Geomorphologic Distribution of Normalized Difference Vegetation Index
Chi-Chung Lau
Water & Ocean Resource Div.,
Energy & Resource Lab., Itri
Bldg. 24, 195-6 Sec. 4, Chung Hsing Rd.,
Chiutung, Hsinchu, Taiwan 310, Republic of China
Tel: (886)-3-591-5468 Fax : (886)-3-582-0038
E-mail :f820027@erlb.erl.itri.org.tw
Abstract
Landsat/TM derived normalized difference vegetation index (NDVI) of for watersheds in Taiwan are used to studied their geomorphologic distribution. Raw data are corrected by Minnaert model to mitigate topographic effect. Results show that slope azimuth (aspect) is major factor affect distribution of NDVI. Inclined surface can be divided to illuminate and non-illuminate sides. The relatively low altitude area in watershed shows low NDVI values, and NDVI rapidly increases to 0.3-0.5 with elevation until 3000 meters. NDVI will decrees with elevation where above 3000 meters. Distributed of NDVI in where slope angle less 60 degree show a curve with a peak value occurring in slopes between 30-40 degrees.
Introduction
Topographic factors affect the vegetation pattern in different ways. Steep slope traps thin layer of soil which hold small amount of moisture, then resulting a low biomass. Aspects in illuminating orientation receive more solar radiometric energy than the aspects in shadowy side. Succeed results is abundant plants growing in there. Considering elevation, there is a common way to divide forest using altitude because it is a natural way for separating climate zone. Lin (1996, personal communication) found that the development of the plan community in Fusan experiment forest is significantly correlated with the altitude, topography, aspect and aspect-topography synthetic index.
Normalized different vegetation index (NDVI) derived by satellite remotely sensed data is a good data source to study the spatial distribution of biomass. The index comes from ratio of near-infrared (NIR) and red bands (RED):
NDVI = (NIR-RED) / (NIR+RED)
This is the most commonly used vegetation index as it retains the ability to minimized topographic effects while producing a linear measurement scale. This property allows us a tool to see the detail of biomass distribution without topographic influence.
Test Sites
Four watersheds in different climate zones were selected to the area without many human activity (Fig.1). The selection was corresponded with a previous project which concern in hydrologic impact due to climate change. 1. Fusan basin located in northern Taiwan. Elevation ranges between 360m to 2130m. Climate zone being to GCWA and GCWB, warm-dry winter and humid summer. The warmest month has a lowest temperature of 22°C. Major plants are deciduous and bamboo, which can be divided into natural forest, secondary forest and grassland. 2. Chichiawan watershed is in alps of middle Taiwan with elevation from 1590m to 3890m. Major forest type is conifer and mixed-type forest. Climate zones are GCWB and GCWC, temperature can be lower than 10C in a short-cold summer. 3. Sandiman basin is in southern Taiwan, south of tropics, with elevation from 110m to 3060m. The climate zone belong to AW, GCWA, and GCWB, humid-warm weather. Rainfall occurred in summer and winter is dry. Major vegetation class is deciduous, conifers are appeared in high altitude area. 4. Lisan basin is in the steep mountain area in eastern Taiwan. Unlike other basin, orientation of major aspect is southeast-east direction. Elevation ranges from 200m to 3450m. Due to the steep terrain, climate zones are belong to AF, GCWA, GCWB, and GCWC, which change from a tropical area to cold-alps area.
Figure 1: Study areas
Data Processes
Landsat 5/TM images acquired on January 9, 1995 at 9:.33 morning were used. Solar zenith angle of the scene was 58.36 and azimuth was 140.36. Original 30-meter-resolution data were rectified and resampled into 25 meter resolution, and were registered to Transverse Mercator (WGS84) projection. Whole data set is a 6000 x 8000-pixel matrix. Four watershed masks derived from Digital Terrain Model (DTM) is applied to extract data.
DTM of entire Taiwan island was produced by author in previous study which was developed by following steps: 1. An incomplete DTM data was collected from Taiwan Province Photogrammetric Service. 2. In the area lack of data, contour lines and labeling points were digitized every 50 meter room 1:5,0000, or 1:10,000 in mountain area, topographic maps. Vector data were converted to raster data by Kriging method using ARC/INFO software.
Once the DTM were prepared, topographic structure was extracted by a flow accumulation approach (Lau, 1993) that 1. created a pit-free DTM, 2. estimate flow direction, 3. calculate flow accumulation, and 4. define stream lines and basin mask. With the basin mask, only the pixels inside study basin were put in the analysis.
