标题:
基于BP人工神经网络模型的沙柳河流域径流模拟后处理研究Runoff Simulation Post-Processing of Shaliu River Basin Based on BP Artificial Neural Network Model
作者:
陈昕, 姚晓磊, 鱼京善
关键字:
BP人工神经网络, 沙柳河流域, 径流模拟后处理, SWAT模型BP-ANN, Shaliu River Basin, Runoff Simulation Post-Processing, SWAT Model
期刊名称:
《Journal of Water Resources Research》, Vol.5 No.4, 2016-08-18
摘要:
基于自主开发的BP人工神经网络模拟应用工具V1.0,将与SWAT模型相同的气象数据与新增的月校正因子和SWAT模型的月径流模拟值作为输入层变量,以实测径流为训练数据,对沙柳河流域径流模拟进行后处理,比较其与单独运用SWAT模型和BP人工神经网络模型进行径流模拟的结果,以此评价BP人工神经网络用于径流模拟后处理的精度与适用性。研究结果表明基于自主开发的BP人工神经网络后处理能显著提高该流域径流模拟精度,且操作简便,实现了模型的率定和验证同步进行,适用于沙柳河流域的径流模拟研究。
Two factors (monthly correction factor and monthly runoff simulation values of SWAT model) and the same meteorological data as SWAT were selected as input layer variables, and observed runoff as data for calibration and validation to do the post-processing of Shaliu River monthly runoff simulation based on self-developed BP Artificial Neural Network Tool V1.0. Meanwhile, the results were compared with SWAT model and BP-ANN model, evaluating the accuracy and applicability of runoff simulation post- processing. The application results in Shaliu River basin indicate that BP post-processing not only improves the simulation accuracy significantly, but also completes calibration and validation at the same time.