Close Window

Forest Structure and Biomass Mapping Using Time Series Landsat Observations, Small Footprint Lidar, and Field Inventory Data in North Carolina

Chengquan Huang, University of Maryland, cqhuang@umd.edu (Presenter)

Although significant efforts have been devoted to deriving forest structure and biomass using optical remote sensing data, relationships between spectral data and forest structure were often very weak, because the former is typically not very sensitive to changes in forest structure. However, optical remote sensing images, especially those produced through a series of 6 Landsat instruments, have been available since 1972. This imagery record makes it possible to detect forest disturbance and calculate age since disturbance for up to four decades. Because forest age is often a good predictor of forest growth and yield, the age since disturbance information derived using time series Landsat observations has been found highly valuable for predicting forest height growth. For forests where no disturbance was observed by the Landsat systems since the 1970s, a spectral record of their growth in multiple decades with the knowledge that they are likely older than 40 years will allow better estimation of their structure. In this study, we will develop an approach for mapping forest structure and biomass using time series Landsat observations. In this approach, forests are first divided into “young” forests that generated following disturbances recorded by the Landsat and “old” forests that did not experience stand-replacement disturbance event since the first available Landsat observation. Height and biomass for these two types of forest are modeled separately using empirical methods trained using lidar or field plot data. We will demonstrate the effectiveness of this approach by using it to develop forest structure and biomass products over North Carolina.

Presentation Type:  Poster

Session:  Poster Session 1-A   (Tue 11:00 AM)

Associated Project(s): 

Poster Location ID: 46

 


Close Window