4.2 Discussion of different vegetation types
Here, we compared vegetation restoration and degradation on the MP, and found that the relative intensity of vegetation change in the three subregions was: FSR > GSR > DSR. For comparison, we considered changes in different types of vegetation cover from previous research over the same time period. Research that focused on the grassland and forest regions of East Asia showed that the mean value of carbon use efficiency (CUE) was grassland > farmland > forest > shrub during the period 2000–2013 in China (Liu et al., 2019). That is, the highest CUE was found in grassland, and the increase of CUE was caused mainly by increases in precipitation, while increases in temperature inhibited the vegetation CUE. Liu et al. (2020) studied the response of the NDVI on different types of vegetation cover influenced by climate change on the Loess Plateau of China. Their results showed that the NDVI of different vegetation covers does not correspond to the size of the occupied area, and that the increasing trend of the NDVI in the forested areas was greater than that in the grassland areas (the grass area is larger than forested area), which is similar to the conclusion of this study.
Other spatiotemporal studies reported different results. De et al. (2013) described the relationship between vegetation and climate over the period between 1981 and 2008 based on meteorological indicators, and found that climate can explain nearly 54% of vegetation change.
The impact of vegetation change on climate is concentrated mainly in areas of desertification and surrounding areas. Charney (1977, 2006) explored the possible impact of vegetation dynamics on climate, and pointed out that desertification increases the surface albedo, reduces the soil moisture content, and reduces the surface roughness. This leads to reduced precipitation, which would be followed by deterioration of the vegetation and soil. Vegetation recovery becomes more difficult under these conditions (Nicholson et al., 1988), and these impacts can then extend to the surrounding areas (Xue, 1993, 1996). The expansion of a desert region may strengthen regional climate change, aggravating land degradation. Grassland has a greater ecological elasticity to climate change compared with desert areas (John et al., 2013). Peng et al. (2012) proposed that, compared with forest ecosystems in humid areas, grassland ecosystems in semi-arid and arid areas are more sensitive to drought. Desert grassland is more sensitive to drought compared with typical grassland and meadow grassland (Bailing et al., 2018). Therefore, there may be a negative feedback relationship between vegetation change and climate in desert areas. Here, we observed that the DSR has the slowest observed vegetation restoration.
Bao et al. (2015) studied vegetation dynamics and its response to regional climate change in Mongolia, and showed that continuous warming is likely to be the main factor driving the reversal of the NDVI trend. Furthermore, climate-related vegetation reduction and related potential desertification may aggravate other environmental problems in East Asia (e.g., sandstorms).
The vegetation restoration associated with human activity in the desert is inseparable from vegetation restoration measures and projects formulated by local governments. We list the policy measures and their impacts on regional vegetation restoration after 2000 in Table 4.
Therefore, we conclude that the desert is the region of the MP most in need of vegetation restoration, although it is also currently the area with the slowest rate of vegetation restoration. Restoration of desert areas is effective, and the desert should be a high priority area for improving vegetation restoration.

5 Conclusions

Our evaluation of vegetation dynamics and the associated driving factors across the MP and its subregions has revealed the following. The proportion of each subregion that experienced vegetation restoration during the study period was greatest in the FSR, followed by the GSR, with the DSR showing the least restoration. From the perspective of driving factors, overall, climate change was conducive to vegetation restoration, whereas the amounts of vegetation restoration and degradation caused by human activity were relatively similar. From the regional perspective, the extent of restoration caused by climate change in the FSR and GSR was far greater than that caused by human activity, but in the DSR, restoration by human activity exceeded that related to climate change. Restoration in the DSR was slower than that in the FSR and GSR. As a result, the DSR should be a priority area for policy and investment. Human activity was the main driving force behind vegetation restoration in the DSR, and thus the restoration effect of human activity is still effective in this region. China’s government should use “the Belt and Road Initiative” (Liu, 2015) to cooperate and establish links with Mongolia and other countries to assist desertification control and curb regional land degradation.
Acknowledgments: This research was supported by National Key R&D Program of China (No. 2018YFC0806900).
Conflicts of Interest: The authors declare no conflict of interests.