Application of Artificial Intelligence in Hydrogeology to Solve Water Related Issues – Case StudiesFfrom South Africa and Malawi

Abstract

Artificial intelligence (AI) a field of computer science aimed at the developing computers capable of solving complex task that are normally done by humans have been used successfully in various spheres of life from medical, gaming, industrial, engineering, science and household. However, little utilisation in the field of hydrogeology has been done in Southern Africa to alleviated water related problems such as water shortage and pollution of aquifers. The key ingredients for AI systems are already available such as huge datasets and high computation resources. This research gives an overview of how this technology can be applied as a possible water-related (groundwater exploration and protection) problem solving tool using case studies from Southern Africa. The field of AI is broad but only three branches are mostly applicable to hydrogeology in this context: the data-driven artificial neural network, knowledge-based fuzzy inference system and hybrid systems. AI systems were built, tested and successful deployed for groundwater potential mapping of Beaufort West and central Malawi. Results from the mapping application are meant to benefit the local water authorities by reducing the groundwater search areas and demarcation of zones for possible wellfields for current and/or future sources of groundwater. The results for central Malawi groundwater potential mapping showed that the groundwater occurrence is fault controlled and localised towards the lower elevation ends of the faults. For groundwater protection, AI systems were used for groundwater vulnerability mapping of Witbank, Ermelo and Highveld coalfields in South Africa. In this research, a new approach for groundwater vulnerability assessment using AI algorithms, specifically for groundwater vulnerability to acid mine drainage (AMD) at a regional coalfield scale was developed. The value of this research lies in the fact that the new approach allows many of the processes of vulnerability assessments to be automated, facilitating rapid groundwater vulnerability assessments and solved the complex problem of assessment of AMD at regional scale specifically for policy and decision making purposes. The research managed to prove that it is possible to utilise the powerful AI tools to solve some of the water-related problems faced by Southern African countries. Applications of such systems can be scaled up to other applications of hydrogeology or fields e.g. modelling and hydraulic mechanics.

AAPG Datapages/Search and Discovery Article #90332 © 2018 AAPG International Conference and Exhibition, Cape Town, South Africa, November 4-11, 2018