HYDROLOGIC DYNAMICS INSIDE HILLSLOPES

Few observations are available to describe the movement of unsaturated and saturated flow through the weathered bedrock that underlies hillslopes. Yet many recent studies point to this flow path as controlling not only dry season base flow of channels, but even peak runoff events, as well as erosional processes. In 2007, we began an intensive hydrologic monitoring program on a ~30^o forested hillslope in northern California (Rivendell) underlain primarily by vertically dipping argillite to explore topographic controls on moisture return to the atmosphere. Here, weathered bedrock varying from about 4 m thick near the base of the hillslope to 24 m at the divide, underlies a thin soil mantle. Seasonally, a perched water table develops on this boundary and directs runoff to the adjacent channel. Through monitoring of 12 wells (drilled to fresh bedrock), sap flow, soil moisture, and rock moisture over a ~4000 m2 area, a picture of the dynamics and diversity of flow in the weathered bedrock zone is emerging. In the spirit of this special session, we focus on three surprises. One simple metric of subsurface dynamics is the lag time between peak rainfall and peak groundwater response. Unsaturated flow travels up to 27 m in our site to the water table, and it would be expected that the greater the travel distance the longer the lag time. Surprisingly, we find that for three of our deeper wells, the response preceded sites with shallower water table depths in the first storms of the season. As the rainy season continued, however, their response systematically slowed- by a factor of 100. The second surprise was the large difference in the groundwater response in each of the wells. Once sufficient rains occurred, most rose rapidly as storm pulses arrived, but the peak height differed greatly: for the same storm, for example, some wells rose 17 m, while others less than 2 m. Some wells rose to a seasonal elevated level and then responded to individual storms while others rose and returned to nearly pre-storm levels. Our observations suggest a highly variable groundwater response conditioned by local fractures that fill, spill and drain. The third surprise, then, is that the stable isotope composition of the groundwater is remarkably invariant despite widely varying meteoric input values, and the large, rapid rise and long, slow decline of the water table.