TAHOPE Project Office

Most of severe rainfall events in the vicinity of Taiwan are resulted from Mei-yu frontal systems and impinging typhoons. Heavy or even severe rainfall, whatever produced over the steep slopes of Central Mountain Range (CMR) or coastal plains, often induces inundation and river runoff, causing significant life loss as well as environmental impacts and economic losses. In particular, severe landslide and debris flow are often associated with intense typhoons impinging Taiwan from every possible direction. There are pronounced improvement on numerical model prediction skill on rainfall recently, due to the advancement of model physics schemes, dynamic core refinement, more robust data assimilations with a variety of available observations, and possibly integrated statistical add-on values with artificial intelligence based on numerous ensemble predictions for varying scenarios (a big data for mining). However, prediction failures also exist occasionally or even become inevitable, most likely owing to insufficient resolvable atmospheric states that dictate the major role in the mechanisms responsible for severe rainfall. Our current radar coverage over Taiwan is world-class leading, together with plentiful surface observations, which should provide reasonable and intensive, although not comprehensive, depiction of the impinging systems, whatever Mei-yu fronts or typhoons (there will be almost no low-elevation observation gaps in 2019). This situation does not puzzle our operational forecasters and academic researchers as we all realize the importance of upgraded radar measurement capability (e.g., dual polarization) in exploring the essential characteristics and nature of cloud microphysics in the convective system as well as effectively collected upstream atmospheric conditions for both upcoming fronts and typhoons. Up to now, there still lack integrated observations near Taiwan or upstream for improving gaps in capability of numerical model prediction on such severe rainfall.

臺灣周邊強降雨事件多為梅雨鋒系統和颱風侵襲。 無論是在中央山脈陡坡或沿海平原上產生的強降水甚至劇烈降水，都經常引發洪水和河川逕流，造成重大生命損失、環境影響和經濟損失。 尤其是嚴重的山崩和土石流，往往伴隨著從各個可能的方向襲擊台灣的強烈颱風。 近期降雨數值模式預報技術有了顯著提高，仰賴於模式物理方案的進步、動力核心細緻化、各種可用的觀測數據進入更可靠的資料同化系統、基於數個不同情境的系集預報與利用人工智慧整合統計附加價值（大數據探勘）。 然而，預報失敗也偶爾發生，甚至不可避免，很可能是由於大氣狀態解析度不足，這決定了造成強降雨機制中的主要因素。 目前台灣的雷達覆蓋率是世界一流的，再加上豐富的地面觀測，雖然不全面，但應該可以提供合理而深入的描述天氣系統，無論梅雨鋒面或颱風（2019年後將幾乎沒有低海拔觀測死角）。 對於即將到來的鋒面和颱風，這種情況並不困擾我們的作業預報員和學術研究人員，因為我們都意識到雷達觀測能力提升（例如雙偏極化雷達）在探索對流系統中雲微物理的基本特徵和性質以及有效收集上游大氣條件很重要。 到目前為止，台灣附近或上游仍缺乏整合觀測，以改善數值模式對此類強降雨的預測能力的差距。

There were field observation experiments during the past targeted on Mei-yu frontal systems and impinging typhoons, paving a way of better understanding the processes of the induced heavy rainfall/convection and/or providing useful upstream data for improving NWP. Then, why is there a need to conduct a joint experiment over Taiwan in the upcoming years? The most important reason is the uniqueness of Taiwan islands as the optimal, if not best, natural mesoscale laboratory in the world that illustrates the possibility of better understanding of various dynamic and physical processes in convective clouds, which could be generally suited to other places, and an arena of various measurements to quantify their contributions to NWP, whatever focusing on severe rainfall or violent wind. These coincident efforts on routinely conducing measurements then will be greatly signified when the joint experiments may additionally collect highly demanding in-situ observations for NWP thru advanced data assimilation to fill the gaps in current forecast skill on convective rainfall and typhoon tracks. These problems range from how to improve initial conditions from different types of observations from surface, land or sea, aircraft or space, different data assimilation systems (ensemble or variation based, or both joint like ensemble/hybrid), how to unveil the detailed processes and nature of convective clouds with dual-polarization measurement, how to cast the improved physics and initial analysis into state-of-the-art numerical models (both regional and global) with updated comprehensive physics and parameterization schemes, and finally also most demanding, how to provide controlling diagrams for the sophisticated interactions of the impinging systems with the topography (steep CMR and surrounding ocean) involved. In brief, we should say that the joint experiment is going to enhance scientific baseline of DA and modeling, provide better understanding of physical and dynamic intricacy, narrow forecast gaps of QPF, and address challenges of key issues and highlight possible opportunities of future movement.

過去曾針對梅雨鋒系統和侵台颱風進行野外觀測實驗，能更好地了解誘發強降雨/強對流的物理過程和/或提供有用的上游資料以提高數值天氣預報。 那麼，為什麼未來幾年還需要在台灣進行聯合觀測實驗呢？最重要的原因是台灣島作為世界上最佳的（如果不是最好的）天然具有獨特性的中尺度實驗室，說明觀測實驗能更好地了解對流雲中各種動力和物理過程的可能性，這些過程通常也適用於其他地方，以及在強降雨或強風下量化各種觀測儀器對數值天氣預報的貢獻。 聯合實驗可以為數值天氣預報模式額外收集珍貴的現地觀測，透過先進的資料同化，填補當前對流降雨和颱風路徑預報的技術空缺，這些短時間密集的努力對往後長期例行的觀測具有重大意義。 這次實驗的科學問題的範圍包括如何透過來自地面、陸地或海洋、飛機或太空的不同類型的觀測、不同的資料同化系統（基於系集、變分或兩者混合）來改善模式初始條件； 如何利用雙偏極化儀器揭開對流雲詳細的物理過程； 如何將改進的物理過程和初始場引入最先進的數值模式（區域和全球）中； 最後也是最困難的，如何提供侵臺的天氣系統與地形（陡峭的中央山脈和周圍的海洋）之間複雜交互作用的機制圖。 簡而言之，我們可以說聯合實驗將提升資料同化和建模的科學水平，更好地理解物理和動力的複雜性，縮小定量降水預報的偏差，解決關鍵科學議題的挑戰，點亮未來實驗行動的機會。