Soil moisture variation estimated from GPS-IR using FFT and LS

Abstract

Accurate and long-term monitoring of soil moisture is of great significance for global water and carbon cycles. The soil moisture variation estimated from GPS-IR (GPS-InterferometricReflectometry) overcomes some drawbacks of traditional ways, and has become an important topic. Changes in the permittivity of the soil, which are associated with fluctuations in soil moisture, affect the effective frequency, phase, and amplitude of signal-to-noise ratio (SNR) data recorded by the GPS receiver. This study used Fast Fourier Transform (FFT) algorithm with equal sinusoidal elevation angle-interval sampling and Least Square (LS) method mainly used in most previous studies to acquire GPS interferogram metrics, by comparing the retrievals with volumetric soil moisture retrieved by PBO H2O group. The values of frequency extracted by these two algorithms linearly and negatively correlate with surface soil moisture, showing correlations of −0.57 for LS and −0.45 for FFT. However, the correlation coefficient for phase extracted by FFT was 0.61, greater than that by LS of 0.31, both positively.