AGU 2011: LIDAR on Ecosystem Structure, Functioning and Services

I am currently attending the AGU 2011 Fall Meeting in San Francisco. Right now, I am inside a session on LiDAR. Here are some interesting studies you may want to check out if you are into LiDAR and its applications to Ecosystem Structure.

Discrete Lidar Pseudo-Waves Versus True Waveform lidar: Are They Fungible for Forest Structural Analyses? by Jordan D. Muss; Naikoa Aguilar-Amuchastegui; Eric Salas; Geoffrey M. Henebry

ABSTRACT BODY: In the past three decades, two types of lidar systems, discrete and waveform, have been used to estimate forest structure, including height, biomass, and basal area. While there are many similarities between these two systems, their fundamental differences are the manner in which the laser backscatter is recorded and, to a lesser degree, how these data are analyzed. Typically, discrete systems record far fewer returns per pulse, but provide fine-scale, three-dimensional positions of reflecting surfaces, whereas waveform lidar captures much more of the backscattered energy at the expense of knowing the precise location of the reflectors. Despite these differences, these data have, overwhelmingly, been analyzed using frequency-based statistics, even though this results in highly correlated metrics and may sacrifice a wealth of structural information that could be latent within the waveforms. Some recent studies have challenged this approach and introduced methods to aggregate discrete lidar data into pseudo-waves, and investigated alternative methods to examine these waves to improve assessments of forest-structure. We propose that the differences between pseudo-waves and true waveforms are minor, and that the same wave-based metrics can be used to analyze either surface. To test these propositions, pseudo-waves are being created from discrete lidar data collected in the spring of 2006, over the La Selva Biological Station in Costa Rica. These pseudo-waves are being geographically co-registered to full-waveform data collected in 2005 over the same area using the LVIS sensor. The two sets of waves will be normalized so that similarities in their shapes can be evaluated, after which we will demonstrate that wave-based metrics, which have been used to relate pseudo-waves to estimate canopy heights, basal area, and biomass in temperate forests can also be used to describe forest structure in the tropics.

Studying the Impact of the Three Dimensional Canopy Structure on LIDAR Waveforms Evaluated with Field Measurements by Liang Xu; Yuri Knyazikhin; Ranga B. Myneni; Alan H. Strahler; Crystal Schaaf; Alexander S. Antonarakis; Paul R. Moorcroft

ABSTRACT BODY: The three-dimensional structure of a forest – its composition, density, height, crown geometry, within-crown foliage distribution and properties of individual leaves – has a direct impact on the lidar waveform. The pair-correlation function defined as the probability of finding simultaneously phytoelements at two points is the most natural and physically meaningful descriptor of the canopy structure over wide range of scales. The stochastic radiative transfer equations naturally admit this measure and thus provide a powerful means to investigate 3D canopy from space. NASA’s Airborne Laser Vegetation Imaging Sensor (LVIS) and ground based data on canopy structure acquired over 5 sites in New England, California and La Selva (Costa Rica) tropical forest were analyzed to assess the impact of 3D canopy structure on lidar waveform and the ability of stochastic radiative transfer equations to simulate the 3D effects. Our results suggest the pair correlation function is sensitive to horizontal and vertical clumping, crown geometry and spatial distribution of trees. Its use in the stochastic radiative transfer equation allows us to accurately simulate the effects of 3D canopy structure on the lidar waveform. Specifically, we found that (1) attenuation of the waveform occurs at a slower rate than 1D models predict; this may result in an underestimation of foliage profile if 3D effects are ignored; (2) 1D model is unable to match simulated waveform and measured surface reflectance, i.e., an unrealistic high value of surface reflectance needs to be used to simulate ground return of sparse vegetation; (3) spatial distribution of trees has a strong impact on the lidar waveform. Simple analytical models of the pair-correlation function will also be discussed.

Here are more:

LiDAR bore-sight calibration: a comparative study
Gil R. Gonçalves; Andre Jalobeanu

Intensity normalization and automatic gain control correction of airborne LiDAR data for classifying a rangeland ecosystem
Rupesh Shrestha; Nancy F. Glenn; Lucas Spaete; Jessica Mitchell

Automated Tree Crown Delineation and Biomass Estimation from Airborne LiDAR data: A Comparison of Statistical and Machine Learning Methods
Colin J. Gleason; Jungho Im

Structural Biomass Estimation from L-band Interferometric SAR and Lidar (Invited)
Robert N. Treuhaft; Bruce D. Chapman; Fabio Goncalves; Scott Hensley; Joao R. dos Santos; Paulo A. Graca; Luciano Dutra

Estimation of Above Ground Biomass in the Everglades National Park using X-, C-, and L-band SAR data and Ground-based LiDAR
Emanuelle A. Feliciano ; Shimon Wdowinski; Matthew Potts; Sang-Wan Kim