太湖流域1995－2010年耕地复种指数时空变化遥感分析

为了重建太湖流域1995-2010年耕地复种指数时空变化过程,结合1995年NOAA-AVHRR的旬NDVI数据、2000-2010年250m16d合成的MODIS-NDVI数据和4期土地利用数据,利用傅里叶变换技术的HANTS方法重建农作物生长NDVI曲线,结合差分算法提取各时相复种指数。结果表明:太湖流域1995-2010年复种指数时空格局呈现以一年两熟为主,但一年一熟比例呈持续增长趋势,复种指数局部格局变化主要集中分布在上海市和浙江省;太湖流域复种指数呈北高南低的空间格局,1995-2010年总体呈较明显的下降趋势,从1995年189.4%持续下降到2010年167.3%,降幅从大到小依次为浙江省、上海市和江苏省;基于遥感和HANTS方法提取太湖流域复种指数在栅格尺度上精度达94.6%;受计算范畴差异及遥感数据空间分辨率影响,在县市行政区划单元尺度上,遥感方法与统计数据计算的复种指数平均误差约15.8%～21.6%。研究结果可为太湖流域农业政策制定及管理提供科学依据。

Spatial and temporal changes of multiple cropping index in 1995-2010 in Taihu Lake basin, China

Abstract: The smooth crops growth NDVI curves from 1995-2010 were rebuilt by the HANTS method of Fourier transform techniques, with the combination of 1 km 10 d NOAA-AVHRR time-series NDVI data in 1995 with the resolution of 1 km, 16 d MODIS time-series NDVI data from 2000 to 2010 with the resolution of 250 m, and land use data in 1995, 2000, 2005 and 2010. Spatial and temporal changes of multiple cropping index in 1995-2010 in Taihu Lake basin were extracted by a difference algorithm, and its pattern was analyzed. The results showed that spatial and temporal pattern of multiple cropping index from 1995 to 2010 in Taihu Lake basin was dominated with the double cropping system, while the proportion for the single cropping system was at the increasing trend. Local changes for multiple cropping index with decreasing or increasing trend mainly concentrated in Shanghai Municipality and Zhejiang province. Spatial pattern of multiple cropping index in Taihu Lake basin presented downward trend from north to south, and average multiple cropping index was generally at the obviously decreasing trend, decreasing from 189.4% in 1995 to 167.3% in 2010. The decreasing extent of multiple cropping index from 1995 to 2010 at the county-level for Zhejiang province and Shanghai Municipality was larger than that in Jiangsu province. The combination of the HANTS method and long time-series NDVI remote sensing data to extract multiple cropping index in Taihu Lake basin has high accuracy with 94.6% at the grid scale. Due to the differences in calculation categories and spatial resolution of remote sensing data, the error between the multiple cropping index extracted by the remote sensing method and the statistics at the county-level ranged from 15.8% to 21.6%. The results can provide a scientific basis for policy-making and management of agriculture in Taihu Lake basin.