Abstract
Mapping the delineation of areas that are flooded due to water control infrastructure failure is a critical issue. Practical difficulties often present challenges to the accurate and effective analysis of dam-break hazard areas. Such studies are expensive, lengthy, and require large volumes of incoming data and refined technical skills. The creation of cost-efficient geospatial tools provides rapid and inexpensive estimates of instantaneous dam-break (due to structural failure) flooded areas that complement, but do not replace, the results of hydrodynamic simulations. The current study implements a Geographic Information System (GIS) based method that can provide useful information regarding the delineation of dam-break flood-prone areas in both data-scarce environments and transboundary regions, in the absence of detailed studies. Moreover, the proposed tool enables, without advanced technical skills, the analysis of a wide number of case studies that support the prioritization of interventions, or, in emergency situations, the simulation of numerous initial hypotheses (e.g., the modification of initial water level/volume in the case of limited dam functionality), without incurring high computational time. The proposed model is based on the commonly available data for masonry dams, i.e., dam geometry (e.g., reservoir capacity, dam height, and crest length), and a Digital Elevation Model. The model allows for rapid and cost-effective dam-break hazard mapping by evaluating three components: (i) the dam-failure discharge hydrograph, (ii) the propagation of the flood, and (iii) the delineation of flood-prone areas. The tool exhibited high accuracy and reliability in the identification of hypothetical dam-break flood-prone areas when compared to the results of traditional hydrodynamic approaches, as applied to a dam in Basilicata (Southern Italy). In particular, the over- and under-estimation rates of the proposed tool, for the San Giuliano dam, Basilicata, were evaluated by comparing its outputs with flood inundation maps that were obtained by two traditional methods whil using a one-dimensional and a two-dimensional propagation model, resulting in a specificity value of roughly 90%. These results confirm that most parts of the flood map were correctly classified as flooded by the proposed GIS model. A sensitivity value of over 75% confirms that several zones were also correctly identified as non-flooded. Moreover, the overall effectiveness and reliability of the proposed model were evaluated, for the Gleno Dam (located in the Central Italian Alps), by comparing the results of literature studies concerning the application of monodimensional numerical models and the extent of the flooded area reconstructed by the available historical information, obtaining an accuracy of around 94%. Finally, the computational efficiency of the proposed tool was tested on a demonstrative application of 250 Italian arch and gravity dams. The results, when carried out using a PC, Pentium Intel Core i5 Processor CPU 3.2 GHz, 8 GB RAM, required about 73 min, showing the potential of such a tool applied to dam-break flood mapping for a large number of dams.
Highlights
Dams play an important role in water resource management by providing essential services, including the provision of potable, industrial and agricultural water, hydroelectricity, pumped storage, flood control and management, ecological outflow management, water table recharge, fishing, and aquaculture
The first method, i.e., 'SanGiulianoDam_upload_data', is composed using a method, called 'set_dam', which defines the dam for which the user would like to perform the dam-break analyses, by 'add_dam' that is used to import the dam geometry data, by 'add_StudyArea' that defines the study area downstream of the dam, by 'add_RiverPath' that reads the information of the vectorial layer of the river path downstream of the dam, by 'add_MainCrossSec' that is used to define the main river cross-section's footprint, and by 'set_DTM', which is able to set the Digital Elevation Model (DEM) to be used
The proposed tool is comprised of three calculation workflows that are capable of computing: (i) the evaluation of the instantaneous breach outflow discharge for masonry dams, (ii) the propagation of the flood in the downstream valley, and (iii) the DEM-based delineation of flood-prone areas
Summary
Dams play an important role in water resource management by providing essential services, including the provision of potable, industrial and agricultural water, hydroelectricity, pumped storage, flood control and management, ecological outflow management, water table recharge, fishing, and aquaculture. Recent studies [15,16] have proposed novel artificial intelligence techniques, which are based on machine learning for dam break flow predictions, to overcome the limits of the above-cited traditional models. These artificial intelligence approaches have the ability to predict output data from a training and testing process, but they requires large data sets that are often not available and, are determined while using analytical results.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
