**5. Conclusions**

Based on the 250, 500, and 1000 m MODIS data during the period 2001–2020 and the ~10 km GIMMS3g data during the period 1982–2015, as well as the rate of change in the curvature of the logistic models, this study investigated the applicability and spatial scaling effects of various remote sensing products with different spatial resolutions on phenology extraction in a complex-terrain region. Our study showed that the MODIS products performed better in phenology analysis across the Loess Plateau, and the phenology results that were derived from the different MODIS products showed only small differences. However, the GIMMS3g-based SOSs that were derived from logistic models had a good performance in the flat region (i.e., the Weihe Plain) but a poor performance in regions with a more heterogeneous topography. Additionally, the performances of the GIMMS3gbased EOSs across the whole Loess Plateau were poor. The phenology results that were derived from the MODIS data presented advanced SOS trends and delayed EOS trends during the period 2001–2020 for the entire Loess Plateau. However, both the SOS and EOS trends that were identified by the GIMMS3g products were the opposite. Our finding emphasized that the 1000 m MODIS product can be used to extract phenology from areas with a complex terrain, such as the Loess Plateau, and almost no difference was found in the phenology extraction between the 500 m MODIS product and the 250 m MODIS products. Furthermore, we also investigated the effects of vegetation and AT10 on the spatiotemporal variability of vegetation phenology, which could help us to understand the driving factors of such phenological changes in the future.

**Author Contributions:** Conceptualization and methodology, F.C., Z.L. and S.W.; software, F.C.; validation, F.C.; formal analysis, F.C. and H.Z.; investigation, F.C. and H.Z.; resources, F.C., Z.L. and S.W.; data curation, F.C.; writing—original draft preparation, F.C.; writing—review and editing, Z.L. and S.W.; visualization, Z.L. and F.C.; supervision, Z.L.; project administration, Z.L.; funding acquisition, Z.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the National Key Research and Development Program of China (grant no. 2017YFC0504701), National Natural Science Foundation of China (grant nos. 41971218 and 41601582), and the Strategic Priority Research Program of the Chinese Academy of Sciences (grant nos. XDA23070302 and XDA28130400).

**Data Availability Statement:** The data that we used in this study can be requested by contacting the corresponding author.

**Acknowledgments:** We appreciate three anonymous reviewers and their valuable suggestions and comments.

**Conflicts of Interest:** The authors declare no conflict of interest.
