An integrated study on change detection and environment evaluation of surface water

Abstract Surface water pollution is one of the serious environment pollution problems, posing threat to human and other creatures. Extraction, change detection and environment evaluation of surface water are prerequisite for water resource management. Undoubtedly, remote sensing data play an important role in these researches because of its large geographic coverage and high temporal frequency. In this study, the Tianjin Binhai New Area was chosen as the study area and the surface water extraction method Modified Normalized Difference Water Index (MNDWI) was used by combining with adaptive dynamic threshold to extract surface water and detect its change. Comparing with single-band threshold, model of multi-band spectral relationship, Iterative Self-organizing Data Analysis Technique Algorithm and MNDWI, MNDWI-based adaptive dynamic threshold method performed better, which considered the influence of background. Analysis on dynamic change of water showed the area of lake and river had increased and the area of seawater had decreased. Meanwhile, the correlation analysis between area change of surface water and impact factors indicated both climatic and anthropogenic factors made positive contribution to the present water environment situation. Finally, an improved model of surface water environment evaluation was established to evaluate water quality by combining genetic algorithm (GA) and backpropagation (BP) neural network model. And the test results proved that the optimized GA-BP neural network is better than the single BP neural network. The environment evaluation indicated that water quality of the Haihe River section in the study area was poor. Therefore, water environment protection should be strengthened in this area. Some suggestions on practical management were given accordingly.