130 
THE DETECTION OF SUBSIDENCE AT PERMANENT FROZEN AREA IN QINGHAI- 
TIBETAN PLATEAU 
Z. Li, C. Xie, Q. Chen 
State Key Laboratory of Remote Sensing Science, Institute of Remote Sensing Application, Chinese Academy of 
Sciences, P.O.Box 9718, Datun 3, Beijing, 100101, China, zli@irsa.ac.cn 
KEY WORDS: Interferometry, Detection, Subsidence, Permanent frozen area, Qinhai-Tibetan Plateau, Railway 
ABSTRACT: 
The surface displacement by seasonally freezing bulge and thawing subsidence are main hazards for engineering construction in 
permafrost regions, especially for the Qinghai-Tibet railway. For detecting the distortion at permafrost area, we try to study the 
interferometric method of monitoring the deformation at permafrost area with time-series EnviSat ASAR data. In this paper, the 
coherence characteristics are analyzed for different baseline, time interval with or without season change and different classes (rock, 
bare soil, vegetation, water), used 13 time-series ASAR data from Jan. 2004 to June 2006 at the Beiluhe test area in the Qinghai- 
Tibetan Plateau. The results showed that the coherence coefficients is lower in summer and fall than in other two seasons as the 
freezing and thawing phenomenon. For pairs of crossed different season, such as from spring to summer or from summer to fall, the 
coherence coefficient decrease for bare soil and vegetation cover, little decrease for rock cover. The deformation of railway roadbed, 
detected by Permanent Scatters Interferometry, mainly appears in the way of thawing subsidence from May to November every year 
with 25 mm maximum subsidence in the two years. 
1. INTRODUCTION 
The permafrost and the seasonal frozen ground cover 
respectively 1,272,709 km2 and l,146,399km2 in the Qinghai- 
Tibetan Plateau, and the highway and railway that passes 
through permafrost areas about 550km. Extensive areas of 
frozen ground hold large quantities of ice. The seasonally 
thawing layer is highly sensitive to temperature changes, and 
thawing and temperature rising has a great influence on railway 
stability. One of the main problems is how to monitor the 
frozen ground’s displacement (Ma , 2006). 
The traditional methods to monitor deformation of plateau 
frozen ground is burying settlement meter, such as inclinometer, 
or constructing time-serial GPS observation station. Using these 
methods, the acquired information of plateau frozen ground’s 
deformation can only be gotten at some limited locations due to 
the limitation of observation condition, especially in the Tibet 
Plateau, although which is with high precision and continuity. 
Differential Synthetic Aperture Radar Interferometry (DInSAR) 
has also been widely used in recent years for monitoring 
ground’s deformation (Li, 2004). DInSAR analysis the phase 
information of radar echo signal, and extract dense deformation 
information in a relative large spatial domain. But temporal and 
geometrical decorrelation often prevents traditional SAR 
interferometry from being an operational tool for surface 
deformation monitoring in the area of frozen ground. To get 
long term deformation information, scientists presented the 
methods of Permanent Scatters (PS) and Small Baseline Subset 
(SBAS), and have successfully used them for monitoring 
subsidence in urban area (Ferretti, 2000; Berardino, 2002). But 
in the area of frozen ground, ground surface changes very 
quickly with time, so the coherence is relative lower in the area 
of frozen ground than in the area of urban. To use the method of 
PS or SBAS to analyze surface subsidence history in the area of 
frozen ground, characteristics of coherence at permanent frozen 
area need to be firstly analyzed. With the analysis of the 
characteristics of coherence, it can be recognized what’s the 
factors that influent object’s coherence, and also be made clear 
when the factors work and what degree they influent the 
object’s coherence. Based on the analysis of coherence 
characteristics, we can determine the SAR data choice for 
detecting the surface deformation and the choice of stable 
scatters. 
2. ACTIVE CHARACTERISTICS OF FROZEN 
GROUND IN THE QINHAI-TIBETAN PLATEAU 
There are a lot of factors that influence freezing and thawing of 
the frozen ground, e.g. season, vegetation, topography, snow 
cover, water, lithology, and moisture. In the zone of meadow in 
the Qinghai-Tibetan plateau, vegetation reduces surface 
temperature and influent thawing and freezing of frozen ground. 
The influence of topography to the frozen ground is mainly 
dues to elevation, gradient and direction of slopes. 
The active process in active layer in the permafrost region of 
the Qinghai-Tibetan Plateau is composed of main four sections 
(Zhao, 2000): 
(1) The process of active layer’s thawing in summer begins at 
the end of April and completes at the middle of 
September. During the process, the frozen ground thaws 
downwards from the surface until reaching the maximum 
depth. 
(2) The process of freezing begins at the middle of September 
and the frozen ground freeze slowly upwards from the 
bottom. From the middle of September to the middle of 
October, the frozen ground freezes from two directions 
(i.e. downwards and upwards). At the end of October, the 
process of freezing ends. 
(3) When the process of freezing completes thoroughly, the