Past Work as Ph.D Candidate

Machine Learning in Water Quality Management

Although machine learning has the potential to reveal the complicated non-linear relationship between built environment and water resources, its application to urban design is still sparse. To address this gap, we used machine learning scenario analysis to investigate the influence of land use and urban development patterns on stream water quality in the Texas Gulf Region. We concluded that stream water quality is best protected by high-density development, with compact development being particularly advantageous. We also incorporate broader socioeconomics factors to investigate their relationships with water quality along with environmental factors in some ongoing projects.



Reference: Wang, R., Kim, J. H., & Li, M. H. (2021). Predicting stream water quality under different urban development pattern scenarios with an interpretable machine learning approach. Science of The Total Environment, 761, 144057.

Low Impact Development

The EPA estimates that $160 billion over the next 20 years in funding will be necessary to conduct stormwater management projects in about 750 cities to correct combined sewer systems. The high cost of gray infrastructure has led to increasing attention to green infrastructure (also known as LID) that is less costly. In one of the past research projects, we have examined both site-scale and regional-scale LID implementation. In a published study in the Journal of Environmental Management, we synthesized previous research findings from 79 bioretention experiments and identified design features that lead to the best pollutant removal performance.

Reference: Wang, R., Zhang, X., & Li, M.-H. (2019). Predicting bioretention pollutant removal efficiency with design features: A data-driven approach. Journal of Environmental Management, 242, 403–414. doi: 10.1016/j.jenvman.2019.04.064