G54A-01 INVITED
Properties of L-band interferograms derived from ALOS/PALSAR radar observations
We present here L-band radar interferograms derived from ALOS/PALSAR satellite observations over several different Earth surface terrains. These InSAR images are representative of data that would be collected by the proposed DESDynI radar mission. We have examined interferograms over Hawaii, Greenland, Egypt, and California, which present surfaces covered by light vegetation, dense tropical forests, vegetation-free desert, and many ice facies. The interferograms exhibit very high correlation when compared to the C-band radar data commonly available from other existing satellites, yielding improved spatial coverage and reliable temporal sampling. In many cases the acquisitions are fully polarimetric, so that we can compare how the polarization state affects the statistics of the interferometric echoes. Specific results to date show that the use of L-band data rather than C-band permits much more comprehensive deformation modeling of rift events in Hawaii and of creep along faults, with little difference between interferograms formed from co- and cross-polarized returns. Ice from the wet snow, percolation, and dry snow zones in Greenland often correlates well (>50%) even with the PALSAR six-week repeat interval. Observable fringe patterns are seen in the Hawaiian rain forest in many locations over this same orbit interval. Interferograms derived from HH and VV data over ice and vegetation are very similar, and HV InSAR data are similarly correlated when allowance is made for the lower signal to noise ratio, even though considerable volume scatter occurs. Cross-interferograms show that the HH return is rather uncorrelated with the VV return, but that a small shift in the mean phase is present in HH-HV difference interferograms. This could be due to a real several cm shift in the phase center of the echo, but we are currently considering whether this apparent phase shift results from cross-polarized contamination of the radar return. With more frequent repeat observations than are possible with ALOS, but with an otherwise similar instrument, the principal science objectives for an InSAR mission as recommended by the National Research Council can be attained.
G54A-02
Assessing ScanSAR Interferometry for Deformation Studies
There is a trend in civil satellite SAR mission design to implement an imaging strategy that incorporates both stripmap mode and ScanSAR imaging. This represents a compromise between high resolution data collection and a desire for greater spatial coverage and more frequent revisit times. However, mixed mode imaging can greatly reduce the number of stripmap images available for measuring subtle ground deformation. Although ScanSAR-ScanSAR and ScanSAR-stripmap repeat-pass interferometry have been demonstrated, these approaches are infrequently used for single interferogram formation and nonexistent for InSAR time series analysis. For future mission design, e.g., a dedicated US InSAR mission, the effect of various ScanSAR system parameter choices on InSAR time series analysis also remains unexplored. Our objective is to determine the utility of ScanSAR differential interferometry. We will demonstrate the use of ScanSAR interferograms for several previous deformation studies: localized and broad-scale urban land subsidence, tunneling, volcanic surface movements and several examples associated with the seismic cycle. We also investigate the effect of various ScanSAR burst synchronization levels on our ability to detect and make quality measurements of deformation. To avoid the issues associated with Envisat ScanSAR burst alignment and to exploit a decade of InSAR measurements, we simulate ScanSAR data by bursting (throwing away range lines of) ERS-1/2 data. All the burst mode datasets are processed using a Modified SPECAN algorithm. To investigate the effects of burst misalignment, a number of cases with varying degrees of burst overlap are considered. In particular, we look at phase decorrelation as a function of percentage of burst overlap. Coherence clearly reduces as the percentage of overlap decreases and we find a useful threshold of 40-70% burst overlap depending on the study site. In order to get a more generalized understanding for different surface conditions, we consider several vastly different study sites. Phoenix, Arizona is an urban area which is located in an arid region with very little vegetation. C-band data over Phoenix is generally coherent over 5+ years. ERS data collected through the 1990s is used to monitor land subsidence in and around the Phoenix metropolitan area. We contrast these measurements with both broad and narrow deformation features in the vegetated Houston, Texas and London, U.K. areas. We find that low resolution ScanSAR data can be used to detect narrow features with small spatial extent. Several additional interferograms demonstrate the general applicability of C-band ScanSAR interferometry to WInSAR community interests, e.g., the Hector Mine earthquake, aseismic fault motion and Long Valley and Yellowstone deformation over time. With the September 2006 implementation of a new burst synchronization strategy for Envisat, 90% of all ScanSAR acquisitions exhibit at least 50% burst overlap. Our results demonstrate that these new data can be successfully used for a number of InSAR applications.
G54A-03
Multi-beam Lidar Instrument Design, Measurement Capabilities, and Technical Readiness
A multi-beam Laser Altimeter has been designed and studied at NASA Goddard Space Flight Center and the critical technologies have developed and tested resulting in a mature and technically ready instrument approach. The instrument consists of three separate beams each providing a near-contiguous profile of 25 m diameter laser footprints. The across-track separation of the beams can be <1 km to as much as ~5 km. A return waveform is collected for each footprint and the system has sufficient Signal-to-Noise Ratio (SNR) to penetrate dense (i.e. 98-99% cover) canopies in relatively clear sky conditions and can penetrate clouds to provide precise topography over unvegetated surfaces. The Multi-beam Laser Altimeter design includes a high-quality GPS receiver for providing precise orbital position information and a state-of-the-art Star Tracker and Inertial Measurement Unit to provide precise and accurate laser beam pointing knowledge. The digitizer-based ranging system will provide ranging to bare surfaces with ~3 cm range precision. The return waveforms will also provide vegetation height measurements with ~1 m of accuracy. The laser transmitters have been fully developed, characterize, and tested. Engineering Test Unit has been built at NASA/GSFC for environmental testing. Test units of the laser demonstrated 5 Billion shots without damage and diode testing indicates lifetimes of ~10 Billion shots per laser can be expected. A new waveform digitizer has been developed with improvements in sampling rate and dynamic range over the ICESat digitizer system, thus allowing higher quality waveforms to be collected, which is critical importantly for vegetation studies. This multi-beam Lidar design is the basis for the Lidar on the DESDynI (Deformation, Ecosystem Structure, and Dynamics of Ice) mission described in the NRC Decadal Survey Report that consists of a Multi-beam Lidar and an L-band InSAR to be launched in the 2010-2013 timeframe.
G54A-04
Laser Altimeter Mission Design Constraints from ICESat Observations of 1064 nm Apparent Surface Reflectance
Performance of laser altimeters, including the probability of detecting a return from the surface and the precision of the resulting range measurement, depends on instrumental and environmental parameters. Environmental conditions can be expressed as apparent surface reflectance at the laser wavelength, combining atmospheric transmission (a function of cloud cover and aerosols) and surface "hot spot" retro-reflectance (i.e., at 0 degrees phase angle, with parallel illumination and view angles). We are using data acquired by the Geoscience Laser Altimeter System (GLAS) on the Ice, Cloud and land Elevation Satellite (ICESat) to establish probability distribution functions for 1064 nm apparent surface reflectance. GLAS was designed to range to the Earth's surface through thin to moderately dense clouds and aerosols, up to an optical depth of 2 (13% one way transmission). Its apparent surface reflectance data, therefore, provides valuable information relevant to the design of future orbital laser altimeters such as those recommended by the National Research Council in their Earth Science Decadal Survey report to NASA and NOAA (the ICESat-II, DESDynI and LIST missions). We are characterizing GLAS-observed apparent surface reflectance as it varies geographically, seasonally, and with the time of day, correcting for several instrumental artifacts (detector response non-linearity, saturation, and field-of- view shadowing). We will present statistics on apparent reflectance and the number of laser pulses that do not yield a surface return due to dense clouds and/or aerosols for the various ICESat observation periods for selected regions with different atmospheric and land cover characteristics. The selected regions include areas where laser altimeter observations of the surface are limited by frequent cloud cover, including polar latitudes important for ice sheet mass balance purposes and equatorial latitudes important for quantifying biomass storage and fluxes in tropical rain forests. This fundamental data on the measurement environment can be used in trade studies of laser altimeter mission designs, operating in the near-infrared, and to quantify expected performance attributes such as the spatial distribution, density and accuracy of surface returns accumulated through time.
G54A-05
Feasibility study on the use of ICESat laser altimetry data for the study of Himalyan mountain glacier mass balance.
We report on a feasibility study on using satellite-based laser altimetry products for profiling selected Himalayan mountain glaciers. ICESat (Ice Clouds and Elevation satellite) is a NASA satellite capable of providing cm-scale elevation measurements. These data have been successfully used for profiling the polar ice sheets (Antarctica and Greenland). Here we report on the feasibility of using this instrument on Himalayan glaciers. There are significant limitations on using ICESat in steep mountain terrain at low latitudes. The laser reflection has poorer accuracy over sloping terrain than over the relatively flat ice sheets. The laser footprint of 70m is large compared to some mountains glaciers, but small for most ice-sheet features. The repeat interval and the pass spacings are large at low latitudes, compared to high latitudes. Finally, ICESat mission length has been compromised by laser failures, reducing the number of passes over the study area. Nevertheless, we identified eight of the largest Himalayan glaciers and extracted the available data. We also extracted data on a number of unnamed glaciers which had good ICESat coverage. These data cover a time range of four years allowing us to assess the stability and repeatability of the data as well as providing some information on mass change of these targets. We conclude that this technique has potential if ICESat were healthy and could be aimed more precisely for mountain glacier work. In addition, considerable ground-truthing by glaciological and geological parties needs to be conducted.
G54A-06 INVITED
Scenarios for Estimating Vegetation Structure and Biomass from InSAR with DESDynI
Two approaches have been used to estimate the vertical stucture of vegetation from InSAR: 1) Use physical models to estimate structural parameters from a set of InSAR observations, with baseline, polarization, or frequency diversity; or 2) apply external calibration to single-observation InSAR. DESDynI will have some baseline and polarization diversity. It will also have lidar and will benefit from calibrations from other missions such as TanDEM-X. DESDynI will therefore be capable of each of the above approaches, as well hybrids involving both approaches. We will show historical and current results of InSAR structure demonstrations of both types, with emphasis on those at L-band, DESDynI's radar frequency. We will propose candidate observation scenarios for DESDynI and use both airborne and spaceborne results to evaluate them. We will highlight the types of observations needed to better specify the performance of the range of possible scenarios. Biomass estimation processes from each scenario will also be described and evaluated.
G54A-07
Quantifying INSAR Temporal Decorrelation and its impact on estimation of vegetation structure
The National Research Council was commissioned by NASA to conduct a decadal survey to assess the strategy NASA should take for Earth science and applications from Space. One of the recommendations of this report is for a mission called 'DESDynI', consisting of an L-band SAR and a laser altimeter, to measure surface and ice sheet deformation, and to measure vegetation structure for ecosystem health. Measuring vegetation structure with an L-band repeat-pass InSAR mission requires that temporal decorrelation between the SAR observations be well understood. In the worst case, temporal decorrelation would exceed the volumetric decorrelation (which may be modeled as a function of the vegetation structure), making accurate vegetation height inversion impossible from the InSAR data. In the best case, non-negligible temporal decorrelation would bias the estimation. In order to quantify InSAR temporal decorrelation over forested areas, the NASA/JPL AIRSAR SAR conducted a comprehensive series of InSAR flightlines in 2004 over the La Selva Biological Reserve, Costa Rica (a mature, well-characterized tropical rainforest), collecting both single-pass and repeat-pass InSAR data at a variety of baselines and temporal separations (0 meters to 200 meters, 20 minutes to 2 weeks). Results of this analysis will be presented. The L-band ALOS PALSAR SAR, launched in 2006, and currently collecting data sets globally, has a 46 day repeat orbit and baselines less than 1km. Data from this mission, as well as from the 1978 Seasat SAR (which had a 3 day repeat for much of the mission) should likewise provide important characterizations of temporal decorrelation over forested regions. Possible observation strategies for DESDynI may have to accommodate temporal decorrelation effects through more frequent observations of areas subject to weather-induced decorrelation, similar to global observation strategies for optical sensors.
G54A-08 INVITED
Estimation of Carbon Stocks, Fluxes and Dynamics using Lidar and Ecosystem Models