Mu Us Sandy Land has an area of approximately 42,200 km², with major regions of desertification, development and reversion in Northern China. Sandy land is easily moved due to the sparse vegetation and unconsolidated sand under harsh dry and windy weather conditions. The desertification process is associated with vegetation degradation, productivity reduction, fragile ecosystems, environmental deterioration and economic loss (Karnieli et al., 2014). Sandy land soil with little or no structure is highly permeable under both wet and dry conditions and it retains little or no water. The soil physical and chemical properties such as soil clay, soil organic matter, and total nitrogen decrease during desertification which affects the regional ecosystems (Yang et al., 2010). The control of sandy land can improve the ecological conditions and increase the amount of farmland (Han et al., 2012). To date, local governments and residents have done much to combat desertification, such as planting trees bushes and grasses; building long windbreaks, shelterbelts and barriers; constructing pastureland enclosures; performing chemical mulching; and developing hydrologic solutions (Jiang, 2002). A monitoring program revealed that the average vegetation coverage in the sandy lands increased from 17.63% in 2004 to 18.69% in 2009 (Karnieli et al., 2014).

The selected area for sandy land rehabilitation in this study is in the Yuyang district, Yulin city, Shaanxi Province, China (Figure 1). The north and south of Yulin city is a wind-sand-grass shoal region and loess hilly-gully region, accounting for 42% and 58% of the total area, respectively. The area of sandy badlands is approximately 5.3 million ha, accounting for 91.65% of the total unused land of Yulin. The agriculture pattern is identified by the coexistence of farm and livestock husbandry (Wang et al., 2013). The typical warm temperate continental monsoon climate dominates this region. The average annual temperature and average annual precipitation are 8.1℃ and 413.9 mm, respectively. In the surface soil (0-20 cm), the total nitrogen content is 0.75 g kg^-1, total phosphorus content is 0.63 g kg^-1, total potassium content is 26.51 g kg^-1 and the organic matter content is 3.32 g kg^-1 (Wang et al., 2016).

Our study area has relatively rich groundwater resources, originating from phreatic water from fissures in the upper Jurassic and Quaternary sediments. Abundant subsurface flow is also found within a 20-m depth below the river bed (Wang et al., 2013). Land resource development and utilization can be supported by the relatively abundant groundwater resources and heat energy. Therefore, Yulin was considered to be the second granary of Shaanxi Province in 2013. However, general and long-term land overuse resulted in severe land degradation and desertification (Wang et al., 2014). Cultivated land quantity and quality were the main factors limiting grain production in Yulin (Ning et al., 2016). Land degradation and desertification limit the improved crop varieties superiority due to the depleted soil fertility. There is an urgent need to improve the sandy land quality and increase agricultural production through sandy land rehabilitation.

Soil is the foundation of the earth’s biosphere. Soil not only produces food and fiber but also sustains ecosystem function and environmental quality (Glanz, 1995). Particles are the basic unit or component of soil. In some cases, particles are randomly oriented, and in other cases, they are arranged in particular forms that can be oriented, or dispersed (Sivakumar et al., 2002). Aggregates originate from the combination of rearranged mineral particles with organic and inorganic substances (Bronick and Lal, 2005). The soil structure is associated with the degree of aggregate stability (Duiker et al., 2003). A favorable soil structure and high aggregate stability are important to improving soil fertility, increasing agronomic productivity, enhancing porosity and decreasing erodibility (Bronick and Lal, 2005).

Complex interactions among various soil properties determine the soil’s potential fitness to produce healthy and nutritious crops (Parr et al., 1992). Sandy soils have low water retention and few macro- and micronutrients due to poor mineral and organic colloids (Weber et al., 2007). Clays are capable of storing and releasing plant nutrients. However, higher clay soils commonly have low hydraulic conductivity and poor tilth and aeration (Dunker and Darmody, 2005). The physical complementarity of sandy and clay soils means that they can be used to construct a new type of soil and increase its productivity. The non-capillary porosity in sandy soil is filled by clay, and aggregates can be formed and cemented with carbonate materials and silicate minerals. Aggregates prevent water and fertilizers from leaching to deeper depths during crop growth periods and avoid useless evaporation in drought periods (Wang et al., 2013).

Soil reconstruction includes soil profile reconstruction and soil structure reconstruction. Good quality soil is defined as having a loose surface layer and tight subsoil that can produce and sustain crops. Supplementary materials are used to improve the structure of degraded land. The structure of the newly mixed soil is formed by first breaking down the particles or aggregate assemblages, and then rearranging them. Water and the soil structure are crucial for soil creation, soil configuration, soil improvement and fertilizer coupling. The soil particle composition and profile structure can be improved in regard to water and nutrient retention through soil reconstruction. Therefore, the innovative microscopic theory, including micro-structure, micro-morphology, and micro-mechanism, was established for agricultural land engineering to couple soil stability, structure rationality and functional efficiency. In research on the development of high-quality farmland, Liu et al., (2016) proposed a theoretical innovation: the soil components, water and soil configuration, and farmland landscape formed a structural chain. Three important issues need to be noted in regard to soil reconstruction. First, suitable materials are widely distributed and convenient for soil construction in the study area. Second, the structure parameters of sandy and clay soils are basic to engineering practice. Third, more chemical and organic fertilizers need to be artificially added to improve the low fertility in the newly mixed soil (Wang et al., 2013). Economic and ecological benefits are also the objectives of sandy soil reconstruction engineering. Various integrative sandy control methods have been developed in recent years. Amendments, such as chemical hydrophilic polymers, viscous pastes, municipal sewage sludge, and plastic-films, can increase the water and nutrient holding capacity of sandy land. However, the above measures have limitations, due to their high cost and potential soil pollution (Han et al., 2012). Sandy soil reconstruction methods use natural, cheap and convenient materials to improve the sandy land structure. Therefore, this economic and ecological concept has much promise and could be very effective.

Red clay and loess are discontinuously distributed and overlaid by sandy land in the central and southern parts of the Mu Us Sandy Land (Zhu et al., 1980). The grain size of red clay is similar to that of loess, with a median size of approximately 4-8 µm. The thicknesses of the red clay and loess deposits at Mu Us Sandy Land are approximately 60 m and 256 m, respectively (Ding et al., 1998). Red clay is mainly used for making bricks and has a price of 6 US dollars per cubic meter. Moreover, loess is generated from rock and soil, forming processes with substantial amounts of fine sandy loam and silt loam. Loess has strong tensile and stress joints, often giving rise to soil collapses that can occur after only 1 to 2 minutes in water (Shi and Shao, 2000). The red clay and loess used in the field experiment were collected from Qinhe town of Yulin city.

Dry mixtures of sand and red clay and dry mixtures of sand and loess at volume ratios of 1:0.2, 1:0.5, 1:1, 1:2, 1:3 and 1:5 were prepared in our Yulin research station in Niujialiang village. Single sand, red clay and loess were treated as the controls. The soil mass, soil profile and soil structure were discussed in the engineering process. Maize, tomato and soybean were planted in the mixed soils. The experimental objective was to sow the improved crop varieties on healthy and fertile soils as determined by coupling soil ecological suitability and crop physiological adaptability. The integrated land optimization configuration with an improved, optimized crop variety selection was suggested to achieve sandy soil oriented consolidation.

4.3.1 Soil reconstruction methods

The comprehensive framework of the experiment included soil reconstruction, crop optimizing selection and farming management measures (Figure 2). The specific procedures are elaborated here using the mixture of sandy land and red clay with a volume ratio of 1:1 as an example. The barren sandy soil was mechanically flattened and compacted to form plough pans. Branches and roots were removed during the flattening period. Barren sandy land was overlaid with red clay at a thickness of 30 cm, and the particle size was less than 4 cm. Then, a 15 cm layer of sand was spread on top of red clay. The surface sand (15 cm) and partial subsurface red clay (15 cm) were thoroughly blended into the plough layer with an excavator and subsoiler. The remaining 15 cm of red clay was considered the plow layer for water and nutrient conservation. Degraded sandy land was transformed into productive new land by improving the soil particle and profile structure.

4.3.2 Crop management measures

The fertility of the newly constructed soil was not enough to support crop growth. Consequently, chemical fertilizers and organic fertilizers were added to accumulate nutrients and improve the mixed soil quality. Organic matter offers more soil cementing agents to support the formation of advanced porosity and prevent soil crusting (Wang et al., 2013). As for farm management, fertigation techniques were used for getting effective fertilization and to improve water capacity and feasibility. Different chemical fertilizers were dissolved and pumped into the same pipeline for fertilization. Water was dripped via three different pipelines to satisfy the demand of the three crops. The water volume was automatically recorded and summed in regard to irrigation to develop a detailed program for each crop.

According to the soil reconstruction process, borrowed red clay or loess and the original sandy land were mixed to create a newly topsoil, which was subjected to ploughing and farming activities. Scenarios on land consolidation and modern agriculture are proposed from soil particle to agricultural system (Figure 3). First, particle materials with different physical characteristics were adopted to determine which was convenient for blending with sand in the Mu Us sandy land. Second, different prescriptions with six mixed ratios were carried out to understand which was more beneficial for forming stabilized aggregates and a favorable soil structure. Third, different, locally grown crops were selected to examine which was suitable for coupling soil ecological suitability with crop physiological adaptability. Last, the different benefits to economic and ecological aspects were studied to determine which was better for improving land productivity and ecosystem services.

Four key issues were studied in the present research according to the above research questions (Figure 4). Crop performance is susceptible to the physical properties of reconstructed soil. The soil particle composition, bulk density, soil porosity and soil pH value were measured after blending sand with red clay or loess to distinguish the better mixing materials. The particle with a 4-cm diameter for red clay and loess was considered in our study, according to the previous experimental results on arsenic stone (Wang et al., 2014). In addition, the organic manure fertilization amount and variety were also important to the structure and formation of the newly constructed soil. Second, after a series of physical blending, chemical cementation and biological processes, the preserving capacity of water and nutrients of the newly constructed soil were gradually improved after sustained fertility treatments (Wang et al., 2013). Soil samples were collected from depths of 0-20 cm and 20-40 cm once per month. The total and available contents of soil nitrogen, phosphorus and potassium were determined to monitor the dynamics of nutrients pool and release rates at different mixed ratios. Third, a desirable microbial group can easily survive if the soil physical and chemical environment is favorable (Dunker and Darmody, 2005). Microorganisms play a vital role in the acquisition and turnover of nutrients in soil. The microbial community structure and functional gene abundances were measured using the phospholipid fatty acid (PLFA) and terminal restriction fragment length polymorphism methods. In view of the increased density of organisms in the mixed soil, it was expected that associative relationships would be particularly pronounced, with favorable effects on the plant. Last, there are significant correlations among leaf photosynthesis, root biomass and crop yield across all cultivars (Elsharkawy et al., 1990). Root penetration into the clay soil and sand was primarily horizontal and vertical respectively. The crop rooting behavior in the newly constructed soils changed due to the extensive soil porosity differences between aggregates. The root biomass and morphology in different soil profiles and the crop yield and quality can provide insight into the improved soil structure and nutrition status.