Search

Summary: The sensitivities of runoff generation to rainfall variability and initial wetness conditions were examined for a major flash flood event that occurred during August 29, 2003 on the upper Tagliamento river basin in the eastern Italian Alps. The Triangulated Irregular Network (TIN)-based Real-time Integrated Basin Simulator (tRIBS) distributed hydrologic model was used to simulate the hydrologic response over a range of sub-basins. The model was calibrated for the single flash flood event based on the observed hydrograph at the outlet of the basin examined (Fella basin) and was validated based on the available observed hydrographs at interior points of the basin. A series of hydrologic simulations were performed for different initial soil moisture conditions and rainfall forcing resolutions in order to evaluate the sensitivity of runoff generation to those variables. Evaluation of the results suggests that both initial wetness and rainfall resolution affect significantly the simulated peak flow and runoff volume during the flash flood event. Sensitivity to initial wetness exhibits a scale dependence with the sensitivity increasing with basin scale. The bias introduced to the basin-averaged rainfall due to aggregation had a significant effect on runoff generation for all basin scales, while the effect of variability smoothing was important only for the larger scale basins. Finally, the sensitivity of the flood hydrograph to rainfall aggregation was shown to be more important for drier initial wetness states. [Copyright &y& Elsevier]

• Regional, machine learning models solely for flood peak prediction. • LSTM based storm classifier for flood warning. • Importance of flood predictors depends on flood severity. • Flood warning framework for early detection and prediction in ungauged basins. Flood prediction across scales and more specifically in ungauged areas remains a great challenge that limits the efficiency of flood risk mitigation strategies and disaster preparedness. Building upon the recent success of Machine Learning (ML) models on streamflow prediction, this work presents a prototype ML-based framework for flood warning and flood peak prediction. The fundamental elements of the proposed system consist of a) a Long-Short Term Memory (LSTM) model for classifying storm events to Flood/no-Flood given a threshold based on the 90th flow percentile and b) the flood peak prediction models. The selected ML-models for flood peak prediction are the Histogram based Gradient Boosting Regressor and the Random Forest. One of the strengths, and reason for selection, of these decision-tree models is their degree of interpretability. This is exploited in the study to help us spatially disentangle the role of both the static and dynamic drivers of flood peak response. Our analysis is presented for 18 distinct hydroclimatic regions across the contiguous US. Results reveal a significant regional dependence on both predictive performance and dominant flood predictors, which emphasize the variability in the complexity of a catchment's hydrologic behavior as well as its impact on forecasting flood response. Evaluation of the drivers of flood peaks noted distinct dependencies among the dynamic and static predictors considered in our models for flood peaks of different severity. Specifically, low-moderate flood events showed a clear preponderance for the static catchment attributes over dynamic predictors like precipitation whereas precipitation was the dominant driver for the highest flood peaks. The proposed flood peak prediction models were compared against a state-of-the-art LSTM model and were shown to outperform in ungauged basins for the majority of basins. Overall, the proposed system classified storms correctly for >80% in all cases and exhibited a percent relative difference in flood peak estimates of <30% in most cases. [ABSTRACT FROM AUTHOR]

Highlights: [•] Spatio-temporal rainfall variability effects on flood modeling. [•] Rainfall moments and storm velocity conceptual frameworks. [•] Experimental framework to assess the conceptual frameworks’ predictability. [ABSTRACT FROM AUTHOR]