中国岛礁区热带气旋不同空间分辨率风场模拟比较研究OA北大核心CSTPCD

China has numerous islands and reefs with complex terrain that are heavily impacted by tropical cyclone disasters.High-resolution tropical cyclone wind-field simulations are beneficial for representing the spatial variations in wind speeds.It is important to conduct high-resolution simulations on relatively small islands and reef areas.To explore the differences in tropical cyclone wind field simulations at various spatial resolutions in the island and reef areas of China,this study compared the modeled wind fields of historical tropical cyclones in China's island and reef areas,which have complex terrains,including plains,peaks,valleys,and cliffs,at three spatial resolutions of 1,000 m,90 m,and 30 m.The wind fields were modeled using land cover and elevation data of the three spatial resolutions as inputs and validated against observed winds at eight stations.Comparisons were made regarding the differences in wind speeds of tropical cyclones with a 100-year return period at three spatial resolutions.The results showed that:(1)the 30 m resolution achieves the best accuracy,with a root mean square error of 4.28 m/s,lower than those of 90 m and 1 km by 0.08 m/s and 1.04 m/s,respectively.(2)Different spatial resolution simulations showed that wind speed errors were related to terrain types.For example,on Zhujiajian Island,located in Zhoushan City,the 30 m resolution captured the spatial heterogeneity of winds better than the other resolutions,especially for mountainous,valley,and cliff terrains.Comparisons between the simulated wind speeds at 90 m and 1,000 m resolutions versus those at 30 m resolution indicate that the differences in the simulation percentages are as follows:6.57%and 7.61%for peak terrain,21.28%and 17.35%for valley terrain,and 22.85%and 23.37%for cliff terrain,respectively.Additionally,the 30 m simulation was more sensitive to transitions between windward and leeward slope terrains.(3)For the 100-year return-period wind speeds,the 30 m resolution produced the highest values and largest spatial variations.On Zhujiajian Island,the maximum wind speeds at 1,000 m,90 m,and 30 m resolutions were 71.13,73.18,and 79.97 m/s,respectively,and standard deviations of 3.88,3.72,and 7.18 m/s.This study demonstrates the importance of using high-resolution data to simulate tropical cyclone winds in complex terrain.However,this study had some limitations.First,the terrain correction factors need to be optimized further.The assessment method provided by the building codes tended to overestimate the impact of the terrain correction factors.In the future,more accurate terrain correction factors could be obtained using computational fluid dynamics and wind tunnel tests.Second,because of the limited types of land cover data used in the calculations,the subdivision of certain land types when calculating the surface roughness is not sufficiently detailed.Additionally,different years of land cover data were not incorporated,making it challenging to reflect the variations in surface roughness.Remote sensing can be used in the future to determine the high-resolution spatial distributions of surface roughness.