Monitoring migration rates of an active subarctic dune field using optical imagery

Abstract

We developed a novel method to quantify subtle rates of landscape evolution using two satellite imaging systems with different viewing angles and spectral sensitivities. We selected the slowly migrating, high-latitude, subarctic Great Kobuk Sand Dunes (GKSD), Kobuk Valley National Park, Alaska (USA), for our study. The COSI-Corr technique was used for precise orthorectification, co-registration, and subpixel correlation of satellite data. ASTER Visible Near Infrared (VNIR) and SPOT Panchromatic images with a 5-year temporal separation were correlated to measure the horizontal velocity of the GKSD. To reduce correlation noise, ASTER VNIR bands were linearly mixed to match the SPOT Panchromatic band, and raw correlation measurements were projected onto a local robust migration direction to estimate unbiased velocity magnitudes. The results show that the most likely migration rate for the GKSD ranges from 0.5 to 1.5 m/year, with peak velocities up to 3.8 m/year, and uncertainty of approximately 0.16 m/year. The unprecedented ability to measure slow migration rates, including those that may occur over a relatively short time interval, illustrates the value of this method to reliably detect and monitor subtle ground movements including dune migration, glacier flow, mass movements, and other small-scale processes.