Application and Validation of LWCI (Leaf Water Content Index) to Tropical Seasonal Forest Region
Michio Anazawa*, Genya Saito* and Haruo Sawada**
*National Institute of Agro-Envionmental Sciences
3-1-1 Kannondai, Tsukuba, Ibaraki 305-8604, JAPAN
Tel + 81-298-38-8192 Fax +81-298-38-8199
E-mail: anazawa@maes.affre.go.jp
**Forestry and Products Research Institure,
1 Matsunosatio, Kukisaki-machi, Ibaraki, 305-8687, JAPAN
Keywords: LECI, Tropical Seasonal Forest, Environmental Gradient Analysis, LAI, Precipitation
Abstract: Leaf Water Content Index derived from Landsat TM is found for detecting the leaf water content for tropical seasonal forest in Thailand. Water availability for detecting the leaf water content for tropical seasonal forest in Thailand. Water availability of plant leaves is recognized as the most important key for forming vegetation types in this forest zone, e.g. shedding of the leaves begins from the bottom of the mountain (which is contrary to that of temperate forest) and the gallery forest along water streams keep leaves all the year. Based on this index, water availability map, which shows the seasonal change of foliage moisture within vegetation area was created fro Kanchana Buri study site in western Thailand. The dominant forest types are dry-dipterocarp forests (DDF), mixed deciduous forest and bamboo forest. Seven Landsat TM data were used to estimate the leaf water contents at each time. Then, the time sequential analysis and filtering process (Modal filter) was applied to delineate the terrain on the basis of differences of water availability.
Integrating the DEM data, environmental gradient analysis was conduct to the water availability map for three watersheds, north side, west side and east side. The result shows that not so many deciduous forests are found even in low elevation at north side, deciduous trees are dominant in al elevation at west side and impact of human activities (agriculture) made up quite different vegetation types and land use at east side.
As for Landsat TM data, the correlation between LWCI and ground truth data at the Mae-Klong watershed research center (the meteorological data, etc.) could be analyzed. It was also found to have a tendency of close relation to precipitation and evaporation with the time lag of evaluating terrain condition and understanding the vegetation types which are found in tropical seasonal forest area.
1.Introduction
Seasonal change is one of the most important information for analyzing growing condition of vegetation and the NDVI derived from satellite remote sensing data is generally used for this purpose. The NDVI, however, is influenced by haze and smoke very much. Therefore, an index is required for monitoring tropical seasonal forest area where smokes caused by fires in agriculture land affect very often.
The change of water condition within forest such as the leaf water content according to seasonal alternation (water stress: wet and dry) seems to play important role in phonological dynamics of tropical monsoon forest, especially in deciduous forest and rain-green forest like teak. As for the monitoring of forest phenology and vegetation activities in this area, it is effective to measure the leaf water content by using remote sensing data. It seems possible to monitor the change of leaf water content by applying LWCI Index to TM processed in advance with radiometric rectification.
In this study, improving upon LWCI proposed by Raymond Hunt and et al, we contrived Mlwci(Modified_LWCI). Applying this index to every pixel, the zoning image was created, which shows the regional feature of the water supply condition in dry season.
We intended to develop the leaf water content index (LWCI) into the practical remote sensing index that can catch the water availability of leaves in tropical seasonal forest without receiving the influence of haze such as clouds and fog.
2. Study Area and Data Used
(1) Study area
North East f Kanchana Buri, Thailand.
UTM Coordinates (N 1602996 - N1630496, E73000 -E00500)
(2) Data used
(a) 7scenes of Landsat TM data in dry season from beginning to the end were selected. These data were obtained from NRCT and NASDA.. The following Table 1 shows acquisition data of Landsat TM.
| Area | Sensor | Path Row | Data |
| Kanchana Buri | Landsat TM | 130 - 50 | 1989/12/5,1990/4/17,1993/1/30,1993/4,
1994/4/17,1994/3/6,195/2/21(scenes) |
Table 1 List of acquired Landsat TM data.
(b) DEM data (UTM) of the study site from the geographic map with the scale 1:50,000(derived from 100m contours, pixel size:30m).
(c) Meteorological data observed at Mae-Klong watershed research station (1993-1997).
