*2.2. Data Collection and Processing*

#### 2.2.1. NDVI Data

Landsat 5, Landsat 7, Landsat 8 TM image data has been collected by the Google Earth Engine. The time resolution is 16 days and the spatial resolution is 30 m. The Maximum Value Composite (MSV) method [38] was used to generate the annual Maximum NDVI Value from 1986 to 2020. This study taken NDVI as an indicator and used the current situation vector data of GH highway to generate a 4 km buffer zone (hereafter referred to as the core area) in ArcGIS to evaluate the direct impact of human activities on the environment of GH highway. To evaluate the indirect effects of human activities on the environment of GH highway. We generated a 4 km buffer zone (hereafter referred to as the contrast area) in the periphery of the core area. The NDVI of the core area was divided into four buffer zones according to 1 km, and the direct impact of GH highway on the NDVI of human activities along the highway was evaluated.

#### 2.2.2. Meteorological Data

Raster data of temperature and precipitation at 1 km resolution were used. The temperature and precipitation from 1986 to 2015 were collected from the Data Center of Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences (https://www.resdc.cn/, accessed on 1 October 2022). The data of annual precipitation and annual temperature from 2016 to 2020 are from the monthly scale data provided by China Meteorological Administration (http://data.cma.cn/, accessed on 1 October 2022). The meteorological interpolation software ANUSPLIN was used to interpolate the temperature and precipitation data with a spatial resolution of 30 m. The digital elevation map (DEM) was introduced in the interpolation process to reduce the effect of topography on climate, thus minimizing the interpolation error and greatly improving the accuracy compared with other interpolation methods [39], which is more suitable for the analysis of meteorological elements of time series [40]. The above monthly synthesized temperature and precipitation data are extracted by ArcGIS to synthesize the annual average precipitation and annual average temperature data, and resampling into 30 m in ArcGIS.

#### 2.2.3. Terrestrial Surface Data

The DEM digital elevation model is derived from the Geospatial Data Cloud (http:// www.gscloud.cn/, accessed on 1 October 2022) with a spatial resolution of 30 m. LUCC data from 2000 to 2020 are obtained from the Globalland30 (http://www.globallandcover.com/, accessed on 1 October 2022), the spatial resolution is 30 m, the dataset includes ten types of land cover [41]. Wetlands are reclassified as water bodies due to its small area in the study region.
