报告题目：Time Series Analysis of Landsat Data for Continuous Monitoring of Land Cover Change and Condition 

报告人：Curtis Woodcock教授 波士顿大学地理系系主任 

Dr. Curtis Woodcock is currently a Professor and the Chair of the Department of Earth and Environment at Boston University. He obtained his PhD from the University of California at Santa Barbara in 1985, and has been on the faculty at Boston University since 1984. Prof. Woodcock has been an international leader in environmental monitoring with remote sensing. He has done remote sensing research for over 35 years on a wide range of topics, including urban expansion in Pearl River Delta, China, land use change in the Nile River Delta, Egypt, agricultural expansion in Turkey, Forest Ecosystem Dynamics in the Pacific Northwest, USA, carbon budgets in New England and the Black Sea region, and global land-cover/land-use changes. He maintains an interest in the theory and methods of remote sensing, particularly as they relate to the spatial and temporal dimensions. He received the Outstanding Contribution Award for Remote Sensing from the Association of American Geographers in 2010, numerous Best Science awards for his scientific papers.   His current research interests remain focused on environmental monitoring with remote sensing. He has been the Team Leader of the USGS/NASA Landsat Science Team since 2007. He also serves as the Co-Chair of the Land Cover Implementation Team of GOFC-GOLD. 

Abstract 

After 40+ years of collection of Landsat imagery, we are finally developing methods that make use off all the available data to monitor change in a continuous fashion on the surface of Earth.  New methods use all “clear” observations for a pixel.  A time series model is fit to each pixel, and change is identified when new observations no longer match “predicted images”, which are based on past observations.  The use of all available data has led to the following: 

• Detection of more subtle disturbance

• More reliable disturbance detection

• Trends in ecosystem health and growth

• Forest phenology – both average and interannual

• Peak greenness – as indications of trends in ecosystem health and climate response

These new methods are made possible by many developments over recent years, including: 

• Consistent data formats and outstanding orthorectification

• Improved computing capabilities

• Consolidation of Landsat holding

• Calibration

• automated atmospheric correction

• automated detection of clouds, cloud shadows and snow

• Free access to the data

There are many applications that benefit from time series analysis of Landsat.  In particular it is becoming easier to monitor land change as it is occurring, which will make the results more helpful for land management.  Another important application is support for greenhouse gas inventories and modeling of the carbon cycle. 

时间：2015年6月4日（星期四）下午2:00

地点：中科院地理资源所2602会议室 

主持人：方红亮研究员