Presenter/Author Information

L. Matejicek

Keywords

multivariate analysis, gis, urban environment, surface water pollution

Start Date

1-7-2006 12:00 AM

Abstract

Integrating modelling of water pollution and GIS enables the connection of environmental process models with geospatial data describing the physical environment. Time series of water pollution data at monitoring profiles are used to complement the spatial database and to interpret the results of data analysis. In this case, the multivariate statistical methods provide an exploratory environment for data analysis and an indication of seasonal changes in the framework of surface water pollution. In addition to a wide range of useful multivariate methods, principal component analysis (PCA) and factor analysis (FA) are used to differentiate seasonal water pollution at monitoring profiles. Integrating modelling of water pollution is demonstrated for a river basin in the urban area of Prague. Data series from long-term measurements (25 years; seasonal measurements for PCA and FA in the period 2001-2004) are used to study the variability of water quality parameters. Consecutively, the PCA and FA are carried out to identify seasonal deviations originating from the time series of water temperature, pH, conductivity, suspended solids, nitrates, phosphates, BOD, COD, etc. The graphs focused on PCAs’ loadings and FAs’ biplots show the standard data outputs. The data from the multivariate seasonal exploratory analysis are transported into GIS to map the changes of FAs’ loadings. The maps of changes are then used to estimate the observed seasonal strengths of the given processes that include simultaneous changes in water pollution parameters. As an example, the seasonal strength of NO3 is mapped and compared together with other strengths of remaining parameters. The significant changes of FAs’ loadings are observed between the winter seasons and the summer seasons (2001-2004), which is in correspondence with the original data. The described exploratory tools are developed to support decision-making processes in the framework of water pollution management.

COinS
 
Jul 1st, 12:00 AM

Modelling of Water Pollution in Urban Areas with GIS and Multivariate Statistical Methods

Integrating modelling of water pollution and GIS enables the connection of environmental process models with geospatial data describing the physical environment. Time series of water pollution data at monitoring profiles are used to complement the spatial database and to interpret the results of data analysis. In this case, the multivariate statistical methods provide an exploratory environment for data analysis and an indication of seasonal changes in the framework of surface water pollution. In addition to a wide range of useful multivariate methods, principal component analysis (PCA) and factor analysis (FA) are used to differentiate seasonal water pollution at monitoring profiles. Integrating modelling of water pollution is demonstrated for a river basin in the urban area of Prague. Data series from long-term measurements (25 years; seasonal measurements for PCA and FA in the period 2001-2004) are used to study the variability of water quality parameters. Consecutively, the PCA and FA are carried out to identify seasonal deviations originating from the time series of water temperature, pH, conductivity, suspended solids, nitrates, phosphates, BOD, COD, etc. The graphs focused on PCAs’ loadings and FAs’ biplots show the standard data outputs. The data from the multivariate seasonal exploratory analysis are transported into GIS to map the changes of FAs’ loadings. The maps of changes are then used to estimate the observed seasonal strengths of the given processes that include simultaneous changes in water pollution parameters. As an example, the seasonal strength of NO3 is mapped and compared together with other strengths of remaining parameters. The significant changes of FAs’ loadings are observed between the winter seasons and the summer seasons (2001-2004), which is in correspondence with the original data. The described exploratory tools are developed to support decision-making processes in the framework of water pollution management.