High-Resolution Ground-Based GPS Measurements Show Intercampaign Bias in ICESat Elevation Data Near Summit, Greenland

The Geoscience Laser Altimeter System (GLAS) aboard the National Aeronautics and Space Administration's Ice, Cloud, and land Elevation Satellite (ICESat) collected data from early 2003 to late 2009 with the specific goal of measuring ice-surface elevation changes. While the precision of GLAS instrumentation has been studied over its intended target (ice), its accuracy has only been robustly estimated using independent (terrestrial nonlaser) methods over salt flats. Here, we perform repeat high-precision Global Positioning System (GPS) surveys under four passes of ICESat track 0412 (campaigns L3I, L3J, L2D, and L2E) to compare directly GLAS elevation data footprints to a coincident GPS ground truth near Summit, Greenland. Analysis and comparison of GLAS data with GPS data show a campaign-dependent elevation bias ranging from -0.112 +/- 0.030 m (L3J) to 0.121 +/- 0.071 m (L2E). Although uncorrected reflectance values and field observations both indicate that forward scattering of the laser signal through the atmosphere accounts for the anomalously negative L3J bias, the biases of all campaigns studied are within the instrument's goal accuracy of +/- 0.15 m. However, our analysis shows a campaign dependence in the bias, which may propagate through estimates of mass balance. The error introduced from intercampaign biases illustrates the importance of long-term independent validation experiments of satellite altimetry data over ice sheets.