Geostatistical and Artificial Intelligence tools for geological and groundwater modeling in Ethiopia.

• Název práce v češtině: Geostatistical and Artificial Intelligence tools for geological and groundwater modeling in Ethiopia.
• Název v anglickém jazyce: Geostatistical and Artificial Intelligence tools for geological and groundwater modeling in Ethiopia.
• Klíčová slova: 3D geological modeling; unconventional groundwater modeling; geostatistical modeling; machine learning tools; Ethiopia.
• Klíčová slova anglicky: 3D geological modeling; unconventional groundwater modeling; geostatistical modeling; machine learning tools; Ethiopia.
• Akademický rok vypsání: 2023/2024
• Typ práce: disertační práce
• Jazyk práce: čeština
• Ústav: Ústav geologie a paleontologie (31-420)
• Vedoucí / školitel: Mgr. Karel Martínek, Ph.D.
• Řešitel: skrytý - zadáno vedoucím/školitelem
• Datum přihlášení: 06.10.2023
• Datum zadání: 06.10.2023
• Konzultanti: doc. RNDr. Kryštof Verner, Ph.D.

Předběžná náplň práce

Preliminary scope of proposed thesis is searching unconventional solutions for groundwater problems in Ethiopia. Thesis will be part of the new project, where applicant will be working on geostatistical groundwater spatial distribution modeling, machine learning/artificial intelligence (AI) tools, 3d geological modeling, hydrogeological and surface data review and re-interpretation in Ethiopia.
The geological 3D models will serve as the basic input for the processing of groundwater spatial variability. Model will be transformed to virtual vertical boreholes in a regular network, whereas the geological pattern is extracted to the virtual boreholes and virtual borehole database is used to transfer the geometry of geological objects to the groundwater flow models.
Geostatistical groundwater spatial distribution modelling will include: advanced processing of the data such as hydrological modeling and topographic wetness index calculations (TWI), evapotranspiration modeling and other relevant derived variables calculations; geostatistical modeling using all available primary field and satellite data and also derived variables focused on identification of areas rich in groundwater. Geostatistical analysis is very effective for processing of complex spatially oriented datasets; tens of parameters characterizing geomorphology, hydrology, precipitation, land-cover, hydrogeology and geology will be evaluated. Input data are also of very different reliability (e.g. ground station precipitation measurements vs. precipitation derived from satellite data) and varying homogeneity (e.g. ground station precipitation measurements are lacking temporally continuous data, there are many significant gaps in the record). Considering complexity and inhomogeneity of input data geostatistical methods using also multivariate statistics will help to understand water cycle. Using machine learning/artificial intelligence (AI) tools, which are being widely used for analysis of complex and inhomogeneous datasets, will provide deeper insights. AI tools will also be used for constructing and solving optimization models for identifying the optimal water drilling locations.

Předběžná náplň práce v anglickém jazyce

Preliminary scope of proposed thesis is searching unconventional solutions for groundwater problems in Ethiopia. Thesis will be part of the new project, where applicant will be working on geostatistical groundwater spatial distribution modeling, machine learning/artificial intelligence (AI) tools, 3d geological modeling, hydrogeological and surface data review and re-interpretation in Ethiopia.
The geological 3D models will serve as the basic input for the processing of groundwater spatial variability. Model will be transformed to virtual vertical boreholes in a regular network, whereas the geological pattern is extracted to the virtual boreholes and virtual borehole database is used to transfer the geometry of geological objects to the groundwater flow models.
Geostatistical groundwater spatial distribution modelling will include: advanced processing of the data such as hydrological modeling and topographic wetness index calculations (TWI), evapotranspiration modeling and other relevant derived variables calculations; geostatistical modeling using all available primary field and satellite data and also derived variables focused on identification of areas rich in groundwater. Geostatistical analysis is very effective for processing of complex spatially oriented datasets; tens of parameters characterizing geomorphology, hydrology, precipitation, land-cover, hydrogeology and geology will be evaluated. Input data are also of very different reliability (e.g. ground station precipitation measurements vs. precipitation derived from satellite data) and varying homogeneity (e.g. ground station precipitation measurements are lacking temporally continuous data, there are many significant gaps in the record). Considering complexity and inhomogeneity of input data geostatistical methods using also multivariate statistics will help to understand water cycle. Using machine learning/artificial intelligence (AI) tools, which are being widely used for analysis of complex and inhomogeneous datasets, will provide deeper insights. AI tools will also be used for constructing and solving optimization models for identifying the optimal water drilling locations.