義的SNAP參數法，取代原本模式中過於簡化的沙塵沉降參數式，來模擬2007年東亞地區全年的沙塵沉降量。同時，由於東亞地區煤的使用量相當大，而煤灰中含的鐵亦為海洋中鐵的來源，因此本研究也發展了一套煤灰計算方法，模擬2007年東亞煤灰的沉降量，以瞭解沙塵與煤灰對海洋營養鹽相對的重要性。模擬結果顯示，鐵年平均沉降量在西北太平洋與南中國海分別有93%、85％來自沙塵；但煤灰中鐵的溶解度遠高於沙塵鐵(大約差了一至兩個數量級)，考慮溶解度影響的情況下，煤灰對東亞邊緣海浮游植物鐵營養鹽的貢獻將高於沙塵。 上述研究中所使用的CMAQ化學與傳輸模式，雖然可有效地長期模擬沙塵的傳送及沉降，但在沙塵對於雲與降水的影響缺乏完整的計算，也高度簡化雲內沙塵洗除過程。因此本研究另外選擇了WRF模式做為工具，將沙塵起沙模組放入模式之中，並以SNAP參數法描述沙塵與雲交互作用的各項過程，包括冰晶異質核化、雲滴活化、碰撞、沉降等。此研究選擇了2007/04/01至2007/04/03的台灣泥雨個案，針對各項沙塵相關機制做敏感度測試，以瞭解各過程的重要性。結果顯示，若不考慮沙塵對雲微物理過程的影響，模式可能會低估冷雲的含量並造成降水略微減少(約3%)，而此影響主要是來自沙塵做為冰核的異質核化過程，其中凝華核化的影響比浸入核化重要。在沙塵濕沉降機制方面，若不考慮雲內洗除過程，則整個大氣中的沙塵濃度及地面沙塵濃度分別會增加87%及5%；若不考慮雲下洗除過程，則整個兩者分別會增加5% 與 1%，顯示過去模式中經常被忽略的雲內洗除過程，比雲下洗除過程更為重要。Airborne mineral dust particles have potentially important influences on cloud and precipitation processes. Dust particles from deserts or semi-arid regions are one of the main sources of natural aerosol particles, and they have been implicated to play the role of ice nuclei which initiate cloud glaciation, and cloud condensation nuclei which initiate cloud drop formation. On the other hand, cloud processes may also have potentially important influences on airborne dust, such as influences dust deposition processes. Furthermore, dust deposition may provide micronutrient such as iron to the ocean and thus influences marine ecosystem. Therefore, this study is aimed to build up a comprehensive dust module to understand the interactions between dust and cloud, and also the input of dust iron into the ocean. The dust emission module was first coupled into the CMAQ (Community Multiscale Air Quality Modeling System) model to understand the input of dust into the ocean. The original empirical scavenge formula was replaced with the SNAP (Statistical-Numerical Aerosol Parameterization) scheme, which was derived from calculations based on the fundamental equations. Beside dust, fly ash is also an important source of iron in the East Asian region, so it is also considered in this study to understand their relative importance comparing to dust. Simulations for the 2007 conditions showed that the iron deposition from mineral dust constituted about 93% of the total iron deposition over the Northwestern Pacific Ocean and 85% of that over the Northern South China Sea. But when considering the much higher solubility of fly ash iron than dust iron, fly ash may have greater influence on the ecosystem over the marginal seas in East Asia. The CMAQ model applied earlier is a chemical transport model, which is useful for a long-term simulation of the long-term dust transportation and deposition, but its lack of coupled cloud processes makes it unsuitable for detailed study on the interactions between dust and cloud, such as the rainout processes. Therefore, a dust emission module was implemented into the WRF model and coupled with the cloud microphysics to understand the issue of dust-cloud interactions. The processes between dust and cloud considered in the model include heterogeneous ice nucleation, cloud drop activation, dust collection, and dry and wet depositions, and they were parameterized with the SNAP scheme. The simulated case is a cold-front leading dust event which caused muddy rain over northern Taiwan during 1st to 3rd of April, 2007. The sensitivity experiments showed that ignoring the dust effect on cloud may cause an underestimation of the cold cloud hydrometeor content and a slight reduction in precipitation (by about 3%). The main mechanisms involved are the heterogeneous ice nucleation processes, with deposition nucleation more important than the immersion freezing. On the other hand, ignored the dust in-cloud scavenging (rainout) process may overestimate total dust loading and surface dust concentration by about 87% and 5%, respectively; for the below-cloud scavenging (washout), the values become 5% and 1% only. This result indicates that the rainout processes, which were commonly ignored in earlier modeling studies, are much more important than the washout process