Interferometric Synthetic Aperture Radar (InSAR) is a powerful tool for monitoring how the Earth's surface changes over time. However, it suffers from a fundamental geometric limitation: a single satellite pass can only measure ground movement along a single dimension—the radar's Line-of-Sight (LOS).
Because satellites look at the Earth from an angle and travel along fixed orbital paths, the raw data they capture is merely a one-dimensional projection of a three-dimensional reality. If a patch of ground moves, a single LOS measurement cannot distinguish whether the land is sinking vertically, shifting horizontally, or undergoing a combination of both. As noted by Chen et al. (2022), failing to decompose this ambiguous measurement makes it incredibly difficult to interpret the true nature, mechanics, and risks of geohazards.
To overcome this limitation, modern geodetic research relies on multi-geometry deformation decomposition. By combining radar observations taken from different angles and flight paths, researchers can reconstruct the true 2D or 3D vector fields of Earth's moving surface.
The Geometry of the Blind Spot
To understand why multi-geometry decomposition works, we must first look at how a satellite views the Earth. Mathematically, the observed LOS velocity ($V_{LOS}$) is the dot product of the true 3D ground velocity vector ($[U_E, U_N, U_U]^T$ representing East, North, and Up components) and the satellite’s unit viewing vector:
$$V_{LOS} = U_E \sin\theta \cos\alpha - U_N \sin\theta \sin\alpha + U_U \cos\theta$$
where $\theta$ is the radar incidence angle and $\alpha$ is the satellite heading angle.
Because most imaging satellites operate in near-polar orbits, they travel almost due North or due South. This trajectory introduces a severe blind spot: the heading angle $\alpha$ relative to the poles is very small, making the coefficient for the North-South component ($U_N$) close to zero. Consequently, standard InSAR is highly sensitive to vertical motion ($U_U$) and moderately sensitive to East-West motion ($U_E$), but virtually blind to North-South movement.
The 2D Solution: Ascending and Descending Orbits
The most common way to break this ambiguity is by combining data from ascending orbits (traveling South to North, looking East) and descending orbits (traveling North to South, looking West).
Ascending Pass Descending Pass
(Moves North, Looks East) (Moves South, Looks West)
\ /
\ /
\ /
▼ ▼
[=== Satellite ===] [=== Satellite ===]
\ /
\ /
\ /
▼ ▼
------------------ Ground Surface ------------------
By looking at the same target area from opposite sides, the satellite generates two distinct geometric projections. This provides a system of two equations. To solve this system without a third independent measurement, researchers frequently introduce a physical simplification: they assume that North-South displacement is negligible ($U_N \approx 0$).
With $U_N$ removed from the equation, the system becomes perfectly determined, allowing for the clean separation of the Vertical and East-West components. This elegant geometric trick allows researchers to isolate distinct physical processes:
- Ice Dynamics: Wang et al. (2020) utilized this 2D decomposition to separate the horizontal drift of landfast sea ice from its vertical tidal flexing, overcoming the constraints of single-orbit data.
- Seismic Faulting: Li et al. (2016) mapped the co-seismic deformation of major earthquakes, separating the vertical ground shifts from the massive horizontal crustal movements. Along slow-moving faults, Yao et al. (2024) combined multi-orbit data to isolate horizontal strike-slip from vertical dip-slip components, revealing complex, branching fault structures.
- Structural and Urban Subsidence: In built environments, isolating pure vertical sinking is critical for risk assessment. Mohamadi et al. (2019) and Mohamadi et al. (2020) integrated cross-heading orbits to isolate vertical motion in urban areas, enabling the identification of specific spatiotemporal deformation patterns that directly precede catastrophic building collapses.
Advanced Inversion and Physical Constraints
When combining data from different satellite sensors, wavelengths (such as C-band and L-band), or acquisition times, simply solving a basic linear system is insufficient due to measurement noise and atmospheric interference.
To resolve this, advanced frameworks like the Multidimensional Small-Baseline Subset (MSBAS) method are used. As demonstrated by Qu et al. (2022), MSBAS simultaneously ingests multi-orbit, multi-sensor datasets to solve a regularized least-squares problem, reconstructing highly stable 2D time-series fields that match independent ground-truth leveling data.
Alternatively, if the underlying physical mechanism of the ground movement is well-understood, researchers can swap generic Cartesian coordinates for custom physical vectors:
- Landslide Slopes: Chen et al. (2020) projected multi-orbit LOS data directly along the true downslope direction of a vulnerable mountainside. This allowed them to successfully isolate mining-induced vertical subsidence from actual gravity-driven landslide acceleration.
- Subsurface Mining Profiles: Liu et al. (2024) developed the Three-Dimensional and Full Parameter Inversion (TDFPI) model, which links the mathematical projections of radar LOS directly to physical models of underground coal extraction, resolving large-gradient ground collapses that traditional InSAR methods fail to track.
When the assumption of zero North-South motion is physically invalid—such as during complex tectonic ruptures—external data can be introduced to anchor the inversion. Luo et al. (2016) achieved this by integrating sparse, 3D GPS measurements with dense 1D InSAR datasets within an elastic theory framework, using the GPS points to constrain the missing dimensions across the entire continuous field.
Achieving True 3D Resolution
To resolve a complete 3D displacement vector purely from radar data without relying on external ground sensors or zero-motion assumptions, the imaging geometry itself must be diversified to gain sensitivity along the North-South axis.
One approach is to introduce highly diverse flight headings. The authors of Garcia-Meza et al. (2024) achieved this by combining standard spaceborne ascending and descending passes with airborne SAR flights flown along orthogonal East-West tracks. Feeding these semi-orthogonal geometries into a weighted least-squares inversion algorithm successfully unlocked the true 3D surface velocity field.
Another method involves altering the radar's looking direction. While most imaging satellites are configured to look only to the right of their flight path, utilizing rare left-looking radar configurations provides an entirely independent perspective. As shown by Yang et al. (2021), combining traditional right-looking data with left-looking descending data provides the missing geometric constraints required to robustly model complex, multi-segment fault systems.
Standard Right-Looking SAR Complementary Left-Looking SAR
[ Satellite ] [ Satellite ]
Moves: ▲ Moves: ▲
Looks: ► (Right) Looks: ◄ (Left)
\ /
\ /
▼ ▼
The future of this field lies in dedicated multi-satellite constellations designed to capture these diverse angles simultaneously. As highlighted by Kääb et al. (2024), upcoming bi-static companion missions like Harmony will place passive companion satellites to fly in formation alongside a primary radar emitter (like Sentinel-1). By recording the same radar echo from multiple lines-of-sight at the exact same moment, these missions will natively deliver true 3D surface deformation maps, permanently resolving the one-dimensional ambiguity of spaceborne radar.
Key Foundational Work
A Model for Complex Subsidence Causality Interpretation Based on PS-InSAR Cross-Heading Orbits Analysis (2019) — contributed 1 concept elements to this direction.
Continent-Wide 2-D Co-Seismic Deformation of the 2015 Mw 8.3 Illapel, Chile Earthquake Derived from Sentinel-1A Data: Correction of Azimuth Co-Registration Error (2016) — contributed 1 concept elements to this direction.
Three-Dimensional Surface Displacement Field Associated with the 25 April 2015 Gorkha, Nepal, Earthquake: Solution from Integrated InSAR and GPS Measurements with an Extended SISTEM Approach (2016) — contributed 1 concept elements to this direction.