4. Time Series Modeling

When in surface water such as rivers, ponds and lakes detailed characteristic are not available, especially in developing countries, application of deterministic model are not applicable. In this regards, stochastic modeling are applied for estimating the future value of water quality parameters. There has been much effort in developing this technique for solving other engineering matters. Time series modeling as a stochastic model is trying to make probabilistic statements about the relation between system components and their future values and is used frequently in water quality management. In this chapter, after a preliminary explaining historical background of the method and introducing of time series modeling, the various methods such as Box-Jenkins methodology including stationary and non-stationary models and seasonal and non-seasonal models, exponential smoothing methods and Winter’s method, was stated and described in detail. Finally the application of time series modeling on Latian Dam, which located in the southeastern part of Tehran province in Iran, was discussed.