Key driving forces of desertification in the Mu Us Desert, China

The temporal trends and key driving forces of desertification in the Mu Us Desert are representatives of most arid regions of Asia with a high risk of desertification. We analyzed the significance of Aeolian transport on desertification in the Mu Us Desert by field investigations, sampling, wind tunnel experiments, particle size and nutrient measurements, and statistics on aeolian transport potentials. The results showed that high intensities of aeolian processes may result in low differences in aeolian transport despite differences in the underlying sediments. When high desertification occurred in the 1970s, the annual losses of the ammonium N, nitrate N, available K, and available P were approximately 116, 312, 46,436, and 1,251 kg km−2, respectively. After 2010, the losses were only 8, 20, 3,208, and 84 kg km−2, which were generally only 6.7% of those in the 1970s. The results showed that although human activity may trigger desertification, the dramatic decline of aeolian transport and low nutrient loss may be the key driving forces for the occurrence of rehabilitation in this region.


Results and Discussion
Variations in particle sizes of surface soils under aeolian processes. Aeolian processes result in great variations in the components of surface soils. Before and after the wind tunnel experiments, little difference in the contents of the fine fraction (<50 µm in diameter) was observed. However, relatively coarser fractions were left after the experiments (Fig. 2). Contents of fractions with diameters ranging between 100 and 250 µm and of >250 µm in surface soils were higher by approximately 2.5% and 10% after aeolian processes, respectively. The  results show that aeolian processes may coarsen the surface soils, leading to nutrient loss, and may decrease the water-holding capacity of surface soils.
Temporal variations in aeolian transport. The average aeolian transport of 75 samples collected in 15 sites is shown in Fig. 3A. The results showed that the average aeolian transports varied from 0.01 to 28.71 g m −2 s −1 , with a coefficient of variation (CV) of 0.37, under wind velocities ranging from 8 to 22 m s −1 . The CV in aeolian transport among the samples decreased with the increase in wind velocity (Fig. 3B). This result indicates that the effects of variation in land use, degradation degrees, and soil properties on aeolian transport under high wind velocities are less than that under low wind velocities. High intensities of aeolian transport may have similar effects on desertification, despite the differences in spatial and temporal variations of the components of the underlying soils in the region.
With the occurrence of aeolian processes, nutrients are lost, mobile dunes and sand sheets develop on the surface, the soil fertility and biomass decrease, and desertification occurs. Wind tunnel experiments and statistical results showed that there were also obvious temporal variations for aeolian transport in the Mu Us Desert (S4). For example, in the 1970s the intensity of aeolian transport was approximately 137 × 10 4 T km −2 , whereas the value was only 9 × 10 4 ton km −2 from 2011 to the present, representing only 6.6% of that in the 1970s (Fig. 4). The dramatic decline of aeolian transport from 2011 to the present showed that there was little development of mobile dunes and sand sheets, which consequently benefited rehabilitation in this region.

Nutrient loss intensities under different wind velocities.
The average contents of ammonium N, nitrate N, available K and available P in the surface soils were 0.08 (±0.05, SD), 0.25 (±0.52), 35.18 (±20.26), and 0.95 (±0.46) mg kg −1 , respectively. There were no significant correlations between the nutrient contents and contents of different particle size fractions, except for the 50 ~ 100 µm and 200 ~ 250 µm fractions that correlated with the contents of available K and available P, respectively (S4). Under a wind velocity of 8 ~ 22 m s −1 , the nutrient loss increased with increasing wind velocity (Fig. 5). For example, under a wind velocity of 8 m s −1 , the contents of ammonium N, nitrate N, available K and available P were only 0.0004, 0.0010, 0.2043, and 0.0053 µg m −2 , respectively. Under a wind velocity of 22 m s −1 , the values were 2.3258, 6.6414, 947.0200, and 26.3171 µg m −2 , respectively. These results suggest that high intensities of aeolian processes may increase nutrient loss and enhance desertification in the region.
Nutrient loss and its importance on desertification. Based on the combined results of wind tunnel experiments, nutrient content analyses, and the statistics of sand-driving winds (S4), there were obvious temporal variations in the nutrient loss that may have played an important role in the desertification or rehabilitation in the region over the past several decades (Fig. 6). The dramatic decline of nutrient loss was mainly because of the significant decrease of aeolian transport potential in the region. For example, in the 1970s the losses of ammonium N, nitrate N, available K and available P were 116, 312, 46,436, and 1,251 kg km −2 , respectively, whereas from 2011 to the present the losses were only 8, 20, 3,208, and 84 kg km −2 , respectively. Nutrient loss of the surface soils from 2011 to the present was generally 6.8% of that in the 1970s.  Some researchers 20 have suggested that human activities such as grazing, reclamation, and deforestation were the key forces driving desertification in the Mu Us Desert. However, over the past several decades there were several fluctuations of desertification in the region. For example, monitoring showed that desertification mainly occurred in the 1970s and 1980s, whereas rehabilitation occurred in most areas of the Mu Us Desert from 1990s to the present 15,16,21,22 . In the mid-1970s, 2000, 2005 and 2010, the areas of mobile dunes and sand sheets in China triggered by desertification were approximately 95,000, 107,000, 117,000, 95,000, and 82,000 km 2 , respectively 23 . From 2010 to 2015, the areas of mobile dunes and sand sheets continuously decreased throughout China 15,16 . Although human activities continuously increased in areas with high risks of desertification in China over the past several decades 5,18 , rehabilitation occurred in these regions. Additionally, drought may be one of the key controllers of desertification in some regions 24 . From 1960 to 2009, annual precipitation decreased at a rate of 11.383 mm 10a −1 in the Mu Us Desert 25 , and the variation in precipitation may have contributed to desertification. The decrease in aeolian transport was therefore the key driving force for the occurrence of rehabilitation, despite the importance of human activities on desertification in the region.

Conclusions
Under high intensities of aeolian processes, there were no obvious differences in aeolian transports despite some variations in the components of the underlying soils. Although human activities played important roles, the temporal trends in desertification showed that the dramatic decline of aeolian transport potentials was the key driving forces for the occurrence of rehabilitation in the Mu Us Desert from 2011 to the present. Desertification in the 1970s led to the loss of ammonium N, nitrate N, available K, and available P at rates of approximately 116, 312, 46,436, and 1,251 kg km −2 . From 2010 to the present, the losses were 8, 20, 3,208, and 84 kg km −2 , respectively, representing only 6.7% of the losses of the 1970s. The results showed that although human activities played important roles in desertification, the distinct decrease of aeolian transport and nutrient loss may be the key driving forces for the occurrence of rehabilitation in the Mu Us desert.

Material and Methods
The selected area for desertification driving force analyses was the Mu Us Desert in Central China ( Fig. 1 and S1), which has been identified as a region of intense desertification 1 . The dominant soil type is aeolian sand, and major landscapes include anchored, semi-anchored and mobile dunes, and arable lands. Dominant natural vegetation in this region includes Salix psammophila C. Wang et Chang Y. Yang, Caragana microphylla Lam., Stipa grandis P.  Smirn., Stipa bungeana Trin., Agropyron cristatum (L.) Gaertn., Thymus mongolicus Ronniger, Caragana tibetica Kom., Oxytropis aciphylla Ledeb., Nitraria sibirica Pall. and Kalidium foliatum (Pall.) Moq., with most species being annual herbaceous 5 . In 2015 and 2016, seventy-five surface soil samples at 15 sites (5 samples per site) were collected for further wind tunnel experiments, particle size distribution analysis, and nutrient level analysis. More details of the regional environments and sampling strategies are provided in S1 and S2.
Wind tunnel experiments were conducted in the Key Laboratory of Desert and Desertification, Chinese Academy of Sciences, China. During the wind tunnel experiments, the samples were air-dried and the relative humidity was between 30 and 50%, similar to the values measured in the field at the sampling sites. More details of the wind tunnel experiments were described in S3. After all wind-tunnel experiments were completed, the aeolian materials collected were weighed using a balance with a precision of 0.001 g and were used for further particle size and nutrient level analyses. Particle-size distribution was measured using a Mastersizer 2000 (Malvern Co. Ltd., Malvern, UK; the sample range was between 0.02 and 2000 μm in diameter). Nutrient level analyses included measurements of the ammonium N, nitrate N, available K and available P, and the measurement methods are described in previous report 26 and in S4.
Additionally, wind data from 1951 to 2015 at 15 stations located in the Mu Us Desert (Fig. 1) were used for further analyses (S5). These data were recorded in accordance with the World Meteorological Organization (WMO) and China National Meteorological Center (CNMC) standards. Because most datasets were complete after 1970, wind data records from 1971 to 2015 were used to evaluate the temporal variation in the aeolian transport potentials. More detailed descriptions of the methods are provided in S5.