A dynamic representation of wetlands for the ISBA land surface model

Wetlands play a critical role in terrestrial hydrology and land–atmosphere exchanges, yet they remain poorly represented in many land surface models. Most approaches rely on static wetland maps, preventing models from capturing hydrological variability and associated feedbacks. Here we introduce a new dynamic wetland scheme in the ISBA land surface model, combining explicit hydrological processes with an annually varying diagnostic of wetland extent.

Wetland extent is computed using a TOPMODEL-based approach that links grid-cell saturation deficit with sub-grid topographic indices, and includes a correction for soil organic content to better represent peat-rich areas. Hydrological properties of wetlands and sub-grid runoff redistribution allow water to accumulate and persist in saturated zones, influencing the overall grid-cell water budget.

Simulated wetland extent shows good spatial agreement with multiple satellite-derived wetland datasets across a range of climate zones. Hydrological evaluation against GRACE-based terrestrial water storage and observed river discharge indicates that dynamic wetlands exert a modest but physically consistent influence on ISBA hydrology: they adjust discharge timing and magnitude without degrading model skill, while increasing grid-cell water storage and associated evapotranspiration. However, regional patterns of simulated evapotranspiration reveal a strong sensitivity to the assumed wetland vegetation type, underscoring the need for improved vegetation representation.

In particular, the dynamic wetland extent opens new opportunities for simulating wetland biogeochemistry, including methane emissions, and for exploring the key role of soil oxygen availability in controlling greenhouse gas fluxes.