1. Effect of vegetation structure on subcanopy solar radiation: a comparative study
    Anand, A; Dubayah, R. H. M. A.  AGU  December 2012

    Vertical structure of vegetation canopy influences spatial variability of radiation regime under forest canopies. A comparison of transmittance profiles and subcanopy radiation regime for two structurally different forest sites is done based on ray tracing and principles of radiative transfer using Lidar data. Medium footprint waveform Lidar data from Laser Vegetation Imaging Sensor (LVIS) was collected from the sites in Sierra National Forest (SNF), California and Smithsonian Environmental Research Center (SERC), Maryland in 2008 and 2003 respectively. Sites in both forest areas have varying vegetation structure with SNF sites representing mixed conifers whereas the sites in SERC represent eastern broadleaf trees. The Lidar waveform is processed to derive canopy gap probability as a function of height which is used to derive transmittance profiles and solar radiation as a function of canopy height using a 3-D light transmittance model. Geostatistics is applied to compare how the vertical and horizontal distribution of solar radiation under sub-canopy surface varies with varying vertical canopy structures such as foliage density, canopy cover and canopy height. This comparison is expected to increase knowledge on vegetation structure effects forest canopies



  2. High-Altitude Laser Altimetry from the Global Hawk UAV for Regional Mapping of Surface Topography
    Blair, J B; Rabine, D. W. S. H. M. A. M. S.  AGU  December 2012

    NASA's Land, Vegetation, and Ice Sensor (LVIS) is a high-altitude, full-waveform, geodetic-imaging laser altimeter system of which a UAV-based version (LVIS-GH) is currently being tested. From 20 km above the surface in the Global Hawk UAV, LVIS-GH images surface topography and roughness (including forest height) across a 4 km wide swath using 15 m diameter footprints. In recent years, the LVIS has been flown at altitudes of up to 14 km over Greenland and Antarctica on flights up to 12 hours in duration, enabling the efficient and precise mapping of large areas from the air. The Global Hawk will extend this capability to up to 32 hours and altitudes approaching 20 km. In order to achieve decimeter level vertical precision and accuracy from high altitude, advanced parameter estimation techniques, based on those implemented in NASA's GEODYN software, are used to estimate the angular, spatial, and temporal biases required to accurately georeference the component lidar data sets. Data from specific in-air maneuvers are utilized in order to isolate the effects of different error sources and to break correlations between biases. Examples of high-altitude data and airborne/spaceborne sensor intercomparison and fusion will be shown. For example, the comparison of data from NASA's ICESat-1 mission with coincident LVIS data collected around 86S (the maximum extent of data collected during ICESat) to quantify inter-campaign biases in Icesat-1 elevation measurements and improve estimates of long -term elevation change rates of ice sheets will be shown. These results illustrate the utility of high-altitude wide swath imaging, particularly from platforms such as the Global-Hawk, for enhancing spacebased data sets



  3. Patterns of Forest Disturbance and Recovery Dynamics on Structure and Carbon Fluxes in New England Forests
    Dolan, K; Hurtt, G. C. H. C. e. a.  AGU  December 2012

    Forest disturbance and recovery strongly influence forest structure, function and services. Forest disturbance and recovery are critical mechanisms for transferring carbon between the land surface and the atmosphere, yet the role of forest disturbance within the terrestrial carbon cycle still remains uncertain and only recently have these events been accounted for within regional-scale and global carbon models. Adding ecological disturbance into biogeochemical models is noted as critical to estimating current and future carbon stocks and fluxes. This study used satellite-based observations of forest change, lidar derived structure data and a height structured ecosystem model to improve knowledge of disturbances role in carbon cycle by quantifying how forest disturbance and recovery vary at different spatial and temporal scales. Annual forest change maps from 1984-2010 were produced using the highly automated Vegetation Change Tracker (VCT) algorithm (Huang et al 2009). Mapped forest change was further broken down into land conversion (forest to non forest), severe disturbance (stand replacing), and non severe (partial clearing/ thinning). Areas of forest change were aggregated at different spatial scales and temporal scales and integrated into the Ecosystem Demography model (ED), a mechanistic model of forest ecosystem dynamics, to calculate changes in biomass and carbon fluxes. Forest structural data derived from NASA's Laser Vegetation Imaging Sensor (LVIS) was used to assess regrowth of forests and compare to ED's height and structure properties. Results in the New England Region show both spatial and temporal variation in area disturbed. The northern region encompassing Northern New Hampshire showed higher and more variable rates with an average annual rate of disturbance of approximately 0.5% (range 0.2- 08%) conversion/ non regeneration forest clearing range 0.02 -0.08%. While the southern averaged annual disturbance of 0.3% (ranged 0.2 - 0.5%) it had a much higher rate of conversion and showed had a positive trend in though time from 0.03- 0.18%.Within the state annual forest disturbance rates were on average 3 times higher outside of federally managed lands, which comprise 20 percent of NH forest lands. Incorporation of canopy height data has constrained model estimates of carbon stocks and annual net ecosystem production considerably with a strong relationship to in field measurements however the current coverage of lidar does not allow the same continuous monitoring provided by imagery thus a combination of the two tools may provide more useful then either alone. Results show that management is a major driver of forest change patterns and more research is needed to separate manmade from natural disturbance events. Results show that future research concerning the proportion of disturbance below VCT detection may be an important aspect of ecosystem modeling



  4. Estimating ICESat-1 Inter-Campaign Elevation Biases at 86S Using LVIS Lidar Data
    Hofton, M A; Luthcke, S. B. B. J. B.  AGU  December 2012

    NASA's ICESat-1 was launched in January 2003 to detect ice elevation changes that are indicative of ice volume changes over time. Over 2 billion elevation measurements were made from 2003 to 2009 during 18, month-long observation periods two to three times a year. Although elevation measurements were precise at the decimeter level, the presence of inter-observation (time-variable) range biases affected the relative accuracy of the elevation measurements, potentially influencing estimates of long term elevation change rates if not accurately accounted for. In 2009 and 2010, NASA's Land, Vegetation and Ice Sensor (LVIS) was flown in NASA's DC-8 aircraft from Punta Arenas, Chile, to and around 86S (the maximum extent of the ICESat-1 orbit) to collect elevation data within a ~2 km-wide data swath at ~25 m resolution. Three separate flights were needed to completely encircle 86S, however the location offered the maximum overlap of ICESat-1 footprints and tracks within an LVIS laser altimeter swath width. Comparison of the differences between the three epochs of LVIS elevations with coincident ICESat-1 elevations from 2003 to 2009 are presented, including analysis of the differences with respect to various data set characteristics (e.g., geographic location). The consistency of the trends with respect to the various subset solutions are summarized and their affect on long-term elevation change rates illustrated.



  5. A new method to extract forest height from repeat-pass polarimetric and interferometric radar data
    Lavalle, M; Hensley, S. D. R.  AGU  December 2012

    The objective of this paper is to present a new remote sensing method and a new physical model that will potentially enable estimating forest height and vegetation 3D structure using radar technology. The method is based on repeat-pass polarimetric-interferometric radar technique; the model is termed random-motion-over-ground (RMoG) model [1, 2]. We will describe a step-by-step procedure that will help the ecosystem community to monitor ecosystems at regional and global scale using radar data available from the forthcoming radar missions. We will show first results of forest height estimated from UAVSAR data and compared against LVIS data. We will quantify the error associated to our method. We will also discuss the improvements that we plan on including in future works. Our ultimate goal is to measure low and large biomass stocks using the large amount of radar data that will be available in the near future. The Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) is a fully polarimetric L-band airborne radar developed at the Jet Propulsion Laboratory (JPL). UAVSAR acquires repeat-pass interferometric data for measuring vegetation structure and monitoring crustal deformations. The UAVSAR team at JPL has acquired and processed several polarimetric-interferometric (Pol-InSAR) datasets over the Harvard Forest in Massachusetts (United States) that allows testing repeat-pass Pol-InSAR technique. Pol-InSAR technique was proposed 15 years ago to estimate vegetation biomass and overcome the inherent saturation of radar backscatter versus biomass [3]. The advantage of Pol-InSAR is the ability to estimate the 3D structure of vegetation using a small number of interferometric acquisitions. In order to extract vegetation properties from Pol-InSAR UAVSAR data, we use a model of temporal-volumetric coherence, the RMoG model, suitable for repeat-pass interferometry. In the RMoG model the vegetation is idealized as a two-layer scattering scenario constituted by a penetrable layer of randomly oriented scattering elements (i.e., the canopy layer) above a dielectric rough surface (i.e., the ground layer). To account for temporal decorrelation, we assume the temporal changes in the canopy layer and on the ground surface to be caused by a Gaussian-statistic motion of the scattering elements. The assumption of first-order Gaussian-statistic motion has been validated with zero-baseline Pol-InSAR UAVSAR data [2]. In this paper we discuss the possibility to estimate the RMoG model parameters based on multiple Pol-InSAR coherence observations. The inversion strategy is based on a set of balanced equations between model parameters and coherence observations. This approach has the potential to provide maps of forest height, canopy structure and temporal decorrelation from single-baseline repeat-pass Pol-InSAR data. [1] M. Lavalle, PhD Dissertation, University of Rome Tor Vergata and University of Rennes 1. 2009. [2] M. Lavalle, M. Simard, and S. Hensley, ``A temporal decorrelation model for polarimetric SAR interferometers,'' IEEE Trans. Geosci. Remote Sens., 2011 (on line). [3] S. Cloude and K. Papathanassiou, ``Polarimetric SAR interferometry,'' IEEE Trans. Geosci. Remote Sens., vol. 36, no. 5, pp. 1551-1565, Sep. 1998



  6. Estimating animal biodiversity across taxa in tropical forests using shape-based waveform lidar metrics and Landsat image time series
    Muss, J D; Aguilar-Amuchastegui, N. H. G. M.  AGU  December 2012

    Studies have shown that forest structural heterogeneity is a key variable for estimating the diversity, richness, and community structure of forest species such as birds, butterflies, and dung beetles. These relationships are especially relevant in tropical forests when assessing the impacts of forest management plans on indicator groups and species. Typically, forest structure and biodiversity are evaluated using field surveys, which are expensive and spatially limited. An alternative is to use the growing archive of imagery to assess the impacts that disturbances (such as those caused by selective logging) have on habitats and biodiversity. But it can be difficult to capture subtle differences in the three-dimensional (3D) forest structure at the landscape scale that are important for modeling these relationships. We use a unique confluence of active and passive optical sensor data, field surveys of biodiversity, and stand management data to link metrics of spatial and spatio-temporal heterogeneity with key indicators of sustainable forest management. Field sites were selected from tropical forest stands along the Atlantic Slope of Costa Rica for which the management history was known and in which biodiversity surveys were conducted. The vertical dimension of forest structure was assessed by applying two shape-based metrics, the centroid (C) and radius of gyration (RG), to full waveform lidar data collected by the LVIS platform over central Costa Rica in 2005. We developed a map of the vertical structure of the forest by implementing a recursive function that used C and RG to identify major segments of each waveform. Differences in 3D structure were related to estimates of animal biodiversity, size and type of disturbance, and time since disturbance---critical measurements for achieving verifiable sustainable management and conservation of biodiversity in tropical forests. Moreover, the relationships found between 3D structure and biodiversity suggests that it may be possible to implement a rapid and robust assessment of forest dynamics and biodiversity at the landscape scale by complementing field surveys with data acquired from active (such as lidar) and passive optical sensors.



  7. Quantifying Ice-sheet/Ice-shelf Dynamics and Variability with Meter-scale DEM and Velocity Timeseries
    Shean, D E; Joughin, I. R. S. B. E. e. a.  AGU  December 2012

    Both the Antarctic and Greenland ice sheets are losing mass at an increasing rate, although loss due to accelerating flow and dynamic thinning remains poorly understood. We are using complementary data from repeat satellite and airborne observations to investigate the relationship between ice-sheet/ice-shelf dynamics and geometry on seasonal to interannual timescales. High-resolution along-track stereo imagery from commercial satellite vendors DigitalGlobe and GeoEye provides unprecedented spatial (~0.5 m/px with ~17 km swath width) and temporal (weekly/monthly) resolution for these efforts. We have developed an automated pipeline using open-source software to produce orthoimage, DEM, and surface velocity products from DigitalGlobe imagery. High-contrast surface texture (e.g. sastrugi, crevasses) visible at sub-meter resolution provides near-perfect image correlation (~99% success rate) during DEM and velocity map derivation. Elevation data from IceBridge ATM/LVIS, ICESat GLAS, and GPS campaigns are used to correct DEMs and perform accuracy assessment. Preliminary tests over exposed bedrock provide relative vertical accuracy estimates of <1-2 m for Worldview-1/2 DEMs. Velocity data from TerraSAR-X and GPS campaigns provide validation for surface velocity products, with horizontal error estimates of <10 m. Velocity and elevation change products with 2-4 m/px spatial resolution allow for unprecedented 3D dynamic characterization of sub-km flow transition zones (e.g. grounding lines, shear margins), capturing both local and regional variations due to melting and dynamic thinning. We present timeseries for West Greenland (Jakobshavn front - 20 observations, Jakobshavn south catchment - 10) and West Antarctica (Pine Island and Thwaites - 5 each) from 2009-2012. These observations complement ongoing efforts to measure and model outlet glacier dynamics, with implications for future ice-sheet mass balance estimates.



  8. Further Studies of Forest Structure Parameter Retrievals Using the Echidna® Ground-Based Lidar
    Strahler, A H; Yao, T. Z. F. e. a.  AGU  December 2012

    Ongoing work with the Echidna® Validation Instrument (EVI), a full-waveform, ground-based scanning lidar (1064 nm) developed by Australia's CSIRO and deployed by Boston University in California conifers (2008) and New England hardwood and softwood (conifer) stands (2007, 2009, 2010), confirms the importance of slope correction in forest structural parameter retrieval; detects growth and disturbance over periods of 2-3 years; provides a new way to measure the between-crown clumping factor in leaf area index retrieval using lidar range; and retrieves foliage profiles with more lower-canopy detail than a large-footprint aircraft scanner (LVIS), while simulating LVIS foliage profiles accurately from a nadir viewpoint using a 3-D point cloud. Slope correction is important for accurate retrieval of forest canopy structural parameters, such as mean diameter at breast height (DBH), stem count density, basal area, and above-ground biomass. Topographic slope can induce errors in parameter retrievals because the horizontal plane of the instrument scan, which is used to identify, measure, and count tree trunks, will intersect trunks below breast height in the uphill direction and above breast height in the downhill direction. A test of three methods at southern Sierra Nevada conifer sites improved the range of correlations of these EVI-retrieved parameters with field measurements from 0.53-0.68 to 0.85-0.93 for the best method. EVI scans can detect change, including both growth and disturbance, in periods of two to three years. We revisited three New England forest sites scanned in 2007-2009 or 2007-2010. A shelterwood stand at the Howland Experimental Forest, Howland, Maine, showed increased mean DBH, above-ground biomass and leaf area index between 2007 and 2009. Two stands at the Harvard Forest, Petersham, Massachusetts, suffered reduced leaf area index and reduced stem count density as the result of an ice storm that damaged the stands. At one stand, broken tops were visible in the 2010 point cloud canopy reconstruction. A new method for retrieval of the forest canopy between-crown clumping index from angular gaps in hemispherically-projected EVI data traces gaps as they narrow with range from the instrument, thus providing the approximate physical size, rather than angular size, of the gaps. In applying this method to a range of sites in the southern Sierra Nevada, element clumping index values are lower (more between-crown clumping effect) in more open stands, providing improved results as compared to conventional hemispherical photography. In dense stands with fewer gaps, the clumping index values were closer. Foliage profiles retrieved from EVI scans at five Sierra Nevada sites are closely correlated with those of the airborne Lidar Vegetation Imaging Sensor (LVIS) when averaged over a diameter of 100 m. At smaller diameters, the EVI scans have more detail in lower canopy layers and the LVIS and EVI foliage profiles are more distinct. Foliage profiles derived from processing 3-D site point clouds with a nadir view match the LVIS foliage profiles more closely than profiles derived from EVI in scan mode. Removal of terrain effects significantly enhances the match with LVIS profiles. This research was supported by the US National Science Foundation under grant MRI DBI-0923389



  9. NASA's Operation IceBridge: using instrumented aircraft to bridge the observational gap between ICESat and ICESat-2 laser altimeter measurements
    Studinger, M.  AGU  December 2012

    NASA's Operation IceBridge images Earth's polar ice in unprecedented detail to better understand processes that connect the polar regions with the global climate system. Operation IceBridge utilizes a highly specialized fleet of research aircraft and the most sophisticated suite of innovative science instruments ever assembled to characterize annual changes in thickness of sea ice, glaciers, and ice sheets. In addition, Operation IceBridge collects critical data used to predict the response of Earth's polar ice to climate change and resulting sea-level rise. IceBridge also helps bridge the gap in polar observations betweenNASA's ICESat satellite missions. Combined with previous aircraft observations, as well as ICESat, CryoSat-2 and the forthcoming ICESat-2 observations, Operation IceBridge will produce a cross-calibrated 17-year time series of ice sheet and sea-ice elevation data over Antarctica, as well as a 27-year time series over Greenland. These time series will be a critical resource for predictive models of sea ice and ice sheet behavior. In addition to laser altimetry, Operation IceBridge is using a comprehensive suite of instruments to produce a three-dimensional view of the Arctic and Antarctic ice sheets, ice shelves and the sea ice. The suite includes two NASA laser altimeters, the Airborne Topographic Mapper (ATM) and the Land, Vegetation and Ice Sensor (LVIS); four radar systems from the University of Kansas' Center for Remote Sensing of Ice Sheets (CReSIS), a Ku-band radar altimeter, accumulation radar, snow radar and the Multichannel Coherent Radar Depth Sounder (MCoRDS); a Sander Geophysics airborne gravimeter (AIRGrav), a magnetometer and a high-resolution stereographic camera (DMS). Since its start in 2009, Operation IceBridge has deployed 7 geophysical survey aircraft, 18 science instruments. All IceBridge data is freely available from NSIDC (http://nsidc.org/data/icebridge) 6 months after completion of a campaign



  10. Deriving Leaf Area Index (LAI) from multiple lidar remote sensing systems
    Tang, H; Dubayah, R. Z. F.  AGU  December 2012

    LAI is an important biophysical variable linking biogeochemical cycles of earth systems. Observations with passive optical remote sensing are plagued by saturation and results from different passive and active sensors are often inconsistent. Recently lidar remote sensing has been applied to derive vertical canopy structure including LAI and its vertical profile. In this research we compare LAI retrievals from three different types of lidar sensors. The study areas include the La Selva Biological Station in Costa Rica and Sierra Nevada Forest in California. We first obtain independent LAI estimates from different lidar systems including airborne lidar (LVIS), spaceborne lidar (GLAS) and ground lidar (Echidna). LAI retrievals are then evaluated between sensors as a function of scale, land cover type and sensor characteristics. We also assess the accuracy of these LAI products against ground measurements. By providing a link between ground observations, ground lidar, aircraft and space-based lidar we hope to demonstrate a path for deriving more accurate estimates of LAI on a global basis, and to provide a more robust means of validating passive optical estimates of this important variable.



  11. Elevation change in Greenland over two decades from cross-platform LiDAR analysis
    Wheelock-Davis, E ; Howat, I. M. B. J. B. H. M. A.  AGU  December 2012

    NASA's Airborne Topographic Mapper (ATM) and the Land, Vegetation, & Ice Sensor (LVIS) are two airborne Light Detection and Ranging (LiDAR) systems that retrieve information about surface elevation and roughness. Both altimeters have been flown in Greenland to measure changes in ice sheet surface elevation through time. ATM surveys in the region have been conducted nearly every year since 1993, with extended, annual coverage by ATM and LVIS since 2009 during Operation IceBridge (OIB). These recent surveys provided repeat coverage of many older, previously unrepeated, ATM flight lines. Combined, these datasets offer a unique, multi-decadal record of ice sheet change. The different beam trajectory technologies on each system require a specific methodology to compare coincident data between the systems. Here, we apply a general slope-fit regression analysis to difference overlapping ATM and LVIS data. We validate this method using, for OIB-era surveys, overlapping data collected at nearly the same time and, for earlier data, overlapping data over ice-free terrain. We then examine the spatial and temporal distribution of ice sheet surface elevation changes since 1993 revealed by cross-platform analysis.



    12. Return to the top of the page