Wang Zhi-wei, Wu Xiao-dong, Yue Guang-yang, Zhao Lin, Wang Qian, Nan Zhuo-tong, Qin Yu, Wu Tong-hua, Shi Jian-zong, Zou De-fu
Guang Pu Xue Yu Guang Pu Fen Xi. 2016 Feb;36(2):471-7.
Recently considerable researches have focused on monitoring vegetation changes because of its important role in regula- ting the terrestrial carbon cycle and the climate system. There were the largest areas with high-altitudes in the Qinghai-Tibet Plateau (QTP), which is often referred to as the third pole of the world. And vegetation in this region is significantly sensitive to the global warming. Meanwhile NDVI dataset was one of the most useful tools to monitor the vegetation activity with high spatial and temporal resolution, which is a normalized transform of the near-infrared radiation (NIR) to red reflectance ratio. Therefore, an extended GIMMS NDVI dataset from 1982-2006 to 1982-2014 was presented using a unary linear regression by MODIS dataset from 2000 to 2014 in QTP. Compared with previous researches, the accuracy of the extended NDVI dataset was improved again with consideration the residuals derived from scale transformation. So the model of extend NDVI dataset could be a new method to integrate different NDVI products. With the extended NDVI dataset, we found that in growing season there was a statistically significant increase (0.000 4 yr⁻¹, r² = 0.585 9, p < 0.001) in QTP from 1982 to 2014. During the study pe- riod, the trends of NDVI were significantly increased in spring (0.000 5 yr⁻¹, r² = 0.295 4, p = 0.001), summer (0.000 3 yr⁻¹, r² = 0.105 3, p = 0.065) and autumn respectively (0.000 6 yr⁻¹, r² = 0.436 7, p < 0.001). Due to the increased vegeta- tion activity in Qinghai-Tibet Plateau from 1982 to 2014, the magnitude of carbon sink was accumulated in this region also at this same period. Then the data of temperature and precipitation was used to explore the reason of vegetation changed. Although the trends of them are both increased, the correlation between NDVI and temperature is higher than precipitation in vegetation grow- ing season, spring, summer and autumn. Furthermore, there is significant spatial heterogeneity of the changing trends for ND- VI, temperature and precipitation at Qinghai-Tibet Plateau scale.
近年来，由于植被在调节陆地碳循环和气候系统中发挥着重要作用，大量研究聚焦于监测植被变化。青藏高原是世界屋脊，拥有最大面积的高海拔地区，该地区植被对全球变暖极为敏感。同时，归一化植被指数（NDVI）数据集是监测植被活动的最有效工具之一，它是近红外辐射（NIR）与红光反射率之比的归一化变换。因此，利用2000 - 2014年MODIS数据集通过一元线性回归得到了青藏高原地区从1982 - 2006年到1982 - 2014年的扩展GIMMS NDVI数据集。与以往研究相比，考虑尺度转换产生的残差后，扩展NDVI数据集的精度再次提高。所以扩展NDVI数据集模型可能是整合不同NDVI产品的新方法。利用扩展NDVI数据集，我们发现1982 - 2014年青藏高原生长季植被呈显著增加趋势（0.000 4 yr⁻¹，r² = 0.585 9，p < 0.001）。研究期间，春季（0.000 5 yr⁻¹，r² = 0.295 4，p = 0.001）、夏季（0.000 3 yr⁻¹，r² = 0.105 3，p = 0.065）和秋季（0.000 6 yr⁻¹，r² = 0.436 7，p < 0.001）的NDVI均呈显著增加趋势。由于1982 - 2014年青藏高原植被活动增强，该地区同期碳汇量也在增加。然后利用气温和降水数据探究植被变化的原因。虽然二者均呈上升趋势，但在植被生长季的春季、夏季和秋季，NDVI与气温的相关性高于与降水的相关性。此外，青藏高原尺度上NDVI、气温和降水变化趋势存在显著的空间异质性。
