A method for simulating wetland hydro-dynamics in regional climate models

Lateral fluxes of surface water and groundwater are not taken into account explicitly
in the terrestrial water balance of most land surface models. This results in an erroneous
representation of the atmospheric feedback from wetlands, which are strongly affected by
these fluxes. However, wetlands are increasingly recognized for their impact on the regional
and global climate, through their share in the methane, carbon dioxide, biomass and water
cycle. Incorporation of dominant wetland hydro-dynamics in land surface models is a first
step towards a proper representation of these feedbacks, which are strongly related to the
wetland open water and soil moisture states.
A new approach is presented for modeling the surface water - groundwater interactions
in wetlands at the resolution of a typical regional climate model. Large wetlands and the
surrounding groundwater reservoirs are both represented by a network of ’interconnected
reservoirs,’ driven by surface water inflow, rainfall and (potential) evaporation. The flow
between these reservoir is considered proportional to the hydraulic gradient.
The method is applied to the Okavango Delta wetland in Botswana, Southern Africa,
where traditional climate models severely underestimate the available water for evaporation.
First order dynamics, i.e., infiltration rates and seasonal variations in inundated area, are
properly simulated after calibration on remote sensing estimates of the monthly flood-extent.
The ratio between infiltration and evaporation observed in seasonal floodplains of the Okavango
Delta is simulated adequately by the calibrated wetland model. This indicates that
the dominant hydrological processes are represented sufficiently in the model structure and
that the proposed method is a promising technique for simulating lateral water movement in
climate models.