Forest harvest effects on streamflow generation have been well described experimentally, but a clear understanding of process-level hydrological controls can be difficult to ascertain from data alone. We apply a new model, Visualizing Ecosystems for Land Management Assessments (VELMA), to elucidate how hillslope and catchment-scale processes control stream discharge in a small Pacific Northwest catchment. VELMA is a spatially distributed ecohydrology model that links hydrological and biogeochemical processes within watersheds. The study site is WS10 of the H.J. Andrews LTER, a 10 ha forested catchment clearcut in 1975. Simulated and observed daily streamflow are in good agreement for both the pre- (1969–1974) and postharvest (1975–2008) periods (Nash-Sutcliffe efficiency = 0.807 and 0.819, respectively). One hundred scenarios, where harvest amounts ranged from 2% to 100% were conducted. Results show that (1) for the case of a 100% clearcut, stream discharge initially increased by ∼29% or 345 mm but returned to preclearcut levels within 50 years, and (2) annual streamflow increased at a near linear rate of 3.5 mm year−1 for each percent of catchment harvested, irrespective of location. Thereafter, to assess the impact of harvest location on stream discharge, 20 harvest scenarios were simulated, where harvest amount was fixed at 20% but harvest location varied. Results show that the streamflow response is strongly sensitive to harvest distance from the stream channel. Specifically, a 20% clearcut area near the catchment divide resulted in an average annual streamflow increase of 53 mm, whereas a 20% clearcut near the stream resulted in an average annual streamflow increase of 92 mm.