Multivariate analysis of sediment properties
To recognize the physico-chemical properties and sediment that controls the variability of the data, the results of a principal component analysis (PCA) of sediment characteristic variables are shown in Table 3. The first and second components together represent 79.3% of the variation. First PC is mainly determined by ACCE, Clay, and silt, and is negatively dependent on Vfs, D50 and Sand. This implies that correlated variables control variation in the data set. The second component showed strong negative correlations with pH and CCE, negatively correlated with OM, and EC. There were large differences between observations from various samples (Figure 3). This may be due to sampling from different rivers representing a wide diversity of properties such as texture, pH, EC, and ACCE. Overall, the PCA analysis revealed that variation of sediments properties is dominated by particle size distribution, and to a slightly smaller extent by EC, ACCE and OM. Particle size distribution varied, which was partly due to differences of geomorphological features of rivers, slope of terrain, and geological structure of region.