融合气象与遥感数据的农作物台风灾情快速评估方法研究

Study on rapid assessment method of crop typhoon disaster based on meteorological and remote sensing data

关键词:  脆弱性曲线  DVDI  降尺度  台风  农田灾情  快速评估

Key words:  Vulnerability curve  DVDI  downscaling  typhoon  farmland disaster  rapid assessment

摘要:

【目的】 通过多尺度数据融合技术结合两种方法的优点，提出一种空间覆盖完整同时分辨率较高的大范围农田台风灾情快速评估方法。【方法】 文章基于2010 —2019年对中国沿海地区农作物损害较大的25个台风样本的省级受灾面积与气象再分析数据（ECMWF Reanalysis v5，ERA5），对农作物的脆弱性模型进行中国区域的参数较准，用以估计粗分辨率（25 km）农田受损率；然后通过遥感数据计算灾害植被损害指数（Disaster Vegetation Damage Index，DVDI）刻画细分辨率的灾情空间细节，并基于该指标对气象数据驱动的脆弱性模型估算受灾率结果进行空间降尺度，最终得到250 m分辨率的农作物灾情评估结果。【结果】 中国区域参数较准的脆弱性曲线模型可以用于评估大尺度上农作物台风灾情的空间分布，其估算的省级受灾面积与统计资料一定程度上相关（R²=0.38）；通过融合遥感DVDI数据对该模型受损率估计结果进行降尺度后，进一步增强灾情评估结果的空间细节，与高分影像目视判读对比显示，融合结果的空间分布细节与实际受灾情况具备较好的一致性。【结论】 该文提出的适用于大区域高分辨率的农作物台风灾情快速评估方法有潜力应用于台风农田灾情的快速全面评估。

Abstract:

[Purpose] In this study，by combining the advantages of the above two methods with multi-scale data fusion technology，we proposed a crop damage assessment method for typhoon disaster，which is expected to produce the assessment result with complete spatial coverage and high spatial resolution.[Method] Firstly，based on the provincial crop damaged area for the 25 typhoons in coastal areas of China from 2010 to 2019 and ECMWF Reanalysis v5（ERA5），the crop fragility model was adjusted and used to estimate the crop damage rate at coarse resolution of 25 km in China. Then，the Disaster Vegetation Damage Index （DVDI） was calculated by Moderate-resolution Imaging Spectroradiometer （MODIS） data to capture the spatial details of the crop damage at 250 m resolution. Based on DVDI，the estimated crop damage rate at coarse resolution was spatially downscaled into a fine resolution of 250 m .[Result] The experimental results show that the fragility model adjusted by the census data of China can be used to evaluate the spatial distribution of typhoon induced crop damage at a large scale，and the estimated provincial crop damage area is positively correlated with the census data （R² = 0.38）. The spatial details of the crop damage rate were further enhanced by the proposed downscaling technique with fusing DVDI data derived from remote sensing. Visually comparison with two high-resolution images acquired before and after “Lekima” typhoon show that the spatial details of the downscaling crop damage rate are well consistent with the actual crop damage.[Conclusion] Therefore，the proposed crop damage assessment method for typhoon disasters has potential for rapidly assessing the typhoon induced crop damage with complete spatial coverage and high spatial resolution.