Analysis of Spatio-temporal Variation on Water Quality using a Statistical Markov Process with the Unobservable States

The mechanism of the water pollution process is becoming more complex due to changes in climate and river environment. There has so far been little effort to explore uncertainty considering these factors in water quality management. The water quality of rivers in Korea has become an issue and even led to a socio-political problem, especially after the environmental changes caused by the development project. We used a machine learning based classification apporoach to investigate the overall pattern of water quality changes over the past 16 years including the construction period. Water quality models are commonly based on a numerical-based deterministic model that has limitations representing stochastic behaviors properly. We employed a statistical Markov process approach to classifying the states of water quality within an unsupervised learning framework. Consequently, the spatio-temporal transition of water quality was accurately identified, and a discussion of the potential causes of the transition is offered.

 

KEYWORDS: Classification, Hidden Markov chain model, Water quality

 

Acknowledgement

This work is supported by the Korea Agency for Infrastructure Technology Advancement(KAIA) grant funded by the Ministry of Land, Infrastructure and Transport (Grant 19AWMP-B121100-04)