基于卫星遥感数据的黄淮海地区植被覆盖时空变化特征

Spatiotemporal variations of vegetation cover in Huang-Huai-Hai area based on remote sensing data

利用1982—2003年GIMMS NDVI遥感数据,在进行合成、重采样和时间序列滤波处理的基础上,采用线性趋势分析、经验正交函数分解等方法,对中国黄淮海地区植被覆盖的时空特征进行了研究。结果表明:22年来黄淮海地区植被覆盖总体上呈略微增加的趋势,且该区域生长季有提前和延长的趋势;黄淮海大部分地区植被活动在增强的同时,局部地区出现了植被退化现象;从季节变化上看,春季上升和夏季下降趋势明显;林地为主的自然植被、草甸类自然植被和所有农业植被未变化类别占主导地位,而草原植被则以增加趋势为主。主要生长季的NDVI距平EOF分析表明,第1模态的主要特征是区域中间、北部和南端为正,四周为负变化;第2模态的主要特征是从东南向西北由正到负变化且正值区明显偏多;第3模态从东南向西北呈现"正-负-正"的空间分布,其中负值区大部分为以林地为主的自然植被区和一年一熟农业植被,正值区大部分为农耕区和草原牧区,该模态大致反映了农牧区和林区的NDVI分布型。

Based on the 1982-2003 GIMMS NDVI data after processed through synthesis, re-sampling, and time-series filtering, and by using the analysis methods of linear trend and empirical orthogonal function （EOF）, this paper studied the spatiotemporal variations of vegetation cover in Huang-Huai-Hai area. In the past 22 years, the vegetation NDVI in the whole area had a slightly increasing trend, and the growth season tended to be advanced and prolonged. In most part of the area, vegetation activity enhanced; but in local parts of it, there existed vegetation degradation. The vegetation NDVI indicated an obvious increasing trend in spring but an obvious decreasing trend in summer. According to the dynamic trend of pixel number for different vegetation types, the unchanged types were dominated by forest, shrub, and agricultural vegetations, while grassland vegetation presented an increasing trend. The EOF analysis for NDVI anomaly during main growth seasons showed that the first mode was characterized by positive variation in the central, south, and north regions but negative variation in edging regions, the second mode was characterized by the variation from positive to negative and dominated by positive variation in the direction from southeast to northwest regions, and the third mode was characterized by the variation of positive-negative-positive in the direction from southeast to northwest regions, in which, the regions with negative variation were dominated by forest and single-season agricultural vegetation, while those with positive variation were dominated by cropland and grassland. The third mode approximately reflected the distribution type of NDVI in the farming-pasturing and forest regions of the Huang-Huai-Hai area.