基于多时相遥感信息可以有效反映区域尺度上作物物候期特征的原理,利用1998—2009年期间的SPOT/VGT NDVI逐旬时间序列数据,采用TIMESAT软件集成的非对称高斯函数拟合方法对数据序列进行平滑重构处理,进而根据拟合曲线变化特征定义并提取了我国东北地区耕地生长季特征参数(包括生长季开始期、峰值期、结束期与生长季长度等)。然后,利用农业物候观测数据(包括出苗期、抽穗期、成熟期与生育期天数等)对上述生长季特征进行了对比验证研究。最后,重点分析了近10年来东北地区耕地典型物候期的空间格局及其时间波动特征。结果表明:基于观测数据的作物物候期特征的时间变化曲线,可以一定程度上反映不同作物类型的生长过程;基于拟合数据提取的耕地生长季特征参数与不同作物的观测物候期均存在显著的相关关系,相对而言,提取的耕地生长季特征能更好地反映区域尺度上作物物候期特征的宏观时空分异。此外,空间网格化的物候特征分析结果还可为研究作物生长过程对外界环境条件变化(诸如区域温度、降水和日照时间等)的时空响应提供重要的数据支持。
The authors investigated spatio-temporal patterns of seasonality parameters of crop growing season in Northeast China,by using the SPOT/VGT NDVI ten day composed time-series data collected from 1998 to 2009.First,to minimize the effects of anomalous values caused by atmospheric haze and cloud contamination,the software TIMESAT was used to generate smooth time series of NDVI based on an asymmetric Gaussian function;second,the seasonality parameters,such as the start date,the end date,the peak date,and length of the growing season,were defined and extracted from the smoothed NVDI time-series dataset;third,each of the extracted parameters and the observed agricultural phenophases(including the stages of seedling,heading and maturity for harvest,length of growth period) were compared and validated by using a scatter plot,respectively;finally,the temporal trends and spatial patterns of the major crop seasonality parameters in Northeast China were illustrated and analyzed over the past 10 years.The results show that the growth process of major crops can be discriminated to a certain extent from the temporal trend of observed crop phenological characteristics.Obvious linear correlations can be found between the extracted seasonality parameters and the observed crop phenophases,which indicates that spatio-temporal variations of crop phenophase can be expressed in details by utilizing the extracted parameters from the smoothed NDVI time series.Meanwhile,all these grid-based crop phenophases can be used as data alternative for studying the spatio-temporal responses of crop growth process caused from fluctuation in external environmental conditions,such as air temperature,precipitation and daylight hours,etc.