STUDY ON RELATION BETWEEN INSAR COHERENCE AND SOIL MOISTURE 
Ting Zhang 3 ’ b , Qiming Zeng a ’ b , Ying Li a ’ b , Yun Xiang 3 ’ b 
institute of RS and GIS, Peking University, Beijing, P.R. China, 100871 
b Key Lab of Spatial Information Integration and Engineering Application of Beijing 
qmzeng@,pku.edu.cn zhangting4012@163.com. 
KEY WORDS: Coherence; Soil moisture; InSAR Interferometry; De-correlation; 
ABSTRACT: 
Coherence is an important parameter to measure InSAR data quality in repeat-pass interferometry and soil moisture change can 
cause de-correlation. Relation between coherence and soil moisture was examined in this paper by using the ERS-1 interferometry 
pairs on Cangzhou area for which there are some heavy rain between the acquisition dates of the two images. The soil moisture is 
obtained based on the distributed hydrological model and the coherence is calculated by reducing the other de-correlation factors as 
much as possible. Correlation and regression analysis are carried out on the two images. It is found that the relation between the two 
may satisfy exponential distribution and is different with the area. Different soil types and land use appear different patterns. 
1. INTRODUCTION 
Interferometrical SAR (InSAR) as a powerful mean for 
topographical mapping, measurement of subtle earth surface 
shift, was chosen to monitor the ground surface subsidence in 
CangZhou, southeast of HeiBei province. However, due to the 
decorrelation between the interferometric pairs, we couldn’t get 
the accurate subsidence measurement of the study area. This 
made us quite puzzle. First, the interferometric pairs were 
acquired on Apr 1995 and May 1995 with about one month 
interval over CangZhou area; then the perpendicular baseline of 
these two images was very small, and the precise orbits data 
were also used to reduce the co-registration errors during 
InSAR processing. In addition, the ground surface subsidence 
in Cangzhou area is quite serious. So all these condition should 
be quite good for Interferometry, but why we got an unqualified 
result. For further research, we find that during the period, there 
was a heavy rainfall in the research area, which might cause the 
soil moisture changed. If this is true, the soil moisture could 
change the surface scattering rate which will promote temporal 
de-correlation and cause an adverse effect on interferometric 
coherence. At present, the specific influence relation is seldom 
studied at home and abroad, so, we carry out the study on the 
relation between InSAR coherence and soil moisture. 
This paper shows the method of calculating the soil moisture 
and obtaining the coherence and the comparison and analysis 
results with the two, and also presents an approach to reduce the 
effect that soil moisture makes on coherence. 2 
2. STUDY AREAS AND DATA 
Cangzhou is situated in the east of North China Plain and is one 
of the areas which is short of water most seriously around the 
Bohai area. The climate is usually dry in Cangzhou area, but 
during our study periods, heavy rain happened for several times. 
The study area covered by the SAR images is located on the 
quadrate area with longitude from 116.143°E to 117.474°E and 
latitude from 37.715°N to 38.831°N. 
The data used in the model contain SRTM elevation data 
provided by NASA, daily meteorological data (rainfall, 
temperature, wind speed, vapor pressure, radiation of sunshine, 
etc) from April to middle of June in 1995 provided by Hebei 
Meteorological Bureau, soil map and land use map. The rainfall 
data of nine meteorological stations are used and the location of 
these stations is shown with green dots in figure 4. The rainfall 
distribution can be seen from figure 1 and the numbers in the 
legend represents Station Number. Besides, two ESA-1 images 
with approximate pixel resolution of 5mx 15m were acquired on 
Apr. 6, 1995 and Jun. 15, 1995 over Cangzhou areas to obtain 
coherence. The perpendicular is 122.53m, which is short and 
very suitable for interferometry. 
Apr. 6. 1995 Jun. 15, 1995 
Figure 1 The rainfall distribution between Apr. 6, 1995 and Jun. 
15, 1995.for different stations. 
3. METHODS 
For the study on the relation between coherence and soil 
moisture, we need coherence and soil moisture data. Coherence 
can be derived in the process of InSAR interferometry from the 
image pair. Because very precise measurement data of soil 
moisture can not be acquired directly, we calculate it through 
meteorological data by building hydrologic model. Our work is 
based on the raster data model of GIS because the distributed 
hydrologic model is popular at present and applied widely. The 
model is built according to water balance principle. All point 
data, such as rainfall, is interpolated to surface raster maps. 
This work has been sponsored by National Science Foundation of China (40571098) and National High Technology 
Program (2006AA12Z150).