1. Gone but not Forgotten? The Use of Airborne Laser Mapping in the Recovery Strategy for the Ivory-billed Woodpecker
    Dubayah, R.; Hofton, M. B. J. B. U. B. K. A.  AGU  December 2006

    The reappearance of the ivory-billed woodpecker in 2004, long-thought to be extinct, has created considerable controversy. Millions of dollars have been spent in the subsequent search to provide confirmation of the sightings, and thus validate what would be one of the great avian rediscoveries of our time. Recent searches for the woodpecker in the Cache and White River National Wildlife Refuges in Arkansas by ground-based teams have failed to locate it. However, the area is immense, with many locations difficult to access on foot. Thus, there is the need for near-term habitat monitoring at the local scale to guide immediate search efforts, and near- and long-term landscape scale characterization and assessment, both to guide further search efforts and to delineate suitable potential habitat should a population be found. A joint research effort by the University of Maryland, the US Fish and Wildlife Service, NASA, and the US Geological Survey was initiated early this year with the goal of using airborne and satellite remote sensing for habitat characterization and monitoring consistent with the ivory-billed woodpecker recovery strategy. A centerpiece of this effort is the use of lidar remote sensing to map management factors such as canopy cover, basal area, tree stocking, height and biomass, all of which are difficult to retrieve by any other means. In June, the Laser Vegetation Imaging Sensor (LVIS) was used to map over 1 million hectares of potential habitat. In this paper we discuss our mapping efforts with particular focus on how these data are used to make maps of forest management factors, and how these maps may then be combined to highlight areas of particularly suitable woodpecker habitat.



  2. Using Lidar-derived 3-D Vegetation Structure Maps to Assist in the Search for the Ivory- billed Woodpecker
    Hofton, M. A.; Blair, J. B. R. D. D. R. G. H.  AGU  December 2006

    Averaging about 20 inches in length, the ivory-billed woodpecker is among the world's largest woodpeckers. It once ranged through swampy forests in the southeastern and lower Mississippi valley states, and until recently was believed to have become extinct in the 1940's when commercial logging destroyed its last known habitat. Recent sightings however, may indicate the birds' survival in remaining bottomland hardwood forest adjacent to the Cache and White Rivers in Arkansas. In June-July 2006, NASA's Laser Vegetation Imaging Sensor (LVIS) was used to map approximately 5000 km2 of the White River National Wildlife Refuge in Arkansas, including sites where recent possible sightings of the bird occurred. LVIS is an airborne, medium- footprint (5- to 25-meter diameter), full waveform-recording, airborne, scanning lidar system which has been used extensively for mapping forest structure, habitat, carbon and natural hazards. The system digitally records the shape of the returning laser echo, or waveform, after its interaction with the various reflecting surfaces of the earth (leaves, branches, ground, etc.), providing a true 3-dimensional record of the surface structure. Data collected included ground elevation and canopy height measurements for each laser footprint, as well as the vertical distribution of intercepted surfaces (the return waveform). Experimental metrics such as canopy structure metrics based on energy quartiles, as well as ground energy/canopy cover and waveform complexity metrics will be derived from each waveform. The project is a collaborative effort between the University of Maryland, NASA, USGS, and the US Fish and Wildlife Service. The LVIS-generated data of the 3- D vegetation structure and underlying terrain will be used as a means to guide local, ground-based search efforts in the upcoming field season as well as identify the remaining areas of habitat suitable for protection should the bird be found.



  3. InSAR/Lidar/Optical Data Fusion for Vegetation Height and Biomass Estimation in Support of the North American Carbon Program
    Kellndorfer, J. M.; Walker, W. B. J. L. E. H. M. W. J.  AGU  December 2006

    A major goal of the North American Carbon Program (NACP) is to develop a quantitative scientific basis for regional to continental scale carbon accounting to reduce uncertainties about the carbon cycle component of the climate system. Accurate area-based estimates of terrestrial biomass and carbon require biophysical measures that capture both horizontal and vertical vegetation structure. Given the highly complementary nature and quasi-synchronous data acquisition of the InSAR 2000 Shuttle Radar Topography Mission (SRTM), the Landsat-based 2001 National Land Cover Database (NLCD), and data sets from the national LANDFIRE project, an exceptional opportunity exists for exploiting data synergies afforded by the fusion of these high- resolution, spatially explicit data sources. Whereas the thematic layers of the NLCD and LANDFIRE are suitable for characterizing horizontal structure (i.e., cover type, canopy density, etc.), SRTM provides information relating to the vertical structure, i.e., primarily height. Currently, a project funded under the NASA "Carbon Cycle Science" program "The National Biomass and Carbon Dataset 2000 NBCD2000" is underway to generate a "millennium" high-resolution ecoregional database of circa-2000 vegetation canopy height, aboveground biomass, and carbon stocks for the conterminous U.S. from these data sets. Generation of the NBCD2000 is performed within the same mapping zones which were defined for the NLCD2001 and LANDFIRE projects. The NBCD2000 data sets will provide an unprecedented baseline against which to compare products from the next generation of advanced microwave and optical remote sensing platforms. In order to develop models for the estimation of height and biomass from the remote sensing based data sources, reference data from the USDA Forest Service Forest Inventory and Analysis (FIA) program are used to develop the regression tree models for canopy height and biomass prediction. Recent research has been conducted incorporating lidar based vegetation height maps from the LVIS sensor to support the model development. A 2003 LVIS lidar acquisition (18x60 km2) in mapping zone 60 was used as height response variable in the regression tree model with predictor variables mean scattering phase center height (= SRTM minus NED), three Landsat Tasseled-Cap layers, land cover, and canopy density form NLCD2001. Independent test results on 377 FIA plot level data for the northern part of the mapping zone show that a robust model can be developed to predict vegetation height with a correlation coefficient of 0.71 between predicted and measured FIA plot basal area weighted heights with an rmse of 4.4 meters. An independent lidar testing population of 2311 samples showed a correlation coefficient between predicted and lidar height of 0.83 with an rmse of 3.0 meters.



  4. Canopy characteristics derived from small-footprint airborne waveform lidar and ICESat/GLAS over non-homogeneous landscapes
    Neuenschwander, A. L.; Gutierrez, R. U. T. J. S. B. E.  AGU  December 2006

    The accurate representation of vegetation structure has advanced over the past decade with the development of waveform lidar systems. Until recently, digitization of the full return waveform from an airborne system was only available on research systems such as LVIS and SLICER or from the spaceborne ICESat/GLAS system. Through an initiative from the University of Texas, Optech Inc. has developed a small-footprint full-waveform digitizer that is integrated into the ALTM system so that both full waveform and the conventional first and last returns are recorded for each transmitted laser pulse. Data from three ICESat ascending passes were acquired over an Oak/Juniper savanna located near San Marcos, Texas on 14 March 2005, 26 October 2005 and again on 26 February 2006 to investigate the use of waveform lidar for determining vegetation structure. In addition, small-footprint airborne lidar data were collected on 12 August 2005 using an Optech ALTM 1225 system with the return full waveform digitization. As a comparison with a different type of land cover, airborne lidar data were also collected over the Starr Forest in east-central Mississippi on 26-27 July, 2006; a research forest composed of upland pine and pine/hardwood forest, bottomland hardwood forest and pine plantation. In addition, data from two ICESat ascending passes were acquired over the Starr Forest on 3 March 2006 and 11 March 2006. Airborne lidar systems provide an excellent opportunity to sample the distribution of light transmittance through the canopy to determine the fraction of photosynthetically active radiation (fPAR) and leaf area index (LAI), which are both critical parameters for ecological modeling. Information derived from airborne systems can be used to calibrate parameters derived from spaceborne systems such as ICESat/GLAS. Previous methods used to compute LAI using large-footprint lidar systems have been applied here for the small-footprint case. LAI is estimated by computing a canopy height profile where the energy through the canopy is normalized by the entire energy from the return pulse. The canopy height profile energy is integrated to provide an estimate of LAI. This research strives to answer what is the critical spatial resolution and lidar point density necessary to accurately represent non-homogeneous land cover. The small-footprint waveforms are combined to simulate footprint diameters of 2m, 5m, 10m, as well as footprint sizes common to LVIS and GLAS. Similar to the airborne analysis, canopy analysis has been conducted on the waveforms collected by ICESat/GLAS and the results are compared.



  5. Analysis of Tropical Forest Vertical and Spatial Structural Dynamics Using Large-footprint Lidar
    Sheldon, S. L.; Dubayah, R. O. C. D. B. H. M. A. B. J.  AGU  December 2006

    In this paper we examine the ability of an airborne lidar, the Laser Vegetation Imaging Sensor (LVIS) to determine changes in the vertical structure of a tropical wet forest. LVIS, a large-footprint scanning lidar, collected data over La Selva Biological Station in Costa Rica, in March of 1998 and March of 2005. The La Selva region contains significant landscapes of old-growth and secondary forests, as well as other vegetation and management types. The specific objective of this study is to analyze the changes in vertical canopy structure and dynamics in secondary forest sites as compared to old-growth forests utilizing waveforms and waveform-derived metrics. Nearly co-incident footprints between years were used to assess structural changes at various spatial scales ranging from individual footprints to landscape level. On average, secondary forests showed significant growth as a function of age/height at all spatial scales. In contrast, old-growth forests were characterized by largely stable lidar heights. At the local (footprint) scale, considerable variability in growth rates for secondary forests, as well as in growth-loss in old-growth areas was observed. The number of footprints with large growth-loss (> 5 m), presumably caused by tree mortality in the old-growth forests, was consistent with expected mortality rates over a 7 year period.



  6. Retrieval of Vegetation Structure and Carbon Balance Parameters Using Ground-Based Lidar and Scaling to Airborne and Spaceborne Lidar Sensors
    Strahler, A. H.; Ni-Meister, W. W. C. E. L. X. J. D. L. C. D.  AGU  December 2006

    This research uses a ground-based, upward hemispherical scanning lidar to retrieve forest canopy structural information, including tree height, mean tree diameter, basal area, stem count density, crown diameter, woody biomass, and green biomass. These parameters are then linked to airborne and spaceborne lidars to provide large-area mapping of structural and biomass parameters. The terrestrial lidar instrument, Echidna(TM), developed by CSIRO Australia, allows rapid acquisition of vegetation structure data that can be readily integrated with downward-looking airborne lidar, such as LVIS (Laser Vegetation Imaging Sensor), and spaceborne lidar, such as GLAS (Geoscience Laser Altimeter System) on ICESat. Lidar waveforms and vegetation structure are linked for these three sensors through the hybrid geometric-optical radiative-transfer (GORT) model, which uses basic vegetation structure parameters and principles of geometric optics, coupled with radiative transfer theory, to model scattering and absorption of light by collections of individual plant crowns. Use of a common model for lidar waveforms at ground, airborne, and spaceborne levels facilitates integration and scaling of the data to provide large-area maps and inventories of vegetation structure and carbon stocks. Our research plan includes acquisition of Echidna(TM) under-canopy hemispherical lidar scans at North American test sites where LVIS and GLAS data have been or are being acquired; analysis and modeling of spatially coincident lidar waveforms acquired by the three sensor systems; linking of the three data sources using the GORT model; and mapping of vegetation structure and carbon-balance parameters at LVIS and GLAS resolutions based on Echidna(TM) measurements.



  7. The importance of heterogeneity: integrating lidar remote sensing and height-structured ecosystem models to improve estimation forest structure and dynamics
    Thomas, R. Q.; Hurtt, G. C. D. R. O. R. K. J. O. S. V. A. J. D.  AGU  December 2006

    Lidar remote sensing data have been shown to effectively represent forest structure, constrain estimates of carbon stocks and, when used to initialize a height-structured ecosystem model fluxes, constrain fluxes. Here we use large-footprint lidar data from the NASA Laser Vegetation Imaging Sensor (LVIS) to initialize the height-structured Ecosystem Demography (ED) model to study forest structure and dynamics at Hubbard Brook Experimental Forest (HBEF). HBEF is in the White Mountains of New Hampshire and includes heterogeneity in elevation dependent abiotic factors (i.e. temperature, precipitation, and soil depth), which are important drivers in a well documented decline in biomass and species change with elevation. We produced an estimate of the forest structure and fluxes in 1999 across all elevations at HBEF by first spinning up the model with elevation dependent climate and soil characteristics and then initializing the model using 1999 LVIS canopy height data. We validated the initialized model estimates against extensive field data and demonstrated that above ground carbon stocks, basal area, and species composition were within 1, 5 and 7%, respectively, of the field data at all elevations. Model projections were validated using data from a second LVIS acquisition obtained in 2003, accounting for the effects of both model uncertainty and lidar uncertainty. The additional constraint provided by including elevation dependent abiotic heterogeneity in the model initialization suggests that the forest is closer to a mature state than when initialized without the heterogeneity. We conclude that together lidar data and a height-structured ecosystem model give more information about forest structure and dynamics because together they account for critical fine scale heterogeneity.



    8. Return to the top of the page