Abstract
The hydrologic conceptualization of virtual condition is relied on the behavior of monitored data at various monitoring gauges but there are frequently missing values in the monitoring database. Reasonably estimating these missing values is important for the complete analysis and modeling of hydrologic cycle. The linear regression and multi-variable interpolation of monitored data were often used to estimate the missing data or check the inadequate data of the hydrologic database. But in many instances, the relationships between the predicting and predicted variables are not truly linear, and the non-linear forms of their relationships are very difficult to be known. In this paper the Artificial Neural Network (ANN) based model is employed using the pre-construction period data to evaluate the inadequate data and missing data in the hydrologic dataset without understanding the detail relation between input and output data elements. Prediction results present that ANN method with highest coefficient of determination (R2) could be the desirable and preferred choice for evaluating the missing hydrologic data.
