CAN WE EXTRACT INTERSEISMIC STRAIN ACCUMULATION SIGNALS FROM HISTORIC AERIAL PHOTOGRAPHS AND MODERN HIGH-RESOLUTION TOPOGRAPHY?

Space-based geodetic techniques have facilitated major advances in our understanding of rates and processes of crustal deformation over the past 20 years. Although we have steadily increased the spatial resolution of these observations, we unfortunately must simply wait for these techniques to gather data into the future to take advantage of this increased resolution. To attempt to take advantage of existing datasets to extend geodetic-quality observations into the recent past, we are testing a workflow that accommodates the measurement of crustal deformation at higher spatial resolution and over longer time spans than possible with space-based techniques through the use of historic aerial photograph collections and recent lidar data. We selected the northern San Andreas fault near Point Arena, CA, as a test locality due to the existing coverage by a publicly available lidar dataset extending 5-7 km on both sides of the fault trace. This region is also covered by multiple small-scale aerial photograph collections extending back to 1957. We produced dense point clouds from the aerial photograph collections using Agisoft Photoscan Pro covering a region corresponding with the lidar coverage. We use the freeware CloudCompare to calculate surface displacements between the point clouds, assuming these displacements should represent the 46 years of strain accumulation between the two datasets. Although the initial 3D displacement vectors can exhibit significant variability, local filtering and manual editing where anthropogenic and vegetative changes produce clearly non-tectonic signals facilitates the extraction of tectonic surface displacements. These methods are computationally intensive; however, the critical challenges lie in scaling and georeferencing the point clouds derived from aerial photographs and testing to ensure that biases are not introduced during this process. Temporal patterns in surface displacement are possible in regions with repeat coverage of high-resolution aerial photography and where the tectonic rates are high enough to produce a measurable signal from shorter time intervals.