This paper shows the results of soil moisture content measurements by GPR tests in a study area in São Paulo City, Brazil and discuss some aspects related to the accuracy of them. Two different data interpretation methods were used: i) lineal adjustment of the arrival times of ground wave for different antenna separation; ii) time differences between the air and ground wave arrival times at a single antenna separation. We used the RAMAC Mala Geosciences system with 50 MHz, 100 MHz and 200 MHz antennas. The GPR survey was constituted of several multi-offset acquisition, using the WARR (Wide Angle Reflection and Refraction) geometry along of the profile in three different time periods (dry, intermediate and wet seasons). The distance increment between each WARR arrangement and traces was 0.1 m for the 100 MHz and 200 MHz antennas and 0.2 m for the 50 MHz. At the same time of each GPR survey, soil samples, from 0.5 m up to 4 m depth, were collected, in 0.5 m intervals, for water content analysis in laboratory tests. The GPR data showed a good correlation with the laboratory data, but the results from the 200 MHz antennas were the best one. The two data interpretation methods showed different accuracies. The time differences method was less precise due to the time shift in the ground wave arrival time at zero antenna offset that occurred in this survey. These results showed that it is very important to apply a correction factor for this effect, when it is present.
GPR; moisture content; soil
