3种不同数据源NDVI的比较分析及其在延河流域的应用研究

Comparison of 3 NDVI Datasets and the Application at Yanhe Basin,China

选取GIMMS、SPOT VGT和MODIS NDVI数据重叠期(2000—2006年)的数据,运用一致性检验和相关分析等方法研究了3种NDVI在延河流域植被变化分析中的一致性与差异性.研究表明,SPOT VGT NDVI和MODIS NDVI的相关性、一致性较强,GIMMS NDVI与前两者存在较大差异.考虑到GIMMS NDVI在较长时间植被变化分析中具有不可替代性,利用其与SPOT VGT NDVI在7、9月份存在较好的一致性这一现象,提出了GIMMS NDVI的修正方法,并将该方法应用于延河流域、延河子流域以及流域内的像元尺度.修正结果显示,在植被覆盖度较低的情形下,该方法对不同空间尺度的GIMMS NDVI均可进行有效的修正,修正后的GIMMS NDVI与SPOT VGT NDVI和MODIS NDVI的相关性显著,可据此进行GIMMS NDVI时间序列的扩展;但在流域植被覆盖较好的地区,修正结果仍然存在较大的误差.

The paper compared consistency and difference among the NDVI datasets of GIMMS, SPOT VGT and MODIS in Yanhe Basin, China. The available overlapping data in 2000--2006 were used to compare their consistency. The comparison was done through consistency and correlation coefficients. It was shown that the SPOT VGT NDVI was significantly correlated with the MODIS NDVI. However,there were some differences between the GIMMS NDVI and SPOT VGT/MODIS NDVI. As the only dataset which contains more than 30 years data, the GIMMS NDVI is irreplaceable in analyzing the long-term should be regulated in accordance with reflecting the vegetation coverage. It was change of the vegetation. Thus, the GIMMS NDVI the other two, which are considered more precise in found that GIMMS NDVI was more closely correlated with SPOT VGT NDVI in July and September than in August in Yanhe River Basin. Therefore, a revision method of the GIMMS NDVI was proposed and was applied in the Yanhe Basin at different spatial scales, which were the whole basin, the sub-basin, and the pixel scale. The results indicate that the GIMMS NDVI of different while the vegetation coverage is low, and the revised spatial scales could be revised effectively GIMMS NDVI shows good consistency with the SPOT VGT NDVI and MODIS NDVI. However, the revise method doesn＇ t work very well in the high vegetation conditions, which required further research.