近年来，在气候变化的大背景下，快速的城市化使得许多城市水问题日益凸显。其中，城市暴雨洪涝造成的经济损失巨大，是当前威胁城市可持续发展的主要问题之一。随着城市雨洪模型和区域尺度数值大气模式的不断发展，通过陆气耦合进行城市暴雨洪水模拟预报能够在保证一定精度的情况下有效延长预见期，在一定程度上解决以往水文预报中存在的预见期和精度难以两全的问题，为城市的防汛应急预案制定提供更加充足的时间，同时还可以为缺乏降水资料地区的相关研究提供一定的数据支撑，是当前实现合理的城市暴雨洪水管理的重要手段。北京是我国暴雨洪涝灾害最为严重的城市之一，为了给北京地区雨洪模拟提供一种新的思路，为洪水预报预警能力的提高提供一定依据，本研究以北京市通州区的城北片区、河西片区和两河片区为研究区，基于天气研究和预报模式（Weather Research and Forecasting Model，WRF模式）和暴雨洪水管理模型（Storm Water Management Model，SWMM）建立了研究区陆气耦合雨洪模型，分析探讨了以大气模式的模拟降水驱动城市水文模型进行暴雨洪水模拟的效果及可行性。研究主要内容与成果如下：

（1）基于WRF模式对研究区强降水过程进行了模拟，并对物理参数化方案进行了优选。研究基于云微物理过程、积云对流过程、行星边界层过程以及陆面过程设置了16种不同的参数化方案组合，结合实测降水资料评估模拟效果。结果表明：大气模式对强降水的模拟具有一定的不确定性，对物理参数化方案进行优选可以有效提高强降水过程的模拟精度；不同的参数化方案及其组合对模拟效果的影响程度各不相同，其中积云对流参数化方案对于研究区强降水模拟的影响最大。当云微物理过程取WSM5，积云对流过程取Grell-Devenyi，行星边界层过程取Mellor-Yamada-Janjic，长、短波辐射过程取RRTMG，表层方案取Monin-Obukhov2，陆地表面取Noah，城市表面取城市冠层模型时，模拟结果最优。

（2）基于SWMM模型建立了研究区的城市雨洪模型。研究依据实测和设计数据，结合当地实际情况对研究区进行了概化，共得到293个节点、314条渠段和145个子汇水区，基于实测洪水数据进行模型模拟效果评估。结果表明：不同的模型参数的调整对模拟结果的影响程度不同，按敏感性从大到小依次为最小下渗率、透水区洼蓄量、透水区曼宁系数、不透水区洼蓄量、不透水区曼宁系数、最大下渗率和渗透衰减常数；经过模型参数率定后，模型在率定期的纳西效率系数、洪峰流量相对误差和峰现时间绝对误差均满足《水文水情预报规范》相关要求，在验证期的综合径流系数也在规定范围内，可以对研究区的产汇流过程进行有效模拟，提供相应的洪水流量数据。

（3）采用单向耦合的方法，构建了研究区陆气耦合暴雨洪水模拟模型。研究将经过处理得到的区域平均模拟降水和分区平均模拟降水作为输入驱动雨洪模型，结合基准流量评估模拟效果，讨论陆气耦合城市暴雨洪水模拟的可行性。结果表明：耦合模型的模拟精度满足实际作业需求，将WRF模式和SWMM模型耦合进行城区暴雨洪水的模拟是可行的；对于流域出口洪水流量过程的模拟，分区平均模拟降水方案与区域平均模拟降水方案差别不明显，后者的运算效率相较于前者会更高一点。在实际应用中，如果是将流域出口的洪水流量过程作为重点关注对象的话，采用区域平均模拟降水方案进行模拟即可达到要求。

In the context of climate change, rapid urbanization has made many urban water problems increasingly prominent. The urban flood due to storm is one of the main disasters that threaten the sustainable development of cities because of the huge economic loss. With the continuous development of urban storm and flood models and regional-scale numerical atmospheric models, the simulation of urban storm and flood based on coupled atmospheric-hydrological modeling can effectively extend the leading time while ensuring a certain accuracy. At the same time, it can also provide data support for related research in areas lacking precipitation data. Thus, using the coupled atmospheric-hydrological modeling is an important means to achieve reasonable urban storm and flood management with providing more sufficient time to formulate flood control emergency plans for cities to reduce losses. Beijing is one of the cities with the most severe storm and flood disasters in China. In order to provide a new view for the simulation of storm and flood in Beijing and a certain basis for the improvement of flood forecasting and early warning capability, this study established a coupled atmospheric-hydrological modeling system based on WRF（Weather Research and Forecasting Model）and SWMM（Storm Water Management Model）by taking the Chengbei, Hexi and Lianghe areas of Tongzhou District, Beijing as the study area. On this basis, the effect and feasibility were discussed finally. The main research contents and conclusions are as follows:
（1）Based on WRF, the heavy precipitation processes in the study area was simulated, and the physical parameterization schemes of the model were optimized. Considering cloud microphysical schemes, cumulus convection schemes, planetary boundary layer schemes and land surface schemes, 16 different parameterization scheme combinations were set in this study. And then evaluated the effect combined with observed precipitation data. The results show that the atmospheric model has a certain degree of uncertainty in the simulation of heavy rain, and the cumulus convective parameterization scheme has the greatest impact on the simulation of precipitation in the study area; Different schemes and their combinations have different effects on the simulation results while the combination of the WSM5 cloud microphysical scheme, the Grell-Devenyi cumulus convection scheme, the Mellor-Yamada-Janjic planetary boundary layer scheme, the RRTMG long-wave and short-wave radiation scheme, the Monin-Obukhov2 surface scheme, the Noah land surface and the urban canopy model is the best in the study area.
（2）Based on SWMM, an urban storm and flood model of the study area was established. This study set 293 junctions, 314 links and 145 subcatchments totally by generalization and division based on measured data, design data and actual conditions. And then evaluated the effect combined with observed flood data. The results show that the adjustments of different model parameters have different effects on the simulation results, and in descending order of sensitivity, they are the minimum infiltration rate, the Manning coefficient of the pervious area, the depression storage of the pervious area, the depression storage of the impervious area, the Manning coefficient of the impervious area, the maximum infiltration rate and the decay constant; After parameter calibration, the Nash efficiency coefficient, the relative error of the peak discharge and the absolute error of the peak present time of the simulation results in calibration period all meet the relevant requirements of Standard For Hydrological Information And Hydrological Forecasting and the comprehensive runoff coefficient in verification period is also within the prescribed range, so that the model can effectively simulate the process of runoff generation and confluence in the study area and provide corresponding flood data.
（3）A coupled atmospheric-hydrological modeling system for storm and flood simulation of the study area was established by one-way coupling method. This study used the processed simulated precipitation for two different schemes as input to drive the storm and flood model. Then evaluated the effect and feasibility combined with base flow. The results show that it’s feasible to simulate the storm and flood in urban area by coupling WRF and SWMM, because the simulation accuracy of the model meets the actual operation requirements; For the simulation of the flood at the outlet, there is no obvious difference between the subregional precipitation scheme and the regional precipitation scheme, and the latter has a higher computational efficiency than the former. So in practical applications, if the flood at the outlet is the focus, the coupled atmospheric-hydrological modeling with regional precipitation scheme can meet the requirements.