Flooding presents a formidable challenge in the United States, endangering lives and causing substantial economic damage, averaging around $5 billion annually. Addressing this issue and improving community resilience is imperative. This project employed machine learning techniques and publicly available data to explore the factors influencing flooding and to develop flood susceptibility maps at various spatial resolutions. Six machine learning algorithms, including Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM), Knearest neighbor (KNN), Adaptive Boosting (Ada Boost), and Extreme Gradient Boosting (XGB) were used. Geospatial datasets comprising thirteen predictor variables and 1528 flood inventory data collected since 1996 were analyzed. The predictor variables are rainfall, elevation, slope, aspect, flow direction, flow accumulation, Topographic Wetness Index (TWI), distance from the nearest stream, evapotranspiration, land cover, impervious surface, land surface temperature, and hydrologic soil group. Five hundred twenty-eight non-flood data points were randomly created using a stream buffer for two scenarios. A total of 2964 data points were classified into flooded (1) and non-flooded (0) categories and used as a target. Overall, testing results showed that the XGB and RF models performed relatively well in both cases over multiple resolutions compared to other models, with an accuracy ranging from 0.82 to 0.97. Variable importance analysis depicted that predictor variables such as distance from the streams, hydrologic soil type, rainfall, elevation, and impervious surfaces significantly affected flood prediction, suggesting a strong association with the underlying driving process. The improved performance and the variation of the susceptible areas across two scenarios showed that considering predictor variables with multiple resolutions and appropriate non-flooding training points is critical for developing flood-susceptibility models. Furthermore, using tree-based ensemble algorithms like RF and XG boost in the stack generalization approach can help achieve robustness in a flood susceptibility model where multiple algorithms are being evaluated.

We present new geochronological data derived from hillslope boulder armor in the Flint Hills in northeastern KS, United States, that provides insights into the rates and timing of lateral retreat in this landscape. Our results show that the surfaces of these limestone boulders date back to the Pleistocene era, well within the last glacial period. We also found that there is a significant increase in the ages of hillslope armor with increasing distance downslope from the modern limestone bench, the source of the boulders. Based on the age-distance relationship of the boulders, we estimate the rate of lateral retreat in this landscape to be 0.02 mm/yr, which falls between the geometrically estimated retreat rates based on calculated denudation rates of the Flint Hills region. We propose that the cooler temperatures and higher effective moisture due to less efficient evapotranspiration during the late Pleistocene period resulted in more effective freeze-thaw and transport processes, such as creep due to soil expansion and contraction. The production and transport of new boulder armor would then have effectively ceased once the climate transitioned to warmer conditions during the Holocene. Our findings suggest that the boulder armor we observe on the soil mantled hillslopes today are relict features from before the LGM-Holocene transition. These results provide important insights into the long-term evolution of these ubiquitous layered sedimentary landscapes.