Uncertainty in LiDAR derived Canopy Height Models in three unique forest ecosystems
Abstract
The National Ecological Observatory Network (NEON) is a continental-scale ecological observation platform designed to collect and disseminate data that contributes to understanding and forecasting the impacts of climate change, land use change, and invasive species on ecology. NEON will collect in-situ and airborne data over 81 sites across the US, including Alaska, Hawaii, and Puerto Rico. The Airborne Observation Platform (AOP) group within the NEON project operates a payload suite that includes a waveform / discrete LiDAR, imaging spectrometer (NIS) and high resolution RGB camera. One of the products derived from the discrete LiDAR is a canopy height model (CHM) raster developed at 1 m spatial resolution. Currently, it is hypothesized that differencing annually acquired CHM products allows identification of tree growth at in-situ distributed plots throughout the NEON sites. To test this hypothesis, the precision of the CHM product was determined through a specialized flight plan that independently repeated up to 20 observations of the same area with varying view geometries. The flight plan was acquired at three NEON sites, each with a unique forest types including 1) San Joaquin Experimental Range (SJER, open woodland dominated by oaks), 2) Soaproot Saddle (SOAP, mixed conifer deciduous forest), and 3) Oak Ridge National Laboratory (ORNL, oak hickory and pine forest). A CHM was developed for each flight line at each site and the overlap area was used to empirically estimate a site-specific precision of the CHM. The average cell-by-cell CHM precision at SJER, SOAP and ORNL was 1.34 m, 4.24 m and 0.72 m respectively. Given the average growth rate of the dominant species at each site and the average CHM uncertainty, the minimum time interval required between LiDAR acquisitions to confidently conclude growth had occurred at the plot scale was estimated to be between one and four years. The minimum interval time was shown to be primarily dependent on the CHM uncertainty and number of cells within a plot which contained vegetation. This indicates that users of NEON data should not expect that changes in canopy height can be confidently identified between annual AOP acquisitions for all areas of NEON sites.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFM.B54B..05G
- Keywords:
-
- 0414 Biogeochemical cycles;
- processes;
- and modeling;
- BIOGEOSCIENCESDE: 0428 Carbon cycling;
- BIOGEOSCIENCESDE: 0439 Ecosystems;
- structure and dynamics;
- BIOGEOSCIENCESDE: 0480 Remote sensing;
- BIOGEOSCIENCES