Frequency and Control of Preferential Flow Occurrence in the Shale Hills Catchment: From the Hillslope to the Catchment Scales.

Understanding temporal and spatial patterns of preferential flow (PF) occurrence in the landscape is important in revealing where and when PF might be important in hillslope/catchment hydrology. Quantitative assessment of the frequency and control of PF occurrence in the field, however, has been largely lacking, especially at the scale of hillslopes and catchments. Our previous work has evaluated the frequency and control of PF occurrence at 10 sites along a hillslope in the forested Shale Hills Catchment using three-years’ (2007-2009) real-time soil moisture response to 175 precipitation events. In this study, we expanded this analysis to include 1) 237 additional events from 2010 to 2013 to test the temporal consistency and predictability of PF occurrence, and 2) 25 additional sites to upscale the spatial coverage from the hillslope to the entire catchment. Our new results showed that considerable temporal consistence existed in both the frequency and the main controls of PF occurrence at the hillslope scale, which was attributed largely to the statistical stability of precipitation pattern over the 6.5-years monitoring period and the relatively stable subsurface preferential pathways. In general, PF tended to occur more often in response to intense rainfall events, and favored the conditions at dry hilltop or wet valley floor sites. When upscaling to the entire catchment, topographic control on the PF occurrence and its spatial pattern became more evident, leading to the identification of a likely hidden subsurface PF network in this catchment. Higher frequency of PF occurrence was observed at the valley floor (average 48%), hilltop (average 46%), and swales/hillslopes near the stream (average 40%), while the hillslopes in the eastern part of the catchment were least likely to experience PF (<1-20%). Soil-terrain attributes selected through stepwise regression provided a moderate but limited predictability (R² = 0.43-0.48) of PF occurrence at the hillslope and catchment scales, suggesting many uncertainties and complexities involved in PF dynamics. This study confirmed that the initiation and persistency of PF were controlled jointly by complex interactions among landform units, soil types, initial soil moisture conditions, precipitation features, and seasons. Further investigations of these five key controls of PF occurrence could lead to improved understanding and modeling of PF at the landscape scale.