MEASURING POST-ERUPTION EVOLUTION OF THE NORTH FORK TOUTLE RIVER, WASHINGTON FROM HISTORICAL AERIAL PHOTOGRAPHY

The eruption of Mount St. Helens on May 18, 1980 and related debris avalanche covered the North Fork Toutle River Basin with about 2.5 km³ of sediment, covering a 14 km reach of the basin with a mean depth of 40 m. During the 37 years since the eruption, extensive monitoring has resulted in a unique collection of data documenting landscape and ecosystem response to disturbance. These data include aerial photography of the North Fork Toutle River Basin, taken annually from 1980 to 1985 by the USGS. Recent advances in structure-from-motion (SfM) photogrammetry have enabled the production of rapid and robust topographic models from photos with > 60% overlap. Here, we demonstrate that these historical aerial photos can produce reasonable topographic data, despite less overlap than most modern SfM photo collections. To correct for low overlap in the photo set, we used a 4-dimensional (i.e., 3D + time) approach in Agisoft Photoscan, whereby photographs from all years were combined to maximize the number of tie points between all photographs. We compared our 1980 – 85 models to a georeferenced set of photos from 2015, identifying stable structures visible in both the archive and 2015 imagery such as parking lots uncovered by debris.. Uncertainty of the models was assessed by analyzing the root mean squared error (RMSE) of tie point positions after a multi-step method used to remove inaccurate and poorly resolved tie points. Significant change between the 1980-1985 models was detected using the M3C2 technique, which calculates distance between two point clouds along a local normal generated from the original surface. This technique accounts for position uncertainty, registration uncertainty between point clouds, and surface roughness related errors. Our dataset documents the transport of volcanic material and development of a post-eruption drainage network in the North Fork Toutle River Basin with unprecedented spatial resolution and geographic extent. Further, the techniques presented here are evidence of our ability to observe geomorphic change directly over longer time scales, even when Lidar and other topographic data are unavailable.