Over the last century, low elevation temperate coniferous rainforests in the Pacific Northwest, USA, have largely been managed for commercial timber production. Intensive management regimes have left a legacy of simplified and fragmented forest habitat, and degraded freshwater systems throughout the region, transforming biological communities and altering ecological processes. Restoration of these landscapes has now become a growing conservation strategy in the region. For example, at the 2,915 ha Ellsworth Creek Preserve, Washington, The Nature Conservancy is aiming to restore ecosystems that support species and ecological processes representative of those found in unmanaged late-successional forest landscapes. However, the most effective approach to forest restoration at this scale is largely unknown. While models and stand-scale experiments indicate that active forest thinning can accelerate the recovery of forest habitat, little information exists concerning the ecological outcomes at a landscape or watershed scale. To address this knowledge gap, the Conservancy is implementing an active adaptive management process to simultaneously compare three alternative restoration treatments. Treatments include forest thinning and vegetation management, road removal, and a control. The study utilizes an unbalanced randomized block design, with treatments applied to eight experimental tributary sub-basins (75-221 ha). Treatment implementation will begin in 2008. Prior to treatment, two years of baseline monitoring is being conducted throughout the watershed. Data collection began in 2005, and includes surveys of forest birds, spawning coho salmon (Oncorhynchus kisutch), stream-associated amphibians, benthic macroinvertebrates, stream hydrology and physical habitat, forest structure and vegetation, and mapping of natural disturbance events.