The main aim of this study was to compare the performances of the hybrid approaches of traditional bivariate weights of evidence (WoE) with multivariate logistic regression (WoE-LR) and machine learning-based random forest (WoE-RF) for landslide susceptibility mapping. The performance of the three landslide models was validated with receiver operating characteristic (ROC) curves and area under the curve (AUC). The results showed that the areas under the curve obtained using the WoE, WoE-LR, and WoE-RF methods were 0.720, 0.773, and 0.802 for the training dataset, and were 0.695, 0.763, and 0.782 for the validation dataset, respectively. The results demonstrate the superiority of hybrid models and that the resultant maps would be useful for land use planning in landslide-prone areas.

Tillage treatments have an important effect on soil microstructure characteristics, water thermal properties and nutrients, but little is known in the newly reclaimed cultivated land. For the reason, a long-term field study was to evaluate the tillage effects on soil physicochemical properties and crop yield in newly reclaimed cultivated land via the macroscopic and microscopic analysis. Three tillage treatments were tested: continuous conventional moldboard plow tillage (CT), sub-soiling/moldboard-tillage/sub-soiling tillage (ST) and no-tillage/sub-soiling/no-tillage (NT). Under CT, the microstructure was dominated by weakly separated plates structure and showed highest bulk density (BD) (1.49 g cm⁻³) and lowest soil organic matter (SOM) (3.68 g kg⁻¹). In addition, CT reduced the capacity of soil moisture retention and temperature maintenance, resulting in aggregate structure deterioration and fragility. Unlike CT, the soil was characterized by moderately separated granular structure and highly separated aggregate structure under conservation tillage practice of ST and NT. NT was associated with the highest soil moisture content (20.42%), highest quantity of macroaggregates (> 0.25 mm) by wet-sieving (34.07%), and highest SOM (6.48 g kg⁻¹) in the surface layer. Besides, NT was better able to regulate soil temperature and improved the values of geometric mean diameter. Under NT and ST, a stable soil structure with compound aggregates and pores was formed, and the maize yield was increased by 12.9% and 14.9% compared with CT, up to 8512.6 kg ha⁻¹ and 8740.9 kg ha⁻¹, respectively. These results demonstrated the positive effects of NT and ST on soil quality and crop yield in newly reclaimed cultivated land.

The purpose of this study is to compare nine models, composed of certainty factors (CFs), weights of evidence (WoE), evidential belief function (EBF) and two machine learning models, namely random forest (RF) and support vector machine (SVM). In the first step, fifteen landslide conditioning factors were selected to prepare thematic maps, including slope aspect, slope angle, elevation, stream power index (SPI), sediment transport index (STI), topographic wetness index (TWI), plan curvature, profile curvature, land use, normalized difference vegetation index (NDVI), soil, lithology, rainfall, distance to rivers and distance to roads. In the second step, 152 landslides were randomly divided into two groups at a ratio of 70/30 as the training and validation datasets. In the third step, the weights of the CF, WoE and EBF models for conditioning factor were calculated separately, and the weights were used to generate the landslide susceptibility maps. The weights of each bivariate model were substituted into the RF and SVM models, respectively, and six integrated models and landslide susceptibility maps were obtained. In the fourth step, the receiver operating characteristic (ROC) curve and related parameters were used for verification and comparison, and then the success rate curve and the prediction rate curves were used for re-analysis. The comprehensive results showed that the hybrid model is superior to the bivariate model, and all nine models have excellent performance. The WoE–RF model has the highest predictive ability (AUC_T: 0.9993, AUC_P: 0.8968). The landslide susceptibility maps produced in this study can be used to manage landslide hazard and risk in Linyou County and other similar areas.

Evapotranspiration (ET) plays a crucial role in hydrological and energy cycles, as well as in the assessments of water resources and irrigation demands. On a regional scale, particularly in the agro-pastoral ecotone, clarification of the distribution of surface ET and its influencing factors is critical for the rational use of water resources, restoration of the ecological environment, and protection of ecological water sources. The SEBAL model was used to invert the regional ET based on Landsat8 images in the agro-pastoral ecotone of northwest China. The results were indirectly verified by monitoring data from meteorological stations. The correlation between ET and surface parameters was analyzed. Thus, the main factors that affect the surface ET were identified. The results show that the SEBAL model determines an accurate inversion, with a correlation coefficient of 0.81 and an average root mean square error of 0.9 mm/d, which is highly suitable for research on water resources. The correlation coefficients of normalized vegetation index, surface temperature, land surface albedo, net radiation flux with daily ET were 0.5830, 0.8425, 0.3428 and 0.9111, respectively. The normalized vegetation index and the net radiation flux positively correlated with the daily ET, while the surface temperature and land surface albedo negatively correlated with the daily ET. The correlation from strong to weak is the net radiation flux > surface temperature > normalized vegetation index > surface albedo. In terms of spatial distribution, the daily ET of water was the highest, followed by woodland, wetland, cropland, built-up land, shrub land, grassland and bare land. However, the SEBAL model overestimates the inversion of daily ET of built-up land.

Soft rock efficiently retains water because it is hard when dry but expands rapidly when wet, thus it can be used as a natural water retention agent. Therefore, soft rock can be blended with sandy soil to enhance the water-holding capacity of sandy land. Understanding how soft rock content in sandy soil affects corn growth (Zea mays L.) is essential for enhancing local agricultural production in the Mu Us Desert. The Root Zone Water Quality Model 2 was calibrated and validated based on 2-years of experimental data for composite soils prepared at three ratios of soft rock to sand (1:1, 1:2, and 1:5 v/v), and then used to estimate the water non-limiting potential yield (Yp) and the potential yield of a rainfed crop [Yp(r)], along with yield gap [YG(r)] of corn using the weather data from 1990 to 2013 in the Mu Us Desert. The mean simulated Yp(r) of corn in the compound soil at a ratio of 1:2 was significantly higher than that of the compound soil at 1:1 and 1:5 ratios, respectively. The mean simulated Yp(r)s of corn were 2.551, 3.527, and 2.924 Mg ha⁻¹ and the YG(r)s were 6.071, 5.096, and 5.698 Mg ha⁻¹ for 1:1, 1:2, and 1:5 ratios, respectively. The crop yield depends on the amount of soft rock added to the sandy soils.

The understanding of land-use and land-cover change (LUCC) is crucial for the study of global climate change and the terrestrial ecological environment. Sandy land is an important component of the terrestrial ecosystem. This study applied the land-use transition matrix and change trajectory analysis to examine the spatiotemporal characteristics of LUCC from 1980 to 2018 in the Mu Us Sandy Land, China. The influences of soil, meteorological conditions, and national policy on LUCC in this region were also investigated. The results of land-use classification of five remote sensing images captured between 1980 and 2018 showed that Grassland was the dominant land-use type. The areas of Woodland and Built-up land in the study area continued to increase, whereas those of Water bodies and Unused land continued to decrease. The area of Cropland first increased and then decreased. The areas where LUCC was driven by natural factors and anthropogenic factors were 25,428 km² and 9,683 km², representing 27.92% and 10.63% of the total area, respectively. The area that experienced no LUCC was 55,950 km ² (61.44% of the total area). These data showed that LUCC was driven by natural and anthropogenic factors, with anthropogenic factors dominant in driving changes in Cropland, Woodland, and Built-up land, whereas natural factors dominated reductions in the areas of Water bodies and Unused land. Although natural factors played an important role in LUCC, this study showed that anthropogenic factors determine the direction of LUCC.