- NH33G-0521: Investigating Hydro-Mechanical Drivers of an Expansive-Clay Highway Slope via SHAP-Enabled Machine Learning
-
Board 0521‚ Hall EFG (Poster Hall)NOLA CC
Author(s):Generic 'disconnected' Message
Shams E Shifat, Jackson State University (First Author, Presenting Author)
A Q M Zohuruzzaman, Jackson State University
Yifei Li, Jackson State University
Sadik Khan, Jackson State University
Suprobha Darothy, Jackson State University
Highway embankments across Mississippi rest on Yazoo Clay, a soil that swells when wet and shrinks when dry, gradually pushing slopes out of line and threatening pavement integrity. To pinpoint the conditions that drive this movement, a comprehensive sensor array was installed on one representative embankment. The system recorded daily readings from 2018 to 2022, capturing horizontal soil movement, moisture content, and pore-water pressure at depths of 5, 10, and 15 feet, along with local weather. Machine-learning analysis revealed three dominant predictors of slope displacement. First is the total rainfall accumulated during the previous seven days. Second is week-to-week volatility in near-surface soil moisture. Third is pore-water pressure (matric suction) measured ten feet below ground. Together these factors explain roughly 80 percent of the observed movement. Rainfall exerts the greatest influence, accounting on average for about 0.2 inches of daily lateral shift, with the most pronounced motion occurring in the upper six feet where infiltrating water softens the clay. These findings show that shallow wetting fronts and perched-water conditions govern Yazoo-Clay slope behavior, guiding transportation agencies toward targeted drainage improvements, surface sealing, and focused suction monitoring to mitigate future embankment deformation.
Scientific DisciplineSuggested ItinerariesNeighborhoodType
Enter Note
Go to previous page in this tab
Session
