Modelling the impact of land use change on floods and drought in a groundwater-dominated catchment

In the UK, land use and land management change are being considered as part of a strategy to tackle flooding; the term natural flood management (NFM) is used in the UK to refer to this approach. Although evidence is limited, it is thought that practices considered as NFM, such as woodland planting or cover cropping, increase soil water storage and infiltration through improved soil structure, potentially reducing flooding within catchments where applied. Moreover, we know that trees and woodland have higher interception than shorter vegetation, increasing soil moisture deficit. We simulate what effect these land use and management changes have in the River Coln catchment (Upper Thames, UK), which is characterised by permeable geology and arable farming. The catchment has experienced both surface water and groundwater-driven flooding in recent history and is also important in maintaining summer low flows in the River Thames.  

The land surface water balance was modelled with the detailed, field-scale hydrological model SWAP. Recharge from the SWAP model was passed to a semi-distributed groundwater model, which produces groundwater baseflow, and direct runoff from SWAP was passed to a linear reservoir model to produce surface runoff. Soil information (such as soil hydraulic parameters) was based on the NATMAP database; vegetation parameters (e.g. typical crop rotations or forest types) were derived from stakeholder workshops, surveys and interviews, and expert elicitation. Soil and farm management decisions were implemented by perturbing the SWAP soil and vegetation parameters. The parameters of the surface water routing and groundwater components were calibrated within a Monte Carlo framework.

Overall, we found that land use and land management measures have very limited potential for reducing flooding in permeable catchments, where the primary driver of high flows is high winter groundwater levels. We found that only large-scale coniferous planting had the potential to reduce winter peak flows and flooding in the Coln (e.g. reduction of 21−27% with two-thirds of catchment as coniferous woodland), but that this would decrease summer mean flows (12% reduction in July mean flow, 8% in August). The levels of coniferous woodland required to achieve these reductions in winter high flows are unrealistic, given the large area of productive arable land that would need to be converted to woodland and the limited ecological benefits of coniferous woodland.