中文摘要：

针对以往生态动力学模型DYRESM-CAEDYM 在应用中, 多采用手动方法确定参数值时较为费时费力的缺点,引入蒙特卡洛随机抽样算法对模型参数进行自动率定。以广东省大沙河水库为例, 选择率定期为2011 年11 月至2012年2 月, 利用2012 年7 月至2012 年11 月实测水质数据对模型进行验证。结果表明, 在率定阶段模型对水温的模拟结果非常好, 表层和底层温度的均方根误差（RMSE）分别为3.53 ℃和2.00 ℃, 相对误差（RE）值均小于0.15; 模型对溶氧的模拟结果良好, 表层和底层的RMSE 值分别为1.23 mg？L？1和1.85 mg？L？1, RE 值分别为0.08 和0.20。在验证阶段, 温度表层和底层的RMSE 分别为2.49 ℃和1.67 ℃, RE 值均小于0.1; 溶氧在表层和底层的RMSE 分别为2.15 mg？L？1和0.79 mg？L？1, RE分别为0.17 和0.11。因此, 经采用蒙特卡洛方法率定后得到的大沙河水库DYRESM-CAEDYM 模型参数值较为可靠。

英文摘要：

In the previous applications of hydrodynamic-ecological model （DYRESM-CAEDYM）, the parameter values were arbitrarily decided by model users. In order to improve the precision of model prediction, the Monte-Carlo sampling method was introduced into auto-calibration process of DYRESM-CAEDYM. The auto-calibration was conducted with monthly water quality measurements （temperature and dissolved oxygen） during November, 2011-February, 2012 at Dashahe Reservoir in Guangdong Province. The calibrated model was validated with the water quality data measured during July, 2012-November, 2012. The model outputs of temperature showed good agreement with the observations during the autocalibriton. The RMSE values of water temperature at the surface and bottom were 3.5℃ and 2.0℃. The RE values were less than 0.15. The model was also good at simulating dissolved oxygen with RMSE values of 1.23 mg.L-1 （surface） and 1.85 mg.L-1 （bottom） and RE values of 0.08 （surface） and 0.20 （bottom）. In the validation process, the RMSE values of water temperature were 2.49 ＂C at the surface and 1.67℃ at the bottom. The RE values were both less than 0.1. The RMSE values of dissolved oxygen were 2.15 mg.L-1 （surface） and 0.79 mg.L-1 （bottom）. The RE values for dissolved oxygen were 0.17 （surface） and 0.11 （bottom）. The successful calibration for DYRESM-CAEDYM at Dashahe Reservoir suggests that the model performance might be improved with automatic calibration based on Monte-Carlo sampling method and high-speed computer.