Integrating geographical information system and multivariate statistical approaches to assess land use impacts on water quality in Nzovwe river catchment in Mbeya, Tanzania

This study aimed to investigate water quality in relation to adjacent land use patterns in the Nzovwe river catchment in Mbeya, Tanzania. An integrated Geographic Information System (GIS) and multivariate statistical analysis approaches for evaluating the correlation between water quality and land use were applied. Electrical conductivity (EC), pH, total dissolved solids (TDS), turbidity, color, dissolved oxygen (DO), temperature, salinity, nitrates, phosphate, total coliform, fecal coliform and Escherichia coli were analyzed in the wet and dry seasons at 15 sampling sites. The geo-processing tool assisted in defining the influence of land use at each sampling site, as well as individual identification of land use in each area (500 and 60 m river buffer zones). Results showed that higher levels of pH, turbidity, DO and temperature were observed during the wet season, while all other parameters showed no significant seasonal variation. Results from Principal Component Analysis (PCA) showed that PC1 contributed 29.505% of the total data variance, with positive loading factors for EC, salinity, TDS and temperature. PC2 explained 20.19% of the total variance and had a high positive loading factor for fecal coliform, E. coli and total coliform. PC3 explained 16.54% of the total variance and had a high positive loading factor for color, turbidity and DO. Furthermore, the findings show that the catchment has poor water quality, suggesting possible pollution from built-up and agricultural land uses. These findings highlight the urgent need for increased localized monitoring stations and improvement in the policy and law implementation in the catchment.