Objectives Soil desertification happens widely in the northwest hilly sandy areas of Shanxi Province, China, because of the severe wind erosion in this area. Some land restoration measures were attempted to improve the farmland quality and to provide a theoretical foundation for the vegetation restoration in the study area.

Methods The study area was located in Wuzhai County, northwestern Shanxi Province, where the soil was belonged to sand and loam. Four land restoration measures were attempted in separate four plots as: plot reminded uncultivated (WL), plot planted with corn (Zea mays L.) and with no-tillage (NTC), plot planted alfalfa (Medicago sativa L.) and with no-tillage (NTA), plot planted corn and with conventional tillage (CTC). Soil samples were collected after the experiment was conducted for 7 years. The soil particle sizes, particle size parameters, soil water contents, soil saturation moisture capacity and capillary moisture capacity at 0−50 cm soil depths were analyzed.

Results 1) The soil was dominated by sand and silt in this area. The sand contents in fields of WL and NTA were higher than those in the other two. The sand contents did not changed obviously with soil depth in WL, while those in 0−20 cm soil layer were significantly lower than further deeper layers in the other three ones. 2) At 0−40 cm soil layer, the soil bulk density was in order of WL>NTA>NTC>CTC. The bulk density in the WL was significantly different from those in the other three ones (P < 0.05). The bulk densities in CTC, NTC and NTA ranged from 1.10 to 1.50 g/cm³, and no significant differences were among them. At 40−50 cm soil layer, the bulk densities were found not significantly different among all the restoration measures. The 0−30 cm soil porosity under the CTC was the highest, which was significantly higher than the other three, and no significant differences were found among the other three measurements. There was no significant difference among all restoration measures at 40−50 cm soil layer. 3) The soil water contents, saturation moisture capacity and maximum water-holing capacity were all significantly different among the measures. The capillary moisture capacity in the 0−20 cm soil layer under CTC was significantly higher than in the other three; no significant differences were found between NTC and NTA, and both were higher than that under CTC. At the 20−30 cm soil layer, the values under NTA were significantly lower than those under CTC and WL, while at the 30−40 cm soil layer, the values under WL and NTC were significantly lower than those under CTC and NTA, and there is no significant difference was found among all restoration measures at 40−50 cm soil layer. 3) The capillary moisture capacity and the saturation moisture capacity were significantly positive correlated with the total capillary porosity, not significant correlated between the capillary porosity and the non-capillary porosity. There were significantly negative correlations between soil water contents and the saturation moisture capacity and the sand contents, and the correlation coefficients were −0.619 and −0.474, respectively, while there were extremely significant positive correlations between soil water contents and the saturation moisture capacity and the silt contents, and the correlation coefficient were 0.639 and 0.584 respectively.

Conclusions The soil physical properties have been improved after 7 years plantation in the tested area. Crop planting under both conventional and no-tillage could decrease the top soil sand contents, increase soil porosity, decrease soil bulk density and increase the water supply capacity to plant. Therefore, leave land uncultivated is not recommended for the land restoration of northwest Shanxi Province.