退耕还林还草工程已经实施近20年，在人工种植和自然封育等措施下我国林草覆盖面积得到有效恢复。草带不仅能够有效减少其覆盖区域土壤侵蚀，还能够拦截上方汇入泥沙和控制非点源污染。但目前对草带拦沙随时间变化过程规律及其水动力驱动机制的认识尚不明晰。本研究开展了不同坡度（5°~20°）、流量（7.5~45 L min^-1 m^-1）、含沙浓度（40~160 g L^-1）、泥沙粒径组成（中值粒径分别为39.9、77.4和207.9 μm）及草带地上部分的组成结构（完整草带、去青草、去青草和枯草）等一系列草带拦沙试验。对坡面水流全过程进行监测，并完成径流样和粒径样的采集与测量。分析了多重影响因素对草带拦沙效果的综合作用，探究了在拦沙过程中泥沙颗粒选择性及其水力学响应，量化了草带地上部分各主要组成结构（青草、枯草和土壤床面）在拦沙不同时段所起的作用。最后，构建了草带拦沙时变过程模型并分析了其不确定性。得到的主要研究成果如下：

（1）试验显示各影响因素对草带次拦沙总量作用的强弱顺序是：坡度 > 草带宽度 > 单宽流量 > 含沙浓度，可见坡度是比较关键的影响因素。含沙浓度越高会引起更低的草带瞬时拦沙效率，但是得到更大的次拦沙总量。入流含沙浓度的作用主要集中在前5 m宽度内的草带，而坡度和流量的作用在整个草带均有体现，且高入流含沙浓度会增强坡度对草带拦沙效果的作用。草带拦沙时空分布呈现出明显的规律性，90%的泥沙沉积发生在达到拦沙稳定状态时间T_stable的前一半时段内，即草带对短历时降雨引发的拦沙过程的效果会更好。前2 m草带的单宽拦沙效果显著优于随后草带，沉积泥沙量沿草带宽度呈幂函数减少。

（2）在草带拦沙过程中，沉积的泥沙会使草带床面变得光滑，逐渐使坡面流具有足够动力把更粗泥沙颗粒输出草带。Darcy–Weisbach阻力系数f和Manning糙率系数n都随拦沙持续时间呈指数函数减小。当草带拦沙持续到一定阶段，坡面水流能量足够大时，部分小尺寸的泥沙颗粒就会出现净流失现象。净流失并不是对原土的侵蚀，也不是对已沉积泥沙的再卷起，而是部分小粒径泥沙颗粒的再卷起率大于沉积率的现象。入流泥沙粒径越粗或坡度越陡都会导致更大粒径范围的泥沙颗粒出现净流失。分析坡面流水动力参数与出口泥沙粒径过程关系，结果显示单位水流功率P比水流剪切力t更适宜用于解释草带拦沙过程中的泥沙颗粒的粒径交互过程。

（3）试验显示土壤床面、青草和枯草对次拦沙总量的贡献强度比约为3:5:2。草带的土壤床面（包括微地形和表面粗糙小砾石颗粒）、青草和枯草三个地上部分组成结构中，土壤床面前5 min内对拦沙量贡献约为90%，在拦沙初始阶段起主要作用。随着拦沙过程的持续土壤床面的作用在减弱，到拦沙稳定阶段起主要作用的是青草。

（4）根据试验结果，发现前期沉积的泥沙会使草带拦沙性能逐渐降低，即草带拦沙过程具有自反馈（自抑制）特征。基于这一特征构建草带拦沙过程微分方程并求得其解析解，建立草带拦沙时间变化过程模型STPMOD-VFS。模型能够较好地预测各条件下（不同入流输沙率、泥沙粒径组成和坡度）草带拦沙时间变化过程。不确定性分析表明该模型表现出显著的“异参同效”现象，不同参数的敏感性程度也存在较大差异。

本研究深化了对草被覆盖坡面水沙过程机理的认识，有助于提升水土模型植被模块功能和推动植被过滤带管理技术更广泛高效应用。

The ‘‘Grain for Green’’ project has been implemented for nearly 20 years in China, and the area covered by forest or grass has been effectively restored by natural enclosure or artificial planting. Grass strips can not only reduce erosion but also trap sediment from uphill silt-laden inflow and play a very effective role in controlling nonpoint source pollution. However, the time-varying process in the performance of sediment trapping and its hydrodynamic driving mechanism are still unclear. In this study, a series of grass strips sediment-trapping experiments have been carried out under the conditions of different slopes (5°-20°), flow rates (7.5-45 L min^-1 m^-1), sediment concentrations (40-160 g L^-1), particle size distribution (the median particle size are 39.9, 77.4, and 207.9 μm, respectively), and the structural components of aboveground parts of grass strips (completed grass strips, remove green grass, and remove green and withered grass). The whole process of overland flow was monitored. The runoff-sediment samples and the particle size samples of the outlet were collected and measured. Comprehensive effect of multiple influencing factors on the performance of grass strips sediment-trapping was analyzed. The particles selectivity and its hydraulic response in sediment-trapping-process was explored. Then the functions of the main structural components of aboveground parts of grass strips (green grass, withered grass and soil bed surface) in different stages of sediment trapping were quantified. Finally, a model for the time-varying process of sediment trapping in grass strips was built and its uncertainty was analyzed. The main findings are as follows:

(1) The experimental results showed that the intensities of the impacting factors on individual total sediment trapping amount were as follows: slope > grass strips width > unit flow rate > sediment concentration, which indicated that slope plays a key role on sediment trapping in grass strips. Greater sediment concentration of silt-laden inflow led to lower instantaneous sediment trapping efficiency, but resulted in greater individual cumulative total amount of sediment trapping. The effect of sediment concentration was concentrated primarily in the first 5-m width, whereas slope and flow rate had similar effect intensity on sediment trapping in each section of grass strips. In addition, the impact of slope on sediment trapping effect was amplified when the sediment concentration increased. The temporal and spatial distribution of sediment trapping in grass strips showed obvious regularity. 90% of sediment deposition occurred within the first half of the time required to reach the stable state, that is, the grass strips has a better effect on the sediment-trapping-process caused by short-duration rainfall. The single-width sediment trapping effect of the first 2-m width grass strips performed better than the subsequent strips significantly, and the deposited sediment decreased along the width of the grass strips in power function.

(2) During the sediment-trapping-process in grass strips, the deposited sediment gradually smoothed the bed surface of the grass strips, and the overland flow gradually had enough power to transport the coarser particles out of the grass strips. Both the Darcy–Weisbach friction factor f and Manning’s roughness n decreased exponentially with time in the process of sediment trapping. When the sediment trapping in grass strips continued to certain stage and the energy of the overland flow was greater enough, the net loss phenomenon of some small-sized particles occurred. Net loss is neither the erosion of original soil in grass strips, nor the re-entrainment of newly deposited sediment, but the phenomenon that the re-entrainment rate of some particles is greater than their deposition rate at a certain time. Coarser sediment input or a steeper slope led a larger range of particle sizes to be net lost. The analysis of hydrodynamic parameters of overland flow and the particle size of the outflow sediment showed that the unit stream power P is more suitable than shear stress t of overland flow to be used to describe the processes of sediment particles transport and exchange during the sediment-trapping-process of grass strips.

(3) The experimental results showed that the contribution intensity ratio of soil bed surface, green grass and withered grass on individual total sediment trapping amount was 3:5:2. Among the three structural components of aboveground parts of grass strips i.e., the soil bed surface (micro-topography and grains on the soil surface), green grass, and withered grass, the soil bed surface played a primary role on sediment deposition in the initial stage of sediment-trapping-process, and its contribution to the total amount of deposited sediment in the first 5 minutes was about 90%. As the sediment trapping process continues, the effect of soil bed surface weakened, and the green grass played the major role at the stable stage of sediment trapping.

(4) According to the experimental results, the deposited sediment can gradually weak the sediment trapping performance of grass strips, which is the self-feedback (self-inhibition) characteristic of grass strips during the sediment trapping process. Based on the self-feedback characteristic, an ordinary differential equation was established with a self-inhibition correction factor and its analytical solution was obtained. A sediment-trapping-process model of grass strips (STPMOD-VFS) was proposed. STPMOD-VFS model can well predict the time-varying process of sediment trapping in grass strips under various conditions (different sediment inflow rates, particle size compositions, and slopes). The uncertainty analysis results showed the model had significant ‘equifinality’ phenomenon. The parameters related to the median particle size of sediment were more sensitive than those related to slope.

These findings can deepen the understanding of the mechanism of sediment-flow process on hillslopes covered by grass, are helpful to improve the vegetation module function of soil and water model, and promote the wider and more efficient application of management technology of vegetation filter strips.