Shrub Cover and Volume Estimates Using Ground and Airborne Laser Scans and Hyperspectral Imagery (2011)

Shrub Cover and Volume Estimates Using Ground and Airborne Laser Scans and Hyperspectral Imagery (2011)

 

Investigators and Affiliations

  • Jessica J. Mitchell, Ph.D. Research Affiliate, Geosciences Department, Idaho State University, Boise, Idaho
  • Nancy F. Glenn, Ph.D., Research Associate Professor, Geosciences Department, Idaho State University, Boise, Idaho
  • Lucas Spaete, Remote Sensing Specialist, Geosciences Department, Idaho State University, Boise, Idaho
  • Rupesh Shrestha, Post-doctoral Research Associate, Geosciences Department, Idaho State University, Boise, Idaho
  • Matt Anderson, Idaho National Laboratory, Idaho Falls, Idaho
  • Ryan Hruska, Idaho National Laboratory, Idaho Falls, Idaho
  • Randy Lee, Idaho National Laboratory, Idaho Falls, Idaho
  • Trent Armstrong, Idaho National Laboratory, Idaho Falls, Idaho

 

Funding Sources:

  • NOAAOAR ESRL/ Physical Sciences Division Grants # NA06OAR4600124 and #NA10OAR4680240
  • National Science Foundation, Idaho EPSCoR Program and National Science Foundation, award number EPS-0814387
  • Idaho National Laboratory - Laboratory Directed Research and Development Grants (Battelle, FY12) and the Idaho National Laboratory and Idaho State University Collaborative Remote Sensing Program.


Background:

Sagebrush vegetation communities cover 1.1 x 106 km2 of North American rangelands and provide food or cover for over 350 wildlife species. Like most vegetation, sagebrush cover and height varies across the landscape. Accurately mapping this variation is important for certain species, such as the greater sage-grouse, where sagebrush percent cover, visual cover and height are important characteristics for habitat selection. Cover, height and volume are important factors when trying to estimate rangeland biomass, which is an indicator of hydrologic function, forage potential, and species dominance. The low stature and spectrally indeterminate vegetation communities in semiarid environments are challenging to quantify with traditional remotely sensed data. Recent research has demonstrated the ability of LiDAR data to extract sagebrush height characteristics and the ability of hyperspectral analysis to quantify foliar nitrogen concentrations from individual sagebrush. We expand upon this body of research by investigating approaches to shrub cover and volume estimation using a combination of field reference data, ground-based terrestrial laser scanning (TLS) plot data and airborne laser (LiDAR) and hyperspectral acquisitions.

Objectives

  • Use object based image analysis techniques to delineate individual shrubs in the TLS plot data
  • Apply volumetric rendering techniques to the TLS plot data to generate bulks estimates of canopy volume
  • Relate ground-based volume estimates to airborne estimates derived from airborne lidar and hyperspectral data fusion products.


Accomplishments through 2011

Terrestrial laser scanning data were collected in September 2011 with a Leica ScanStation C10. The TLS dataset were collected at the highest resolution (0.02-m spacing at 100-m) using a defined field of view. A Topcon HiPer lite real time kinematic global positioning system (Topcon Positioning Systems Inc., California, USA) was used to record horizontal and vertical position of the control points.

Twenty 7-m x 7-m plots (n=20) were randomly located in the study area. The north and south boundaries of each plot were established and a tape was used to sample at 1-m intervals, for a total of eight, 7-m transects per plot. Vegetation information was recorded at each plot using both the line intercept and point intercept methods. Shrub metrics (i.e., height and intercepted length) were recorded for each transect for the line intercept method. Presence and vegetation type including ground type were recorded every 1-m along each transect for the point intercept method. Percent shrub cover was determined for the line intercept method by summing the total length of intercepted shrubs per plot and dividing by total plot length (n = 56-m). Likewise, shrub percent cover was determined for the point intercept method by dividing the total number of shrub hits per plot by total number of sampling locations (n = 64).

Preliminary results for estimating vegetation cover and volume using TLS and LiDAR data and hyperspectral imagery were presented in poster format at the 2011 American Geophysical Union (AGU) Fall Meeting in San Francisco, California. Exploratory results for fusing LiDAR data and hyperspectral imagery were also presented in poster format at the 2011 AGU Fall Meeting in San Francisco, California.

Results

Preliminary results show a potential relationship between field measured sagebrush vegetation cover and LiDAR and field measured sagebrush vegetation cover and hyperspectral indices. This relationship also appears to be strengthened when the two are combined; however, work is still needed to integrate object based image analysis and TLS data. It is apparent that TLS data have high potential in aiding volumetric estimations across landscapes (Figure 1).
Figure 1
Figure 9-9. Terrestrial Laser Scanning (TLS) Offers the Ability to Make Very Accurate Volume Estimates. Top view of individual sagebrush (left) and side view of individual sagebrush (right) were used to derive height (0.67 m), major axis (0.69 m), minor axis (0.53 m), and volume (0.128 m3).

Plans for Continuation

Final shrub volume estimation results using TLS data will be submitted to Photogrammetric Engineering and Remote Sensing or other journal for publication. Final LiDAR and hyperspectral data fusion techniques for shrub canopy volume estimation will be submitted to Advances in Ecological Research or other journal for publication.

Publications, Reports, and Theses

Spaete, L., Mitchell, J., Glenn, N., Shrestha, R., Sankey, T., Murgoitio, J., Gould, S., Leedy, T., and Hardegree, S. Vegetation Cover and Volume Estimates in Semi-arid Rangelands using LiDAR and Hyperspectral Data. 2011 AGU Fall Meeting, 5-9 December, San Francisco, CA.

Moore, C., Olsoy, P., Gertman, V., Mitchell, J., Glenn, N., Joshi, A., Shrestha, R., Spaete, L., Norpchen, D., Pernice, M., Whiting, E., Grover, S., Lee, R., Anderson, J., and Busche, C. 3D Immersive Environments: Discovering New Methods for Data Fusion, Visualization, and Analysis of Hyperspectral and Laser Altimetry Data. 2011 AGU Fall Meeting, 5-9 December, San Francisco, CA.

Mitchell, J., Glenn N., Spaete, L., Shrestha, R., Lee, R. and Armstrong, T. Shrub canopy delineation and volume estimation using Terrestrial Laser Scanning (TLS) data, in preparation. Target journal: Photogrammetric Engineering and Remote Sensing.

Mitchell, J., Glenn N., Spaete, L., Murgoitio, J., Anderson, M., and Hruska, R. Explorations in LiDAR and hyperspectral fusion for shrub canopy volume estimation, in preparation. Target journal: Advances in Ecological Research.