基于投影寻踪回归技术的塔里木河水质模拟分析

PROJECTION PURSUIT REGRESSION TECHNIQUE BASED MODELLING AND ANALYZING ON THE WATER QUALITY OF TARIM RIVER

利用投影寻踪回归技术（PPR）进行水质评价是一种新方法．本文在现有水文观测资料基础上，通过PPR分析，建立了塔里木河水质模型，并利用PPR对影响河流水质的因素作了分析，可为流域水质管理和控制提供依据．

The water quality models by using computer technique have been widely applied in water quality control and evaluation because of their specialities of saving time and cost as well as the convenience.It is a new method to use PPR(projection pursuit regression) model to evaluate water quality. PPR is to use projection pursuit technique to deal witb and analyze the interre1ations between multidimension observated data, Y and X, and then a regression model Y=F(X) is relevantly extablished. Based on the analyses on teh data of routine hydrologic observation, two water quality indexes, aminonitrogen and chloride are selected as the analogue targets bccause they have the representativeness and univcrsal overstandard. The 11 variable factors are selected, and all these data are input the PPR software. The water quality models of the conccntration of aminonitrogen and chloride are established at Xinquman Hydrometric Station of Tarim River. The analogic results are ideal. Meanwhile, by using these two water quality models, the influence of each variable on the concentration of aminonitrogen and chloride is analyzed. It is pointed out that these two water quality indexes not only have the close relation to discharge, but also to the concentration of aminonitrogen and chloride at the upper reaches, which belong to point-source pollution.PPR water quality model is different from the others. It can be used for researching the movement law or pollutants in water and their interrelations without any supposition just laying stress on the obaerved data directly so as to analyze and obtain the various informahon and practical law from the Firsthand data. Therefore. the results are reliable.