25
Advanced Ground-Penetrating Radar for Soil Moisture Retrieval
Figure 2.9c shows the soil moisture inferred from time-lapse GPR measurements.
The top few centimeters of the soil were sensitive to evaporation and dried more
rapidly. This effect can be observed by the faster decrease in the GPR-derived soil
moisture. During the night, a slight increase in the surface soil moisture occurred,
most likely because of the dew, which can be observed in the GPR-estimated soil
moisture. For instance, in Figure 2.9a, four gray patches represent the periods when
there was almost no evaporation. These periods are highlighted in Figure 2.9c, showing the corresponding slight increase in the GPR-derived water content. Three undisturbed cylindrical samples of 100 cm 3 were extracted near the time-lapse GPR
setup. The mean saturated soil moisture estimated from the three soil samples was
0.412 m 3 /m 3 . At the time of the precipitation events, the mean of the three maximum soil moisture estimated by GPR was 0.426 m 3 /m 3 , which is very close to the
saturated soil moisture inferred from the soil samples. The slight difference in the
saturated soil moisture obtained by the two methods may be due to the different
characterization scales, soil spatial variability, and the petrophysical model relating
dielectric permittivity to water content, or a combination of these three major factors. Figure 2.9d depicts the maximum peak-to-peak (PtP) amplitude of the signal
recorded between 10.5 and 11.0 ns in Figure 2.9b. The trend of the PtP amplitude
corresponded well to the evaporation and precipitation events. The maximum PtP
amplitude can be observed during precipitation events, and the decreasing trend
shows the effect of evaporation.
2.4.6 teMPoRal StaBility of Soil MoiStuRe PatteRnS
Soil moisture is an ephemeral variable characterized by a high spatial and temporal
variability. When installing soil moisture point measurement devices (e.g., TDR
and capacitance probes), representative locations of a field or catchment in terms of
soil moisture would be preferred. In that respect, several authors have investigated
the temporal stability of soil moisture pattern (e.g., Guber et al. 2008). Using timelapse GPR measurements, we characterized the spatiotemporal soil moisture distribution in a 2.5-ha agricultural field in Vieusart, Belgium, using five high-resolution
acquisitions of GPR data in March and April 2010 (Minet et al. in preparation;
Figure 2.10).
The first three dates were characterized by dry conditions, whereas rainfalls were
observed the day before the fourth date. In Figure 2.10, zones where interpolated soil
moisture values are equal to the field average (±0.01 m 3 /m 3 ) are outlined by black
hatched areas. These zones intersect between the five dates (orange areas), indicating time-stable locations for the field average soil moisture. There was a remarkable
temporal stability of soil moisture patterns for the first three and the two last dates,
respectively. However, due to moderate rainfalls (24.8 mm), the soil moisture pattern
largely changed between the third and fourth dates. In particular, the zones indicating the spatial-average soil moisture shrank from dry to wet conditions as the standard deviation of soil moisture increased. Finally, the time-stable zones indicating
the field average appeared to be located in mid-slopes areas, as already noticed by
Jacobs et al. (2004). Nevertheless, field acquisitions in other seasons are needed to
Advanced Ground-Penetrating Radar for Soil Moisture Retrieval
Figure 2.9c shows the soil moisture inferred from time-lapse GPR measurements.
The top few centimeters of the soil were sensitive to evaporation and dried more
rapidly. This effect can be observed by the faster decrease in the GPR-derived soil
moisture. During the night, a slight increase in the surface soil moisture occurred,
most likely because of the dew, which can be observed in the GPR-estimated soil
moisture. For instance, in Figure 2.9a, four gray patches represent the periods when
there was almost no evaporation. These periods are highlighted in Figure 2.9c, showing the corresponding slight increase in the GPR-derived water content. Three undisturbed cylindrical samples of 100 cm 3 were extracted near the time-lapse GPR
setup. The mean saturated soil moisture estimated from the three soil samples was
0.412 m 3 /m 3 . At the time of the precipitation events, the mean of the three maximum soil moisture estimated by GPR was 0.426 m 3 /m 3 , which is very close to the
saturated soil moisture inferred from the soil samples. The slight difference in the
saturated soil moisture obtained by the two methods may be due to the different
characterization scales, soil spatial variability, and the petrophysical model relating
dielectric permittivity to water content, or a combination of these three major factors. Figure 2.9d depicts the maximum peak-to-peak (PtP) amplitude of the signal
recorded between 10.5 and 11.0 ns in Figure 2.9b. The trend of the PtP amplitude
corresponded well to the evaporation and precipitation events. The maximum PtP
amplitude can be observed during precipitation events, and the decreasing trend
shows the effect of evaporation.
2.4.6 teMPoRal StaBility of Soil MoiStuRe PatteRnS
Soil moisture is an ephemeral variable characterized by a high spatial and temporal
variability. When installing soil moisture point measurement devices (e.g., TDR
and capacitance probes), representative locations of a field or catchment in terms of
soil moisture would be preferred. In that respect, several authors have investigated
the temporal stability of soil moisture pattern (e.g., Guber et al. 2008). Using timelapse GPR measurements, we characterized the spatiotemporal soil moisture distribution in a 2.5-ha agricultural field in Vieusart, Belgium, using five high-resolution
acquisitions of GPR data in March and April 2010 (Minet et al. in preparation;
Figure 2.10).
The first three dates were characterized by dry conditions, whereas rainfalls were
observed the day before the fourth date. In Figure 2.10, zones where interpolated soil
moisture values are equal to the field average (±0.01 m 3 /m 3 ) are outlined by black
hatched areas. These zones intersect between the five dates (orange areas), indicating time-stable locations for the field average soil moisture. There was a remarkable
temporal stability of soil moisture patterns for the first three and the two last dates,
respectively. However, due to moderate rainfalls (24.8 mm), the soil moisture pattern
largely changed between the third and fourth dates. In particular, the zones indicating the spatial-average soil moisture shrank from dry to wet conditions as the standard deviation of soil moisture increased. Finally, the time-stable zones indicating
the field average appeared to be located in mid-slopes areas, as already noticed by
Jacobs et al. (2004). Nevertheless, field acquisitions in other seasons are needed to