Ming Li

This study investigates the fusion of GNSS-R reflectivity from Spire Global Inc. and the CYGNSS constellation to enhance SMAP soil moisture retrievals at 3-km resolution. Bias corrections were applied to align Spire BDS and Galileo to GPS and to reconcile differences with CYGNSS, yielding a consistent +2.41 dB offset and overall correlation of 0.81, with land cover-dependent variability. After fusion and spatial completion using the Previously Observed Behavior Interpolation (POBI) method, the combined reflectivity was used to downscale SMAP brightness temperature (TB) from 9 km to 3 km. Analysis showed that TB decreased with increasing soil moisture but was influenced by vegetation and surface conditions, whereas surface emissivity exhibited a stronger relationship to soil moisture (R 2 = -0.84). Validation against Sentinel/SMAP 3-km products, SMAP 9-km products, and in-situ measurements from 85 ground stations demonstrated improved performance of the Spire+CYGNSS/SMAP retrievals, with higher correlations (up to R 2 = 0.89) and lower RMSE (as low as 0.038 cm 3 /cm 3 ) relative to SMAP alone, particularly over sparsely vegetated areas. The fused product retains SMAP's 2-3 day revisit cycle while enhancing spatial resolution, underscoring the complementarity of multi-mission GNSS-R for high-resolution hydrological applications.

This study develops semi-empirical and linear regression algorithms to estimate near-surface soil moisture (SM) using reflectivity observations of Global Navigation Satellite System reflection (GNSS-R) signals collected by Spire Global commercial satellites. The semi-empirical method employs a forward model of GNSS-R reflectivity considering soil dielectric properties, vegetation attenuation, and surface roughness effects. As direct roughness measurements are unavailable globally, we obtain effective roughness attenuation maps using Cyclone GNSS (CYGNSS) reflectivity and SMAP SM and vegetation optical depth (VOD) data. Systematic biases between Spire and CYGNSS observations are also characterized and corrected. Daily gridded corrected Spire reflectivity observations at a 36-km resolution are used to invert soil dielectric permittivity and then SM. Linear regression parameters are also determined by correlating corrected Spire reflectivity measurements to SMAP SM data for each grid. Both algorithms are applied to six months of Spire data from late Jan to Jul 2024 and evaluated over Australia. The linear regression approach achieves better overall agreement with SMAP (correlation coefficient: 0.85, root mean squared difference (RMSD): 0.05 cm3/cm3) compared to the semi-empirical method (correlation: 0.64, RMSD: 0.07 cm3/cm3). The semi-emprical method demonstrates relatively better accuracy over sparsely vegetated regions with RMSD smaller than 0.06 cm3/cm3. Both methods show degraded performance over dense vegetation. Improved characterization of vegetation and surface roughness effects is essential to enhance retrieval accuracy. This work contributes to positioning GNSS-R as a complementary technology for independent SM observations.

Presents corrections to the paper, Corrections to “Spire Near-Nadir GNSS-R for Sea Ice Detection: First Results”.

Global Navigation Satellite System Reflectometry (GNSS-R) has emerged as a valuable tool forEarth observations, providing low-cost measurements with high spatiotemporal resolution. This study presents the first application of Spire Global Inc. near-nadir (NN) GNSS-R observations for the sea ice detection and sea ice extent mapping, covering the Arctic and Antarctic. The delay and Doppler maps (DDMs) of Spire NN GNSS-R show significant difference over various surfaces. For a test the period from February to July 2024, results show that the Arctic’s sea ice extent decreases from 584,600 km₂ to 333,400 km₂, and Antarctic’s sea ice extent increases from 142,825 km₂ to 584,750 km₂. Compared to Advanced Microwave Scanning Radiometer 2 (AMSR2) estimates, Spire NN GNSS-R retrievals achieve a probability of detection of 94–96%, and the probability of false alarm and Probability of Error below 0.23%. These results demonstrate the high sensitivity, accuracy and operational potential of Spire NN GNSS-R for the sea ice remote sensing.

Measurements have been made at P-band (0.707 GHz) to construct a model for the dielectric constant of sea water and extend the model for the dielectric constant to high salinity (S > 50 pss). The measurements are part of research to develop a model for the dielectric constant suitable for future wide-bandwidth remote sensing of salinity and for application to water bodies such as the Great Salt Lake with salinity significantly above that found in the open ocean. Measurements have been made at temperatures from 2 to 30 °C and salinity from 0 to 138 pss. The data have been fit to a Debye model for the dielectric constant with a single relaxation mode as has been employed at L-band (1.413 GHz) where remote sensing of salinity is currently done. Comparison with contemporary models developed from data at L-band indicates that the L-band model and new P-band model do well at both frequencies for salinity less than 50 pss but at higher values of salinity the L-band models diverge from the data. The data has also been used to test at high salinity the mathematical relationship between salinity and conductivity which is the basis for the practical salinity scale (pss).

This study explores the potential of P-band and L-band radiometry for estimating soil moisture and temperature profiles. A total of 3977 hourly in-situ soil data were collected at depths ( d ) of 5-60 cm in Beltsville, MD, USA. Using the data, a coherent model was utilized to generate synthetic brightness temperatures at an incidence angle of 40 degrees for frequencies of 0.8, 0.9, 1.1, and 1.4 GHz. These synthetic brightness temperatures facilitated the estimation of soil moisture ( m_v ) and temperature ( T ) profiles, which were modeled as quadratic functions with three coefficients. The inversion problem was formulated as a least squares problem and optimized by the Adaptive Simulated Annealing (ASA) algorithm. Regression analysis highlighted the sensitivity of the soil moisture and temperature function coefficients on the brightness temperature at 1.4 GHz and the differential between 1.4 GHz and 0.8 GHz (R ² = 0.62 − 0.97), yielding the best-fit models to describe the relationships. The application of the ASA algorithm incorporating the best-fit models to constrain the search spaces, showed the RMSE of soil moisture m_v ≤ 0.04 cm ³ / cm ³ and soil temperature T ≤ 1.35 for depths d ≤ 20 cm , and m_v ≤ 0.10 cm ³ / cm ³ , T ≤ 1.60 for d ≤ 60 cm . A streamlined inversion method was also investigated which used the best-fit models as the retrieval formula and solely relied on the V-pol data. The streamlined inversion method achieved comparable retrieval accuracy but reducing the runtime from 770s to 0.54s for one inversion. This approach simplifies the data collection process by eliminating the need for H-pol data, potentially broadening its flexibility and applicability.

The range of salinity and temperature measurements used to develop models for the dielectric constant of sea water have in the past been limited to values appropriate for the open ocean. But there are important water bodies such as the Great Salt Lake in Utah with salinities much larger than those encountered in the open ocean. Extrapolating existing models to values of salinity beyond the limits of the data used to create the model can result in unrealistic predictions in remote sensing applications. This can be prevented by using a model for conductivity based on the definition of salinity and ensuring that polynomials used to model the other unknown parameters result in bounded behavior at high salinity. New laboratory measurements with high salinity (up to 150 pss) are reported and used to test a model with these adjustments.

We develop the extended boundary condition method (EBCM) to solve the 3-D vector electromagnetic scattering problem from natural bare soil surfaces. Natural surfaces are modeled by the Weierstrass–Mandelbrot fractal surface (FS) that is widely recognized as the most accurate model to describe the natural surfaces properties. The EBCM-FS is developed based on the equivalence theorem that has no limitation on surface roughness in principle. The EBCM-FS is applied to simulate the bistatic scattering cross sections of Mars religoth surfaces. The numerical performance of the EBCM-FS has an equation system much smaller than the numerical method. Solving an equation system including 105 Floquet modes takes 5.4 s for one realization. These properties are expected to make the EBCM-FS as a competitive forward solver to give a better understanding of the measurements of L-band radar returns from Martian surfaces.

This article proposes a closed-form asymptotic solution for the bistatic coherent scattering of a multilayered rough surface structure based upon its effective dielectric constant (EDC) in specular direction. The EDC is modeled by establishing an equivalence of the coherent scattering between the multilayered rough surface structure and a half-space homogeneous medium. The scattering solution is then solved using the scalar Kirchhoff approximation (SKA) method. This new method is referred to as the SKA-EDC method, and it is applied to analyze the sensitivity of the EDC and coherent reflectivity to bare soils with realistic parameters at P- and L-bands. The result indicates that EDC can give different responses to the soil moisture variations with respect to the coherent reflectivity, enabling the potentials of root-zone soil moisture retrieval. At incidence angle smaller than 35°, EDC gives the same value for both polarizations, and the coherent reflectivity can show a significant response to soil at depths <15–50 cm at 0.80 GHz and <5–15 cm at 1.57 GHz, depending on soil moisture. The simulation also demonstrates that subsurface roughness has a trivial effect on the EDC and coherent reflectivity for three rough surfaces separated by continuously dielectric profiles. Subsurface can, thus, be assumed to be flat for reducing uncertainty in soil moisture inversion algorithms when the subsurface roughness is unknown.