Integrated Threshold Development for Parametric Insurance Solutions for Guangdong Province China (INPAIS)

气象极端事件对中国经济发展、社会福利状况的影响非同寻常。强热带气旋（TCs或台风）的影响至关重要，并导致中国南方，特别是广东省的重大损失。减轻对社会不同部门的严重负面影响的一种方法是开发和应用金融工具来转移风险和作出适当的反应。除了经典的（私人组织的）（再）保险解决方案外，自最近以来，参数保险解决方案已在某些领域开发用于测试案例，例如我们的执业合作伙伴瑞士再保险北京分公司。瑞信北京分公司目前面临的挑战是，为广东省政府为TC指数触发器设计和承保的参数保险计划实现可靠的结构。第一个主要挑战是台风的区域化影响，导致难以准确估计广东省每个分区或地级市的损失，这将导致对特定地级市的补偿不足或过高。第二个主要挑战是极端台风的气象资料（以及过去的损失）非常有限和不稳定。因此，参数保险的设计一方面在评估实际灾害的频率和强度方面存在缺陷，另一方面又受到历史损失（保险损失和经济损失）的有限可用性的影响。INPAIS旨在改进灾害风险评估，从而改进广东省的响应触发点。最终，这将提高台风袭击后的快速反应和恢复。为了解决挑战一，即缺少对受灾地区损失的区域评估的问题，INPAIS将应用并进一步开发基于伯明翰大学WiTrack算法的成功工具来客观测量和量化风暴严重性（风暴严重性指数，SSI）。这种危害评估将允许在事件的基础上进行综合描述。将提供每个台风系统的详细资料（足迹、受影响的地区、相对于气候背景的风速资料、路径位置等）。我们的合作研究者叶教授将从中国气象局的档案中收集地级损失信息，这些档案只在当地网站和中文中提供。为了解决挑战二，即物理上一致的气象数据样本较少的问题，从而导致对实际危害风险及其不确定性的评估不那么可靠，客观工具将以气气学方法应用于业务预报档案TIGGE （THORPEX互动大全球集合）。TIGGE数据集包括来自10个全球数值天气预报中心的近10年多模式综合预报数据。这将使我们能够从相当于4万到5万年的样本中估计台风的返回水平。INPAIS的产出将用于：A)科学地：在区域范围内提供关于综合灾害台风（干湿）的更现实的频率强度分布的信息。这将导致对产出所需的季节性至十年时间尺度的灾害不确定性的改进估计。B)金融工具：改进广东省现有参数台风保险的触发点。这将改善损害/损失与现有保险的匹配。这也将解决特定区域补偿不足或补偿过度的问题，以及可能发展改进的分配机制，最终使更快的反应和恢复成为可能。

Meteorological extreme events affect China's economic development, society and welfare condition in an extra-ordinary way. The impact of strong tropical cyclones (TCs or Typhoons) is of crucial importance and leading to major losses in southern China, especially in Guangdong Province. One way of mitigating severe, negative impacts on different sectors of society is the development and application of financial instruments for risk transfer and adequate response. Beside classical (privately organised) (re-)insurance solutions, since recently, parametric insurance solutions have been developed for test cases in some areas, as e.g. by SwissRe Beijing Branch, our practitioner partner. The challenge SwissRe Beijing Branch is now facing, is to achieve a reliable structure for the parametric insurance programmes designed and underwritten for TC index triggers for Guangdong Government. The first major challenge is the regionalised effect of Typhoons, leading to difficulties in the robust estimate of losses per sub-region or prefecture level in Guangdong province, which will lead to under- or overcompensating for specific prefectures. The second major challenge is the very limited and instationary nature of meteorological information (as well as for past losses) available for extreme Typhoons.Consequently, the design of parametric insurance has shortcomings in the assessment of the real hazard frequency and intensity on one side and on the other side suffers from limited availability of historic loss (insured and economic loss) to calibrate the impact. INPAIS aims to improve the hazard risk assessment and thus to improve the response trigger points for Guangdong province. Ultimately, this will lead to increased rapid response and recovery after a Typhoon strikes. To address challenge one, the problem of missing regional assessment of losses in affected prefectures, INPAIS will apply and further develop the successful tool to objectively measure and quantify storm severity (Storm Severity Index, SSI) based on the WiTrack algorithm from University of Birmingham. This hazard assessment will allow for an integrated characterisation on event basis. Detailed information per Typhoon system (footprint, area affected in relation to damages; wind speed information relative to climatological background; track location; etc.) will be provided. Our co-investigator, Prof. Ye, will collect information of losses on prefecture level from archives of the Chinese Meteorological Agency, only available in local sites and in Chinese. To address challenge two, the problem of small samples of physically consistent meteorological data available and thus leading to less robust assessments of the real hazard risk and its uncertainty, the objective tool will be applied to the operational forecast archive TIGGE (THORPEX Interactive Grand Global Ensemble) in a climatological approach. The TIGGE dataset consists of multi-model ensemble forecast data from 10 global Numerical Weather Prediction centres, available for the last 10 years. This will allow to estimate return-levels of Typhoons out of a sample equivalent to some 40,000 to 50,000 years. The INPAIS outputs will be used: A) Scientifically: To inform about more realistic frequency-intensity distributions of the integrated hazard Typhoon (dry - wet) on a regional scale. This will lead to improved estimates of uncertainties of hazards on seasonal to decadal time scales necessary for output:B) Financial Instrument: to improve the trigger points of the existing parametric Typhoon insurance for Guangdong province. This will led to improved matching of damages/losses and cover existing. This will also tackle the problem of under- or overcompensating in specific regions and the potential development of improved distribution mechanisms, ultimately enabling increased rapid response and recovery.