Remote sensing of Japanese beech forest decline using an improved Temperature Vegetation Dryness Index (iTVDI)

Abstract

The Tanzawa Mountains, which cover parts of Kanagawa, Yamanashi and Shizuoka prefectures in Japan, are known for their natural beech forests. Since the 1980s, decline of the beech forests, probably caused by air pollution, water stress and insect infestation, has become a serious problem. We estimated the natural beech forest mortality rate in the mountains by using multi-temporal 8-day composite data recorded at the MODIS instrument aboard the Terra satellite, daily air temperature data at meteorological stations (AMeDAS) in 2007, and a global digital elevation model obtained from ASTER aboard the Terra satellite. For the estimation, we used a Normalized Difference Vegetation Index (NDVI) indicating the vegetation density, a Temperature Vegetation Dryness Index (TVDI), and an improved TVDI (iTVDI) indicating the differences in transpiration rates between areas of similar vegetation density. We compared the NDVI, TVDI, and iTVDI maps with an existing mortality map of beech forests in the study area to verify their accuracy. To produce iTVDI maps, we calculated maps of air temperature by using ambient air temperature and elevation data. By interpolation using an environmental lapse rate, we calibrated air temperature maps with good accuracy (RMSE = 0.49 °C). The iTVDI map could detect mortality more accurately than the NDVI and TVDI maps in both spring and summer. Use of iTVDI enabled us to detect forest decline caused by air pollution and water deficits, inducing a reduction in transpiration rates. This index should be useful for monitoring vegetation decline.