Analysis About The Temporal And Variation Characteristics Of Water Quality And Its Influence By Land Use Composition And Pattern In Xitiaoxi Watershed

With the control of point source pollution,non-point source pollution is becoming a key factor for water pollution in Xitiaoxi watershed play an important role in protecting Taihu lake as it's the upstream of Taihu lake.It will make contribution for the continuous improvement of water quality in Taihu Iake by making certain the condition of non-point source pollution of Xitiaoxi watershed.Land use condition playing an important part in non-point source pollution had been proved in numerous studies.Therefore,this study want to do some analysis in this field by water quality data collected and analyzed from June 2009 to July 2010.First,get a better information about temporal and spatial variation of Xitiaoxi watershed.Second,research the influence mechanism of water quality by land use structure and pattern which was obtained by TM remote sensing images in 2010 and the using of geographic information technology.In the end introducing the spatial no stationarity,make a deep study about their relationship using the area weighted function and geographical weighted regression model.The results are as follows:(1)The watershed can be classified into four parts,the upstream covered with forest,the main river through the city,the downstream main river and the downstream through irrigation district from the spatial dimensionally.Only N element show the sensitivity in the entire study period from the time dimensionally.We found that all parts showing little difference between flat period and dry period but significant difference with wet period.The main river through the city show the minimal difference in the all of the watershed,which show the persistent anthropogenic interference.(2)By using the partial redundancy analysis,the individual contribution rate of each factor affecting the water quality is analyzed from the whole.The results showed that the land use factors,other natural geographical factors and their interaction make different contribution to water quality.But land use factors lead much higher than other natural geographical factors in every period.(3)The vast majority of water quality parameters showed significant or extremely significant correlation with land use types,but their relationship changed in different period.Land use intensity fit very well with composite pollution index(CPI),which showed that increasing land use intensity has direct correlation with water quality.(4)Among the land use types,the stronger connectivity,the greater the dominance of a land use type,the larger the area of the main patch,the better the quality of the water;The more the land use pattern is broken,the more uniform distribution of land use types,the worse the water quality.Land use pattern accounted the highest for water quality in low water period and the lowest in flat water period.(5)Land use composition predicted much better for TN and TP after adding weight decay function.Meanwhile,linear attenuation weight function showed better prediction than exponential decay weight function.(6)Different water quality parameters showed different spatial autocorrelation,including TN,NH4+,COD and PO43-show extremely significant spatial positive correlation(p<0.01),TP and NOx-show significant spatial positive correlation(0.01<p<0.05),SS and OM did not show any spatial autocorrelation.Using GWR and OLS models to do regression analysis and compared their advantages with TN as dependent variable factor and land use pattern indexes as independent variables.All regression results show that GWR models are better than OLS models.The results show that GWR model can be very good solution of spatial autocorrelation in entire watershed,which can more accurately analyze the impact of water quality by land use pattern.