The Characteristics of Vegetation Indices Using Landsat Data of Land Cover Species

Abstract

　This paper presents a reflectance-based method to quantify the characteristics of vegetation cover from TM and ETM+ data more accurately. The study was conducted on the vegetation cover of a Kake-cho area, which is situated on the northern side of Hiroshima City in Hiroshima Prefecture. The normalized difference vegetation index (NDVI), one of the most popular vegetation indices, was used to assess and evaluate the vegetation indices of four species.
　The vegetation indices appear different because the image acquisition times were different in every year and the different stages of green leaf density of the plants had different heights. The lower values of vegetation indices were showed in the years 1985, 1987 and 1996 due to the acquisitions of vegetation images at the early stage of green leaf density and the lowest values came into view in 2003. The rate of change (1985 to 1987) of vegetation index was higher for “Q” species because early growing of leaves enhanced higher density of the plant canopy. The vegetation index was relatively less in the year 1990 and the lowest index was found in 2003 for Q species due to its lack of vigor as well as being a deciduous species. It was discovered that the values of vegetation indices are higher for taller plants in the cases of Q, Ca and R species. The rate of change of vegetation index was higher (1990 to 2002) for Ca species compared to other species, and the yearly changes were nearly uniform in both R and C species. It was revealed from the study that NDVI is highly correlated with green leaf density and may be used as a representative for the status of ground surface biomass.