Nan Fang Du Shi Bao·2025-10-16 06:52Core Insights - China Ping An's subsidiary, Ping An Property & Casualty, launched the "Ping An Typhoon Risk Map" and "Ping An Typhoon Catastrophe Model" to enhance disaster prevention and control using AI technology [1][2] - The typhoon season has been particularly impactful this year, with 23 typhoons generated in the Northwest Pacific and South China Sea, 9 of which made landfall in China, highlighting the need for improved disaster prediction and defense capabilities [1] Group 1: Typhoon Risk Map - The Typhoon Risk Map consists of three layers: typhoon wind hazard, typhoon rainstorm hazard, and comprehensive typhoon hazard [1] - It utilizes various models, including the "N-year encounter" wind field model and extreme precipitation model, with data covering 76 years of historical typhoon paths and over 2,600 national meteorological station observations [1] - The map has a resolution of 1km x 1km and an accuracy rate exceeding 70%, providing quantifiable typhoon risk assessments for insurance underwriting and urban disaster prevention [1] Group 2: Typhoon Catastrophe Model - The Typhoon Catastrophe Model comprises four core models: random event model, vulnerability model, loss estimation model, and typhoon probability model, using data from national GDP, population distribution, and over 20 years of typhoon claims data [2] - The model can estimate the probability of typhoon catastrophes and the potential loss range, supporting catastrophe insurance pricing [2] - A trial analysis conducted in Haikou City indicated that the city faces a maximum risk level of 10 for wind and rain, with an annual probability of 1.77% for encountering super typhoons above level 17, potentially leading to economic losses in the hundred billion range [2] Group 3: Industry Implications - The integration of AI tools like the Typhoon Risk Map and Catastrophe Model represents a new momentum for disaster prevention and mitigation in the AI era, enhancing urban safety and public governance [2]