The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Voi. XXXVII. Part B7. Beijing 2008 
88 
atmospheric variations at pixel (x,r) , and the last term 
(x,r) for temporal decorrelation, orbital errors and 
thermal noise effects. 
In Equation (1), it is clear that the smaller the perpendicular 
baseline, the less the impact of DEM errors on a deformation 
map. Assuming a nominal incidence angle of 23° and a 
perpendicular baseline of 100 m, the typical topographic errors 
in the SRTM DEM (c. 8.7m in Eurasia (Farr et al., 2007)) might 
lead to a phase error of 0.61 rad, which is well below the typical 
phase noise level of the InSAR pairs, on the order of -0.70 rad 
(Hanssen, 2001). Therefore, the topographic contribution can be 
considered negligible for InSAR pairs with a perpendicular 
baseline of 100 m or shorter. 
Atmospheric effects were first identified in repeat-pass InSAR 
measurements by (Massonnet et al., 1994) when they studied 
the 1992 Landers earthquake. Zebker et al. (1997) suggested 
that a 20% spatial or temporal change in relative humidity could 
result in a 10-14 cm error in deformation measurement 
retrievals, independent of baseline parameter. It should be noted 
that the atmospheric phase component in Equation (1) also 
includes ionospheric effects (Hanssen, 2001). However, 
because the occurrence of phase scintillation due to small-scale 
ionospheric disturbances is limited in the equatorial and auroral 
regions, and extremely rare in mid-latitude regions; and 
ionospheric effects on InSAR measurements should be -17 
times less at C-band than at L-band, an assumption is made in 
this paper that ionospheric effects will not significantly affect 
phase variations in C-band SAR images, although they may 
lead to long wavelength gradients which can be removed using 
a best-fit plane (Massonnet and Feigl, 1998). 3 
3. MERIS WATER VAPOUR CORRECTION MODEL 
FOR A SINGLE INTERFEROGRAM 
3.1 MERIS near IR water vapour product 
The MEdium Resolution Imaging Spectrometer (MERIS) was 
launched together with the Advanced Synthetic Aperture Radar 
(ASAR) on the ESA ENVISAT spacecraft on 1 March 2002. 
MERIS is a push-broom passive imaging instrument and 
measures the solar radiation reflected from the Earth’s surface 
and clouds in the visible and near IR spectral range during the 
daytime with a swath width of 1150 km for the entire 68.5° 
field of view (ESA, 2004). MERIS uses two out of fifteen 
narrow spectral channels in the near IR for the remote sensing 
of Precipitable Water Vapour (PWV) either above land or ocean 
surfaces under cloud free conditions (Bennartz and Fischer, 
2001) or above the highest cloud level under cloudy conditions 
(Albert et al., 2001). The wavelengths of these near IR channels 
have been chosen by the MERIS science team to minimize the 
effects of different spectral slopes of the surface. The two 
channels (i.e. channel 14: 885 nm and channel 15: 900 nm) are 
only 15 nm apart and spectral variations of surface reflectance 
are generally small (Bennartz and Fischer, 2001). In order to 
further minimize the spectral surface effect, a simple correction 
algorithm based on the ratio between MERIS channel 10 
(753.75 nm) and channel 14 (885 nm) has been implemented in 
the ESA MERIS water vapour retrieval algorithm (Fischer and 
Bennartz, 1997). MERIS near IR water vapour products are 
available at two nominal spatial resolutions: 0.3 km for full 
resolution (FR) mode, and 1.2 km for reduced resolution (RR) 
mode. 
Only measurements with wavelengths between 0.4 pm and 0.9 
pm are used in the MERIS cloud mask algorithm to estimate 
whether a pixel is cloudy or not, and very valuable thermal 
information and information on liquid and ice water absorption 
at 1.6pm and 3pm are not available, therefore the MERIS cloud 
mask is not ‘perfect’, particularly in mountainous areas. A 
‘conservative’ cloud mask using the relationship between 
surface pressure and topography can detect and mask out thin 
clouds to obtain better water vapour than those obtained using 
the ESA standard product (Li et al., 2006a; Ramon et al., 2003). 
Spatio-temporal comparisons show c. 1.1 mm agreement 
between MERIS and GPS/radiosonde water vapour products in 
terms of standard deviations (Li et al., 2006b). 
3.2 MERIS water vapour correction 
When MERIS water vapour data collected at times t s and t M 
are available, Zenith Path Delay Difference Maps (ZPDDM) 
can be derived as follows: In step 1, MERIS-PWV needs to be 
converted into Zenith Wet Delay (ZWD) using surface 
temperature measurements, which were obtained from 
radiosonde data in this study; In step 2, a ZPDDM is calculated 
by differencing two 2D ZWD fields (i.e. 
ZPDDM = ZWD 2 - ZWD, ); In step 3, an improved inverse 
distance weighted interpolation method (IIDW) (Li, 2004) can 
be applied to fill in the missing pixels due to the presence of 
clouds and/or problematic reflectance values; In step 4, to 
suppress the inherent noise of MERIS-PWV, a low-pass filter 
can be applied to the ZPDDM with an average width of c. 0.6 
km for FR MERIS data and c. 2.4 km for RR MERIS data. 
Assuming pixel to pixel PWV errors are uncorrelated, the 
accuracy of the ZPDDM increases by a factor of 2 at the 
Figure 1. Two-pass differential InSAR processing flowchart 
with MERIS water vapour correction.