SEVEN YEARS OF MINING SUBSIDENCE DETECTED BY D-InSAR TECHNIQUE 
IN FUSHUN CITY, CHINA 
Y. L. Chen“, X. L. Ding*', C. Huang 1 Z. W. Li , b 
a. Department of Land Surveying & Geo-Informatics, The Hong Kong Polytechnic University, Hong Kong 
b. Shanghai Astronomical Observatory, Chinese Academy of Sciences, Shanghai 200030, China 
c. Graduate School of the Chinese Academy of Sciences, Beijing 100039,China 
ylchen@shao.ac.cn, lsxlding@polyu.edu.hk 
KEY WORDS: D-InSAR ; Two-pass; Mining Subsidence 
ABSTRACT: 
Continuous mining will make the rock mass near the minery lose balance, which will then result in different geological hazards, such 
as landslide, rock movement, earthquake, ground fissures, etc. After more than 100 years’ mining, Fushun city, Chinese mine capital, 
have seriously suffered from these kinds of damages. It is essential to carry out long-term and real-time monitoring of the progress of 
the mining subsidence in Fushun city. Such monitoring will help to prevent the geological hazards on one hand and improve the 
planning and designing of mining activities and systems on the other. D-InSAR (Differential Synthetic Aperture Radar Interferometry) 
provides a new remote sensing tool to monitor land motion with high precision, wide coverage, and under nearly all weather 
condition. This paper uses 6 ERS-1 and ERS-2 SAR images acquired from May 22, 1993 to Jun 13, 2000 and covering Fushun city, 
and the 3 arc second SRTM DEM to conduct two-pass differential SAR interferometry. They can form 15 differential interferograms 
in total, but only 5 with satisfactory baseline and coherence. By investigating the 5 interferometric pairs, we obtain the subsidence 
rate of Fushun city. The case study verifies that D-InSAR technique is able to detect cm-level subsidence of the mining area. 
1 INSTRUCT ON I 
Traditional deformation monitoring system in most mines 
includes GPS surveys, conventional precise levelling and 
theodolite surveys, Electronic Distance Measurement surveys, 
remote electronic monitoring, etc. However these techniques 
are costly, time-consuming, high influence of field and weather 
conditions, and obviously inefficient compared with remote 
sensing technique such as Differential Synthetic Aperture 
Radar Interferometry (D-InSAR), which provides a new 
advanced tool for monitoring the ground movement over large 
areas of the mine and long interval from months to years, 
what’s more, whose deformation accuracy has been up to 3mm 
level theoretically. These obvious advantages make scientists 
investigate the motions of the land surface using D-InSAR 
technique. Timmen (1996), Perski (2001), Ge (2001, 2004), and 
Jarosz (2003) et al. have ever carried out studies on mining 
subsidence monitoring with D-InSAR technique and have got a 
lot of encouraging achievements, while related application have 
been seldom carried on in China . 
In general, there are three main D-InSAR techniques: two-pass; 
three pass; and four pass. This paper employs 2 -pass D-InSAR 
technique with 5 pairs of ERS SAR images and SRTM DEM to 
monitor 7 years’ average subsidence rate. The result shows that 
D-InSAR technique is able to detect cm-level subsidence of the 
mining area. 
2 PRINCIPLE OF D-InSAR TECHNIQUE 
Mathematically, the phase measurements of the repeat-pass 
InSAR, ^ and ^ , can be written as y/ = + n , 
A 
^2=y^2+«2 (1) 
Where R ] and R 2 are the slant ranges between a ground 
resolution cell and the SAR platform; A is the radar 
wavelength; Yl x , Tl 2 are noises generated by different 
scattering features and other noises. Then the interferometric 
phase theoretically is: 
4it 
0 = i//\-y /l =—(R x -R 2 ) + (n l -n 2 ) (2) 
A 
Meanwhile, (f) can be physically divided into two parts as 
follows, which is corresponding to the two components in 
function (2) respectively: 
</> = (fiflat + </>topo +( />def) 
t . . . . (3) 
+ (0ork +( Patm +0noi) 
where (f)ji at , <j) topo , (¡) def , (j) orb , (j) atm , (j) noi represent the phase 
due to earth curvature, topographic phase, orbit 
error,atmospheric delay, and noise respectively. 
The destination of D-InSAR is to extract (f> de j , so we need 
remove the reluctant phase. (J) f]al can be removed by using 
the known baseline and earth model geometry functions; 
0topo can be subtracted by InSAR pairs of short interval or 
external DEM , which is the most essential difference between 
two pass , three pass, and four pass D-InSAR. Two-pass 
D-InSAR adopts the external DEM to obtain (f> topo , but 
three-pass and four- pass use the DEM from interferometric 
SAR image pairs. To reduce the phase error (f) orb due to the