This paper presents a review of papers specifically focused on the use of both numerical and machine learning methods for groundwater level modelling. In the reviewed papers, machine learning models (also called data-driven models) are used to improve the prediction or speed process of existing numerical modelling. When long runtimes inhibit the use of numerical models, machine learning models can be a valid alternative, capable of reducing the time for model development and calibration without sacrificing accuracy of detail in groundwater level forecasting. The results of this review highlight that machine learning models do not offer a complete representation of the physical system, such as flux estimates or total water balance and, thus, cannot be used to substitute numerical models in large study areas; however, they are affordable tools to improve predictions at specific observation wells. Numerical and machine learning models can be successfully used as complementary to each other as a powerful groundwater management tool. The machine learning techniques can be used to improve calibration of numerical models, whereas results of numerical models allow us to understand the physical system and select proper input variables for machine learning models. Machine learning models can be integrated in decision-making processes when rapid and effective solutions for groundwater management need to be considered. Finally, machine learning models are computationally efficient tools to correct head error prediction of numerical models.
The map shows the result of a procedure for pluvial flood hazard (PFH) mapping in urban areas, developed by using easy to find data, usually available from local authorities. Data were processed using a sequence of tools in a GIS environment. Two parameters have been evaluated: (1) susceptibility, defined as the probability of a flood occurring in a certain area (‘flood-prone areas’) which depends on the ground morphology (i.e. presence of depressions, fill volume of depressions) and spatial density of previously observed floods and (2) potential impact, formed from all factors influencing the damage (e.g. value of exposed heritage or number of people potentially involved), as well as the induced hazard due to damage. Susceptibility and potential impact were each divided into five classes and a score matrix was constructed; the final PFH is then defined by the summation of the scores within the matrix. The methodology used is suitable for a comprehensive, mostly automatic, first-level analysis of PFH in urban areas, and it is easily replicable. The obtained flood hazard map could provide a useful tool for civil protection purposes, that is, for hazard evaluation and emergency planning.
This paper concerns a stepwise modelling procedure for groundwater flow simulation in a folded and faulted, multilayer carbonate aquifer, which constitutes a source of good quality water for human consumption in the Apennine Range in Central Italy. A perennial river acts as the main natural drain for groundwater while sustaining valuable water-related ecosystems. The spatial distribution of recharge was estimated using the Thornthwaite–Mather method on 60 years of climate data. The system was conceptualized as three main aquifers separated by two locally discontinuous aquitards. Three numerical models were implemented by gradually adding complexity to the model grid: single layer (2D), three layers (quasi-3D) and five layers (fully 3D), using an equivalent porous medium approach, in order to find the best solution with a parsimonious model setting. To overcome dry-cell problems in the fully 3D model, the Newton–Raphson formulation for MODFLOW-2005 was invoked. The calibration results show that a fully 3D model was required to match the observed distribution of aquifer outflow to the river baseflow. The numerical model demonstrated the major impact of folded and faulted geological structures on controlling the flow dynamics in terms of flow direction, water heads and the spatial distribution of the outflows to the river and springs.
Geological models are very useful tools for developing conceptual schemes owing to their capacity to optimize the management of stratigraphic information. This is particularly true in areas where archaeological heritage is exposed to hydrogeological hazards; 3D models can constitute the first step toward the construction of numerical models created to understand processes and plan mitigation actions to improve visitor safety and preserve archaeological heritage. This paper illustrates the results of a 3D hydrostratigraphic model of the site of the Colosseum in the Central Archaeological Area of Rome. In recent years, this area has experienced numerous floods caused by intense meteorological events. A new borehole survey provided the opportunity to update previous maps and cross sections and build a local scale 3D model. The resulting conceptual model was used to identify primary gaps in existing knowledge about the groundwater system and to optimize the planning of a piezometer monitoring network. Further studies can then focus on the development of groundwater numerical models to verify hypotheses regarding inflow-outflow dynamics and facilitate the optimization of water management.