Bentonite has important applications in curbing soil heavy metal pollution. Jinchang city is an important industrial city in western China, and the mining activities in this area inevitably lead to the heavy metal pollution of soil; in particular, the maximum concentrations of Ni and Cu in the soil exceeded the standard by 30 and 25 times, respectively. In this study, we conducted soil cultivation experiments to study the effects of bentonite addition (BA) on the fractions of the heavy metals Ni and Cu in an old tailings area of Jinchang city. Surface soil (0–20 cm) samples were collected, and Ni and Cu fractionation was performed using the Tessier sequential extraction method. The results showed that residual (R) was the main fraction of Ni and Cu, which accounted for 53% and 57% of their total amounts, respectively. The fraction bound to organic matter (BM), bound to Fe-Mn oxides (BO), bound to carbonates (BC), and the exchangeable (E) accounted for 20% and 16%, 18% and 12%, 6%, and 7%, 3% and 4% of the total amounts, respectively. Their contents ranked in the order: R > BM > BO > BC > E. Each fraction of Ni and Cu decreased with increasing levels of BA. The proportion of E of Ni and Cu was considerably reduced, while the proportion of BO and R increased significantly following the BA. BA can significantly reduce the mobility of Ni and Cu. Our findings indicated that BA can significantly reduce the biological toxicity and mobilization of heavy metals in polluted soil, which can be recommended as a safe stabilizer for heavy metal pollution in soil.

Runoff and nutrient loss in drylands are closely related to vegetation cover. Simulated rainfall experiments with an intensity of 80 mm/h were conducted in sandy grassland and shrubland in the Tengger Desert, China, to investigate the responses of runoff and associated carbon (C) and nitrogen (N) losses to the replacement of grassland by shrubland. Times to ponding and to generate runoff, and the amount of rainfall for runoff commencement in bare, inter-shrub plots were significantly smaller than in shrub (ST) and grassland (GT) plots; no statistical differences were found for these parameters between ST and GT. Overall, this indicated a higher soil water infiltration rate in grassland than in shrubland. The volume-weighted concentrations of organic C (OC) and total N (TN) in runoff from shrubland (0.083 and 0.011 g/L, respectively) were lower than those from grassland (0.103 and 0.012 g/L, respectively). The cover-weighted runoff coefficients, and sediment, OC, and TN losses from shrubland (34.46%, and 44.95 1.72, and 0.23 g/m2, respectively) were greater than from grassland (15.22%, and 15.91, 0.94, and 0.11 g/m2). Vegetation degradation was accompanied by reduced nutrient retention capacity; both soil OC and TN of grassland (8.97 and 0.62 g/kg, respectively) were greater than those weighted values for shrubland (4.18 and 0.26 g/kg). Understanding of these processes suggests that decline or loss of vegetation cover, with the appearance of biological soil crust patches, inevitably leads to increases in runoff and induced soil loss, further accelerating desertification.