Forecasting and Trend Analytics of Water quality parameters using ARIMA series Models in Kahman river watershed

Kahman is the most beneficial river in Alashtar city for agriculture and aquiculture. As hydrology processes have random nature, statistics and probability are base of analysis of these processes and time series are used for this purpose. The first step in time series analysis includes parameters variation through time. Second step is to stationary data, third is normalization and forth is model parameters recognizing. Finally, for model accuracy evaluation in prediction, the root-mean square standardized error and Akaike information criteria are used. In this research, time variations trend of three parameters, PH, HCO3- and Na that are effective parameters on drinking and agriculture water, were studied at Darreh-tang station of Kahman river from 1366 to 1392 and then Based on the time-series graphs, as well as autocorrelation and partial autocorrelation plot, Multiplicative seasonal ARIMA model select and in XLSTAT and MINITAB software’s was used. For HCO3 and PH, ARIMA4(1¡1¡1)*(1¡1¡1) was recognized for prediction optimum model but for NA any of ARIMA models weren’t recognized suitable. HCO3 always have an ascending trend. On the base of time series diagram, optimum variations limit of PH is from 6.5 to 8.5. Because the formation in Kahman river watershed is calcareous and because of nonexistence of salty formations in studied region, Na has a constant trend and isn’t ascending.