Abstract: It is important to estimate forest coverage for forest management and conservation in a certain region. However, forests in
China are mainly distributed in remote and mountainous regions. It is costly to investigate forest coverage due to the complex
topography. In this study, we developed a cost-effective approach to estimating forest coverage based on free high-resolution images in
Google Earth and applied this approach in Yunnan Province, which is a mountainous region. According to the interpretation of 42 880
sample plots based on Google Earth, from 2013 to 2015, forest coverage rate is 51.3%（95% CI：50.8%-51.8%）in Yunnan, and dense
forest coverage rate is 27.3%（95% CI：26.8%-27.8%）. Through the analysis of 1 000 times of return samplings under different sample
sizes, the relationship between the estimation accuracy of forest coverage and the sample size is a power function . The precision of
estimated forest cover rate showed a power law dependence on sample size. If the sample size is higher than 20 000, the precision
would be improved to 1%. As the results indicate, it is feasible and cost-efficient to measure forest coverage rate in mountainous region
based on visual interpretation of sample plots on Google Earth.

Key words: Google Earth, remote sensing, visual interpretation, forest coverage, sampling