International Soil and Water Conservation Research最新文献

The Water Erosion Prediction Project (WEPP) model code was modified extensively to support the simulation of nonpoint source (NPS) pollutant sourcing and transport in nonuniform hillslopes based on NPS science from the Soil and Water Assessment Tool (SWAT). This was accomplished utilizing WEPP's overland flow element (OFE) in place of SWAT's hydrologic response unit (HRU) construct which enabled more physically plausible routing within a hillslope. In addition, several improvements to the NPS code base were implemented. These include: free-source format, modern-Fortran conventions, minor enhancements to NPS model science, and code refactoring. This manuscript documents all model development activities, presents a comparison of relevant WEPP and WEPP-WQ code bases, and performs a local sensitivity analysis of the final model code for the most important input parameters and processes. Sensitivity results indicated that the model performed as expected according to its design and provided important insights for potential subsequent validation studies.

Soil erosion is mainly affected by the rainfall characteristics and land cover conditions, and soil erosion modelling is important for evaluating land degradation status. The revised Universal Soil Loss Equation (RUSLE) have been widely used to simulate soil loss rate. Previous studies usually considered the general rainfall characteristics and direct effect of runoff with the event rainfall erosivity factor (R_e) to produce event soil loss (A_e), whereas the fluctuation of rainfall intensity within the natural rainfall profile has rarely been considered. In this study, the relative amplitude of rainfall intensity (R_am) was proposed to generalize the features of rainfall intensity fluctuation under natural rainfall, and it was incorporated in a new R_e (R_e=R_amEI₃₀) to develop the RUSLE model considering the fluctuation of rainfall intensity (RUSLE-F). The simulation performance of RUSLE-F model was compared with RUSLE-M1 model (R_e=EI₃₀) and RUSLE-M2 model (R_e=Q_REI₃₀) using observations in field plots of grassland, orchard and shrubland during 2011–2016 in a loess hilly catchment of China. The results indicated that the relationship between A_e and R_amEI₃₀ was well described by a power function with higher R² values (0.82–0.96) compared to Q_REI₃₀ (0.80–0.88) and EI₃₀ (0.24–0.28). The RUSLE-F model much improved the accuracy in simulating A_e with higher NSE (0.55–0.79 vs −0.11∼0.54) and lower RMSE (0.82–1.67 vs 1.04–2.49) than RUSLE-M1 model. Furthermore, the RUSLE-F model had better simulation performance than RUSLE-M2 model under grassland and orchard, and more importantly the rainfall data in the RUSLE-F model can be easily obtained compared to the measurements or estimations of runoff data required by the RUSLE-M2 model. This study highlighted the paramount importance of rainfall intensity fluctuation in event soil loss prediction, and the RUSLE-F model contributed to the further development of USLE/RUSLE family of models.

Better soil structure promotes extension of plant roots thereby improving plant growth and yield. Differences in soil structure can be determined by changes in the three phases of soil, which in turn affect soil function and fertility levels. To compare the quality of soil structure under different conditions, we used Generalized Soil Structure Index (GSSI) as an indicator to determine the relationship between the “input” of soil three phases and the “output” of soil structure. To achieve optimum monitoring of comprehensive indicators, we used Successive Projections Algorithm (SPA) for differential processing based on 0.0–2.0 fractional orders and 3.0–10.0 integer orders and select important wavelengths to process soil spectral data. In addition, we also applied multivariate regression learning models including Gaussian Process Regression (GPR) and Artificial Neural Network (ANN), exploring potential capabilities of hyperspectral in predicting GSSI. The results showed that spectral reflection, mainly contributed by long-wave near-infrared radiation had an inverse relationship with GSSI values. The wavelengths between 404-418 nm and 2193–2400 nm were important GSSI wavelengths in fractional differential spectroscopy data, while those ranging from 543 to 999 nm were important GSSI wavelengths in integer differential spectroscopy data. Also, non-linear models were more accurate than linear models. In addition, wide neural networks were best suited for establishing fractional-order differentiation and second-order differentiation models, while fine Gaussian support vector machines were best suited for establishing first-order differentiation models. In terms of preprocessing, a differential order of 0.9 was found as the best choice. From the results, we propose that when constructing optimal prediction models, it is necessary to consider indicators, differential orders, and model adaptability. Above all, this study provided a new method for an in-depth analyses of generalized soil structure. This also fills the gap limiting the detection of soil three phases structural characteristics and their dynamic changes and provides a technical references for quantitative and rapid evaluation of soil structure, function, and quality.

Dust emission caused by wind erosion of soil is an important surface process in arid and semi-arid regions. However, existing dust emission models pay insufficient attention to the impacts of aerodynamic entrainment of particles. In addition, studies of wind erosion processes do not adequately account for the dynamics of wind erosion rates and dust emission fluxes, or for the impact of soil texture on dust emission. Our wind tunnel simulations of wind erosion and dust emission showed that the soil texture, wind erosion duration, and shear velocity are major factors that affect the dynamics of wind erosion and dust emission. Because of the limited supply of surface sand and the change in surface erosion resistance caused by surface coarsening during erosion, the wind erosion rate and the flux of particles smaller than 10 μm (PM₁₀) caused by aerodynamic entrainment decreased rapidly with increasing erosion duration, which suggests that surface wind erosion and dust emission occur primarily during the initial stage of wind erosion. The PM₁₀ emission efficiency decreased with increasing shear velocity following a power function, and finer textured sandy loam soils had greater PM₁₀ emission efficiency than loamy sand soils.

Heterogeneous karst surfaces exerted scaling effects whereby specific runoff decrease with increasing area. The existing RUSLE-L equations are limited by the default implicit assumption that the surface-runoff intensity is constant at any slope length. The objective of this study was to modify the L-equation by establishing the functional relationship between surface-runoff intensity and karst slope length, and to evaluate its predictive capability at different resolution DEMs. Transfer grid layers were generated based on the area rate of surface karstification and considered the runoff transmission percentage at the exposed karst fractures or conduits to be zero. Using the multiple flow direction algorithm united with the transfer grid (MFDTG), the flow accumulation of each grid cell was simulated to estimate the average surface-runoff intensity over different slope lengths. The effectiveness of MFDTG algorithm was validated by runoff plot data in Southwestern China. The simulated results in a typical peak-cluster depression basin with an area rate of surface karstification of 6.5% showed that the relationship between surface-runoff intensity and slope length was a negative power function. Estimated by the proposed modified L-equation ((al_x^(b+1)/22.13)^m), the L-factor averages of the study basin ranged from 0.35 to 0.41 at 1, 5, 25 and 90 m resolutions respectively. This study indicated that the modified L-equation enables an improved prediction of the much smaller L-factor and the use of any resolution DEMs on karst landscapes. Particular attention should be given to the variation of surface-runoff intensity with slope length when predicting L-factor on hillslopes with runoff scale effect.

Predicting and allocating surface water resources are becoming increasingly important tasks for addressing the risk of water shortages and challenges of climate change, especially in reservoir basins. However, surface water resource management includes many systematic uncertainties and complexities that are difficult to address. Thus, advanced models must be developed to support predictive simulations and optimal allocations of surface water resources, which are urgently required to ensure regional water supply security and sustainable socioeconomic development. In this study, a soil and water assessment tool-based interval linear multi-objective programming (SWAT-ILMP) model was developed and integrated with climate change scenarios, SWAT, interval parameter programming, and multi-objective programming. The developed model was applied to the Xinfengjiang Reservoir basin in South China and was able to identify optimal allocation schemes for water resources under different climate change scenarios. In the forecast year 2025, the optimal water quantity for power generation would be the highest and account for ∼60% of all water resources, the optimal water quantity for water supply would account for ∼35%, while the optimal surplus water released from the reservoir would be the lowest at ≤5%. In addition, climate change and reservoir initial storage would greatly affect the surplus water quantity but not the power generation or water supply quantity. In general, the SWAT-ILMP model is applicable and effective for water resource prediction and management systems. The results from different scenarios can provide multiple alternatives to support rational water resource allocation in the study area.

Magnetic powder is regarded as an effective and economical tracer for estimating soil erosion. However, the principle and application for using magnetic powder to estimate soil erosion are still not fully developed. In this study, magnetic powders with mean diameters of both 30 and 1 μm were mixed into three soils at different mass proportion. The relationship between magnetic susceptibility and the mass proportion of the introduced magnetic powder in the tagged soil, and the binding ability of magnetic powder to soil particles after both dry and wet sieving were investigated. The accuracy of tracking soil loss by using magnetic powder as a tracer was assessed. The results showed that there was a significant linear relationship between the magnetic susceptibility and the mass proportion of the introduced magnetic powder in the tagged soil. The relationship between the amount of soil captured by a magnet and the mass proportion of magnetic powder in the tagged soil indicated that soils were readily magnetized by magnetic powder, especially fine fractions. The magnetic susceptibility of magnetic powder in different sizes of soil aggregates was variable. A majority of magnetic powder of both 30 and 1 μm diameters was strongly bound with fine particles <0.05 mm in dry and wet sieving. Using the estimated tracer mass proportions, the relative errors between measured and estimated soil losses with enrichment correction factor were less than 18.3% under the simulated rain events. This study not only reveal the principle of Fe₃O₄ powder in soil erosion, but also improve its estimated precision of soil loss, which can make the tracing method by Fe₃O₄ magnetic powder more useable in future.

Gully erosion can lead to the destruction of farmland and the reduction in crop yield. Gully mapping from remote sensing images is critical for quickly obtaining the distribution of gullies at regional scales and arranging corresponding prevention and control measures. The narrow and irregular shapes and similar colors to the surrounding farmland make mapping erosion gullies in sloping farmland from remote sensing images challenging. To implement gully erosion mapping, we developed a small training samples-oriented lightweight deep leaning model, called asymmetric non-local LinkNet (ASNL-LinkNet). The ASNL-LinkNet integrates global context information through an asymmetric non-local operation and conducts multilayer feature fusion to improve the robustness of the extracted features. Experiment results show that the proposed ASNL-LinkNet achieves the best performance when compared with other deep learning methods. The quantitative evaluation results in the three test areas show that the F1-score of erosion gully recognition varies from 0.62 to 0.72. This study provides theoretical reference and practical guidance for monitoring erosion gullies on slope farmland in the black soil region of Northeast China.

Gully erosion is one of the most severe types of land degradation, hindering food production and sustainable agricultural development. However, the historical evolution process and the impact of land use change on gully erosion remain unclear. To address this issue, we conducted a field investigation on gully erosion in 2018 and interpreted land use and gullies using historical remote sensing images in 1968 and 1978 over an area of 84.48 km². The study found that from 1968 to 1978 to 2018, all gully morphological parameters including gully length density and gully areal density increased significantly. The main origin of gully erosion found was from dry farmland. The annual soil loss rate induced by gully erosion was 1.46 mm during 1968–2018. Gully erosion rates were higher during 1968–1978 than during 1978–2018. Furthermore, the length, areal and volumetric erosion rates in gullies formed by multiple gullies merging was greater than that of newly formed gullies (NFG) and gullies developing continuously from a single pre-existing gully, while the widening rate of NFG was highest. The susceptibility of land use types to gully erosion was in the order of woodland < dry farmland < degraded land. The annual average increase in gully area was 871.09 m² km^-2 year^-1 for parcels that were converted from woodland to dry farmland, which was 5.56 times and 1.78 times greater than that of woodland and dry farmland maintenance, respectively. Therefore, urgent implementation of ecological land use plans and gully erosion control practices is suggested for this region.