Accurate soil moisture content measurements are vital to precision irrigation management. Remote sensing using the microwave spectrum (such as GPS signals) has been used for measuring large area soil moisture contents. In our previous work, we estimated surface soil moisture contents for bare soil using a GPS Delay Mapping Receiver (DMR) developed by NASA. However, the effect of vegetation was not considered in these studies. Hence the objectives of this study were to: 1) investigate the feasibility of using DMR to determine soil moisture content in cotton production fields; 2) evaluate the attenuation effect of vegetation (cotton) on reflected GPS signal. Field experiments were conducted during the 2013 and 2014 growing seasons in South Carolina. GPS antennas were mounted at three heights (1.6, 2.7, and 4.2 m) over cotton fields to measure reflected GPS signals (DMR readings). DMR readings, soil core samples, and plant measurements were collected about once a week and attenuation effect of plant canopy was calculated. Results showed that DMR was able to detect soil moisture changes within one week after precipitation events that were larger than 25 mm per day. However, the DMR readings were poorly correlated with soil volumetric water content during dry periods. Attenuation effect of plant canopy was not significant. For irrigation purpose, the results suggested that the sensitivity of reflected GPS signals to soil moisture changes needed to be further studied before this technology could be utilized for irrigation scheduling in cotton production. Refinement of this technology will expand the use of advanced remote sensing technology for site-specific and timely irrigation scheduling. This would eliminate the need to install moisture sensors in production fields, which can interfere with farming operations and increase production costs.
Competition for limited water resources is one of the most critical issues currently affecting civilization. The ability to make more water available for domestic, agricultural, industrial and environmental uses will depend on better management of water resources, watersheds, and storm water runoff. Accurate soil moisture measurements are vital for improving agricultural management practices. The large contrast between the dielectric constants of water (εw ≈ 80) and dry soil (εs ≈ 3 - 5) at microwave frequencies has provided an opportunity to estimate soil moisture contents using microwave signals [
where ε' is permittivity of soil medium, ε" is the loss factor, and j is the imaginary unit
Soil moisture can be estimated from microwave frequencies using active or passive remote sensing. The active remote sensing technique utilizes radar, which transmits electromagnetic waves to the ground surface and measures the backscattered signals. The passive remote sensing on the other hand, utilizes a radiometer, which measures thermal microwave emission from the soil surface [
Various active or passive systems are currently mounted on ground carriers, aircrafts, or spacecrafts for the purpose of soil moisture monitoring with different spatial resolutions. Over ground level, [
Recently, utilization of GPS signal for remote sensing of soil moisture has been demonstrated over various systems. Geophysical and geodetic intended GPS receivers were utilized for soil moisture retrieval with resolution of 300 m2 [
Field experiments were conducted at the Edisto Research and Education Center (EREC) of Clemson University near Blackville, South Carolina (Latitude 33˚21'55.34N, Longitude 81˚19'48.22"W), during the 2013 and 2014 cotton growing seasons. In 2013, the experiment was conducted in a cotton field equipped with a lateral variable rate irrigation system (Lateral Field, LF). In 2014, in addition to LF field, a separate field equipped with drip irrigation system was used (Drip Field, DF). Field LF was planted with cotton variety DP 1050 on May 22, 2013 and May 7, 2014. The DF field was planted to cotton variety PHY333WRF on May 27, 2014. During both years, plant density was 82,145 seeds per hectare with 96-cm row spacing. The DF field allowed testing the effect of different levels of vegetation on GPS reflectivity since cotton at this field was planted at a later date. The plot size at both fields was 7.6 m by 15.2 m. Both fields had the same soil type, Varina loamy sand (Fine, kaolinitic, thermic Plinthic Paleudults).
The DMR is a bi-static system, which uses two antennas with a zenith Right Hand Circularly Polarized (RHCP) antenna (3G15A-XS-1, Antcom, U.S.A.) viewing the sky and a nadir Left Hand Circularly Polarized (LHCP) antenna (3G15L-A-XS-1, Antcom, U.S.A.) viewing the ground. The DMR system utilizes a Zarlink GP2010/ 2021 chipset (ZarlinkSemiconductors, Ottawa, Canada) which has two antenna inputs and, therefore, both the direct and reflected signals can be acquired.
In 2013, the antenna platform (including RHCP and LHCP GPS antennas) was mounted on the lateral VRI at the LF field (
The DMR records the signal strength of direct and reflected GPS signal from the RHCP and LHCP antennas, respectively. The instrument output, which is the ratio of strength of reflected signal to direct signal, can be used to estimate surface reflectivity. Hence, it is important to make sure the output difference of the two DMR channels is only due to ambient environment, rather than the circuitry noise. The two input channels inside the DMR were evaluated using a balanced RF power divider, as explained by Privette et al. [
ratio (SNR) of each channel was recorded. In 2014, the system was also calibrated over a surface with known reflectivity, which was a circular plastic container (2.4 m diameter and 0.3 m deep) filled with water. The reflectivity of a water body is relatively constant (Γs = 0.61) at a temperature range from 20˚C to 30˚C and GPS satellite elevation angles between 50 and 90 degrees [
where Gc is reflectivity from DMR. This calibration factor accounts for noises such as variation in satellite power, instrument temperature, cables, and cable connectors [
Before taking appropriate soil and plant samples in the field, it was necessary to calculate the resolution of the DMR signal. This resolution was calculated using equation 3 [
where A is area (m2) of First Fresnel Zone, h is height (m) of nadir antenna (LHCP antenna) above surface, γ is the elevation angle (degree) of satellite signal.
The satellites traverse an arc in the sky; therefore, during a data acquisition event, the apparent reflection point must then describe an arc on the ground. Since the repeat time for the satellites and their illumination is approximately twice per day (exactly twice per sidereal day), arc paths can be constructed based on the location of the GPS reflectometer [
In both years, data was collected about once a week, during cotton growing season. In 2013, the antenna platform was moved to a predetermined sampling location (reference point) in each experimental plot and the DMR data was collected for 20 minutes. Soil core samples and plant measurements were also collected from the predicted satellite reflection paths. Soil samples were taken at 10 and 20 cm depths and were analyzed for volumetric water content (VWC). Plant parameters such as leaf water potential and stomatal conductance were measured using Model 600 Pressure Chamber (PMS Instrument Company, USA) and SC-1 Leaf Porometer (Decagon Inc., USA), respectively. Similar procedure were followed in 2014, however, the DMR data was collected for 10 minutes per plot, and soil samples were taken from three depths (5, 10, and 15 cm). In addition, two randomly selected plant samples were taken and dried in oven at 65˚C for 72 hours to obtain plant gravimetric moisture content (wet-basis), which was calculated as [
where GW is mass of wet material and GD is mass of dry material. Attenuation of GPS signal caused by plants was calculated as [
where
The dielectric constant of plant (
For each successful data acquisition event in each plot, the DMR stored a series of txt files, including: DCOERR.txt (delay timing information), DEBUG.txt, DELYSUM.txt, DELYTIC.txt (reflected signal strength), DIRECT.txt (direct signal strength), NAVSTATE.txt (DMR position), RINEX2N.txt, and RINEX2O.txt. Satellite information was stored in the two RINEX files. Data was stored at every 0.1 second, named “tics”. For each series of txt file obtained at each plot, a Matlab program developed by NASA was used to calculate average reflectivity during the sampling cycle (20 min in 2013 and 10 min in 2014). Linear regression was used to determine the correlation of reflectivity, either adjusted (in 2014) or not adjusted (in 2013), with volumetric water content of soil and crop stress parameters (stomatal conductance and leaf water potential).
The resolution of the DMR system changes with satellite elevation angle and antenna height (Equation (3)). The DMR software is set to accept only high elevation angle satellites (greater than 60 degrees) to make the area of reflection specific to the area directly below the receiver [
During 2014, the DMR was calibrated over a circular water body on different dates for different satellites. The calibration factor (fc) ranged from 1.03 to 1.21 with standard deviation of 0.06. The daily average calibration factor was used to scale reflectivity collected on the corresponding date. It was reported by [
the DMR was not able to detect changes in soil moisture (
During 2014, reflectivity measurements were adjusted by average calibration factors obtained on different dates. For antenna height of 1.6 m, reflectivity correlated with VWC collected at 5, 10, and 15 cm with R2 of 0.30 (P = 0.0015), 0.24 (P = 0.0053), and 0.27 (P = 0.0031), respectively. For antenna height of 2.7 m, reflectivity correlated with VWC at 5, 10, and 15 cm with R2 of 0.17 (P = 0.0236), 0.08 (P = 0.129), and 0.17 (P = 0.0226), respectively.
During the period of data sampling in 2014, there was heavy rainfall on Sep 9th (27.4 mm), Sep 17th (36.9 mm), and Sep 21st (21.4 mm). Data collected between Sep 9th and Sep 26th was separated as “After Precipitation” points. Similar to 2013, correlation of reflectivity with VWC improved after separating data points in two groups (
0.0266), and 0.45 (P = 0.0335), respectively. Similar to 2013, reflectivity was less correlated with VWC collected during “Dry” periods (
The measured gravimetric moisturecontent of cotton plant (MCW) was in the range of 0.6 to 0.85. According to [
The effect of attenuation to GPS reflectivity caused by cotton canopy was accounted for by multiplying the original reflectivity by the loss factor due to leaves. It should be pointed out that the attenuation of GPS signal due to cotton stalks would be insignificant in this study, since only satellites at high elevation angles were selected
by the DMR software, which makes the signals almost parallel to plant stalks [
This study was conducted to investigate the feasibility of using DMR to determine soil moisture content in cotton production fields and to evaluate the attenuation effect of vegetation (cotton) on reflected GPS signal. Results showed that DMR was able to detect soil moisture changes within one week after precipitation events that
were larger than 25 mm per day. However, the DMR readings were poorly correlated with soil volumetric water content during dry periods. The ability of this advanced remote sensing technology to differentiate between wet and dry soils has potential applications in large scale hydrological modeling and watershed management. The DMR was evaluated using a RF power divider and the result showed that the top and bottom channels of the system produced similar signal to noise ratio. In addition, calibration results over water showed a small variation in calibration factor (1.14 ± 0.06), which indicated that noise of the system was stable throughout the experiment. The dielectric constant of cotton, both real and imaginary components increased with gravimetric water content of cotton. In addition, the loss factor due to leaves increased with volumetric water content of cotton. However, attenuation effect of plant canopy was not significant. After applying loss factors to reflectivity, the correlations between reflectivity and actual soil moisture content improved only for measurements made for antenna height
Year | Height1 (m) | Depth2 (cm) | Original Reflectivity | Adjusted Reflectivity | ||
---|---|---|---|---|---|---|
After Precipitation | Dry | After Precipitation | Dry | |||
2013 | 4.2 | 10 | 0.46 | 0.22 | / | / |
2013 | 4.2 | 20 | 0.01 | 0.08 | / | / |
2014 | 1.6 | 5 | 0.54 | 0.01 | 0.51 | 0.00 |
2014 | 1.6 | 10 | 0.42 | 0.01 | 0.41 | 0.00 |
2014 | 1.6 | 15 | 0.45 | 0.02 | 0.42 | 0.02 |
2014 | 2.7 | 5 | 0.68 | 0.14 | 0.65 | 0.16 |
2014 | 2.7 | 10 | 0.15 | 0.09 | 0.24 | 0.11 |
2014 | 2.7 | 15 | 0.45 | 0.19 | 0.61 | 0.21 |
1Height refers to distance of antenna platform to ground surface. 2Depth refers to distance of soil core samples took from ground surface.
of 2.7 m. For irrigation purpose, the results suggested that the sensitivity of reflected GPS signals to soil moisture changes needed to be further studied before this technology could be utilized for irrigation scheduling in cotton production. Refinement of this technology will allow growers to install the DMR systems on top of an overhead irrigation system for site-specific and timely irrigation scheduling. This would eliminate the need to install moisture sensors in production fields, which usually can interfere with farming operations and increase production costs.
Technical Contribution No. 6433 of the Clemson University Experiment Station. This material is based upon work supported by NIFA/USDA, under project number SC-1700498. The authors acknowledge the support of the South Carolina Space Grant Consortium as well as technical support from NASA Langley Research Center, Hampton, VA, United States.
Mention of a trade name does not imply endorsement of the product by Clemson University to the exclusion of others that might be available.
Xin Qiao,Ahmad Khalilian,Jose O. Payero,Joe Mari Maja,Charles V. Privette,Young J. Han, (2016) Evaluating Reflected GPS Signal as a Potential Tool for Cotton Irrigation Scheduling. Advances in Remote Sensing,05,157-167. doi: 10.4236/ars.2016.53013