The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008
424
Figure 3-1 Chongqing NDVI on August 7 th , 2006
Figure 3-2 Chongqing soil moisture on August 7 th , 2006
Figure 3-3 Chongqing soil moisture on August 7th, 2006
Figure 3-4 Meteorological Observation of Chongqing Drought
in Summer of 2006 (July 21 st —August 10 th )
Figure 3-4 shows the drought distribution observed by
Chongqing Municipality Climate Centre.
Comparing the monitoring results from Chongqing Climate
Centre with remote sensing retrieval results, we find that they
are basically consistent. However, remote sensing retrieval
result is more objective duo to its regional characteristic. It can
be seen that political borders of counties are obvious in that of
that meteorological observation data, because the
meteorological observations are point data. Therefore,
quantitative remote sensing retrieval soil moisture is more
macro, objective, real-time in monitoring the actual local
drought status.
4. FOREST FIRE MONITORING IN CHONGQING
4,1 Research Methods
According to Wilhelm Wien's Displacement Law (Price J C,
1983), MODIS bands 20, 21, 22 and 23 are designed for
monitoring forest fire, which temperature is about 600 K and
atmospheric window (8-14/jm) bands 31 and 32 are used to
retrieval land surface temperature which temperature is about
300 K. For the forest fire with smaller size, especially when
forest fire area is smaller than that of a pixel, band 31 and band
32 have important significance for the fire points monitoring.
Furthermore, Channel 7 of MODIS has 500m spatial resolution.
Our experiment showed that it is sensitive to high temperature,
especially 600 ~ 1000K, though this channel has often been
neglected in monitoring forest fires. We believe band 7 will
greatly helpful for forest fire identification.
Figure 4-1 and Figure 4-2 are brightness temperature of band 21
and band 31 respectively.
Figure 4-3 Brightness temperature of band 21 on August 11 th
Figure 4-4 Brightness temperature of band 31 on August 11 th
4.2 The Process and Results
4.2.1 Direct Method for Detecting High Temperature
Fire Points
We use band 7 to detect high temperature fire points. Figure 4-3
is the composite image of MODIS bands 7, 4 and 3. Figure 4-4
is the enlarged one. The arrows point to two fire points. The left
one is a high temperature open fire point. The right one is
suspected a hidden fire point. According to reality information,
at 12:00 on August 11 th , Bishan county, Aokang Industrial Park,
Tiger hill occurred forest fire (fire should be 1). At 11:30 on
August 11 th , Shapingba, Zengjia county occurred forest fire
(fire should be 2). And this image’s acquisition time was
August 11 th , MODIS (TERRA) crossing at local time 10:30, it
was the very likely reason why lire point 2 has no smoke plume.
In addition, some fires can also be found in Banan.
Figure 4-3 RGB composite image of bands 7, 4, 3