信息熵在新安江模型和NAM模型比较中的应用

Application of Information Entropy in Comparing Xinanjiang Model with NAM Hydrological Model

选取资水流域新宁水文站以上集水面积为研究区域,基于空间分辨率为30s的数字高程模型数据构建数字流域水系,运用新安江模型和NAM模型分别对新宁站的日径流过程进行模拟。采用确定性系数和信息熵对两个模型的模拟结果进行比较分析。在信息熵计算过程中,以实测流量为标准序列,计算两种日模型模拟结果与实测序列残差,然后进行归一化,计算潜在信息测度和信息熵相对测度。信息熵越大表明流量计算序列与实测序列过程线在垂向上相似性强,但这并不代表确定性系数大、径流深相对误差小。

The area controlled Based on the 30s-resolution g by the Xinning Hydrometric Station on the Zishui River was selected as the target region. rid-based digital elevation model data, the digital drainage network was generated. Then Xinanjiang Hydrological Model and NAM Hydrological Model were used to simulate the daily streamflow processes at the Xinning Hydrometric Station. The simulated streamflow was analyzed and compared by using the Nash-Sutcliffe Model efficiency coefficient and information entropy. When the information entropy was computed, the residual square between the simulated results and observed streamflow series were calculated and normalized, and the potential information and the relative information entropy of the simulated streamflow series were derived. The results showed that the higher value of relative information entropy coulci indicate larger similarity of simulated and observed discharge series in the vertical direction, and would not mean higher Nash-Sutcliffe coefficient and lower relative error of runoff depth.