Remote sensing of forest LAI from multitemporal optical satellite images over mountain area

Abstract

Leaf area index (LAI) is one of the most important structural parameters in terrestrial ecosystems. In this study, multitemporal Landsat TM images covering experiment station in the northwest mountain areas of Beijing were acquired and in-situ forest LAI was measured. By correlation analysis of three vegetation indexes (NDVI, EVI and TGDVI) and LAI, it is found that the correlation between LAI and NDVI in exponential form behaved a good relativity. This model was applied in mapping multitemporal forest LAI. Further, the sensitivity of the models between vegetation indices and LAI were tested for the broadleaf forest, coniferous and mixed forests, respectively. The results show that the accuracy was improved both in broadleaf and mixed forests compared with the advanced findings which didn't take care of different vegetation types, except for conifer stands with little accuracy decreasing. However, the accuracy of all models has reached a significant level. © 2013 IFIP International Federation for Information Processing.