*2.1. Processing*

The flow chart of research ideas for this paper is as follows (Figure 1). First, we calculated the plant phenology according to the following steps: (1) the NDVI of the TRHR was calculated from MOD09A1 datasets in Google Earth Engine; (2) next, bare soil, sparse vegetation, and evergreen forest pixels were eliminated according to certain requirements; (3) then, the NDVI datasets were smoothed by harmonic analysis of time series (HANTS); (4) we used relative and absolute rates of change to calculate plant phenology (SOS and EOS) based on the NDVI datasets smoothed by HANTS; (5) the phenological data obtained by remote sensing monitoring were verified by using the observation data of phenological stations. Then, we analyzed the spatiotemporal dynamic pattern of plant phenology on different types of terrain and basins. Finally, we explored the potential influence mechanism of climate and soil factors on the phenology of the TRHR based on the structural equation model (SEM) and Pearson correlation coefficients.

**Figure 1.** Flow chart of research ideas for this paper. NDVI, HANTS, SOS, EOS, LOS, DEM, MMT, MMP, MMH, MMR, MMST, MMSM, pH, and TN indicate the normalized difference vegetation index, harmonic analysis of time series, start of the growing season, end of the growing season, length of the growing season, digital elevation model, monthly mean temperature, monthly mean precipitation, monthly mean relative humidity, monthly mean shortwave radiation, monthly mean soil temperature, monthly mean soil moisture, pH (H2O), and total N, respectively.
