Artificial Intelligence as an Efficient Alternative to the Conventional Hydrological Modelling for Groundwater Forecasting
Abstract
Accurate prediction of the groundwater, as one of the vital resources of water supply for domestic, agricultural and industrial users, has emerged as a dominant strategy to forecast and reduce the potential unpropitious consequences such as the saltwater intrusion in residential water-supply wells. It also helps to ensure a timely-manner formulation of a prudent groundwater resource management. Additionally, with a precise estimation of groundwater change, the farmers would be able to plan to meet the crop water necessities for their ranches. The current hydrological models such as MODFLOW can be harnessed to approximate the groundwater flow, but due to the complex nature of these physical models and the potential deficiencies in their performance, caused by the limited applicable data and the tedious computational time, equivalent sophisticated and time-efficient mathematical models, including but not limited to the Artificial Neural Networks (ANN), have been exploited as a powerful alternative for these conventionally-used hydrological models. Fine-tuning the hyper-parameters of these mathematical models guarantees the tackling of the afore-mentioned shortcomings of a physical-based hydrological model. The present study aims to evaluate the capability of other hydrological components such as precipitation, temperature, Irrigation, water usage .etc. as the potential drivers to determine the groundwater level at a certain geographical point. For this purpose, the existing machine learning techniques as well as a novel fully mathematical similarity-based predictive model are examined to unravel the inherent inter-dependencies of these drivers and the groundwater level.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.H21J1782S
- Keywords:
-
- 1869 Stochastic hydrology;
- HYDROLOGYDE: 1895 Instruments and techniques: monitoring;
- HYDROLOGYDE: 1914 Data mining;
- INFORMATICSDE: 1942 Machine learning;
- INFORMATICS