Climate changes in the Pacific Northwest, USA, may cause both retreat of alpine glaciers and increases in the frequency and magnitude of storms delivering rainfall at high elevations absent significant snowpack, and both of these changes may affect the frequency and severity of destructive debris flows initiating on the region's composite volcanoes. A better understanding of debris-flow susceptibility on these volcanoes’ slopes is therefore warranted. Field mapping and remote sensing data, including airborne light detection and ranging (LiDAR), were used to locate and characterize initiation sites of six debris flows that occurred during an “atmospheric river” event (warm wet storm) on Mount Rainier, Washington, in November 2006, and data from prior studies identified six more debris flows that occurred in 2001–2005. These 12 debris flows had initiation sources at the heads of 17 gullies distributed over seven distinct initiation zones near the termini of glaciers, and all debris-flow initiation sites were located within areas exposed by glacier retreat in the past century. Gully locations were identified by their steep walls and heads on a 1-m digital elevation model (DEM) from LiDAR data collected in 2007–2008. Gullies in which debris flows initiated were differentiated from numerous non-initiating gullies primarily by the greater upslope contributing areas of the former. Initiation mechanisms were inferred from pre- and post-2006 gully width measurements from aerial photos and the LiDAR DEM, respectively, field observations of gully banks, and elevation changes calculated from repeated LiDAR, and these data indicate that debris flows were initiated by distributed sources, including bank mass failures, related to erosion by overland flow of water. Using gully-head initiation sites for debris flows that occurred during 2001–2006, a data model was developed to explore the viability of the method for characterization of debris-flow initiation susceptibilities on Mount Rainier. The initiation sites were found to occupy a restricted part of the four-dimensional space defined by mean and standard deviation of simulated glacial meltwater flow, slope angle, and minimum distance to an area of recent (1994–2008) glacier retreat. The model identifies the heads of most gullies, including all sites of known debris-flow initiation, as high-susceptibility areas, but does not appear to differentiate between areas of varying gully-head density or between debris-flow and no-debris-flow gullies. The model and field data, despite limitations, do provide insight into debris-flow processes, as well as feasible methods for mapping and assessment of debris-flow susceptibilities on periglacial areas of the Cascade Range.