泾河上游固原地区的NDVI变化与降水的相关性研究

The Relationship between NDVI Change and Precipitation in Guyuan City

我国自2000年开始实行退耕还林还草政策。本文使用2000年与2004年250m分辨率的MODIS植被指数数据（NDVI），分析了泾河上游固原地区植被和降水利用效率变化，结果表明该区植被覆盖情况有较大好转，NDVI在0．2～0．4之间的地区减少了31％，而NDVI在0．4～0．6的地区增加了67％；同时该地区的降水利用效率也有所增加，从平均0．53增加到0．65。原州、彭阳和西吉3县NDVI与降水的相关系数分别从0．27、0．58、0．50增加到0．83、0．92和0．95。研究认为这种增加主要是由于降水在年内分布与植被生长配合较好，而不是由于退耕还林政策所致。分析表明当生长季内降水小于450mm时，NDVI随降水增加而增加较慢，当降水大于450mm时．NDVI增加较快。生长季降水量小于450mm的地区以草地为主，大于450mm的地区以林地为主。所以生长季450mm降水可以看作是降水影响森林和草地一个临界点，在这一阈值之上退耕还林政策实施应以还林为主。在这一阈值之下以还草为主。

The relationship between precipitation and vegetation is an important part in the field of earth surface processes. Here MODIS （Moderate-resolution Imaging Spectroradiometer） NDVI time series data with 250m resolution was used to study the NDVI change from 2000 to 2004 in Guyuan city which lies in west semiarid China. Grain for green policy was carried out in this area from 2000 to reduce the severe soil erosion and improve the ecological environment. The dada showed that NDVI between 0.2 and 0.4 decreased 31 percent and NDVI between 0.4 and 0.6 increased 67 percent. This change indicated an improvement of vegetation cover. Rainfall data derived from six weather stations in Guyuan City was used to calculate rainfall use efficiency and corelation coefficient between NDVI and rainfall. The result showed that the average rainfall use efficiency increased from 0.53 to 0.65. The corelation coefficient between NDVI and rainfall increased from 0.27, 0.58, 0.50 to 0.83, 0.92,0.95 in Yuanzhou, Pengyang and Xiji county respectively, while Jingyuan and Longde county showed a small change. It was considered that this increase was mainly due to the accordance between rainfall and NDVI, not grain for green policy. This implied that the variance of annual rainfall is bigger in semiarid areas than that in semiwet areas and the vegetation in semiwet areas has a higher stability than that in semiarid areas. Here it is also found that when the rainfall in growing season is less than 450ram, NDVI increased slowly with rainfall, while when the rainfall is greater than 450ram it is observed that NDVI increased quickly. The main vegetation is grass in areas with less than 450mm rainfall, and forest in areas with bigger than 450mm rainfall. So 450mm annual rainfall can be considered as a threshold to decide the suitability of forest and grass plantation in this area. The nonlinearity between forest cover and rainfall is based on NDVI, if a new vegetation index such as NDWI or EVI is used, maybe this nonlinearity will changed. So more research is still needed to verify.