Testing of MODFLOW related packages to simulate flow patterns to horizontal wells in shallow porous aquifers for effective catchment-scale groundwater modelling

Horizontal wells are an interesting alternative to vertical wells for water supply in shallow aquifers and have been successfully implemented at many different water supply sites worldwide. However, the flow patterns to horizontal wells are much more complicated than those to vertical wells, making their mathematical treatment more demanding. In previous studies, some numerical approaches have been introduced and used to represent horizontal wells in numerical groundwater models, but there is no "one size fits all" solution that can be identified. In order to develop a practice-oriented guideline for the appropriate implementation of horizontal wells in MODFLOW based groundwater models, we designed a comprehensive numerical experiment with four different synthetic aquifer models of increasing complexity in terms of hydraulic properties and boundary conditions in the current work. Four different MODFLOW related packages, including WEL, Drain, MNW2, and CFPy, were tested and evaluated for their performance under steady and transient conditions at catchment scale. The advantages and limitations of these four tested MODLFLOW packages were systematically analyzed and compared in terms of performance metrics, computational efficiency, and numerical stability. The results indicate that while both CFPy and MNW2 deliver hydraulically representative simulations, CFPy demands 40–80% higher computational effort compared to MNW2. In contrast to the WEL package, which oversimplifies flow dynamics, and the Drain package, which struggles to represent lateral flow patterns effectively, MNW2 captures variable inflow rates along the well and incorporates essential factors such as skin effects and wellbore storage. Consequently, MNW2 emerges as the preferred choice for practical and more complex applications due to its ease of use, reliable simulation of well-aquifer interactions, and sufficient accuracy for flow modelling at the catchment scale. The outcomes of this study provide actionable guidelines for selecting appropriate modelling approaches for horizontal wells based on specific project requirements. Furthermore, the methodology used in this study and its combination with synthetic aquifer experiments, multiple complexity levels, and comparative package evaluations is transferable to other regions and applications. By offering insights into horizontal well representation, the findings support improved groundwater management and water supply planning in both research and operational contexts.