Prediction method for Tianjin coastal water quality

According to the periodicity and self-similarity,a new method for predicting the change characteristics of the coastal water quality is proposed through the study of the fractional theory based on the environmental monitoring data completely.Taking the coastal marine of Tianjin as an example,firstly,according to the fractional collage theory,the fractal interpolation based on affine transform is used to find the iterated function systems of the historical water quality.Secondly,the weighted summation method is adopted to find the iterated function system of the predicting period according to the above iter ated function systems based on the periodicity.Then the fractal predicting rmdel is established according to the iterated function system of predicting period.Finally,the random iterated algorithm is used to find the attractor of each prediction periods which can provide the prediction data according to the time values.The predicted results show that the average prediction error of the dry,rainy and mean water period are 29.6%,27.5% and 16.1%,respectively,and the average prediction error of the three periods is 24.4%.Through application it can be found that the fractional prediction method has high prediction preciseon.It is practicable and can povide the decision-making for the environment management of the coastal marine.