Effective forest governance is central to the efficient, sustainable, and equitable use of forest resources, yet challenges in assessing forest governance impede efforts to improve it. Contemporary forest governance involves decisions by multiple stakeholders across multiple sectors of economy and society, from local to global scales – making forest governance assessments inherently complex. Here, we assess forest governance in the context of federal forest management in the U.S. Pacific Northwest (PNW), a region whose history is intertwined with that of a storied and significant federal forestry institution – the USDA Forest Service. Our assessment included (1) a review of literature on the emergence of forest governance and existing methodologies for assessing it, (2) a characterization of forest governance in the PNW by way of a synoptic history of federal forestry in the region, (3) the application of a novel remote sensing-based method to detect patterns of timber harvesting in the Western Cascades of Oregon, a physiographic ‘sub-region’ of the PNW defined in the Northwest Forest Plan (NWFP). We quantified the volume of timber extracted by two harvest methods (‘regeneration harvest’ and ‘partial harvest’) in >7,000 individual harvest events which occurred over a 19-year period (1991-2012). Total annual harvest volumes extracted by each harvest method were aggregated within spatially explicit land use allocations where management objectives are prescribed by the NWFP. For three land use allocation types (‘Matrix’, ‘Late-Successional Reserves’, and ‘Adaptive Management Areas’) we developed hypothetical expectations for timber harvest volumes. Observed patterns of timber harvest volume were evaluated relative to the expected outcomes. Results indicate that some expectations have been met (e.g. timber harvesting methods shifted from regeneration harvests to primarily partial harvests, and total harvest volume declined steeply across all land use allocations), while other expectations have not been met (e.g. timber harvest volume was significantly lower than expected, especially in ‘Matrix’ lands where timber extraction is an intended management priority). Additionally, by objectively analyzing timber harvest volume at the scale of individual harvest events, we demonstrate our method’s utility in locating ‘outlier’ harvests – i.e. those that counter general patterns or are otherwise distinct – which serve as points of reference for further research at local scales.