Abstract: [Objective] Monitoring bamboo forest resources and their changes using time series remote sensing has a great significance in achieving scientific management and efficient utilization of bamboo forest resources.[Method] This study extracted the information and analyzed spatiotemporal evolution of bamboo forest based on the Landsat5 TM and Landsat8 OLI time series data in 2000, 2004, 2008, and 2014 throughout Zhejiang Province to reveal the change of bamboo forest area and its response to land use change.Firstly, the land utilization information and spatiotemporal distribution of bamboo forest was extracted by maximum likelihood classification based on time series Landsat data, which was preprocessed by atmospheric correction and geometric correction. Secondly, the spatiotemporal evolution of bamboo forest during 2000-2014 using two indexes (variation amplitude and dynamic degree). Finally, the relationship between spatiotemporal dynamic of bamboo forest and land utilization pattern was analyzed based on the established provincial land use transition matrix.[Result] The result showed that:1) The extracted bamboo forest information based on time series Landsat data achieved a high accuracy, which yield more than 75% of classification accuracy and 91% of user's accuracy, and the extracted bamboo forest area was highly correlated with investigation data, with area accuracy of beyond 96%. 2) The bamboo forest in Zhejiang Province was increased from 2000 to 2014, the variation amplitude and annual change rate were 16.55% and 1.18% respectively. 3) The bamboo forest area accounted for 7.33% in 2000, and increased to 8.56% in 2014, wherein the land use change of coniferous forest, broad-leaved forest and cultivated land have a great contribution to the increased of bamboo forest area, with the contribution rate of 28.62%,37.23% and 16.15%, respectively.[Conclusion] The bamboo forest resources could be accurately monitored based on the Landsat time series data, and the decreasing of coniferous forest, broad-leaved forest and cultivated land has significant influences on the spatiotemporal evolution of bamboo forest.

Key words: bamboo forest, spatiotemporal evolution, Landsat dataset, dynamic degree, land use change