| 摘要: |
| 本文以北京市北运河通县水文站为例，使用新安江模型和降雨径流（NAM）模型进行通县断面的洪水预报。选取1980-2018年共18场历史洪水进行参数率定和验证，考虑到下垫面变化等因素，分为1980-2000年和2001-2018年两个阶段进行模拟。结果表明：（1）使用新安江模型1980-2000年的洪峰总体合格率达到75%，洪量总体合格率达到50%；2001-2018年间的洪峰总体合格率达到67%，洪量总体合格率78%，明显优于1980-2001年间的预报结果；（2）使用NAM模型1980-2000年间的洪峰总体合格率达到89%，洪量总体合格率达到40%；2001-2018年间的洪峰、洪量总体合格率均为78%；（3）NAM模型对于洪峰的模拟效果在两个时间段均优于新安江模型，但对洪量的模拟优势并不明显；NAM模型和新安江模型对于2001-2018年的洪峰和洪量模拟效果相近；（4）新安江模型中的不透水面积Imp、地表径流调蓄系数Cks、土壤含水总量Wm、壤中流调蓄系数Cki和自由水蓄量Sm对模拟结果较为敏感；NAM模型中的地表蓄水层蓄水容量Umax、浅层蓄水层土壤蓄水量Lmax和地表径流调蓄系数CQOF对模拟结果影响较大；（5）新安江改进模型的参数数量要多于NAM模型，对于产汇流过程的描述划分较为详细；NAM模型参数较少，设置相对简洁。建议根据流域、资料具体情况，综合选用模型进行模拟。 |
| 关键词: 北运河，NAM模型，MIKE 11，新安江模型，洪水预报 |
| DOI： |
| 分类号:P338 |
| 基金项目: |
|
| Flood Forecasting research of Tongxian based on Xinanjiang and NAM model |
|
Wang Yichen, Du Longgang
|
|
Beijing Hydrology center
|
| Abstract: |
| The flood forecasting research was focused on Tongxian which was located on Bei Channel riverside of Beijing. Xinanjiang Model and NAM model were used in the process of forecasting. Significant floods in the year of 1980-2018 were chosen to simulate. There were 18 historical floods. The results showed: 1) In the period of 1980-2000, the qualification rate of flood peak and flood total aggregate predicted by Xinanjiang model has reached 75% and 50% generally. Besides, in the stage of 2001-2018, the qualification rate of flood peak and flood total aggregate value were reached at 67% and 78% also predicated by Xinanjiang model, which was obviously better than the results of 1980-2000. 2) The qualification rate of flood peak and flood total aggregate predicted by NAM model in the stage of 1980-2000 were 89% and 40%, perspectively. In the stage of 2001-2018, the qualification rate of flood peak and flood total aggregate both reached at 78%. 3) NAM Model was more suitable than Xinanjiang model to simulate flood peak in two different stages, but it was weak in simulating flood total aggregate. In the period of 2001-2018, the two models performed similarly. 4) Parameters of Imp, Cks, Wm, Cki and Sm in Xinanjiang model were sensitive to the simulating results. In NAM model, parameters of Umax, Lmax and CQOF influenced the simulating results positively. 5) The number of parameters in Xinanjiang model was more than NAM model’s. The describing and dividing process in Xinanjiang of runoff water was quite detailed. On contrary, NAM model has less parameters and more simple than Xinanjiang model. Choosing the simulating model was decided according to the detail situation of drainage area and the history data. |
| Key words: Bei Channel, NAM Model, MIKE 11, Xinanjiang Model, Flood Forecasting |
