Every day, hundreds of Global Navigation Satellite System (GNSS) satellites—such as GPS and Galileo—bathe the Earth in microwave signals. While these signals are primarily designed for navigation, they also act as an opportunistic, planetary-scale radar system. When a GNSS signal strikes the Earth's surface, it bounces off the terrain before scattering into space.
By capturing these reflections using specialized satellite receivers, a technique known as Global Navigation Satellite System Reflectometry (GNSS-R), scientists can infer the physical properties of the ground below. This approach has emerged as a powerful, cost-effective tool for tracking the movement and accumulation of water across the globe.
The Intuition: Earth as a Mirror
To understand why GNSS-R works, imagine shining a flashlight onto two different surfaces: a rough concrete floor and a puddle of water. The concrete scatters the light in every direction, making the reflection diffuse and faint. The puddle, acting as a mirror, reflects a concentrated beam of light directly back at you.
GNSS signals behave the same way. Dry soil is rough and has a low dielectric constant, causing the microwave signals to scatter widely. Open water or highly saturated soil, however, acts as a specular (mirror-like) reflector. Because water has a remarkably high dielectric constant, it bounces the signal back with immense power. By measuring the strength of this reflected signal—termed surface reflectivity (SR)—satellites can easily differentiate between dry land and water.
Dry Soil (Low Dielectric) Flooded/Wet Soil (High Dielectric)
| |
| (Incident Signal) | (Incident Signal)
v v
/ \ \ (Diffuse Scattering) / (Strong Specular Reflection)
/ \ \ /
From Raw Signals to Flood Maps
The primary data structure captured by a GNSS-R receiver is a Delay-Doppler Map (DDM). When a reflected signal is collected, it does not arrive all at once. Variations in the terrain cause portions of the signal to arrive at slightly different times (delay), while the relative motion of the satellite introduces frequency shifts (Doppler shift). A DDM plots the power of the reflected signal across these two dimensions.
When a region experiences flooding, the distribution of power within the DDM changes dramatically. The diffuse energy concentrates into a sharp, high-intensity peak.
To turn these raw power maps into actionable flood boundaries, researchers rely on empirical thresholding. In studies of extreme weather events, such as those conducted by Liwen Xu and Wei Wan, data from NASA’s CYGNSS satellite constellation—originally designed to measure hurricane wind speeds—was repurposed for land hydrology. In their work, the authors demonstrated that setting a simple empirical threshold (such as 17 dB) on derived surface reflectivity could reliably isolate inundated areas during intense precipitation events.
Similarly, Shuangcheng Zhang et al. utilized CYGNSS DDMs to map the devastating 2021 floods in Henan Province, China. In their paper, the authors showed that tracking the sudden inflation of surface reflectivity values before, during, and after a storm provides a highly responsive, cloud-penetrating narrative of flood progression.
Beyond Binary: Soil Moisture Inversion
The utility of GNSS-R extends beyond detecting open floodwaters; it can also capture subtle gradients of moisture trapped within the soil. As soil absorbs water, its dielectric properties increase proportionally, scaling the surface reflectivity smoothly between the "dry land" and "open water" extremes.
However, translating a specific reflectivity value into an exact soil moisture percentage is challenging on complex terrain, where slope angles and vegetation scatter the signal unpredictably. To solve this, Qingfeng Hu et al. developed a deep learning framework to monitor the expansive soil slopes of China's South-to-North Water Diversion Project. As detailed in their study, the authors fused signals from multiple GNSS frequencies and trained neural networks to invert the reflectivity data, achieving highly accurate soil moisture estimates that serve as early warning signs for slope failures and landslides.
The Broader Sensing Landscape
While GNSS-R offers exceptional temporal frequency by recycling existing infrastructure, it exists alongside a broader ecosystem of environmental sensing techniques, each optimized for different scales and environments:
- Oceanic Scale: Beyond land hydrology, GNSS-R is widely used over the ocean. As noted by Ana B. Villas Bôas et al. in their overview of marine requirements, analyzing how waves distort the specular reflection allows scientists to map global surface winds, ocean currents, and sea state conditions.
- Sub-Pixel Urban Scale: In dense cities, tall buildings cast shadows and block satellite lines-of-sight, creating noise in microwave data. To map urban water, bo du et al. proposed a complementary optical approach in their paper. By applying spectral unmixing to Landsat imagery, they isolated a "Low Albedo Fraction" to map surface water at a sub-pixel level, bypassing the structural obstructions that complicate radar methods.
- Ground-Based Mesoscale: For localized agricultural management, satellite footprints can sometimes be too coarse. Addressing this gap, Yifei Jiang et al. investigated Cosmic-Ray Neutron Sensing (CRNS) in their research. CRNS operates on the ground, measuring ambient neutron intensity—which is inversely proportional to the number of hydrogen atoms in the area—providing a non-contact measurement of soil water content across entire farmlands.
By turning the planet's existing communication overhead into a scientific instrument, GNSS-R bridges the gap between these localized ground sensors and traditional active satellites, offering a continuous, real-time pulse of the global water cycle.
Key Foundational Work
Using CYGNSS Data to Monitor China's Flood Inundation during Typhoon and Extreme Precipitation Events in 2017 (2019) — contributed 4 concept elements to this direction.
Quantifying Sub-Pixel Surface Water Coverage in Urban Environments Using Low-Albedo Fraction from Landsat Imagery (2017) — contributed 2 concept elements to this direction.
Integrated Observations of Global Surface Winds, Currents, and Waves: Requirements and Challenges for the Next Decade (2019) — contributed 1 concept elements to this direction.