HR: 0800h
AN: H41F-0473    [Abstracts]
TI: Non-Destructive Geophysical Monitoring of Near-Surface Water-Content: Comparison Between GPR Velocities Obtained From Surface and Well Experiments
AU: * MANGUE, M
EM: madeleine.mangue@univ-pau.fr
AF: MIGP FRE 2639 UPPA CNRS TOTAL, BP 1155 CEDEX, PAU, 64013 France
AB: Pollution of sub-surface aquifers is a major environmental problem. An important step would be the validation of a non-destructive methodology able to monitor water tranfers from the soil surface to the water table in quality and quantity. The study zone is situated in the alluvial plain of the Adour river in the northen Pyrénées piemont (France). It is located at the border of corn field where fertilizers and irrigation are applied. The alluvial deposits are composed of pebbles, sand, silt which overly an impermeable layer of clay with pebbles. From february 2002 to january 2003, multi-offset Ground Penetrating Radar (GPR) profiling was conducted, completed with electrical conductivity measurements. From the obtained data, total porosity was determinated and the monitoring of vertical transport from soil surface to the water table was performed each month. To process GPR data, original technique were used, Common Refexion Surface Stack (CRS) for the RMS velocity estimation, followed by kriging for interpolation. In order to convert RMS velocities in time into interval velocity in depth, the classical Dix formula was used. In january 2004 and in june 2005, crosswell transmission GPR data were recorded so as to compare with the results obtained from the multi-offset GPR acquisition. The purpose is to validate the non-destructive surface methodology from different comparisons between the two datasets. Ten holes were drilled along the same profile every five meters with a six meters depth. Transmitter and receiver are in different holes and measurements were recorded from wells pair. From the crosswell GPR dataset, in a first step, only horizontal raypathes were considered, i.e., transmitter and receiver at the same depth and at a five meters distance. For such a subset, assumption of straight raypathes is reasonnable due to the weak dip of the soil layers that we can consider as sub-horizontal. Comparisons were made between velocities obtained from multi-offset GPR data and crosswell GPR data for dry (june to july) and moist period (december to january). In dry period, the water table varies from 2 m to 3.5 m and in moist period, it varies from 1 m to 1.5 m. In each case, RMS and interval velocities were calculated from both experiments. In the moist period, both interval and RMS velocity maps reveal the same range of velocity variation. In the dry period, interval velocities for both acquisitions seem to have similar variation but discrepancies appear for RMS velocities. If we consider only the moist period, the non-destructive surface methodology can be validated. The differences in RMS velocities in the dry period, could be due to the volumetric water content which was not exactly similar in july 2002 and in june 2005. The differences could also be due to the processing techniques for multi-offset GPR profiling. To use the CRS technique, a constant surface velocity has to be chosen, and this latter could generate discrepancies in RMS velocities comparisons. A more accurate and perhaps variable, estimation of this surface velocity could provide closer RMS velocities for both acquisitions, in the dry period.
DE: 1835 Hydrogeophysics
SC: Hydrology [H]
MN: Fall Meeting 2005