Abstract
It has proved difficult to monitor the state of plantations by field observations due to the high costs involved, although the extent of unsuccessful plantations has been increasing. The objective of this study was to develop a method to estimate the survival rates of planted trees in young Sugi (Cryptomeria japonica) and Hinoki (Chamaecyparis obtusa) plantations by analyzing LANDSAT TM (Thematic Mapper) data. Our hypothesis was that the survival rate of planted trees, which was regarded as closely correlated with canopy coverage of planted trees, could be represented by the digital number of bands or intensity of some index of remote sensing data. The study sites were young plantations (4 to 12 years old) in Kyushu University Forests in Miyazaki which is located in southern Kyushu, Japan (32°22′10″N, 131°10′40″E), within the cool-temperate deciduous zone. Plantations were divided into aspect classes as the minimum units of this study. The method had two main parts. 1) Interpreting survival rates of planted trees on a pair of aerial photographs. 2) Analyzing the relationships between TM data and the survival rates of trees under two models, MODEL I using NDVI (Normalized Difference Vegetation Index) and MODEL II using the TM band 3 digital number (DN) for detecting the difference in spectra between planted trees and grass. The results showed that: 1) there were big differences in the survival rates according to aspect, and northern slopes had relatively high survival rates compared with southern slopes. 2) It was difficult to estimate the survival rates of planted trees by NDVI. 3) The relationship between survival rates and DN in TM band 3 showed strong negative correlation with a coefficient of determination of 0.643. Thus, DN in TM band 3 appears to be preferable for estimating the survival rates of planted trees. This study thus indicated that TM band 3 approach would be the most effective monitoring method for plantations in inaccessible mountainous regions.
