The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
152 
Some conventional implemented stacking methods showed that 
the derived data could obtain the acceptable accuracies (Zebker 
et al., 1997) but, the results are not ideal in most cases (Li et al., 
2005). On the other hand, an extensive parametric model is 
needed to achieve higher accuracies due to the various 
atmospheric effects on the InSAR data. Furthermore, the model 
should be flexible enough to be extended for special cases. 
In this paper, some calibration methods will be considered in 
order to reduce the errors in some scenes acquired in September 
and October of 2005 from Mashhad in North East of Iran which 
is a semi mountainous area. Therefore, model estimations and 
data acquiring processes were determined to sustain the 
climate’s requirements. 
In many cases, using a dense GPS network could be simple, 
accurate and appropriate for atmospheric modeling (Li et. al, 
2005). These kinds of networks are not available in the existing 
area of interest so, a calibrated spacebom water vapor product 
was chosen as the next suitable choice due to its extensive 
coverage. Moreover, due to the smooth changes of atmospheric 
parameters, the resolution of the optical water vapor products is 
suitable either. Since we have used Advanced Synthetic 
Aperture Radar (ASAR) data for interferometry purpose, 
MERIS seemed to be an appropriate data source due to the 
exact similarity of the acquisition times of MERIS and ASAR. 
As water vapor products which derived from optical Spacebom 
sensors are significantly sensitive to the clouds (Li et., al., 2005; 
Hanssen, 2001), a cloud extraction algorithm was issued and an 
interpolation method was utilized to fill the empty pixels of the 
product. To recover these values, it is needed to have some 
information about the geometrical depth and the liquid water 
content of the clouds (Hanssen, 2001). As such data are difficult 
to obtain, a simple method was used to mask or repair data. 
MERIS Air pressure is another data source which could be 
helpful for the atmospheric correction of InSAR data. This layer 
is appropriate for different weather condition of two acquisition 
times. In this study, due to the minute changes of air pressure 
gradient, synoptic data was used due to its better accuracy. 
Ionospheric total electron content was the other critical 
parameter which influences the Radar single ray (Hanssen, 
2001). As the changes of total electron content in the narrow 
area like the interested region of fine mode images of ASAR, is 
spatially mitigating, the difference error map of Ionosphere 
seems to be planar. This error plane is difficult to separate from 
the baseline effect of the interferogram (Hanssen, 2001). Hence, 
by applying the flattening strategy after the error reduction 
process, this error and some residuals of other errors may be 
reduced. 
In this paper and in the next section, the characteristics of 
utilized MERIS images will be introduced. Then the details of 
error removal strategies will be considered and in the other 
session, the results and validation methods will be noted. A 
short discussion will be done at the end of this paper. 
2. MERIS 
MERIS (MEdium Resolution Imaging Spectrometer) is a useful 
optical sensor of European ENVISAT for ocean color and 
atmospheric studies (Kramer, 2002). The images of this sensor 
consist of 15 spectral bands in visible and near infrared regions 
of electromagnetic waves. MERIS image is an appropriate tool 
for atmosphere monitoring and extraction of atmospheric 
parameters (ESA, 2006). One of the most important capabality 
of these images is the column water vapor estimation (ESA, 
2006). 
To estimate the total précipitable water vapor content of the 
atmosphere in the earth-sensor direction, a quadratic model 
between the band 14 and 15 could be used. These two bands are 
suitable for water vapor estimation because one of them is a 
water absorption channel and the other one is a non absorbing 
band (Fischer and Bennartz, 1998). Moreover, closeness of 
these two bands in the spectral pattern will result in the small 
difference between the surface albedo in two channels. 
The column water vapor content is calculated in level 2 data of 
MERIS. In this research the MERIS Reduced resolution product 
was chosen to form the error maps (ESA, 2006). 
These dataset consists of extracted cloud features and column 
water vapor. The nominal spatial resolution of data is 1200 
meter and the accuracy of water vapor amount is 20% (ESA, 
2006). 
3. ERROR REDUCTOION 
An interferogram was formed from SLC images of ASAR. Two 
single phase images were acquired in May 30, 2005 and August 
08, 2005 from north east of Iran. Forming method and noise 
reduction strategies result in an interferogram with pixel spacing 
of 90 meters. 
5 main signals were taken into account in this research (Hanssen, 
2001). 
1- Topographic term 
2- Deformation signal 
3- Noise 
4- Baseline phase ramp 
5- Atmospheric term 
In this research, Topographic term considered as the additional 
signal and as the first step, topographic effect was removed by 
using DEMs. 
Deformation is the required signal and all the considered 
strategies attempt to maintain this signal. 
Noise reduction strategies were utilized in forming the 
interferogram and the remained noise was neglected. 
Two remained terms was considered simultaneously. As the 
baseline error and atmospheric effect is hard to separate 
(Hanssen, 2001), a plane fit algorithm was chosen to flatten the 
interferogram. This algorithm reduced the remained 
atmospheric errors and baseline phase ramp at once. 
As some parameters like water vapor have a fixed term in its 
total error, flattening before error reduction will cause the 
removal of this fixed term. Implementing the error map after 
flattening may reduce the influences more than theirs real 
amount. 
As a result, the influences divided into two categories. First 
category includes the nonlinear effects like water vapor and 
second one consists of linear effects. Depends on the weather 
condition, air pressure and Ionospheric electron content may 
categorize in first or second group.