Channel 1 and channel 2 are sensitive to the vegetation, waterbody and clouds. So these 2 channels can reflect burnt area and smoke from the fire.

Using the image processing system IVAS, IDRISI for Windows, PCI we have created methodology to detect the fire sources and burnt area, and estimated the threshold values, as follows:

I. Active fire: a) CH3³ ³ 45^oC
b) CH1 (or CH2) =6 - 12 II. Burnt area: a) CH3 > 35 - 45^oC
b) CH1 (or CH2) = 3-6 Channel 4 or 5 are used for the cloud masking

By above mentioned methodology using the daily NOAA satellite data we can monitor and estimate total burnt area, as showed on Figures 1-6.

Figure 2-5 are showing the total burnt area of forest and steppe fire over territory of Mongolia, of Spring 1996-1999 respectively. Total burnt area calculated as 10778.3 and 12448.2, 3981.2 and 3104.5 thousand hectares of forest and steppe respectively.

In figure 6 have been illustrated the trends of steppe fire over Dashbalbar sum (Sun is smaller unit of aimag) of Dornod aimag (North Eastern part of Mongolia).


Figure 1: Steppe fire of Sumber and Khalkh gol sums Dornod aimag.


Figure 2: Fire map, Spring 1996


Figure 3: Fire mao, Spring 1997


Figure 4: Fire map, Spring 1998


Figure 5: Fire map, Spring 1999


Figure 6. Steppe fire monitoring (Dashbalbar country of Dornod aimag)

3.2 Fire estimation by JERS data
Landsat, SPOT, JERS high resolution satellite data are very useful tools to estimate the burnt area and assess the fire impact to the ecology. But timely we can not buy thiskind of data, because of their high price. In 1996 NADSA of Japan had provide us by JERS satellite data for the fire estimation.

For study purpose we have used the JERS composed data from 21-24 May, 1996. The JERS data has 18 m of ground resolution which is suitable to calculate the burnt area and to integrate with GIS data.

However we can monitor the forest and steppe fire status by daily NOAA data as showed above. For the accurate assessment of the burnt area, of small area, as presented by curcle with letter A in Figure 2, of Tsagaan-Uur, Tarialan and Teshig sums of Kubsugul and Bulgan aimags, we have processed JERS composed data of channels 1,3,2 (Figure 7).