Soil Moisture Retrieval Over Bare Soil Surface From Single-Polarization SAR Data by Combining Neighborhood Pixels

• Published in: IEEE geoscience and remote sensing letters Vol. 19; pp. 1 - 5
• Main Authors: Tang, Pinjun, Zhao, Rong, Zhu, Jianjun, Xie, Qinghua, Hu, Jun
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Agricultural land; Backscatter; Bare soil surface; Correlation; Correlation coefficient; Correlation coefficients; Dielectric constant; Distribution; Height; IEM; In situ measurement; Integral equations; Methods; Moisture content; Moisture effects; neighborhood pixels; Neighborhoods; Permittivity; Pixels; Polarization; Probability distribution; Probability theory; Radar data; Rough surfaces; SAR; SAR (radar); single-polarization; Soil; Soil moisture; Soil moisture content; Soil surfaces; Surface roughness; Synthetic aperture radar; Water content
• ISSN: 1545-598X; 1558-0571
• DOI: 10.1109/LGRS.2022.3197832

The objective of this letter is to extend a method proposed by Kweon et al. to retrieve soil moisture (<inline-formula> <tex-math notation="LaTeX">m_{v} </tex-math></inline-formula>) over bare soil surface by combining neighborhood pixels of single-polarization synthetic aperture radar (SAR) data. This letter uses single-polarization (HH, VV) SAR data to simultaneously retrieve the root-mean-square (rms) height (<inline-formula> <tex-math notation="LaTeX">h_{\mathrm {rms}} </tex-math></inline-formula>) and the real part of the relative dielectric constant (<inline-formula> <tex-math notation="LaTeX">\varepsilon _{s} </tex-math></inline-formula>) which can be converted to soil moisture content. For the copolarization SAR data, the letter first uses the integral equation model (IEM) and the semiempirical calibration of the correlation length (<inline-formula> <tex-math notation="LaTeX">L </tex-math></inline-formula>) to obtain the probability distribution curve of rms height and the real part of the relative dielectric constant for each neighborhood pixel. Then, these probability distribution curves are placed on the <inline-formula> <tex-math notation="LaTeX">\varepsilon _{s} </tex-math></inline-formula>-<inline-formula> <tex-math notation="LaTeX">h_{\mathrm {rms}} </tex-math></inline-formula> plane, and the juxtaposition model is applied to obtain the average value of estimations of neighborhood pixels. The average soil moisture estimations of neighborhood pixels in farmlands are compared with the in situ measurements with the RMSE equal to 0.036 cm 3 /cm 3 and the correlation coefficient equal to 0.84 at VV polarization in the L-band, which demonstrates that the proposed method is suitable to invert soil moisture with acceptable accuracy and high resolution. However, volume scattering contribution from crops can decrease the performance of the proposed method.