Current Projects
Flood Modeling
Advanced flood modeling using sub-grid sampling techniques for computational efficiency while preserving critical channel detail. Our approach enables rapid analysis of flood scenarios across large geographic areas.
Key Features:
- Models can be calibrated to specific storms (Eastern Kentucky, North Carolina)
- Mitigation implications and climate scenario testing capabilities
- High-resolution channel preservation with efficient computation
- GPU-accelerated processing for rapid results
Products to Date:
- 1st SP&M-enabled publication in review, 2nd in preparation
- 2 training workshops scheduled (UK-KGS-Morehead collaboration)
Monte Carlo Analysis
Comprehensive assessment of machine learning approaches for landslide susceptibility modeling in Kentucky's diverse geological settings. This research focuses on developing robust, statistically validated models for landslide hazard prediction.
Research Focus:
- Detailed analysis of relationship between inventory and susceptibility results
- Thousands of models generated within sensitivity assessment framework
- Comprehensive evaluation of different machine learning algorithms (SVM, LR, NB, BT)
- Statistical validation using AUC and model performance metrics
- Regional adaptation for Appalachian terrain characteristics
The project employs multiple machine learning algorithms including Support Vector Machine (SVM), Logistic Regression (LR), Naive Bayes (NB), and Boosted Trees (BT) to create the most accurate susceptibility models for Kentucky's unique geological conditions.
Products to Date:
Ongoing Activities:
- Statewide landslide inventory compilation and validation
- Multi-algorithm performance comparison study
- Climate-informed susceptibility modeling development
Status: active
Future Directions: Stochastic Storm Analysis
Research Goal: Develop probabilistic frameworks for extreme weather event analysis to improve flood and landslide hazard prediction.
Approach:
- Aggregating precipitation from historical storms (tropical cyclones, stalled fronts)
- Stochastic transposition of storm patterns across different regions
- Probabilistic basis for iterating flood and landslide models
- Integration with climate change scenarios
Foundation Research: Lawler, S., Deshotel, M., Dietrich, A.H. et al. Application of stochastic storm transposition for hydrologic modeling in the mountainous western US. Stoch Environ Res Risk Assess (2024).