Back-Propagation Neural Network on Characterization of Groundwater Quality in Pingtung Champaign, Taiwan
Four clusters were classified according to the similarities and dissimilarities of water quality data collected from 1998 to 2008 of the 76 monitoring wells in Pingtung Champaign. The results showed that the groundwater quality of hinterland was better than that of coastal area. Some wells along the coastal areas have already been influenced by seawater intrusion incurring aquifer salinization. The overall grouping accuracy of cluster analysis (CA) for rainy and dry seasons was 95.8% and 97% according to the discriminant analysis, respectively. The results from the back-propagation neural network (BPNN) model showed no significant variations of nitrate (NO3-N) between the estimated and observed data, however most of the estimated total organic carbon (TOC) data were higher than the observed ones. It meant that the groundwater quality may be susceptible to the impacts of organic pollutants in the future within 90% confidence intervals. The results of this study show that application of CA and BPNN can simplify complex non-linear systems, moreover, groundwater quality can be estimated and modeled via monitoring data over the years. The modeling results can be utilized to frame the corresponding strategies to reduce the monitoring cost and to enhance the cost-effective benefits. The proposed analysis methods can be referred as a management plan for groundwater resources and pollution prevention to achieve the goal of early warning.
Y. C. Huang et al., "Back-Propagation Neural Network on Characterization of Groundwater Quality in Pingtung Champaign, Taiwan", Applied Mechanics and Materials, Vols. 58-60, pp. 2456-2461, 2011