**6. Conclusions**

In this study, six widely used drought indices (PCI, VCI, TCI, VHI, SDCI, and TVDI) were used for monitoring drought in the GCM from the viewpoints of temperature, precipitation, and vegetation condition. Unlike the remote sensing products such as LST, precipitation, and vegetation indices, which provide absolute information, the aforementioned indices are computed for a certain area based on one or more of the remote sensing products, to reflect drought conditions. The spatiotemporal variations of drought were examined using the annual series of these indices from 2001 to 2018. The drought trends may be different for different indices. This indicates that the applicability of different indices differs with the location within the study area itself. Both PCI and SDCI show similar patterns, indicating that the central region of the GCM is getting wetter and the southwestern area is getting drier. VCI and VHI exhibit similar patterns with more drying trends. The correlations between these drought indices and meteorological factors were discussed to reveal that different indices are affected differently by precipitation and temperature variations. This is because these indices focus on different aspects of drought causes and symptoms, namely, precipitation, LST, and vegetation health. A comparison of the time series of the indices with precipitation and temperature showed that some drought indices cannot be explained by meteorological observations probably because of the time lag between meteorological drought and vegetation response. In particular, VHI and SDCI were generally employed for agricultural drought monitoring based on empirical weights. Note that the weights of these indices are adjustable. They may help us gain a better judgment about the drought conditions in di fferent study areas. For this purpose, in situ observations can be employed in the future.

An examination of the slope of changes in drought for the di fferent land cover types showed that the evergreen needleleaf forests in the GCM experienced increasingly severe drought conditions in recent years. Similar patterns with weaker changes were obtained for savannas and deciduous needleleaf forests. VCI and VHI exhibited similar variation patterns across di fferent land cover types. The slopes of VCI and VHI for all land cover types are negative, indicating an overall drying trend. Note that the trend of vegetation-based indices may be a ffected by the long-term physical changes in vegetation.

The terrain is regarded an important factor for drought conditions. Based on the statistical analysis of drought patterns in di fferent landforms, it was found that although di fferent indices indicate distinct drought conditions, the relative drought situation of di fferent landforms is consistent regardless of the index. This implies that the landform type may be important ancillary information for drought monitoring.

Note that many drought indices and drought-monitoring methods were not considered in this study. From the perspective of drought consequences, soil moisture is a direct indicator for drought and has been adopted in several drought indices. Our future work is to adopt the indices related to soil moisture for a more comprehensive study in order to propose the most appropriate method for monitoring drought in the GCM. Nevertheless, the results of this study so far have preliminarily demonstrated the convenience of using remote sensing product-based indices for drought monitoring in the GCM as well as the di fferences between these indices. The results are expected to provide guidance for drought monitoring to help understand and monitor the ecosystem conditions and the environment in this region.

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

**Funding:** This research was supported by the Fundamental Research Funds for the Central Universities (2412018ZD012); National Natural Science Foundation of China (41301364; 41630749).

**Acknowledgments:** The authors are grateful to three anonymous reviewers for their helpful comments. We also want to thank Y. Wang, H. Du, S. Liu and Z. Liu for the constructive suggestions.

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