<p>Climate change poses a growing threat to water resources, particularly in arid and semi-arid regions where both surface and groundwater are essential for ecological stability and human livelihoods. The Zagros Mountains of Iran host extensive karst aquifers and springs that serve as critical sources of drinking water and sustain regional hydrological systems. Despite their importance, the response of these karst water systems to future climate change remains poorly quantified. This study aims to evaluate the long-term impacts of climate change on karst spring discharge and water resource sustainability in the Zagros region. To achieve this, we integrated downscaled outputs from three CMIP6 General Circulation Models (GCMs) with the Long Ashton Research Station Weather Generator (LARS-WG) to generate high-resolution projections of temperature, precipitation, and potential evapotranspiration for the period 2021–2100. The XGBoost machine learning model was calibrated and tested using time-aware modeling to simulate and forecast monthly discharge from 17 major karst springs under three shared socioeconomic pathways (SSP1-2.6, SSP2-4.5, and SSP5-8.5). The results indicate a significant warming trend and a pronounced decline in precipitation (P), leading to an increase in potential evapotranspiration (PET) and a substantial reduction in the aridity index (AI = P/PET). Under the high-emission SSP5-8.5 scenario, localized maximum reduction of precipitation up to 83% are projected in the most vulnerable climate zones, while the aridity index is expected to decline by up to 85%. Consequently, spring discharge is anticipated to decrease by approximately 40%, particularly during post-monsoon periods (July–September). These findings underscore the high vulnerability of the Zagros karst aquifers to climatic shifts, with serious implications for water security, ecosystems, and regional socio-economic stability. The study highlights the necessity of adaptive water management strategies and provides a transferable framework for assessing climate impacts on groundwater-dependent systems in other arid and semi-arid environments.</p> Graphical Abstract <p></p> <p>This graphical abstract presents a comprehensive visual synthesis of the study “Climate Change Threatens Karst Water Resources in the Zagros Mountains: Insights from CMIP6 and Machine Learning” It illustrates the methodology, datasets, and key hydro-climatic outcomes for the period 2021–2100. The upper section summarizes projected changes in major climatic variables including precipitation (P), temperature (T), potential evapotranspiration (PET), and aridity index (AI) across four representative time slices (2020, 2040, 2060, and 2100) under the high-emission scenario (SSP5-8.5). These maps depict a progressive shift toward hotter and drier conditions, marked by declining precipitation, increasing temperature and evapotranspiration, and a substantial reduction in AI. The central panel outlines the modeling workflow, coupling CMIP6-based climate projections with statistical downscaling using the LARS-WG model and hydrological simulation through the eXtreme Gradient Boosting (XGBoost) algorithm. This integration enabled simulation of monthly discharge for 17 major karst springs, linking climatic drivers to hydrological responses. The lower section presents a time-series plot of projected total spring discharge, revealing a consistent downward trend and an estimated 40% reduction in cumulative flow by 2100. This decline aligns with the modeled decreases in precipitation and AI, emphasizing the cascading impacts of climate change on groundwater resources. Overall, the graphical abstract encapsulates the cause-effect chain between climate forcing and hydrological response. It highlights the heightened vulnerability of karst aquifers in the Zagros Mountains to future climatic stress and underscores the need for adaptive and sustainable water management strategies to mitigate climate-induced risks in this critical water-supply region.</p>