Solutions for Water Quality in Northern Iran Using Multivariate Statistical Techniques.

Deforestation in northern Iran is a great challenge these days. The percentage of deforestation in Guilan, Mazandran, and Golestan Provinces was 21, 49, and 69%, respectively, between 1988 and 2004, and continues today. The aim of this paper is to investigate the catastrophic effect of deforestation on water quality (a thirty-year period) in Golestan Province (Northern Iran) and present the best approach to ameliorate the problem. Total dissolved solids (TDS) were used as an indicator of poor water quality. Principal component analysis (PCA), partial least squares (PLS) and principal component regression (PCR) were used to predict total dissolved solids (TDS). The results indicated that TDS has increasingly significant year by year due to increasing deforestation in the area. An exponential semivariogram using PLS predictions revealed stronger spatial dependence among TDS [r2 = 0.90, and RMSE= 0.082] than PCR [r2 = 0.91, and RMSE= 0.091]. Therefore, PLS method indicated the best fit for the data. The results show that water quality will decrease annually if the current situation continues. The best solution to stop or at least alleviate the problem is to recommend appropriate management program, such as afforestation, in the area.