The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008
1344
2. APPROACHES
2.1 Study area and ground truth data collection
There are three provinces in western part of Mongolia. These are
Bayanulgii, KJhovd, Uvs provinces.
where L is the canopy background adjustment that addresses
nonlinear, differential NIR and Red radian transfer through a
canopy, and Cl, C2 are the coefficients of the aerosol resistance
term, which uses the blue band to correct the aerosol influences
of the red band. The coefficients adopted in the EVI algorithm
are, L=l, Cl=6, C2=7.5, and G (gain factor)=2.5 (Kaurivi, Jorry
Z.U., 2003,).
Figure 1 Land cover classification for the study area using EVI
data from MODIS , June 2006
Collection of good ground truth data is a key issue for reliable
land cover mapping. Ground truth data were collected mainly
from field trip to the study area and existing thematic maps from
Joint Russian-Mongolian complex biological expedition Map
Ecosystems of Mongolia, Moscow 2005, Soil Mapping of
Mongolia. Ground observation data from Khovd Agricultural
University was used for the project work in western provinces of
Mongolia. The ground truth collection was held from June 2007 to
September 2007. Totally 57 sample areas chosen for ground truth
collection. Land cover class code is taken from Land Cover
Working group, Asian Association of Remote Sensing 2001).
The land cover classification map was done using Enhanced
Vegetation Index (EVI) data from MODIS , June 2006 (Figure
1).
2.2 Methodology and data
Different land cover/use maps were developed for the west part
Mongolia; land cover, vegetation change, land surface
temperature, snow coverage, forestry, crop and socio
economic information maps using multi spectral data from PAL
NDVI NOAA 8km, SPOT VEGETATION (www.free.vgt.vito.be),
MODIS and LANDSAT (http://glcf.umiacs.umd.edu) images.
For the land cover classification map we used MODIS 250 m
Enhanced Vegetation Index (EVI) data. The Land-use GIS data,
Landsat TM/ETM, were employed as reference data. Enhanced
vegetation index (EVI) was developed to optimize the vegetation
signal with improved sensitivity for high biomass regions and
improved monitoring through de-coupling of the canopy
background signal and reduction in atmospheric influences. The
EVI is represented by the following equation 1 :
Land Surface Temperature:
The MODIS Land Surface Temperature (LST) products are
created through spatial and temporal transformations, to daily,
eight-day and monthly global gridded products. For this research
work we downloaded MODIS Land Surface Temperature Daily
daytime 1 km grid data from the MOD 11 production
(http://edcdaac.usgs.gov/includes/edg_bridge.php). After
downloading MODIS LST data we made mosaic and reprojection
and converted it to C degree units. LST data were applied in the
study area in the time period of 2001 to 2006 from April to
September . As evident in figure 2 LST of study area is
increasing.
EVI = G
NIR-Red
NIR + C, Re d - C 2 Blue + L
(1)
Figure 2. Change of surface temperature using LST data between
the years 2001-2006