Assessing the Relative Role of Climate and Human Activities on Vegetation Cover Changes in the Up-down Stream of Danjiangkou, China


 Danjiangkou Reservoir is water source of Middle Route Project of the South-to-North Water Diversion (SNWD) Project, research on the dynamic changes in vegetation cover and its influencing factors is of great significance for understanding the ecological environment of the water diversion area and formulating protection measures. In this study, The normalized difference vegetation index (NDVI) was used to analyze the dynamic changes and influencing factors of vegetation in the upstream and downstream of Danjiangkou Dam from 1982 to 2018. The results show that the NDVI shows an upward trend of 0.017 year-1 (P < 0.05), and the significantly increased area is located near the valley upstream of the dam, while the significantly decreased area is mainly distributed in the basin downstream of the dam and around the central city. The comprehensive contributions of climate and human activity factors to NDVI changes were 92.03% and 7.97%, respectively. The human activities in the upstream of the dam are mainly reflected in the ecological measures such as returning farmland to forest; the human activities at the downstream of the dam are mainly reflected in urban expansion, occupation of cultivated land and forest land by construction land.


22
As climate change and human activities continue to intensify, the global natural 23 ecosystem has been seriously threatened1. Vegetation constitutes a major part of to monitor vegetation cover change 10 . NDVI 11 has a linear or near-linear relationship 35 with green leaf density, photosynthetically active radiation, vegetation productivity 36 and cumulative biomass, which is recognized as an effective index to reflect the 37 vegetation coverage and growth status on a large scale [12][13][14] . With increasing concern 38 over global climate change, there is much interest in the relationship between 39 terrestrial ecosystems and global climate change [15][16] . According to previous studies,    The meteorological data are based on site data pertaining to the monthly mean    The equation of the significance test was as follows: In the equation above, t is the test statistic, n is the number of samples, and K is 222 the number of control variables (1 here). In this study, we predict the impact of human activities by establishing 232 regression relationships among precipitation, temperature and NDVI. The expression 233 of residual analysis is as follows: where α and β are regression coefficients of the NDVI for precipitation and air 237 temperature; λ is the regression constant term; P and T represent precipitation and 238 temperature, respectively; NDVIreal is the true NDVI value; NDVIpre is the 239 predicted NDVI value; ε>0 indicates that human activities have a positive impact; ε < 240 0 indicates that human activities have a negative impact; and ε = 0 indicates that 241 human activities have a slight impact. The specific steps are as follows:

242
(1) Use the least squares method to calculate slopes α and β.

243
(2) Calculate regression constant λ using the multiyear average NDVI, 244 precipitation, and temperature.  contribution of human activities to the NDVI is: The overall contribution of natural factors to the NDVI is:

256
Long-term NDVI data 257 Based on MODIS NDVI observed monthly data and predicted 1 km resolution EOT  heterogeneous, but the main NDVI change trend was an increase (Fig. 4). A total of 277 75.88% of the pixels showed an increasing trend (68.54% significantly with p < 0.05),    The annual precipitation and annual mean temperature values were spatially 303 heterogeneous (Fig. 6). The annual average temperature decreases gradually from 304 south to north, the highest annual average temperature is located in the southeast, and 305 the lowest is located in the west and north. The annual precipitation also shows a 306 decreasing trend from south to north, the highest annual precipitation is located in the 307 southeast and the southwest, and the lowest is located in the northern.

315
The correlation of annual NDVI, annual climate data and monthly NDVI and 316 monthly climate data were analyzed. The results show that the annual NDVI has no 317 significant positive correlation with annual temperature and precipitation, and the 318 correlation coefficients are 0.125 and 0.320, respectively. There was a significant 319 positive correlation between monthly NDVI and monthly average temperature, with a 320 correlation coefficient of 0.708 (P < 0.05). There was no significant negative 321 correlation between monthly NDVI and monthly average precipitation, with a 322 correlation coefficient of 0.199. Overall, the influence of temperature on NDVI is 323 greater than that of precipitation.

324
Based on the pixel scale, partial correlation coefficients of monthly average 325 temperature, monthly precipitation and monthly average NDVI were calculated and 326 tested for significance (Fig. 7). The results show that 95.78% of the pixel temperature 327 has a positive correlation with NDVI (81.90% significantly with p < 0.05), and 4.22%    343 Based on regression analysis of NDVI and precipitation and temperature, the 344 residuals of the NDVI are calculated on the pixel scale (Fig. 8). Positive values 345 indicate that other factors have a positive effect on regional NDVI changes, while      458 Between 1982 and 2018, the precipitation showed a decreasing trend, while the 459 temperature showed an increasing trend, indicating that the regional climate gradually 460 became warmer and drier. There is a significant positive correlation between regional 461 temperature and NDVI and a negative correlation between precipitation and NDVI.

462
Therefore, the climate change trend is conducive to the growth of regional vegetation.

463
The overall growth trend of vegetation in the study period is closely related to the 464 climate change trend.

465
The study area has a developed water system, rich water resources, relatively have gradually become an important factor affecting regional vegetation changes, and 506 the degree of influence has gradually increased.

508
Climate change is a long-term process, and the length of meteorological and 509 hydrological data used for analysis to a large extent determines the trend results. At

571
(2) The influence of temperature on NDVI is greater than that of precipitation, which 572 is the main climatic factor affecting NDVI change. The influence of temperature and 573 precipitation on NDVI shows obvious spatial heterogeneity. In the upstream valley 574 area and downstream farming area, the correlation among NDVI, precipitation and 575 temperature is significant; in the basin area with higher terrain in the upstream and 576 downstream, the correlation among NDVI, precipitation and temperature is not increasing trend in recent years and has become an important factor affecting NDVI.

592
Data availability: The data used in current study are available from the corresponding author 593 upon reasonable request. 594