The Impact of Upland Land Management on Flooding: Results from a Multi-Scale Experimental and Modelling Programme

In response to a growing awareness of flood risk and questions as to the impact of land use management on this, an extensive dataset is currently being collected within the Pontbren catchment in Wales. This catchment, predominately covered with heavy clay soil, has a long history of land drainage and intensifying, but otherwise unchanging, land use. Focus on Pontbren was driven by local farmers noting increased water runoff, and the mitigating effect on this of tree-planted areas. A preliminary study within the catchment indicated that strategically placed, small scale planting of trees could improve the infiltration capacity of grazed permanent pasture. Following this, an intensive experimental programme was initiated to further examine how changes in land use might impact at different spatial scales. Stream flow, soil water potentials, overland flow and drain flow, precipitation, and other climatic variables are continuously monitored, and soil hydraulic properties and runoff processes are being investigated under different land use treatments including woodland buffer strips and no grazing management. An associated modelling programme is using these data to inform development and calibration of models examining the effects of land use change over differing spatial scales and levels of process representation. A multi-dimensional Richards' equation soil water model with macropore and overland flow representations is used to examine dominant processes at the hillslope scale. These processes, with associated parameters conditioned on experimental data, are then used to examine appropriate model representations and parameterisation at larger scales. Preliminary results from both the experimental dataset and the physically based modelling are presented, with a focus on the potential of localised strategic changes to land-use for reduction of flood risk at hillslope and catchment scales. Specifically, the implications of introducing a small tree strip to a grassed, clay hillslope is investigated through both analysis of current data and interrogation of a physically based model partially conditioned on measurements from a hillslope site. Results are preliminary, but indicative of the sensitivity of the system to the vegetative effects. Changes to soil infiltration rate, hillslope storage of water and interception properties are shown to be significant, and the strips appear to be capable of substantially perturbing the hillslope hydrological response. Issues of scaling and the need for further supporting data are discussed.