Use of stochastic model for a long-term forecast of water quality from groundwater aquifer of Khorramabad and Shiraz intermountain basins

Abstract

Long-term prognosis of groundwater aquifers water quality in Khorramabad and Shiraz basins of western Iran is an important part in evaluation of groundwater operational reserves. Such a forecast (for 1 year or longer) has been proposed based on the study of the laws of fluctuations in water quality in the past, reflecting the impact of the main factors on the water regime and the interaction in the system

Interaction between water-rock occurs during movement of groundwater. In this connection, it seems appropriate to look for in the ranks of such fluctuations legitimate (harmonic) components, which are possible to predict.

The investigated process can be random, i.e. the action is determined by a variety of factors that cannot be currently analytically accounted, and is subject to the normal distribution law. Such a process is characterized by rapidly decreasing normalized correlation function.

The investigated process in addition to the random component may contain one or more periodic components. The normalized correlation function of the number of decays acts much more slowly than in the first case, and must contain the harmonic components with one or another correlation time.

In the first case, the original number of observations does not contain regular components, i.e. is determined by the random component in the second – in the initial data series, there are legitimate components that can be extrapolated to the future.

For the preparation of long-term forecast a stochastic model of hydro-chemical process is proposed, which can be represented by a sum of several harmonic components, complicated by the random component. The model is implemented on a PC using mathematical calculations package MathCAD 2000 Professional. Information is given in the form of discrete data of hydrochemical observations, for example, per diem, monthly average or average value of the mineralization, any chemical element or substance.

Long-term prognosis has been compiled using data on changes in salinity and hardness of groundwater basins of Khorramabad and Shiraz for the past 10 years.

These results show that if the original number of oscillations of mineralization and water hardness is decomposed into periodic components entirely or with a small amplitude residue, the forecast of the changes in the groundwater is not difficult. If after selecting there is a random component with considerable amplitude, it is possible to make the forecast of the general direction of the process.