Soil Moisture Remote Sensing using GPS-Interferometric Reflectometry

Chew, Clara Christabel

Graduate Thesis Or Dissertation

Soil Moisture Remote Sensing using GPS-Interferometric Reflectometry Public Deposited

Analytics

Download PDF

Citations

Citeable URL: https://scholar.colorado.edu/concern/graduate_thesis_or_dissertations/12579s30p

Abstract

Ground-reflected Global Positioning System (GPS) signals can be used opportunistically to infer changes in land-surface characteristics surrounding a GPS monument. GPS satellites transmit at L-band, and at microwave frequencies the permittivity of the ground surface changes primarily due to its moisture content. Temporal changes in ground-reflected GPS signals are thus indicative of temporal changes in the moisture content surrounding a GPS antenna.

The interference pattern of the direct and reflected GPS signal for a single satellite track is recorded in signal-to-noise ratio (SNR) data. Alternating constructive and destructive interference as the satellite passes over the antenna results in a noisy oscillating wave at low satellite elevation angles, from which the phase, amplitude, and frequency (or reflector height) can be calculated.

Here, an electrodynamic model that simulates SNR data is validated against field observations. The model is then used to show that temporal changes in these SNR metrics may be used to estimate changes in surface soil moisture in the top 5 cm of the soil column. Results show that changes in SNR phase are best correlated with changes in soil moisture, with an approximately linear slope. Surface roughness decreases the sensitivity of SNR phase to soil moisture, though the effect is not significant for small roughness values (<5 cm).

Modeling experiments show that all three SNR metrics are affected by changes in the permittivity and height of a vegetation canopy. SNR amplitude is the best indicator of changes in vegetation. An increase in either canopy permittivity or height will cause a corresponding decrease in SNR phase. Seasonal changes in vegetation must be removed if soil moisture is to be estimated using phase data.

An algorithm is presented that uses modeled relationships between canopy parameters and SNR metrics to remove seasonal vegetation effects from the phase time series, from which soil moisture time series may be estimated. Results indicate that this algorithm can successfully estimate surface soil moisture with an RMSE of 0.05 cm³ cm^-3 or lower for many of the antennas that comprise the Plate Boundary Observatory (PBO) network.

Creator