266 
In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
There are strong regional climatic differences within the area 
and a diverse range of vegetation and land surface including 
large areas of desert, desert steppe, steppe, forested steppe, and 
forest. The elevation also has large range from more than 
4000m at high mountain areas and Tibetan Plateau to near 0 at 
the east beach. Vegetation degeneration and extreme 
environmental events (e.g. sandstorm) are also undergoing or 
frequently happened (Tsunekawa, et al., 2005; Zhang, et al., 
2008). Its various conditions in middle latitude areas make the 
essential of understanding springtime soil thaw event in this 
area. 
3. MATERIALS AND PREPROCESSING 
3.1. Microwave remote sensing records 
A practical active microwave remote sensing data, QuikSCAT 
Ku-band measurement, was acquired from NASA’s Jet 
Propulsion Laboratory as their level 2A product. First, all 
backscatter measurements at the first 180 days of 2004 were 
extracted and averaged on a daily basis; second, a grid format 
time series with 25x25 km spatial resolution was produced by 
spatially interpolating from the averaged backscatter daily times 
series. And the out-beam record was adopted in this work 
because of its wider swath. 
Brightness temperature data from Special Sensor 
Microwave/Imager (SSM/I), provided by National Snow and 
Ice Data Center, was adopted in this work, as the passive 
microwave remote sensing data. We used the 19 and 37 GHz 
vertically polarized brightness temperature (T BA9V and T Bilv ) 
data from both ascending and descending tracks at the first 180 
days of 2004. This data also have a spatial resolution of 25x25 
km. Because daily SSM/I data did not cover the whole study 
area, we used time series of brightness temperature by 7-day 
maximum/minimum combined with considering about the 
revisit period of SSM/I in our study area. 
3.2. Soil temperature and moisture data 
We obtained soil temperatures and moisture recorded at five 
reference sites used in the Coordinated Energy and Water Cycle 
Observation Project from the Centralized Integrated Data 
Archive (http://www.ceop.net/) and selected three of the five 
stations on the basis of their location and the period covered by 
their data. Soil temperatures and moistures recorded at Shenmu 
in northern China and Bayan-Unjuul in Mongolia were also 
included in our study. We used soil temperatures measured at 5 
cm depth because most microwave radiation emitted at the 
surface emerges from the top layer of soil (Zwally and Gloersen, 
1977). 
3.3. Meteorological records 
The Global Summary of the Day product from the National 
Climate Data Center of United States was applied in our study. 
Daily air temperature, precipitation, and snow depth 
observations were used to support interpretation of springtime 
thaw events from the QuikSCAT backscatter time series and to 
estimate the soil moisture by using a water balance based model. 
4. METHODOLOGY 
4.1. Active microwave remote sensing method 
The method in this work is based on our previous approach 
(Han, et al., 2010b), which studied the typical backscatter 
signatures when springtime soil thaw occurred and suggested 
the method for the thaw event deriving from QuikSCAT time 
series. 
Geographical boundary is firstly detected by Equation 1. 
Slope = 
n ( n Y n \ 
nxŸ j ixS i - X* ÏL Ô i 
i=1 \i=l J\i=\ J 
n ( n V 
wx Z /2 ~ 
(=1 V «=1 ) 
(1) 
Where Slope is the trend of QuikSCAT backscatter with the 
unit dB serving as the indicator for boundary clarification; 
n represents analyzed range in every 5-day step; i is the each 
5-day period, /=1 for the 5-day period from the 1st day of year 
(DOY) to the 5th DOY, i = 2 for the 5-day period from the 6th 
DOY to the 10th DOY, ...; and is the average QuikSCAT 
backscatter of the each 5-day i . Slope larger than 0 identifies 
an area of no thaw event, and Slope less than 0 identifies an 
area in which a thaw event occurred. 
a primary thaw date, which defines as the middle day of a short 
period in which backscatter data are most diverse which is 
deduced as the most significant soil dielectric constant change 
driving by dominated amount soil water’s state changing from 
ice to liquid water, is estimated by equations as shown below. 
PI(i) = Stdev(sigmaO i _ n : sigmaO i+n ) (2) 
Ptd = i, whenPI (/) = max(P/(/)) (3) 
Where PI (/) represents the primary thaw indicator on day i, 
which is a short period’s standard deviation of QuikSCAT 
backscatter (sigmaO)n days (n = 4 was set in this research, 
plus the current day it also means 5-day’s step) before and after 
day i ; Ptd represents the primary thaw date when 
PI(i) reaches its maximum. The primary thaw date is only 
calculated in the areas where soil thaw events happened as 
estimated in Equation 1. 
4.2. Passive microwave remote sensing method 
The soil freeze-thaw algorithm was adopted in this research, 
which requires two parameters: a negative spectral gradient 
between T B 19v and T B37V , as described by equation (4), and a 
threshold T B 37V as described by equation (5): 
dT, 
T, 
B <0 
< T 
B.YIV — 1 37 ’ 
(4) 
(5)