2.2 Experimental design
The experiment was initiated at the Daqinggou Ecological Station in 1992. The single-factor completely randomized design was used for comparing the three treatments or approaches aimed at restoring degraded grasslands, namely (1) enclosing the grassland within a shelter belt ofPopulus × beijingensis , referred to as FG, in which the dominant species were Artemisia scoparia , Setaria viride ,Agropyron cristatum , and Cleistogenes chinensis ; (2) afforesting small but well-distributed patches within the grassland using Pinus sylvestris var. mongolica , referred to as MG, in which the dominant species were Cleistogenes chinensis ,Setaria virides , Artemisia scoparia , andCalamagrostis arundinacea ; and (3) similar afforestation usingUlmus pumila , referred to as UG, in which the dominant species were Setaria virides , Cleistogenes chinensis ,Temisia scoparia , and Artemisia lavandulaefolia . All the restoration sites were established in 1992. After tree planting, the land was left without any intervention or disturbance. A plot of desertified sandy grassland in which the dominant species wereAgriophyllum squarrosum (Linn.) Moq. and Cenchrus echinatus Linn. served as a control or a check, referred to as CK, which was also established in 1992 and has remained unmanaged since then.