基于MODIS数据的冬小麦雹灾空间分布信息提取研究——以河南省平顶山市为例

Spatial Distribution Information Extraction of Winter Wheat Hail Disater Based on MODIS Data:A Case Study of Pingdingshan City,Henan Province

以河南省平顶山市为研究区,利用2015年雹灾前后MODIS数据的红波段、近红外波段、植被指数变化以及当年植被指数与过去近3 a(2012—2014年)同时期正常植被指数均值的变化等数据,统计分析正常生育的冬小麦与发生重度雹灾、中度雹灾和轻度雹灾的冬小麦间的分割阈值,提取重度雹灾、中度雹灾和轻度雹灾空间分布信息,并与实地调查数据验证对比。结果表明:(1)利用雹灾前后归一化植被指数(NDVI)变化更能有效提取雹灾空间分布信息,其精度与实际吻合较好;基于比值植被指数(RVI)变化提取中度雹灾和轻度雹灾的效果不理想。(2)雹灾呈西北—东南走向的长条状分布,且主要集中在坡度5°以下中部低山平原区。(3)雹灾分布区呈以重度雹灾为中心、向外依次为中度雹灾和轻度雹灾的环状分布。

Taking Pingdingshan City of Henan Province as study area, Eight bands based on MODIS data are employed such as red bands, near infrared bands, the difference of vegetation index（NDVI and RVI） during the period of hail disaster in current year, and the difference between vegetation index in current year and the mean of vegetation index in the last three years. Based on that, the segmentation threshold of normal winter wheat, severe hail disater winter wheat,moderate hail disater winter wheat and mild hail disater winter wheat are calculated by the sample, the spatial distribution information of different grades hail diaster is identified. And then, the spatial distribution accuracy of different grades hail diaster is verified in comparison with the actual survey results. The results show that： ① It is effective to extract the spatial distribution information of different grades hail diaster using NDVI difference, and the accuracy agrees well with the actual survey results, it is inefficient to extract the spatial distribution information of the moderate hail disater and the mild hail disater using RVI difference. ② The spatial distribution area of hail diaster occurs from the northwest to the southeast,and is mainly located at low mountain plain area below 5° slope. ③ The spatial distribution area is separated into three annular areas, namely the inner annular area of severe hail disater, the middle annular area of moderate hail disater and the outer annular area of mild hail disater.