1. Demonstration of centimeter-level precision, swath mapping, full-waveform laser altimetry from high altitude on the Global Hawk UAV for future application to cryospheric remote sensing
    Blair, J. and Wake, S. and Rabine, D. and Hofton, M. and Mitchell, S.  AGU  December 2013

    The Land Vegetation and Ice Sensor (LVIS) is a high-altitude, wide-swath laser altimeter that has, for over 15 years, demonstrated state-of-the-art performance in surface altimetry, including many aspects of remote sensing of the cryosphere such as precise topography of ice sheets and sea ice. NASA Goddard, in cooperation with NASA's Earth Science Technology Office (ESTO), has developed a new, more capable sensor that can operate autonomously from a high-altitude UAV aircraft to further enhance the LVIS capability and extend its reach and coverage. In June 2012, this latest sensor, known as LVIS-GH, was integrated onto NASA's Global Hawk aircraft and completed a successful high-altitude demonstration flight over Death Valley, Owens Valley, and the Sierra Nevada region of California. Data were collected over a wide variety of terrain types from 58,000' (> 17 km) altitude during the 6 hour long test flight. The full-waveform laser altimetry technique employed by LVIS and LVIS-GH provides precise surface topography measurements for solid earth and cryospheric applications and captures the vertical structure of forests in support of territorial ecology studies. LVIS-GH fully illuminates and maps a 4 km swath and provides cm-level range precision, as demonstrated in laboratory and horizontal range testing, as well as during this test flight. The cm range precision is notable as it applies to accurate measurements of sea ice freeboard and change detection of subtle surface deformation such as heaving in permafrost areas. In recent years, LVIS has primarily supported Operation IceBridge activities, including deployments to the Arctic and Antarctic on manned aircraft such as the NASA DC-8 and P-3. The LVIS-GH sensor provides an major upgrade of coverage capability and remote access; LVIS-GH operating on the long-duration Global Hawk aircraft can map up to 50,000 km^2 in a single flight and can provide access to remote regions such as the entirety of Antarctica. Future applications of LVIS-GH could include comprehensive mapping of cryosphere targets over large regions such as Alaska, Greenland, and Antarctica as well as an opportunity for seasonal mapping of sea and land ice. Data from the test flight will be presented along with accuracy assessment and specific examples of the cm-level range precision and wide swath mapping ability relevant to cryospheric remote sensing. Cite as: Author(s) (2013), Title, Abstract C21D-0655 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  2. Classification and Characterization of Neotropical Rainforest Vegetation from Hyperspectral and LiDAR Data
    Crawford, M. and Prasad, S. and Jung, J. and Yang, H. and Zhang, Y.  AGU  December 2013

    Mapping species and forest vertical structure at regional, continental, and global scale is of increasing importance for climate science and decision support systems. Remote sensing technologies have been widely utilized to achieve this goal since they help overcome limitations of the direct and indirect measurement approaches. While the use of multi-sensor data for characterizing forest structure has gained significant attention in recent years, research on the integration of full waveform LiDAR and hyperspectral data for a) classification and b) characterization of vegetation structure has been limited. Given sufficient labeled ground reference samples, supervised learning methods have evolved to effectively classify data in a high dimensional feature space. However, it is expensive and time-consuming to obtain labeled data, although the very high dimensionality of feature spaces from hyperspectral and LiDAR inputs make it difficult to design reliable classifiers with a limited quantity of labeled data. Therefore, it is important to concentrate on developing training data sets which are the most “informative” and “useful” for the classification task. Active learning (AL) was developed in the machine learning community, and has been demonstrated to be useful for classification of remote sensing data. In the active learning framework, classifiers are initially trained on a very limited pool of training samples, but additional informative and representative samples are identified from the abundant unlabeled data, labeled, and then inducted into this pool, thereby growing the training dataset in a systematic way. The goal is to choose data points such that a more accurate classification boundary is learned. We propose a novel Multi-kernel Active Learning (MKL-AL) approach that incorporates features from multiple sensors with an automatically optimized kernel composite ¬function, and kernel parameters are selected intelligently during the AL learning process. The high dimensionality of full waveform LiDAR and hyperspectral data is also problematic for predicting structural variables such as leaf area index (LAI) from full waveform LiDAR and hyperspectral data. A new approach based on nonlinear multi-sensor feature extraction is applied to HyMap and LVIS remotely sensed data acquired over old-growth neotropical rainforests in the La Selva Biological Station, Costa Rica. Prediction models are developed based on a stepwise multiple linear regression analysis using the low dimensional features derived from the integrated data and field measured LAIs. Experimental results indicate that the best classification and prediction models are achieved when multi-sensor features are incorporated into the model. Experimental results also indicate that synergism between full waveform LiDAR and hyperspectral data is greater when vegetation structure is complex. Cite as: Author(s) (2013), Title, Abstract IN21A-1387 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  3. Increasing the Impact of High-Resolution Lidar Topography Through Online Data Access and Processing
    Crosby, C. and Nandigam, V. and Baru, C. and Arrowsmith, R.  AGU  December 2013

    Topography data acquired with lidar (light detection and ranging) technology are revolutionizing the way we study the Earth’s surface and overlying vegetation. These data, collected from satellite, airborne, tripod, or mobile-mounted scanners have emerged as a fundamental tool for research on topics including earthquake hazards, hillslope processes, and cyrosphere change. The U.S. National Science Foundation-funded OpenTopography (OT) Facility (http://www.opentopography.org) is a web-based system designed to democratize access to earth science-oriented lidar topography data. OT provides free, online access to lidar data in a number of forms, including the point cloud and associated geospatial-processing tools for customized analysis. The point cloud data are co-located with on-demand processing tools to generate digital elevation models, and derived products and visualizations which allow users to quickly access data in a format appropriate for their scientific application. The OT system is built using a service-oriented architecture (SOA) that leverages cyberinfrastructure resources at the San Diego Supercomputer Center at the University of California San Diego to allow users, regardless of expertise, to access these massive lidar datasets and derived raster data products for use in research and teaching. OT hosts over 600 billion lidar returns covering more than 120,000 km2. These data are provided by a variety of partners under joint agreements and memoranda of understanding with OT. Partners include national facilities such as the NSF-funded National Center for Airborne Lidar Mapping (NCALM), as well as non-governmental organizations and local, state, and federal agencies. OT has become a hub for high-resolution topography resources. Datasets hosted by other organizations, as well as lidar-specific software, can be registered into the OT catalog, providing users a “one-stop shop” for such information. OT is also a partner on the NASA Lidar Access System (NLAS) project, collaborative research funded by the NASA ACCESS program, that makes NASA airborne and space based laser altimetry data (GLAS and LVIS) available through OT using federated web services. With several thousand active users, OT is an excellent example of a cyberinfrastructure-based airborne science data system that is enabling access to challenging data for research, education and outreach. OT has demonstrated that by democratizing access to lidar topography, the impact of these expensive research datasets is greatly increased, through reused in research, education, and commercial applications beyond their original scope. This presentation will highlight the OT system and lessons learned during its development. We will also highlight ongoing work related to creation of a more flexible and scalable high-performance environment for processing of large datasets; creation of a “pluggable” infrastructure for third-party programs and algorithms to be run against the OT data holdings; and interoperability of OT with other earth science data systems. Cite as: Author(s) (2013), Title, Abstract IN31A-1499 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  4. Greenland Ice Sheet Mass Loss and Outlet Glacier Dynamics from Laser Altimetry Record (1993-2013)
    Csatho, B. and Schenk, A. and Duncan, K. and Babonis, G. and Sonntag, J. and Krabill, W. and van den Broeke, M. and van Angelen, J. and Blair, J. and Hofton, M.  AGU  December 2013

    Comprehensive monitoring of the Greenland Ice Sheet (GrIS) by satellite observations has revealed increasing mass loss since the late 1990s. Dynamic processes have been responsible for as much as half of this estimated loss, including ice flow adjustments to past climate variations and contemporary atmospheric and oceanic forcings. Dynamical processes act on different spatial and temporal scales and can cause non-linear changes, even on short, sub-decadal time scales. Quantitative investigation of these processes is imperative for improving ice sheet models and sea-level predictions. Our 1992-2011 altimetry record has shown that dynamic thinning substantially contributes to mass loss. The large spatial and temporal variations of dynamic mass loss and widespread intermittent thinning indicated the complexity of ice sheet response to climate forcing and points to the need of continuing monitoring of the GrIS at high spatial resolution. Airborne Topographic Mapper (ATM) and Laser Vegetation and Ice Sensor (LVIS) airborne laser altimetry measurements, acquired by NASA’s IceBridge mission, allowed us to extend the altimetry record to 2013. We generated a record of ice thickness and mass change of the GrIS spanning the period of 1992-2013, reconstructed at several thousand locations using the Surface Elevation Reconstruction and Change detection (SERAC) approach. Elevation changes are corrected for Glacial Isostatic Adjustment and partitioned into climate and ice dynamics induced components. We present the evolution of ice dynamics and climate induced mass loss of the major GrIS drainage basins in 2003-2013 to investigate their contributions to sea-level change. The detailed record of outlet glacier elevation change is consistent with the propagation of dynamic thinning or thickening initiated at lower elevations. We focus our attention to SE and NE Greenland. In SE Greenland we investigate if thinning continued on fast flowing SE Greenland glaciers (e.g., Koge Bugt, A.P. Bernstorff Glacier) after the brief period of thickening observed in 2006-2009. Increasing thinning would imply that earlier decrease in dynamic mass loss represented a short-term trend, rather than a new balance state of the ice sheet. In NE Greenland, continuing thinning at the grounding zones of Ryder, Zachariae and Nioghalvfjerdsfjorden glaciers could lead to retreat, flow acceleration and increasing mass loss as the ice plains buttressing these glaciers will become ungrounded in a region where the bed is under sea level. Cite as: Author(s) (2013), Title, Abstract C53D-02 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  5. The RISCO RapidIce Viewer: An application for monitoring the polar ice sheets with multi-resolution, multi-temporal, multi-sensor satellite imagery
    Herried, B. and Porter, C. and Morin, P. and Howat, I.  AGU  December 2013

    The Rapid Ice Sheet Change Observatory (RISCO) is a NASA-funded, inter-organizational collaboration created to provide a systematic framework for gathering, processing, analyzing, and distributing consistent satellite imagery of polar ice sheet change for Antarctica and Greenland. RISCO gathers observations over areas of rapid change and makes them easily accessible to investigators, media, and the general public. As opposed to existing data centers, which are structured to archive and distribute diverse types of raw data to end users with the specialized software and skills to analyze them, RISCO distributes processed georeferenced raster image data products in JPEG and GeoTIFF formats, making them immediately viewable in a browser-based application. Currently, the archive includes 16 sensors including: MODIS Terra, MODIS Aqua, MODIS Terra Bands 3-6-7, Landsat MSS, Landsat TM, Landsat ETM , Landsat 8 OLI, EO-1, SPOT, ASTER VNIR, Operation IceBridge ATM and LVIS, and commercial satellites such as WorldView-1, WorldView-2, QuickBird-2, GeoEye-1 and IKONOS. The RISCO RapidIce Viewer is a lightweight JavaScript application that provides an interface to viewing and downloading the satellite imagery from predefined areas-of-interest (or “subsets”), which are normally between 10,000 and 20,000 sq km. Users select a subset (from a map or drop-down) and the archive of individual granules is loaded in a thumbnail grid, sorted chronologically (newest first). For each thumbnail, users can choose to view a larger preview JPG, download a GeoTIFF, or be redirected back to the original data center to see the original imagery or view metadata. There are several options for filtering displayed including by sensor, by date range, by month, or by cloud cover. Last, users can select multiple images to play back as an animation. The RapidIce Viewer is an easy-to-use, software independent application for researchers to quickly monitor daily changes in ice sheets or download historical satellite data in one interface. Future plans include increasing the number of sensors (including radar), completing the historical archive, and adding more community-driven subsets. The web application can be found at http://www.rapidice.org/viewer. Cite as: Author(s) (2013), Title, Abstract IN34A-02 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  6. Evaluation of ICESat-1 Inter-Campaign Elevation Biases From the Comparison of Airborne and Space-based Lidar Data in Antarctica
    Hofton, M. and Luthcke, S. and Blair, J.  AGU  December 2013

    We estimate the Ice, Cloud, and land Elevation Satellite (ICESat-1) inter-campaign elevation biases from an analysis of ICESat-1 and LVIS (airborne lidar) intra- and inter-sensor elevation differences from two regions of the Antarctic ice sheet with minimal surface elevation change. The effects of various elevation corrections are investigated, including the ICESat-1 Gaussian-Centroid correction, firn densification and glacial isostatic adjustment. The inter-campaign biases determined from the two regions agree well. Relative to ICESat-1 campaign L3I, biases are less than ~8 cm, except for campaign L2E at ~15 cm. Campaign L2E (March 2009) occurred during a significant accumulation anomaly event in East Antarctica. We present trends estimated from our preferred bias solution, evaluate differences between the solutions, and estimate the impact on ICESat-1 derived Antarctica mass balance. For example, the trend estimated for L2A to L2F excluding L2E (Sep. 2003 to Oct. 2009) is 1.04 -0.48 cm/yr, and represents a correction to ICESat-1 derived Antarctica mass balance on the order of 117 -53Gt/yr. Cite as: Author(s) (2013), Title, Abstract C21D-0657 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  7. Greenland Ice sheet mass balance from satellite and airborne altimetry
    Khan, S. A. and Bevis, M. and Wahr, J. and Wouters, B. and Sasgen, I. and van Dam, T. and van den Broeke, M. and Hanna, E. and Huybrechts, P. and Kjaer, K. and others  AGU  December 2013

    Ice loss from the Greenland Ice Sheet (GrIS) is dominated by loss in the marginal areas. Dynamic induced ice loss and its associated ice surface lowering is often largest close to the glacier calving front and may vary from rates of tens of meters per years to a few meters per year over relatively short distances. Hence, high spatial resolution data are required to accurately estimate volume changes. Here, we estimate ice volume change rate of the Greenland ice sheet using data from Ice, Cloud and land Elevation Satellite (ICESat) laser altimeter during 2003-2009 and CryoSat-2 data during 2010-2012. To improve the volume change estimate we supplement the ICESat and CryoSat data with altimeter surveys from NASA's Airborne Topographic Mapper (ATM) during 2003-2012 and NASA’s Land, Vegetation and Ice Sensor (LVIS) during 2007-2012. The Airborne data are mainly concentrated along the ice margin and therefore significantly improve the estimate of the total volume change. Furthermore, we divide the GrIS into six major drainage basins and provide volume loss estimates during 2003-2006, 2006-2009 and 2009-2012 for each basin and separate between melt induced and dynamic ice loss. In order to separate dynamic ice loss from melt processes, we use SMB values from the Regional Atmospheric Climate Model (RACMO2) and SMB values from a positive degree day runoff retention model (Janssens & Huybrechts 2000, Hanna et al. 2011 JGR, updated for this study). Our results show increasing SMB ice loss over the last decade, while dynamic ice loss increased during 2003-2009, but has since been decreasing. Finally, we assess the estimated mass loss using GPS observations from stations located along the edge of the GrIS and measurements from the Gravity Recovery and Climate Experiment (GRACE) satellite gravity mission. Hanna, E., et al. (2011), Greenland Ice Sheet surface mass balance 1870 to 2010 based on Twentieth Century Reanalysis, and links with global climate forcing, J. Geophys. Res., 116, D24121 Janssens, I., and P. Huybrechts (2000). The treatment of meltwater retention in mass-balance parameterisations of the Greenland ice sheet. Annals of Glaciology 31, 133-140



  8. Spatial pattern of mass loss processes across the Greenland Ice Sheet from the Little Ice Age to 2010
    Kjaer, K. and Korsgaard, N. J. and Kjeldsen, K. K. and Bjork, A. and Khan, S. A. and Funder, S. and Nuth, C. and Larsen, N. K. and Vinther, B. and Andresen, C. S. and others  AGU  December 2013

    The Greenland Ice Sheet loses mass through surface meltwater runoff and discharge from marine terminating outlet glaciers. The spatial variability and magnitude of these processes have been studied and described in detail for the past decades. Here, we combine the mass loss between the LIA to 2010 with a SMB model extending back to ~1900 in order to investigate the spatial distribution of mass loss processes. We use high quality aerial stereo photogrammetric imagery recorded between 1978 and 1987 to map morphological features such as trim lines and end moraines marking the maximum ice extent of the LIA, which enables us to obtain vertical point-based differences associated with former ice extent. These point measurements are combined with contemporary ice surface differences derived using NASA's Airborne Topographic Mapper (ATM) from 2003-2010, NASA's Ice, Cloud, and land Elevation Satellite (ICESat) from 2003-2009, NASA's Land, Vegetation, and Ice Sensor (LVIS) from 2010, and ASTER (Silcast AST14DMO) co-registered to ICESat, to estimate mass loss throughout the 20th and early 21st Century. The mass balance estimates of the GrIS since retreat from maximum LIA is combined with a SMB model for the period for three intervals, LIAmax (~1900) - 1978/87, 1978/87 - 2003, and 2003 - 2010. Across the GrIS the total mass loss if found to be spatially- and temporally variable. However, when assessing the mass loss due to SMB and mass loss due to dynamic ice loss, we find that that the ratios between these components are variable between the different sectors of the GrIS, e.g. in the southeast sector of the GrIS we find substantial mass loss, possibly driven by high precipitation rates but also the presence of a large number of marine terminating glaciers. Furthermore many areas currently undergoing changes correspond to those that experienced considerable thinning throughout the 20th century. Consequently, comparing the 20th century thinning pattern to that of the last decade, and assuming a similar warming pattern, we argue that the present sensitivity distribution will hold also for future ice sheet mass loss until marine outlet glaciers become grounded. Cite as: Author(s) (2013), Title, Abstract C31C-03 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  9. Mass loss of the Greenland Ice Sheet since the Little Ice Age, implications on sea level
    Kjeldsen, K. K. and Kjaer, K. and Bjork, A. and Khan, S. A. and Korsgaard, N. J. and Larsen, N. K. and Long, A. J. and Woodroffe, S. and Milne, G. A. and Wahr, J. and others  AGU  December 2013

    The impact of mass loss from the Greenland Ice Sheet (GrIS) on 20th Century sea level rise (SLR) has long been subject to intense discussions. While globally distributed tide gauges suggest a global mean SLR of 15-20 cm, quantifying the separate components is of great concern - in particular for modeling sea level projections into the 21st Century. Estimates of the past GrIS contribution to SLR have been derived using a number of different approaches, e.g. surface mass balance (SMB) calculations combined with estimates of ice discharge found by in correlating SMB anomalies and calving rates. Here, we adopt a novel geometric approach to determine the post-Little Ice Age (LIA) mass loss of the GrIS. We use high quality aerial stereo photogrammetric imagery recorded between 1978 and 1987 to map morphological features such as trim lines (boundary between freshly eroded and non-eroded bedrock) and end moraines marking the ice extent of the LIA, which thereby enables us to obtain vertical point-based differences associated with changes in ice extent. These point measurements are combined with contemporary ice surface differences derived using NASA's Airborne Topographic Mapper (ATM) from 2002-2010, NASA's Ice, Cloud, and land Elevation Satellite (ICESat) from 2003-2009, and NASA's Land, Vegetation, and Ice Sensor (LVIS) from 2010, to estimate mass loss throughout the 20th and early 21st Century. We present mass balance estimates of the GrIS since retreat commence from the maximum extent of the LIA to 2010 derived for three intervals, LIAmax (1900) - 1978/87, 1978/87 - 2002, and 2002 - 2010. Results suggest that despite highly spatially- and temporally variable post-LIA mass loss, the total mass loss and thus the contribution from the GrIS to global SLR has accelerated significantly during the 20th Century. Cite as: Author(s) (2013), Title, Abstract C33A-0664 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  10. Greenland Ice Sheet Surface Elevation at Summit, Greenland: 2007-2013
    Lutz, E. and Hawley, R. and Herring, T.  AGU  December 2013

    Quantifying surface elevation change is essential for ice sheet mass balance estimates. We assessed surface elevation and elevation change of the Greenland Ice Sheet at a range of spatial scales using six years of monthly GPS surveys conducted near Summit between 8/2007 and 3/2013. The ~11 km route consists of 15 transects that run parallel and orthogonal to IceSat’s groundtrack 412 and includes 121 repeat locations spatially distributed along the route (Figure 1). Horizontal velocities and velocity gradients derived from base station and transect positions agree closely with previous studies. At the survey scale, no significant linear elevation trend is evident over the study period. However, local- and transect-scale time series revealed significant elevation increases of 1—2 cm per year in the central and southern regions of the survey that spatially and temporally may correlate with wind transport events from Summit station. This finding illustrates how sample scale (e.g., density, location and extent) affects surface elevation estimates critical to remote sensing validation and mass balance estimation. Spectral time series analysis showed that the expected annual elevation cycle was dwarfed by a two-year periodicity that dominated nearly all time series. The elevation maximum of Winter 2012-2013 fell short of the expected elevation peak, possibly due to accelerated compaction forced by high temperatures in the preceding Summer of 2012. We also highlight spatial comparisons with elevation products from three NASA altimeters, including the Airborne Topographic Mapper (ATM), the Land, Vegetation, and Ice Sensor (LVIS), and the Multiple Altimeter Beam Experiment Lidar (MABEL). This unique long-term GPS dataset is valuable for assessing ice sheet elevation change at a range of spatial-temporal scales, and for validating remote sensing products. With continued effort this survey will provide invaluable ground-based observations linking ICESat, IceBridge and ICESat-2 data products at Summit, Greenland.



  11. Rich Support for Heterogeneous Polar Data in RAMADDA
    McWhirter, J. and Crosby, C. and Griffith, P. and Khalsa, S. and Lazzara, M. and Weber, W.  AGU  December 2013

    Difficult to navigate environments, tenuous logistics, strange forms, deeply rooted cultures - these are all experiences shared by Polar scientist in the field as well as the developers of the underlying data management systems back in the office. Among the key data management challenges that Polar investigations present are the heterogeneity and complexity of data that are generated. Polar regions are intensely studied across many science domains through a variety of techniques - satellite and aircraft remote sensing, in-situ observation networks, modeling, sociological investigations, and extensive PI-driven field project data collection. While many data management efforts focus on large homogeneous collections of data targeting specific science domains (e.g., satellite, GPS, modeling), multi-disciplinary efforts that focus on Polar data need to be able to address a wide range of data formats, science domains and user communities. There is growing use of the RAMADDA (Repository for Archiving, Managing and Accessing Diverse Data) system to manage and provide services for Polar data. RAMADDA is a freely available extensible data repository framework that supports a wide range of data types and services to allow the creation, management, discovery and use of data and metadata. The broad range of capabilities provided by RAMADDA and its extensibility makes it well-suited as an archive solution for Polar data. RAMADDA can run in a number of diverse contexts - as a centralized archive, at local institutions, and can even run on an investigator's laptop in the field, providing in-situ metadata and data management services. We are actively developing archives and support for a number of Polar initiatives: - NASA-Arctic Boreal Vulnerability Experiment (ABoVE): ABoVE is a long-term multi-instrument field campaign that will make use of a wide range of data. We have developed an extensive ontology of program, project and site metadata in RAMADDA, in support of the ABoVE Science Definition Team and Project Office. See: http://above.nasa.gov - UNAVCO Terrestrial Laser Scanning (TLS): UNAVCO’s Polar program provides support for terrestrial laser scanning field projects. We are using RAMADDA to archive these field projects, with over 40 projects ingested to date. - NASA-IceBridge: As part of the NASA LiDAR Access System (NLAS) project, RAMADDA supports numerous airborne and satellite LiDAR data sets - GLAS, LVIS, ATM, Paris, McORDS, etc. - Antarctic Meteorological Research Center (AMRC): Satellite and surface observation network - Support for numerous other data from AON-ACADIS, Greenland GC-Net, NOAA-GMD, AmeriFlux, etc. In this talk we will discuss some of the challenges that Polar data brings to geoinformatics and describe the approaches we have taken to address these challenges in RAMADDA. Cite as: Author(s) (2013), Title, Abstract IN41C-1632 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  12. Lidar and Ground Assessment of Diversity, Wood Density, and Aboveground Biomass Along an Elevation Gradient in Tropical Montane Forest of Costa Rica
    Robinson, C. and Saatchi, S. and Clark, D. and Andelman, S. and Gillespie, T.  AGU  December 2013

    This research seeks to understand how tree diversity relates to three-dimensional vegetation structure along environmental gradients in the tropical montane forest of Braulio Carrillo National Park in Costa Rica. Elevation gradients along mountains provide landscape-size scales through which variations in topography and climatic conditions can be tested as drivers of biodiversity. In this study we report on the effectiveness of relating patterns of tree alpha diversity to three-dimensional structure of a tropical montane forest using remote sensing observations of forest structure. The study was utilized forest inventory and botanical data from nine 1-ha plots ranging from 100m-2800m above sea level and remote sensing data from an airborne lidar sensor (NASA’s Land, Vegetation, and Ice Sensor [LVIS]) to quantify variations in forest structure. In addition to calculating alpha diversity, we report on the variations in wood density with elevation, important for biomass and carbon estimations. Tree cores were analyzed for wood density and compared to existing database values for the same species, often collected only in the lowlands. In this manner we were able to test the effect of the gradient on effective wood density. Through the comparison to the lidar, our results show that there is a strong relationship between forest 3D structure and alpha diversity controlled by variations in abiotic factors along the elevational gradient. Using spatial analysis with the aid of remote sensing data, we found distinct patterns along the environmental gradients defining species composition. Wood density values with elevation change were found to vary significantly from database values for the same species. These wood density values are directly tied to biomass estimates, and it is possible that carbon storage has been overestimated along this gradient using prior methods. This variation in individual tree growth has repercussions on overall forest structure, as well as subsequent carbon estimates extrapolated from field measurements.



  13. Fusion of multi-sensor surface elevation data for a better characterization of rapidly changing outlet glaciers in Greenland
    Schenk, A. and Csatho, B. and McCormick, D. and Van der Veen, C.  AGU  December 2013

    During the last two decades surface elevation data have been gathered over the Greenland Ice Sheet (GrIS) from a variety of different sensors such as spaceborne and airborne laser altimetry (ICESat, ATM and LVIS) as well as from stereo imaging systems, most notably from ASTER and Worldview. The spatio-temporal resolution, the accuracy, and the spatial coverage of all these data differ widely. For example, laser altimetry systems are much more accurate than DEMs derived by correlation from imaging systems. On the other hand, DEMs usually have a superior spatial resolution and extended spatial coverage. We have developed the SERAC (Surface Elevation Reconstruction And Change detection) system to cope with the data complexity and the computation of elevation change histories. SERAC simultaneously determines the ice sheet surface shape and the time-series of elevation changes for surface patches whose size depends on the ruggedness of the surface and the point distribution of the sensors involved. By incorporating different sensors, SERAC is a true fusion system that generates the best plausible result (time-series of elevation changes)-a result that is better than the sum of its individual parts. We present detail examples of Kangerlussuaq and Helheim glaciers, involving ICESat, ATM and LVIS laser altimetry data, together with ASTER DEMs. ASTER DEMs are readily available but notorious for their accuracy behavior. The nominally stated accuracy of ~15 m may occasionally reach much higher values. By embedding ASTER DEMs into the SERAC time-series of elevation changes, we are able to determine plausible corrections. Thus, we can use ASTER DEMs to temporally and spatially densify the elevation change record. This is especially important on rapidly changing outlet glaciers where laser altimetry data might only be available sporadically To investigate the mechanisms controlling their behavior, we reconstructed elevation change histories along the central flowlines of these outlet glaciers. Elevation changes were partitioned into climate and ice dynamics related components using SMB estimates from RACMO2/GR. Our record shows decreasing thinning rates towards the ice sheet interior on Kangerlussuaq Glacier, consistent with the propagation of dynamic thinning initiated at lower elevation. Helheim Glacier exhibits a different behavior, as short term thinning was followed by rapid dynamic thickening, reaching rates of more than 10 m/yr. Kinematic wave modeling is used to interpret the thickness change record and to estimate the magnitude and timing of perturbation explaining the observed elevation changes. The relative importance of different forcing mechanisms, i.e., perturbations near the grounding line versus along the shear margin, was also evaluated. This study demonstrates how the fused elevation record allows a quantitative evaluation of ice sheet thickness changes by providing detailed and accurate information at or near the grounding line of very dynamic outlet glaciers. Cite as: Author(s) (2013), Title, Abstract C43C-0684 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  14. Analysis of the seasonal and interannual evolution of Jakobshavn Isbrae from 2010-2013 using high spatial/temporal resolution DEM and velocity data
    Shean, D. and Joughin, I. and Smith, B. and Moratto, Z. and Alexandrov, O. and Floricioiu, D. and Morin, P. and Porter, C. and Beyer, R. and Fong, T.  AGU  December 2013



  15. NASA's Operation IceBridge: using instrumented aircraft to bridge the observational gap between ICESat and ICESat-2 laser altimeter measurements
    Studinger, M.  AGU  December 2013

    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 between NASA'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 8 geophysical survey aircraft and 19 science instruments. All IceBridge data is freely available from NSIDC (http://nsidc.org/data/icebridge) 6 months after completion of a campaign. Cite as: Author(s) (2013), Title, Abstract C51A-0503 presented at 2013 Fall Meeting, AGU, San Francisco, Calif., 9-13 Dec.



  16. Forest Biomass Mapping from Stereo Imagery and Radar Data
    Sun, G. and Ni, W. and Zhang, Z.  AGU  December 2013

    Both InSAR and lidar data provide critical information on forest vertical structure, which are critical for regional mapping of biomass. However, the regional application of these data is limited by the availability and acquisition costs. Some researchers have demonstrated potentials of stereo imagery in the estimation of forest height. Most of these researches were conducted on aerial images or spaceborne images with very high resolutions (~0.5m). Space-born stereo imagers with global coverage such as ALOS/PRISM have coarser spatial resolutions (2-3m) to achieve wider swath. The features of stereo images are directly affected by resolutions and the approaches use by most of researchers need to be adjusted for stereo imagery with lower resolutions. This study concentrated on analyzing the features of point clouds synthesized from multi-view stereo imagery over forested areas. The small footprint lidar and lidar waveform data were used as references. The triplets of ALOS/PRISM data form three pairs (forward/nadir, backward/nadir and forward/backward) of stereo images. Each pair of the stereo images can be used to generate points (pixels) with 3D coordinates. By carefully co-register these points from three pairs of stereo images, a point cloud data was generated. The height of each point above ground surface was then calculated using DEM from National Elevation Dataset, USGS as the ground surface elevation. The height data were gridded into pixel of different sizes and the histograms of the points within a pixel were analyzed. The average height of the points within a pixel was used as the height of the pixel to generate a canopy height map. The results showed that the synergy of point clouds from different views were necessary, which increased the point density so the point cloud could detect the vertical structure of sparse and unclosed forests. The top layer of multi-layered forest could be captured but the dense forest prevented the stereo imagery to see through. The canopy height map exhibited spatial patterns of roads, forest edges and patches. The linear regression showed that the canopy height map had a good correlation with RH50 of LVIS data at 30m pixel size with a gain of 1.04, bias of 4.3m and R2 of 0.74 (Fig. 1). The canopy height map from PRISM and dual-pol PALSAR data were used together to map biomass in our study area near Howland, Maine, and the results were evaluated using biomass map generated from LVIS waveform data independently. The results showed that adding CHM from PRISM significantly improved biomass accuracy and raised the biomass saturation level of L-band SAR data in forest biomass mapping.



  17. Arctic sea ice freeboard from NASA's Airborne Topographic Mapper (ATM) and Land, Vegetation, and Ice Sensor (LVIS)
    Yi, D. and Harbeck, J. and Manizade, S. and Kurtz, N. and Studinger, M. and Hofton, M.  AGU  December 2013

    Data from an IceBridge Arctic campaign on April 20, 2010 with both the Airborne Topographic Mapper (ATM) and the Land, Vegetation, and Ice Sensor (LVIS) in operation were used in this study. The LVIS data were collected first on the outgoing track at a higher altitude. The ATM data were collected on the same track on the way back at a lower altitude. The transmitted and received ATM and LVIS lidar waveforms were fitted with Gaussian curves to calculate pulse width, peak location, pulse amplitude, and noise level. For each transmitted and received waveform, centroid, skewness, kurtosis, and pulse area were also calculated. Received waveform parameters, such as pulse width, pulse amplitude, pulse area, skewness, and kurtosis show geographically correlated patterns along an ATM or LVIS swath. These parameters, combined with elevation, were used to identify leads in ATM and LVIS sea ice freeboard calculation. The relationship between these parameters and sea ice freeboard and surface features were studied by comparing the parameters with ATM and LVIS derived freeboard and coincident Continuous Airborne Mapping By Optical Translator (CAMBOT) and Digital Mapping System (DMS) images which have been used to classify sea ice surface types such as leads, thin ice, grey ice and thick ice. A scan-angle-related elevation bias was found in the ATM data and an empirical correction (peak to peak is about 15 cm) as a function of scan angle is applied to the ATM elevations. The newly derived ATM freeboard is compared with the current ATM freeboard product at NSIDC. The ATM freeboards were also compared with the freeboard derived from LVIS data. Over the studied area, the mean freeboard for the ATM product at NSIDC is 0.535 m, for the ATM after empirical elevation correction is 0.551 m, and for LVIS is 0.512 m. The details of the differences of ATM and LVIS in flight altitude, footprint size, scan pattern, and their impact on waveform parameters and measured freeboard will also be discussed.



    18. Return to the top of the page