Recent large flood events in the UK have focused attention on the effects of agricultural intensification and de-intensification on flood risk. Possible detrimental factors for these floods are thought to be the degradation of soil structure through compaction from grazing animals and agricultural machinery; while possible ameliorating factors may be the re-introduction of natural landscape features such as woodland and wetlands.
In order to address this issue in an upland setting a multi-scale experimental programme has been implemented at Pontbren, a sheep-farmed headwater catchment of the River Severn, UK. The research aimed to evaluate the impact of different land management regimes on hydrological response. An integral component of this was to determine the role that native, broadleaf trees, planted as shelterbelts on intensively grazed pasture, may have in the reduction of flooding. Effects of trees on soil properties, soil moisture status, rainfall interception and overland flow were investigated
Comparisons of soil hydraulic properties show significant increases in saturated moisture content and hydraulic conductivity of soil under trees compared to adjacent improved pasture. These effects, combined with increased interception and water use by the trees, have led to the soil under the trees being considerably drier than that under the pasture. The influence of the tree strips on rainfall patterns in the wider landscape is determined by the wind direction and inclination angle of the rain during rainfall events. The results show seasonal variation. Overland flow results show substantial differences in surface runoff between the grazed pasture and the ungrazed tree-planted area. Manipulation experiments reveal that, excluding sheep has a significant influence on surface hydrology at the plot scale, which is enhanced by planting trees. Data were used to underpin the conceptual understanding of the hydrological response, to support model development and thus make predictions under speculative land management scenarios.