Monitoring of forest stand condition in Thailand
– Tropical; seasonal forest-
(1) NDVI images of Landsat TM
The Landsat series polar-orbiting satellites with TM (thematic Mapper) sensor and other satellite, such as SPOT and MOS, provide frequent observations of Earth’s surface. These remote sensing data have been received at the ground receiving station since 1986 in Thailand.
The NDVI is the deference of near-infrared and visible red reflectance values normalized over total reflectance of the two channels.
Near_IR (TM. Channel 4)-Red (TM. Channel_3)
Near_IR (TM, Channel 4)+Red(TM. Channel_3)
This equation Produces NDVI values in the ranges of -1.0 to 1.0, where negative values generally represent clouds, water and other non-vegetated surfaces, and positive values represent vegetated surfaces, The NDVI relates to photosynthetic activity of living plants. The higher the NDVI value, the more “green” tge cover type (Deering et al.), That is , the NDVI increases as the quantity of green biomass increases (Burgan and Hartford, 1993).
To interpret the NDVI values for field use, we have devised methods to convert the NDVI data into more easily understandable representations of vegetation greenness as is used for AVHRR (Burgan and Hartford, 19930. These are called “Visual greenness” and “relative greenness”.
Visual greenness (VG) indicates how green each pixel is in relation to a standard reference such as a highly green and densely vegetated evergreen forest. It is calculated as:
NDV = observed NDVI value, V_dens=NDVE value of densely
Vegetated evergreen forest.
An image is produced that portrays vegetation greenness as we would expect to see it if we were flying over the landscape. In this context, normally deforiaged,sparsely vegetated area will look cured compared to fully vegetated area such as the evergreen forest.
Because the visual greenness image may indicate rather limited changes over time, a second measure of vegetation greenness is expresses how green each pixel currently is in relation to the range of greenness observations for that pixel among the data collected. It is calculated as:
RG =( NDV-ND min ) /(Ndmax-ND_min )
NDV=observed NDVI value
ND_min =minimum NDVIvalue observed historically for that Pixel
ND_max=maximum NDVI value observed historically for that pixel
Historical maximum and minimum NDVI maps for the study area are necessary to be produced by searching all the TM NDVI values of collected data and saving the largest and smallest values observed for each pixel. Pixels affected by clouds and noises are excluded. These NDVI values are then composed into maximum and minimum maps and used with current NDVI maps to perform the visual and the relative greenness calculations.
(2) Forest Type Classification and shifting cultivation in chiang Mai
Typical forest types are tropical seasonal forests in chiang Mai province. The highest mountain in Thailand, Doi Intanon, is located at the west side of the study site. Four Landsat TM data of the same dry season, 31 Dec. 1990, 17 FEB. 1991, 5 March 1991 and 21March 1991 (path-Row: 131-47), were selected and geometrically corrected to be overlaid with each other using several ground control points obtained on UTM maps.
The forest type classification was executed using the normalized differential vegetation index (NDVI) images derived from the four Landsat TM data by the following formula.
NVDI =(TM4 -TM3 )/(TM4+TM3)
TM4 and TM3 indicate the channel 3 and the channel 4 of Landsat TM, respectively.
The steps of the analysis are as follows;
- the minimum value is subtracted from each channel data separately to value is subtracted from each channel data separately to reduced atmospheric effect before calculating the NDVI.
- The clustering analysis is applied to make forest cover map based on the NDVIs. The NDVI values are considered to be correspond to terrestrial (forest) condition and vegetation cover types in each image
- After the clustering , the vegetation type of each cluster is checked with existing vegetation maps , aerial photography and ground survey .
- digital elevation data are also collected by digitizing a present map and overland on TM data to analyze the correlation between the classification result and elevation that indicates some environmental condition , especially , such as temperature and water availability to vegetation .
Digital elevation data were extracted from an existing map with the scale on 1:50,000. the contour lines of every 100 meters of the map were input into a GIS system and elevation of each pixel correspond to TM data was generated .
(3) Field check system
Field checks were conducted to compared the results of clustering analysis and to check changes of forest types according to the elevation and shifting cultivation activities and land conditions in forest area.
Especially , the relations between forest types and terrestrial condition was the one of the main concerns in chiang mai study area . We compared the classification maps derived from the clustering method using satellite NDVI images with elevation and slope direction data.
A GPS (Global Positioning System ), a Polaroid camera , a video camera with fisheye lens , a rotator system for the video camera , a video camera with fisheye lens , a rotator system for the video camera and a field note were used for the field survey.