Measuring snow water equivalent from common-offset GPR records through migration velocity analysis

Abstract

Abstract. Many mountainous regions depend on seasonal snowfall for their water resources. Current methods of predicting the availability of water resources rely on long-term relationships between stream discharge and snowpack monitoring at isolated locations, which are less reliable during abnormal snow years. Ground-penetrating radar (GPR) has been shown to be an effective tool for measuring snow water equivalent (SWE) because of the close relationship between snow density and radar velocity. However, the standard methods of measuring radar velocity can be time-consuming. Here we apply a migration focusing method originally developed for extracting velocity information from diffracted energy observed in zero-offset seismic sections to the problem of estimating radar velocities in seasonal snow from common-offset GPR data. Diffractions are isolated by planewave-destruction (PWD) filtering and the optimal migration velocity is chosen based on the varimax norm of the migrated image. We then use the radar velocity to estimate snow density, depth, and SWE. The GPR-derived SWE estimates are within 6 % of manual SWE measurements when the GPR antenna is coupled to the snow surface and 3-21 % of the manual measurements when the antenna is mounted on the front of a snowmobile ∼ 0.5 m above the snow surface.

Quantifying snow water equivalent using terrestrial ground penetrating radar and unmanned aerial vehicle photogrammetry

“…The RMSE of both transect and raster SWEs are ±63 mm and ± 69 mm, respectively. The RMSE for the transect SWE and raster SWE relative to the mean SWE of the snow tube are 17% and 19% St. Clair and Holbrook (2017).…”

Detection of Underground Anomalies Using Analysis of Ground Penetrating Radar Attribute

“…If the migrated velocity is not correct, over-migration or under-migration effects can occur with higher velocity or lower velocity than the environment velocity, respectively (Figure 1). In the wrong velocity case, the GPR migrated amplitudes are still stretched in curved shapes upward or downward, causing highly entropy data (De Vries and Berkhout, 1984, Wiggins, 1978, Prego et al, 2017, Fomel et al, 2007, Levy and Oldenburg, 1987, Clair and Holbrook, 2017. For these research works, the varimax and its inverse, entropy, within window zone containing the peak of diffraction hyperbola can also reach maximum or minimum, respectively.…”

“…Building GPR attribute diagram needs a set of migrated GPR sections from a velocity band in advanced. The extremes of the GPR attribute (i.e., entropy, energy, and varimax) can reflect to the suitable environment velocity (Fomel et al, 2007, Clair andHolbrook, 2017). Each GPR attribute is formed using two variables, velocity and window zone through its equation as followed:…”

“…We have developed a new workflow of calculating varimax section for defining environment velocity and size of an object. For small objects, the conventional workflow (Fomel et al, 2007, Clair andHolbrook, 2017) expresses that maximum varimax value in a full rectangular window can relate to environment velocity. Apparently, maximum varimax values can lead to smallest number of spikiness in the window area of the GPR migrated data.…”

“…The diagram can show its extreme value corresponding to the environment velocity. Entropy and its inverse, varimax, are of great indicator for velocity analysis in both geophysical data, seismic and GPR data (De Vries and Berkhout, 1984, Wiggins, 1978, Prego et al, 2017, Fomel et al, 2007, Levy and Oldenburg, 1987, Clair and Holbrook, 2017. For these research works, diffractions and reflections can be separated and the diffracted hyperbolae are migrated with suitable velocity band.…”