Prediction of flood quantiles at ungauged catchments for the contiguous USA using Artificial Neural Networks

In this study, we utilise Artificial Neural Network (ANN) models to estimate the 100- and 1500-year return levels for around 900,000 ungauged catchments in the contiguous USA. The models were trained and validated using 4,079 gauges and several selected catchment descriptors out of a total of 25 available. The study area was split into 15 regions, which represent major watersheds. ANN models were developed for each region and evaluated by calculating several performance metrics such as root-mean-squared error (RMSE), coefficient of determination (R2) and absolute percent error. The availability of a large dataset of gauges made it possible to test different model architectures and assess the regional performance of the models. The results indicate that ANN models with only one hidden layer are sufficient to describe the relationship between flood quantiles and catchment descriptors. The regional performance depends on climate type as models perform worse in arid and humid continental climates. Overall, the study suggests that ANN models are particularly applicable for predicting ungauged flood quantiles across a large geographic area. The paper presents recommendations about future application of ANN in regional flood frequency analysis.

Hydrological and Hydraulic Flood Hazard Modeling in Poorly Gauged Catchments: An Analysis in Northern Italy

Flood hazard is assessed for a watershed with scarce hydrological data in the lower plain of Northern Italy, where the current defense system is inadequate to protect a highly populated urban area located at a river confluence and crossed by numerous bridges. An integrated approach is adopted. Firstly, to overcome the scarcity of data, a regional flood frequency analysis is performed to derive synthetic design hydrographs, with an original approach to obtain the flow reduction curve from recorded water stages. The hydrographs are then imposed as upstream boundary conditions for hydraulic modeling using the fully 2D shallow water model PARFLOOD with the recently proposed inclusion of bridges. High-resolution simulations of the potential flooding in the urban center and surrounding areas are, therefore, performed as a novel extensive application of a truly 2D framework for bridge modeling. Moreover, simulated flooded areas and water levels, with and without bridges, are compared to quantify the interference of the crossing structures and to assess the effectiveness of a structural measure for flood hazard reduction, i.e., bridge adaptation. This work shows how the use of an integrated hydrological–hydraulic approach can be useful for infrastructure design and civil protection purposes in a poorly gauged watershed.