Limiting factors of
SAR interferometry
Rüdiger Gens
Causes of decorrelation
Limiting factors of InSAR
γ = γ baseline γ SNR γ Doppler γ volume γ temporal γ atmosphere
γ = (1 − δ )
γ = coherence
δ = decorrelat ion
y baseline decorrelation
y thermal noise
y non-overlapping Doppler spectral energy
GEOS 639 – InSAR and its applications (Fall 2006)
2
Causes of decorrelation
Limiting factors of InSAR
γ = γ baseline γ SNR γ Doppler γ volume γ temporal γ atmosphere
γ = (1 − δ )
γ = coherence
δ = decorrelat ion
y volume scattering
y temporal changes
y atmospheric phenomenon
GEOS 639 – InSAR and its applications (Fall 2006)
3
Limiting factors of InSAR
Geometrical decorrelation
y occurs as the separation between the two
orbital trajectories (baseline) increases
y correlation coefficient is inversely proportional
to the perpendicular baseline component
GEOS 639 – InSAR and its applications (Fall 2006)
4
Limiting factors of InSAR
Non-overlapping
Doppler Spectral Energy
Source:
Carande
y Doppler spectral width determined by azimuth beam
width of SAR
y Doppler center frequency determined by attitude,
antenna and earth rotation
y for max coherence, filter out non-overlapping energy
GEOS 639 – InSAR and its applications (Fall 2006)
5
Limiting factors of InSAR
Volume scattering
y scattering from different
(random) heights within
each resolution cell,
such as in volume
scattering, produces
decorrelation
GEOS 639 – InSAR and its applications (Fall 2006)
6
Temporal decorrelation
Limiting factors of InSAR
y change of scattering within a resolution cell or
their electrical properties
x function of time between the first and the second
data acquisition
x caused by vegetation, wind effects, soil moisture,
snow fall, farming activities etc.
GEOS 639 – InSAR and its applications (Fall 2006)
7
Atmospheric perturbation
Limiting factors of InSAR
y ionospheric path delay
x variations in the total electron content along the path
(depends on the time of day and influences the
whole scene homogeneously)
x traveling ionospheric disturbances (cause localized
artifacts)
y tropospheric path delay
x dominant dry part
x small but highly variable wet part which is caused by
the strong temporal and spatial variability of the
water vapor concentration
GEOS 639 – InSAR and its applications (Fall 2006)
8
Limiting factors of InSAR
Tropospheric water vapor
y probably most limiting factor for differential
SAR interferometry
y very localized, heterogeneous effect
x can cause misinterpretation of deformation fields
derived by InSAR in the order of centimeters
GEOS 639 – InSAR and its applications (Fall 2006)
9
Limiting factors of InSAR
Example: the Netherlands
Effect of precipitation
y A: slant delay variation, mapped to zenith-integrated precipitable
water differences
y B: weather radar rain rate; surface wind velocity was 4.1 m/s
Hanssen, R.F., Weckwerth, T.M., Zebker, H.A. and Klees, R., 1999. High-resolution water
vapor mapping from interferometric radar measurements. Science, 283(5406): 1297-1299
GEOS 639 – InSAR and its applications (Fall 2006)
10
Example: the Netherlands
Effect of a cold front
Limiting factors of InSAR
y A: cold front is
visible as the line
between the arrows
y B: weather radar
image
x two weather radar
stations are
indicated by the
yellow circles
x superposed are
surface observations
Hanssen, R.F., Weckwerth, T.M., Zebker, H.A. and Klees, R., 1999. High-resolution water
vapor mapping from interferometric radar measurements. Science, 283(5406): 1297-1299
GEOS 639 – InSAR and its applications (Fall 2006)
11
Limiting factors of InSAR
Orbit errors
y orbital error vector is usually expressed in the
coordinate system rotating with the satellite and
consists of three components
x along-track
x across-track
x radial
GEOS 639 – InSAR and its applications (Fall 2006)
12
Limiting factors of InSAR
Orbit errors:
Filtering approach
y orbital filtering approach can significantly
improve precise orbit estimates of short arcs of
ERS trajectories in an absolute sense
y approach restricted to those orbital passes in
which SAR data were acquired
y limited to determining the across-track and
radial components of the orbital trajectory
x could use timing errors to estimate the along-track
error
Kohlhase, A.O., Feigl, K.L. and Massonnet, D., 2003. Applying differential InSAR to orbital
dynamics: a new approach for estimating ERS trajectories.
Journal Of Geodesy, 77(9): 493-502
GEOS 639 – InSAR and its applications (Fall 2006)
13
Orbit errors:
Filtering approach
Limiting factors of InSAR
y solving the weighted least-squares orbit
determination problem
x
lsq
(
)
−1
= A WA AT Wd
T
δri lsq = ri lsq − ri prior , δri lsq ⊆ x lsq
(
) (
n12 = r2lsq − r1lsq − r2prior − r1prior
= B lsq − B prior ≈ Δ(δr12 )
)
Kohlhase, A.O., Feigl, K.L. and Massonnet, D., 2003. Applying differential InSAR to orbital
dynamics: a new approach for estimating ERS trajectories.
Journal Of Geodesy, 77(9): 493-502
GEOS 639 – InSAR and its applications (Fall 2006)
14