黄土高原是世界上水土流失最严重和生态环境最脆弱的区域之一，其泥沙主要来源于坡沟系统这一基本地貌单元。梯田建设作为重要的水土保持实践，在黄土高原地区水土流失治理中发挥着关键作用。然而，以往研究主要关注坡面或流域尺度上梯田减水减沙效益评价，忽视了梯田对其下方沟道侵蚀发育的影响，使得梯田对作为联系坡面和流域关键的坡沟系统的水沙调控作用机制尚不清楚，梯田减水减沙效益得不到全面系统的评估。此外，黄土高原以往梯田建设主要考虑粮食生产因素，而很少从梯田水土保持功能角度关注流域梯田建设位置和规模，导致梯田的减水减沙作用得不到充分发挥。

为厘清梯田对坡沟系统产水产沙的调控作用机制，优化流域梯田建设以实现生态效益最大化，本文以孤山川和延河安塞流域为研究区，并选取固原和浑源作为实验区，构建了多尺度沟道侵蚀监测方法和改进优化了LCM-MUSLE模型，通过开展低空无人机遥感观测实验和数字实验，分析了梯田空间配置与规模变化对坡沟系统和流域产水产沙的影响，揭示了黄土高原梯田与坡沟系统-流域减水减沙之间的关系与规律，并应用减水减沙与梯田措施变化之间的响应关系曲线对无监测坡沟系统梯田效益进行了有效评估，同时提出可有效遏制黄土高原丘陵沟壑区流域产沙的梯田规模阈值。

通过研究，形成的主要成果和结论包括：

（1）构建多尺度沟道侵蚀监测方法，评价黄土高原长期沟道侵蚀发育状况，分析集水面积变化对沟道侵蚀发育的影响，结果表明，集水面积是影响沟道侵蚀的主要因素，通过增加梯田面积可减小沟道集水面积，能够有效减轻沟道侵蚀强度。

基于无人机遥感技术，结合历史航空摄影测量和地面调查，构建多尺度沟道侵蚀监测方法，监测固原和浑源实验区众多沟道长期沟道形态参数和环境因子等关键指标变化。在固原实验区，1976-2017年，19条沟道的侵蚀体积变化监测量在133.70~8938.44 m³之间，平均侵蚀模数为4364.69 t/(km²·a)。在浑源实验区，1959-2018年，19条沟道侵蚀体积变化在1034.89~49192.33 m³之间，平均侵蚀模数为5762.73 t/(km²·a)。沟头后退侵蚀体积对沟道侵蚀体积的贡献是沟壁后退贡献的6.1倍，表明沟头后退是浑源实验区长期沟道发育的主要贡献因素。

浑源实验区沟道起始地形阈值关系为S=0.0478A^-0.4915，分析表明研究区沟道侵蚀的主导过程是坡面流侵蚀。该阈值可作为预测黄土高原边缘地带相似环境条件下沟道发生潜在位置的指标。回归分析结果表明，与平均坡度梯度相比，现状集水面积（A_p）能够更好地解释长期沟头后退体积速率（V_e），二者之间存在幂函数关系：V_e = 0.09A_p^0.60 (R² = 0.51, n = 30)。对比本研究与其它研究报告的S-A和A_p-V_e关系，发现集水面积在黄土高原边缘地带沟道发生和长期沟头后退过程中扮演了关键作用。

以开展过坡改梯工程的固原实验区为例，分析集水面积变化对沟道侵蚀发育的影响。监测结果表明，2017年沟道集水面积大小与侵蚀模数存在较好的相关性；2017年集水面积与1976年相比，沟道集水面积有不同程度的减小，减小比例在0.71%~77.88%。集水面积减小是引起沟道侵蚀模数变化的主要原因。集水面积减小幅度越大，沟道侵蚀模数越小，二者呈指数相关关系。因此，减小沟道集水面积能够有效控制沟道发育。进一步对比40年下垫面情况变化，结果表明实验区梯田面积增加是导致沟道集水面积减小的主要原因。梯田建设，尤其是将坡耕地转化为水平梯田，增加了其上方坡面来水的拦截量，改变了原有沟道集水面积，减少了进入沟道的径流量，从而使沟道侵蚀强度减弱。

（2）构建分析坡沟系统侵蚀产沙对梯田措施变化的响应关系曲线，并在黄土高原典型流域推广。结果表明，梯田措施不仅可以减少自身产沙，还具有异地减沙作用；同时关系曲线具有较好的适用性和推广性。

耦合沟道侵蚀监测方法与LCM-MUSLE模型估算梯田建设前后的坡沟系统侵蚀量。固原实验区梯田建设对各坡沟系统年均侵蚀量的消减比例在30.93%~96.31%之间，平均消减比例为62.53%。浑源实验区梯田建设对各坡沟系统年均侵蚀量的消减比例在13.69%~99.04%之间，平均消减比例为51.60%。

以固原和浑源实验区坡沟系统作为样本，构建坡沟系统梯田面积比例-减水减沙幅度关系曲线，结果显示，梯田面积越大，对坡沟系统的减水减沙效果越好，二者之间均存在显著相关的对数函数关系，相关系数达0.9以上。进一步分析该关系曲线，发现坡沟系统减水减沙幅度均大于相应的梯田面积比例，结果表明梯田措施不仅可以减少自身的产沙，还具有异地减沙作用。

进一步将坡沟系统梯田面积比例-减水减沙幅度关系曲线应用至延河安塞流域和孤山川流域，实现了18686个坡沟系统减水减沙幅度计算，结果表明关系曲线具有良好的适用性和推广性。

（3）分析梯田空间配置对流域场次“降水-径流-输沙”过程的影响，结果表明，梯田配置在坡面下部时减水减沙作用更好。等面积梯田配置在坡面下部时，延河安塞和孤山川流域的减水幅度分别可提升16.60%和7.80%，减沙幅度可提升15.35%和13.30%。

分析空间配置变化对梯田减水减沙效果的影响。发现梯田空间配置对其减水减沙效果影响很大。结果显示，等面积梯田位于坡面下部时的减水减沙幅度最大，位于坡面中部时次之，位于坡面上部时最小。对比现实空间配置与虚拟空间配置情形下的等面积梯田减水减沙幅度，发现延河流域和孤山川流域减水减沙幅度仍存在较大提升空间。现实梯田情景下，延河流域平均减水幅度为15.28%，平均减沙幅度为15.97%。孤山川流域平均减水幅度为8.68%，平均减沙幅度为7.83%。当等面积梯田位于下坡时，减水减沙幅度提升最大，减水幅度分别增加了16.60%和7.80%，减沙幅度分别增加了15.35%和13.30%。结果表明，在提升梯田防洪设计标准并加强田埂定期维护避免梯田发生水毁的前提下，在坡面可修梯田之处，梯田应尽可能靠下部修建，以提升其减水减沙效益。

（4）分析考虑空间配置的梯田规模变化对流域场次“降水-径流-输沙”过程的影响，结果表明，任意配置梯田可有效遏制黄土高原丘陵沟壑区流域产沙的梯田比阈值为35%；从下坡至上坡配置梯田这一阈值可进一步降至25%。

通过情景设置的方法，计算相同降水和植被条件下不同规模梯田的减水减沙幅度，构建了两种梯田极端空间配置和梯田任意空间配置情形的梯田比-减水减沙幅度响应关系曲线。分析发现继续增大流域梯田规模，可进一步提升延河安塞和孤山川流域的减水减沙幅度，且减水减沙幅度均表现出随梯田比增大而增大的趋势，且增长速率逐渐放缓，最终趋于稳定。基于构建的梯田比-减沙幅度响应关系研究表明，任意配置梯田时，可有效遏制黄土高原丘陵沟壑区流域产沙的梯田比阈值约为35%，即在梯田比大于35%以后，继续增加梯田规模对减沙效果提升不明显。考虑梯田空间配置从下坡至上坡修建梯田时，可有效遏制黄土高原丘陵沟壑区流域产沙的梯田比阈值约为25%。

    综上，本文的研究方法和结果可为黄土高原坡沟系统梯田减水减沙效应评价与流域梯田措施优化配置以实现最大生态效益提供科学依据和方法支撑。

The Loess Plateau is one of the regions with the most serious soil erosion and the most fragile ecological environment, and the sediment generated mainly comes from the basic geomorphic unit-slope gully system. As an important practice of soil and water conservation, terrace plays a crucial role in soil and water loss control in the Loess Plateau. However, previous studies were mainly focused on the evaluation of water and sediment reduction benefits of terraces on the slope or watershed scale, the impact of terraces on gully erosion below them was ignored. As a result, the regulatory mechanism of terraces on the runoff and sediment of the slope gully system, which is the key link between the slope and watershed, is not clear. Additionally, previous terrace was constructed mainly to benefit grain production in the Loess Plateau, and the soil and water conservation was rarely taken into consideration. Consequently, the benefits of terraces on water and sediment reduction cannot be comprehensively and systematically evaluated and released.

In this study, Gushanchuan basin and Yanhe basin were taken as the study areas, Guyuan and Hunyuan were selected as the experimental areas in order to clarify the regulatory mechanism of terraces on water and sediment production in slope gully system and to optimize the construction of terraces in watershed to obtain the maximum ecological benefits. And a multi-scale method to monitor the gully erosion was developed, the LCM-MUSLE model was also optimized and improved. Thus, the impact of terrace spatial configuration and scale change on water and sediment yield of slope gully system and watershed was analyzed through conducting low altitude UAV remote sensing observation experiment and digital experiment. Results revealed the relationship and law between terraced fields and water and sediment reduction of the slope gully system and watershed in the Loess Plateau. Finally, the response curves between the change of terrace measures and water and sediment reduction were applied to effectively evaluate the benefits of terraced fields in slope and gully systems without monitoring. Meanwhile, the threshold of terrace scale that can effectively curb sediment production in the watershed of the hilly and gully regions of the Loess Plateau was proposed. 

The main results and conclusions obtained in this paper were described as follow.

 (1) A multi-scale gully erosion monitoring method was established to evaluate the long-term gully erosion development in the Loess Plateau, and the impact of catchment area change on development of gully erosion was analyzed. The results indicate that catchment area is the main factor affecting the gully erosion. Increasing the terrace area can reduce the gully catchment area and effectively reduce the gully erosion intensity.

Combing UAV technology with historical aerial photographs and field survey, a multi-scale gully erosion monitoring method was established to monitor long-term changes in key indicators of morphological parameters and environmental factors of many gullies in Guyuan and Hunyuan experimental areas. In Guyuan, volume change range of 19 selected gullies is 133.70~8938.44 m³, and the mean erosion modulus is 4364.69 t/(km²·a) in the period of 1976-2017. In Hunyuan, volume change range of 19 gullies is 1034.89~49192.33 m³, and the mean erosion modulus is 5762.73 t/(km²·a) in the period of 1959-2018. Eroded volume of gullyhead retreat contributing to total eroded volume of gully is 6.1 times that of gully sidewall retreat, showing that primary mechanism of gully development is gullyhead retreat in Hunyuan during the study period.

The relationship of S-A topographic threshold in Hunyuan is S=0.0478A^-0.4915, indicating that dominant process of gully erosion is overland flow erosion. This threshold could be taken as indicator for the prediction of gully initiation potential location in marginal zone of Loess Plateau with similar environmental conditions. Regression analysis indicate that, long-term volumetric gullyhead retreat rate (V_e) can be better explained by present drainage basin area (A_p) compared with average slope gradient, and a relationship with power function form is obtained between them: V_e = 0.09A_p^0.60 (R² = 0.51, n = 30). Drainage basin area played a crucial role in gully initiation and long-term gullyhead retreat in marginal zone of Loess Plateau comparing with the relationships of S–A and A_p–V_e from other previous studies.

Taking Guyuan experimental area as an example, the influence of catchment area change on gully erosion development was analyzed. The monitoring results showed that there is a good correlation between gully catchment area and erosion modulus in 2017; The range of decreases in gully catchment areas by 2017 is 0.71%~77.88% compared with 1976. Main reason for changes in gully erosion moduli is the decrease in catchment area. The larger the decrease in catchment area, the smaller the gully erosion modulus is, and there is a relationship of exponential correlation between them. Thus, gully development can be effectively controlled by reducing gully catchment area. Further, this study showed that the main reason for decrease in gully catchment area is an increase in terraced areas in the study area comparing changes in the underlying ground surface over time. Terrace construction has enhanced runoff interception from upslope, changed original gully catchment area, reduced runoff into gully, and thus reduced gully erosion intensity.

 (2) The response relationships of water and sediment yield of slope gully system to the change of terrace measures were constructed and analyzed, and were further popularized in the typical watershed of Loess Plateau. The results showed that terrace measures can not only reduce local sediment production of themselves, but also have the effect of reducing sediment off-site; at the same time, the relationship curve has good applicability and generalization.

Coupling the gully erosion monitoring method and LCM-MUSLE model can estimate the erosion amount of slope gully system before and after terrace construction. In Guyuan, the reduction rate to the average annual erosion of each slope gully system is 30.93%~96.31% by terrace construction, with an average reduction ratio of 62.53%. In Hunyuan, the reduction rate to the average annual erosion of slope gully systems is between 13.69% and 99.04% by terrace construction, with an average reduction ratio of 51.60%.

Taking the slope gully systems of Guyuan and Hunyuan as samples, the relationship curve between terrace area proportion and sediment reduction rate of slope gully system was constructed. The results showed that the larger the terrace area is, the better the sediment reduction effect of slope gully system is. There is a significant logarithmic function relationship between them, and the correlation coefficient is more than 0.9. Further analysis of the relationship curve revealed that the sediment reduction rate of slope gully system is greater than the corresponding terrace area proportion, indicating that terrace measures can not only reduce the local sediment yield of themselves, but also have the effect of sediment reduction in other places.

Furthermore, the relationship curve between terrace area proportion and water and sediment reduction rates of slope gully system was applied to Yanhe basin and Gushanchuan basin, and the calculation of water and sediment reduction rates of 18686 slope gully systems was realized. The results showed that the relationship curve has good applicability and popularization.

(3) The impact of terrace spatial configuration on the “precipitation-runoff-sediment” process in the watershed was analyzed. The results indicated that the water and sediment reduction effect is better when terraces are configured on the lower part of the slope. When equal area terraces are configured in the lower part of the slope, the water reduction rate of Yanhe and Gushanchuan watershed can be increased by 16.60% and 7.80%, respectively, and the sediment reduction rate can be increased by 15.35% and 13.30%.

In this part, the effect of spatial configuration change on water and sediment reduction of terrace was analyzed, and it was found that the spatial configuration of terraces has a considerable influence on reducing the amount of water and sediment lost. The results showed when equal area terrace is located in the lower part of the slope, the water and sediment reduction rate is the largest, followed by the midslope and finally the upslope. Comparing the water and sediment reduction rate of equal-area terraces in realistic spatial configuration and virtual spatial configuration, it found that there is still much room for improvement in water and sediment reduction rate in Yanhe basin and Gushanchuan basin. Under the realistic terrace scenario, the average water reduction rate in Yanhe basin is 15.28%, and the average sediment reduction rate is 15.97%. The average water reduction rate in Gushanchuan basin is 8.68%, and the average sediment reduction rate is 7.83%. When equal area terraces are located on downslope, the rate of water reduction and sediment reduction increases the most. The rate of water reduction increases by 16.60% and 7.80%, and the rate of sediment reduction increases by 15.35% and 13.30%, respectively. Therefore, under the premise of improving the design standards for flood control of terraces and strengthening the regular maintenance of embankments to avoid flood damage to the terraces, terraces should be constructed as low as possible on hillsides to improve their water and sediment reduction rate.

(4) In this part, the impact of terrace scale change considering spatial configuration on “precipitation-runoff-sediment” process in the watershed was analyzed. The results indicate that the threshold of terrace scale that can effectively curb the sediment yield of watershed in hilly and gully region of Loess Plateau is 35% when random spatial configuration of terrace was set, and it can be further reduced to 25% when terraces are configured from downslope to upslope.

Through scenario setting, the water and sediment reduction rates of different scale terraces under the same precipitation and vegetation conditions were calculated, the response relationships between water and sediment reduction rate and terrace ratio under extreme and random spatial configurations of terraces were constructed. It was found that increasing terrace scale can further improve the water and sediment reduction in Yanhe and Gushanchuan basins, and with the increasing of terrace ratio, water and sediment reduction rates would firstly increase considerably before gradually slowing and eventually stabilizing. The threshold terrace ratio that effectively contains sediment yield in the hill and gully regions of the Loess Plateau is approximately 35% when random spatial configuration of terrace was set. This means that continuous increase of the scale of terraces does not significantly improve the sediment reduction effect when the terrace ratio is greater than 35%. The terrace ratio threshold will be 25% when considering the spatial configuration of terraces and constructing terraces from the downslope to upslope.

In conclusion, this paper can provide scientific basis and method support for the evaluation of water and sediment reduction effect of terraces in slope gully system of Loess Plateau and the optimal allocation of watershed terraces to achieve the maximum ecological benefits.