Rockfalls and rockslides are often dominant geomorphic processes in steep bedrock landscapes, but documenting their occurrence can be challenging, requiring frequent monitoring and well resolved spatial data. Repeat application of remote sensing methods such as Terrestrial Laser Scanning (TLS) and Structure-from-Motion (SfM) photogrammetry can detect even very small rockfalls, but typically these acquisitions span only years and may not record rockfall activity representative of longer-term rates of cliff erosion. Inventory databases can extend rockfall records, but are commonly incomplete and prone to observation bias. We employed TLS and SfM on two adjacent cliffs (El Capitan and Middle Brother) in Yosemite Valley, integrating semi-annual data collections from 2010 to 2017 with “historical” (archival) SfM models derived from oblique photographs taken in 1976. Comparing the 1976 SfM models against more recent data allows for more accurate and precise rockfall detection and volume measurement over a 40-year period. Change detection indicates that 235 rockfalls occurred from the two cliffs, more than twice as many events as are recorded in Yosemite's inventory database. Although individual rockfall volumes reported in the inventory database vary from those measured by SfM-TLS, reported cumulative volumes are similar to measured volumes, likely because the large-volume events that account for most of the cumulative volume tend to be widely observed and well-documented. Volume-frequency relationships indicate that the cliffs erode predominantly by less frequent, larger-volume rockfalls, at rates of 0.9 to 1.7 mm/yr. Our study demonstrates how integrated SfM and TLS measurements, especially utilizing SfM models derived from historical imagery, allow detection and quantification of rockfalls spanning several decades, complementing and improving inventory databases, informing rockfall hazard assessment, and providing longer-term rates of cliff erosion.