The impact of climate change, land use change and groundwater extraction on groundwater-dependent wetlands extent worldwide

Wetlands are recognised as one of the most important ecosystems in terms of their unique biodiversity. Yet since 1900, the world has lost more than 50% of its wetland area (Davidson, 2014; Dugan, 2005; Winkler & DeWitt, 1985). Groundwater dependent wetlands (GDWs) provide critical ecological functions but face increasing pressure from climate variability and groundwater abstraction, while their global extent and dynamics remain poorly quantified. Here we present the first spatially and temporally explicit global assessment of GDW extent using a high resolution, physically based groundwater modelling framework.

We developed a dynamic GDW mapping framework based on GLOBGM global groundwater model (Verkaik et al. (2022), a 30 arc second, two-layer MODFLOW model, coupled offline with PCR-GLOBWB (Sutanudjaja et al., 2018). We include improved recharge estimates through bias correction and enhanced groundwater-surface water coupling via dynamically updated drainage elevation using saturated area fraction to better represent shallow groundwater processes.

Following ISIMIP3a and ISIMIP3b protocols, we simulated monthly groundwater conditions from 1960 to 2014 and projected changes up to 2050 under SSP1-2.6, SSP3-7.0 and SSP5-8.5 using five CMIP6 global climate models per scenario (Lange & Büchner, 2021). Evaluation of groundwater heads against more than 15000 wells from the IGRAC dataset (IGRAC, 2024) show strongest performance for shallow water tables (average depth less than 5m) with about 67 % of wells with Kling Gupta Efficiency KGE ≥ −0.41.

Following earlier work (Otoo et al., 2025), we identified GDWs as areas with a saturated area fraction greater than 0.5 and water table depth less than or equal to 5 m, capturing both core groundwater fed wetlands and peripheral drought adapted systems while accounting for sub-grid variability. We exclude areas where groundwater results remain unreliable due to spin up issues, karst, permafrost and mountain areas (in total, only about 19 % of global areas). This approach shows strong spatial agreement with independent datasets, achieving a global hit rate of 76 % against the Global Lakes and Wetlands Database 2 (GLWD 2) (Lehner et al, 2025), and exceeding 80 % against the Australian Atlas (Doody et al., 2017).

We also account for land conversion when simulating area changes by excluding potential GDWs that coincide with agricultural areas in PCR-GLOBWB (rainfed and irrigated agriculture and pasture) for both past and future periods. For the historical period, we attributed groundwater level trends to climate-driven recharge changes, human-driven groundwater abstraction or a combination of both and quantified GDW area changes aggregated by WWF biome realm units (Olson et al., 2001). Preliminary results show strong reductions of past and future wetlands in Afrotropical, Indo-Malayan and Neotropical regions with distinct areas where either groundwater level decline or land use change are the dominant drivers. This framework addresses a major gap in global wetland assessments and provides a physically groundwater basis for evaluating past and future GDW dynamics in support of conservation planning, climate impact assessment and policy development aligned with Ramsar Convention, the Sustainable Development Goals and global biodiversity targets.