土壤侵蚀破坏土壤结构，降低土壤持水能力，影响大豆的生长和产量的形成。东北黑土区由于其肥沃的土壤和较高的土地生产力成为我国重要的粮食生产基地，但随着百余年毫无保护措施的开垦，黑土地已出现严重的侵蚀现象。这种侵蚀已经危胁到国家的粮食生产安全。东北黑土区以旱作“雨养农业”为主，旱灾在当地时有发生，侵蚀的发生直接影响土壤对降水的利用。因此研究该区域不同侵蚀土壤在不同干湿年景对大豆产量和生长过程的影响尤为重要，可为农业生产提供指导和预警作用。

目前，在东北黑土区土壤侵蚀研究多集中在影响土壤理化性质、产流产沙、秸秆还田与免耕对侵蚀阻控等方面，侵蚀影响土地生产力研究侧重与影响结果分析，缺乏侵蚀影响机理研究。由于侵蚀明显改变土壤性质和持水能力，在雨养农业的情况下，这种改变是否影响大豆生长-水分关系？遇到不同的降水年景会有何影响？回答这些问题对维护黑土地生产力，维持黑土资源可持续利用具有重要意义。因此本研究主要开展三方面研究：（1）通过盆栽试验，研究不同侵蚀强度土壤的大豆生长过程和产量对水分的响应机理，进行作物生长模型公式修订和参数率定；（2）利用田块尺度土壤剖面数据和多年大豆生长过程生态、水分和产量等指标的连续动态监测，对修订和率定后的作物生长模型进行验证，并根据模拟过程指标变化，揭示不同侵蚀强度土壤影响大豆生长和产量、及其在不同降水年景的变化机理；（3）将率定后的作物生长模型用于黑龙江省嫩江鹤山农场面积约30km²的鹤北小流域，用修订和率定后的模型模拟1951-2019年共计69年流域产量，并结合流域内黑土厚度数据，评价土壤侵蚀对产量的影响及不同降水年景下的产量变化。主要结论如下：

（1）不同侵蚀强度土壤的大豆产量-水分响应关系存在明显差异。当土壤水分含量达到田间持水量FC（Field Capacity）的80%时（80%FC），轻度侵蚀土壤的大豆生长各生态指标为最优，而中度和重度侵蚀土壤的大豆生态指标到最优的土壤充分供水时的水分含量分别为100%FC和130%FC，表明侵蚀导致土壤持水能力明显下降，需要维持更高的含水量实现充分供水。值的注意的是，在最优供水条件下，中度侵蚀土壤的大豆产量最大，为3.35 t/ha，轻度和重度侵蚀土壤的大豆产量分别为3.10 t/ha和2.50 t/ha。本试验所取中度侵蚀使土壤通气性增强而使产量有所提高，其侵蚀上限为多少需要进一步深入研究。土壤水分胁迫会影响大豆的光合器官叶面积的生长。在缺水胁迫下，轻、中、重度侵蚀的水分胁迫造成的叶面积衰减速度增大。土壤水分胁迫还会影响大豆的光能利用率，在缺水胁迫下，轻、中、重度侵蚀的水分胁迫造成的光能利用率衰减速度减小。渍水胁迫主要对轻度和中度侵蚀土壤的大豆叶面积和生物量生长有影响，重度侵蚀土壤基本不存在大豆的渍水胁迫。

（2）优化了水分胁迫函数的大豆生长模型模拟精度明显提高。基于上述研究结果，对大豆生长模型进行了优化，模拟精度明显提高：优化前模型模拟的轻度和中度侵蚀土壤的大豆产量偏低6.97 %和9.61 %，重度侵蚀土壤的大豆产量偏高13.54 %。优化后模型模拟的轻度和中度侵蚀土壤的大豆产量偏低0.90 %和4.04 %，重度侵蚀土壤的大豆产量偏高0.82 %，原因是轻度和中度侵蚀强度土壤水分的模拟结果偏低，导致对叶面积指数的模拟结果偏低，而重度侵蚀强度土壤水分的模拟结果偏高。模型对地下生物量模拟精度较高。

（3）不同侵蚀强度地块的大豆产量及其对干湿年份的响应有明显差异。9年大豆产量在不同降雨年景的产量变化结果显示：轻度和中度侵蚀土壤的大豆产量在偏涝年份最大，重度侵蚀土壤的大豆产量在涝年最大，即轻度和中度侵蚀土壤相较重度侵蚀土壤不耐涝，由其持水能力差异所致。当降水减少时，轻度侵蚀土壤由于持水能力强，其大豆产量较中度和重度侵蚀土壤降低缓慢。

（4）流域不同侵蚀强度土壤的大豆生长和水分特征在干湿年景有明显差异。湿年景下大豆各生态指标在侵蚀强度强的薄层和中层黑土呈增产趋势，而侵蚀强度小的厚层黑土有减产趋势，前者水分利用率低于后者。偏湿和正常年景下，大豆的各指标均为增产趋势，但增产幅度有别：薄层黑土地下生物量增产幅度高于其他四个指标，水分利用率小于其他四个指标；中层黑土大豆产量和水分利用率增加最为明显；厚层黑土地下生物量变化不大。偏干和干年景土壤侵蚀对产量影响的差异不显著，各种侵蚀强度下的大豆生态指标均减少，但水分利用率明显增强。

Soil erosion destroys the soil structure, reduce the water holding capacity of the soil, thus affecting the growth and yield formation of soybeans. The black soil area in Northeast China has become an important food production base in China, due to its fertile soil and high productivity. But with predatory cultivation for more than a hundred years, serious erosion has occurred in the black soil; which has threatened the country's food production security. The black soil area in Northeast China is mainly characterized as dry farming and rain-fed agriculture. Drought has occurred in this region. The occurrence of erosion would directly affect the utilization of precipitation by the soil. Therefore, it is particularly important to study the impacts of different eroded soils on soybean yield and growth in different dry and wet years in this region, which can provide theoretical guidance and warning for agricultural production.

At present, research in the black soil area of Northeast China mainly focuses on physical and chemical properties of soil, runoff and sediment production, straw returning to the field, no tillage, and erosion rate, etc. Researches on soil productivity mainly focuses on results analysis, lacking researches on the mechanism of soybean yield impact caused by erosion. As dry farming and rain-fed agriculture, has erosion changed the growth water relationship of soybeans under the influence of soil properties and water holding capacity? How do different precipitation years affect it? Answering these questions are of great significance to maintain the productivity and sustainable utilization of black soil resources. Therefore, this doctoral dissertation conducts researches from three aspects: (1) Three types of erosion intensity soils were collected for pot experiments to study the response patterns of soil water stress with different erosion intensities to soybean growth indicators, and the parameters of soybean growth models were determined; (2) Collect soil profile data from 19 plots and soybean growth indicator data monitored for many years. Based on the validation of soybean growth models, use the model simulation results to elucidate the reasons for the impact of erosion on soybean growth process; (3) The crop growth model was applied to the Hebei small watershed,  which covers an area of 30km2. The revised and calibrated model was used to simulate the basin yield for 69 years from 1951 to 2019. Furthermore, combined with the black soil thickness data in the basin, the effects of soil erosion on the yield and the yield changes in different precipitation years were evaluated. The chief conclusions are as follows:

(1) There are significant differences in the yield water response relationship of soybean in soils with different erosion intensities. When the soil moisture content reaches 80% of the field capacity FC (Field Capacity), the ecological indicators of soybean growth in lightly eroded soil are optimal. However, when the ecological indicators of soybean in moderately and severely eroded soil reach the optimal level, the water content is 100% FC and 130% FC, respectively. This indicates that erosion leads to a significant decrease in soil moisture capacity and requires maintaining a higher water content to achieve sufficient water supply. It should be noted that under the optimal water supply conditions, the soybean yield in moderately eroded soil is the highest, at 3.35 t/ha, while the soybean yield in lightly and severely eroded soil is 3.10 t/ha and 2.50 t/ha, respectively. The moderate erosion in this experiment enhances soil aeration and increases yield, and further research is needed to determine the upper limit of erosion. Soil water stress can affect the growth of leaf area in soybean photosynthetic organs. Under water scarcity stress, the rate of leaf area decay caused by light, medium, and severe erosion water stress increases. Soil water stress can also affect the light energy utilization efficiency of soybeans. Under water scarcity stress, the decay rate of light energy utilization efficiency caused by light, medium, and severe erosion of water stress decreases. Waterlogging stress mainly affects the leaf area and biomass growth of soybeans in mild and moderate eroded soils, while waterlogging stress of soybeans is not present in heavily eroded soils.

(2) The simulation accuracy of the soybean growth model optimized for water stress function has significantly improved. Based on the above research results, the soybean growth model was optimized and the simulation accuracy was significantly improved: the soybean yield simulated by the model before optimization was 6.97% and 9.61% lower in mild and moderate erosion soil, while the soybean yield in severe erosion soil was 13.54% higher. After optimization, the soybean yield simulated by the model for mild and moderate erosion soil was 0.90% and 4.04% lower, while the soybean yield for severe erosion soil was 0.82% higher. The reason is that the simulation results for mild and moderate erosion intensity soil moisture were lower, resulting in lower leaf area index simulation results, while the simulation results for severe erosion intensity soil moisture were higher. The model has high accuracy in simulating underground biomass.

(3) There are significant differences in soybean yield and its response to dry and wet years in soils with different erosion intensities at the plot scale. After dividing the soybean yield in the field monitored for 9 years into drought and flood levels based on rainfall, the analysis results of soybean yield showed that the soybean yield in lightly and moderately eroded soil was the highest in the year of partial flooding, while the soybean yield in severely eroded soil was the highest in the year of flooding. That is to say that slightly and moderately eroded soil are less tolerant to flooding compared to severely eroded soil, which is due to differences in their water holding capacity. When precipitation decreases, the soybean yield of lightly eroded soil decreases slower than that of moderately and severely eroded soil due to its strong water holding capacity.

(4) There are significant differences in the growth and water characteristics of soybean in different erosion intensity soils at the watershed during dry and wet years. Under wet year conditions, the ecological indexes of soybean are generally higher under thin and moderate layer black soil, while decreased under light erosion intensity, and the water use efficiency of the former was lower than that of the latter. Under wetter year and normal conditions, all indexes of soybean showed an increasing trend, but the increasing range was different: the increasing range of underground biomass under thin layer black soil was higher than the other four indexes, and the water use efficiency was lower than the other four indexes. The increase of soybean yield and water use efficiency under moderate layer black soil was the most obvious. There was little change in underground biomass under thick layer black soil. Under dry year conditions, the ecological indexes of soybean are generally decreased under all erosion intensities, but the water use efficiency increased significantly.