1. Global Ice Sheet Interferometric Radar NASA ESTO IIP Ohio State Univ., JPL, Univ. Kansas, VEXCEL Corp., E.G&G Corp., Wallops Flight Facility

2. GISIR Instrument GISIR mid-year review –Science –Spaceborne Instrument Concept –Strawman Design –Simulations –Aircraft Experiment –Management GISIR Team –K.Jezek, E. Rodriguez, P. Gogineni, J. Curlander, X. Wu, J. Sonntag, W. Krabill, P. Kanagaratnam, C.Allen, A. Freeman. T. Akins –D. MacAyeal, R. Forster, S. Tulazek, M. Fahnestock, S. Clifford

3. Science

4. Glaciers and Ice Sheets ‘Grand Challenges’ Understand the polar ice sheets sufficiently to predict their response to global climate change and their contribution global sea level rise What is the mass balance of the polar ice sheets? How will the mass balance change in the future?

5. Reservoirs of Fresh Water Ice Thickness Average: 2500m Maximum: 4500m Fresh Water Resource Polar Ice Sheets and Glaciers 77% East Antarctica 80% West Antarctica 11% Greenland 8% Glaciers 1% National Geographic Magazine

6. Glacier/Ice Shelf Change In last 4 years, numerous observations of glacier/ice shelf reduction –Interior Antarctic Peninsula glaciers (Scambos, 2004) –Larsen ice shelf (many) –Jacobshaven glacier in Greenland (Thomas, 2003) –Pine Island and Thwaites (Rignot, 2001) None of these observations are predicted by climate models 0.8% decrease in Ice Shelf extent between 1963 and 1997 1963-1997 Kim and Jezek

7. Mass Balance Ice sheet mass balance is described by the mass continuity equation Altimeters InSAR Act/Pass. Microwave No spaceborne technique available Evaluations of the left and right hand sides of the equation will yield a far more complete result

8. Ice Dynamics and Prediction Terms related to gradients in ice velocity (InSAR) integrated over thickness Understanding dynamics coupled with the continuity equations yields predictions on future changes in mass balance Force Balance Equations Basal Drag, Inferred at best Satellite Altimetry No Sat. Cover

9. Glaciers and Ice Sheets Mapping Orbiter Key Measurements: –Determine total global volume of ice in glaciers and ice sheets –Map the basal topography of Antarctica and Greenland –Determine basal boundary conditions from radar reflectivity –Map internal structures (bottom crevasses, buried moraine bands, brine infiltration layers)

10. Key Scientific Drivers –Ice thickness (with other data) needed to map derived stresses and stress gradients: force balance equations –Ice thickness to estimate mass balance when combined with interferometrically derived ice velocity –Internal layering of the ice sheet to unravel climatic history of the polar ice sheet –Ice thickness to predict response of glaciers and ice sheets to changing climate and the impact of changing ice volume on global sea level –Contribute technologies to planetary studies; understanding the phenomenology of radar sounding Glaciers and Ice Sheets Mapping Orbiter

11. Global Ice Sheet Mapping Orbiter Prioritized Science Requirements Measure ice thickness to an accuracy of 20 m or better Measure ice thickness every 1x1 km (airborne < 250 m) Measure ice thickness ranging from 100 m to 5 km Measure radar reflectivity from basal interfaces (rel. 2 dB) Obtain swath data for full 3-d mapping Measure internal layers to about 20 m elevation accuracy Pole to pole observations; ice divides to ice terminus One time only measurement of ice thickness Repeat every 5-10 years for changes in basal properties

12. GISMO Concept

13. A New Technical Approach Required Nadir sounding ‘profiler’ cannot meet science requirements: –Beam limited cross-track spatial resolution (1km) requires antenna size beyond current capabilities (420 m at P-band) –Available bandwidth (and range resolution) insufficient for desired 10m height accuracy: 6Hz gives a range resolution of 13.8 m in ice. Worse at VHF –Full spatial coverage requires years of mission lifetime (high costs)

14. Other Conventional Approaches Limited Conventional Interferometry is Insufficient –Coverage, spatial resolution, and height accuracy suggest a swath SAR interferometer might meet concept –Ambiguous returns from surface clutter and the opposite side basal layer make this approach not feasible An alternate approach: multiple baselines to resolve subsurface/surface, expensive and hard to implement: –Minimum of 3 antennas are required to get necessary baselines –The technique is sensitive to calibration and to SNR –Data rate is high

15. Primary Technical Challenge Separate basal return from surface clutter Surface Clutter Weak Echos Strong Attenuation

16. Interferometric Sounding Concept Additional information contained in the spatial frequency of the phase: –Because of the difference in incidence angles, the near nadir interferometric phase spatial frequency from the basal return is much larger than the equivalent frequency for surface clutter –Opposite side ambiguities have opposite interferometric frequencies: while the phase in one side increases with range, it decreases with range in the opposite side (+/- spatial frequencies of complex interferogram) Conventional interferometry uses phase information one pixel at a time IFSAR sounding concept: spatially filter interferogram to retain only basal returns from one side Satellite height (H); ice surface height (h); Depth of the basal layer (D); topographic variations of the basal layer (d); cross- track coordinate of the basal layer point under observation (x b ); and, x s is the cross-track coordinate of the surface point whose two-way travel time is the same as the two-way travel time for x b.

17. Interferometric Phase and Phase Frequency Surface interferometric phase difference Basal interferometric phase difference Surface interferogram slope as a function of range. h x is the surface topography cross-track slope Basal Interferogram slope as a function of range. d x is the basal topography cross-track slope Complex interferogram (From E. Rodriguez, 2004)

18. Surface vs Basal Cross-Track Distance Surface cross-track distance x s as a function of basal cross-track distance x b for a platform height of 600km. Notice that near nadir x s is nearly independent of x b, while for x b > 50 km, the two are close to being linearly dependent.

19. Fringe Spectrum Interferogram spectra for first 50 km of x b - signal to clutter ratio - 1 - radar freq - 430 MHz - bandwidth of 6 MHz Basal spectrum is colored orange Surface spectra for - D = 1 km (black), - D = 2 km (red), - D = 3 km (green), - D = 4 km (blue). Note the basal fringe spectrum depends very weakly on depth

20. Clutter Effect on Interferometric Error Height error as a function of signal to clutter (  ) for a baseline of 45 m and a center frequency of 430~MHz (solid lines) and 130 MHz (dashed lines). The values of (  ) are: 0 dB (black), -10 dB (red), and -20 dB (blue).

21. Height Noise vs SNR Height error as a function of SNR for a baseline of 45m and a center frequency of 430 MHz (solid lines) and 130 MHz (dashed lines). The values of SNR are: 0 dB (blue), -5 dB (red), and -10 dB (black). The number of looks is assumed to be 100.

22. Height Error After Interferometric Filtering: 45m Baseline Assumed clutter to signal ratio (same side): 0dB Assumed clutter to signal ratio (opposite side): -3dB Assumed SNR: 0dB Assumed number of spatial looks: 100

23. Strawman Design

24. Desired Swath/Altitude A 50 km swath will enable a mission lifetime < 4 months A 50 km swath is consistent with the interferometric sounder approach for a height of 600 km and a 6 MHz bandwidth Heights lower than 600 km will require a decrease in swath and a consequent increase in mission lifetime (although this is not a strong constraint)

25. Ground Resolution vs Bandwidth

26. Radar Frequency Selection Select P-band (430MHz) to minimize baseline length, antenna size and range resolution –Antenna length of 12.5 m consistent with demonstrated space technology (TRL 9) –Baseline range between 30m to 60m consistent with SRTM mast (TRL9) –6MHz bandwidth possible at P-band –Higher clutter mitigated by interferometric sounder technique VHF only feasible using repeat pass interferometry, although antenna size (~40m) may be problematic. –Can 1 MHz remote sensing band be increased over the polar regions?

27. Pulse Length Selection In order to minimize contamination from nadir surface return, use short 20usec chirp –surface nadir return sidelobes stops after 1.7 km of ice depth Much longer pulses will contaminate basal returns significantly

28. PRF Selection Required PRF for SAR synthetic aperture: ~1KHz Use significantly higher PRF (~7KHz - 10KHz) together with onboard presum to improve signal to noise ratio Instrument duty cycle: ~20%

29. Mission Concept Two antennas, 45 m baseline, off-nadir boresight - 1.5 degrees –mesh dishes, SRTM-like boom, 50 km swath from 10 to 60 km cross-track –use conventional nadir sounding for layering studies Fully polarimetric for ionospheric effects 600 km altitude, 1 year minimum mission lifetime P-band (430 MHz), 6 MHz bandwidth attenuation is essentially same at from 100 MHz to 500 MHz along-track resolution from SAR processing cross-track resolution from pulse bandwidth

30. Key Instrument Parameters Polarimetric4 channels Center frequency 430 MHz Bandwidth 6 MHz Pulse length 20  s Peak transmit power 5 kW System losses -3 dB Receiver noise figure 4 dB Platform height 600 km Azimuth resolution 7 m PRF 10 kHz Duty cycle 20 % Antenna length 12.5 m Antenna efficiency -2. dB 1-way Antenna boresight angle 1.5 deg Wavelength 0.7 m Baseline 45 m Swath 50 km Minimum number of looks 500

31. Estimated Performance

32. Assumed Antenna Pattern Assume uniform circular illumination Antenna diameter 12.5m –consistent with available space qualified antennas Antenna boresight: 1.5deg Assumed antenna efficiency: -2dB 1-way

33. Model Backscatter Cross Section The backscatter model consists of two contributions: –Geometrical optics (surface RMS slope dependent) –Lambertian scattering assumed Lambertian contribution at nadir: - 25dB

34. Observed Surface Sigma0 Angular Dependence at 120 MHz Data obtained with the JPL Europa Testbed Sounder in deployment with the Kansas U. sounder over Greenland Angular decay near nadir (>15 dB in 5 degrees) consistent with very smooth ice surface Change in behavior at P- band is still unknown, but probably bounded by 1-3 degree slope models

35. Estimated One-Way Absorption Absorption losses present severe constraints on system design if an SNR > 0dB is desired. Together with pulse length considerations, this leads to a desired peak power of ~5kW, which is technologically feasible at P-Band

36. Sounder Performance 1 km Ice

37. Sounder Performance 2 km Ice

38. Sounder Performance 3 km Ice

39. Sounder Performance 4 km Ice

40. Spaceborne Simulations

41. Phase History Simulation DEM for surface and basal region from Greenland Assume homogeneous ice volume –permittivity 3.4 for ice, 9 for bedrock –intermediate layers are weakly scattering at off-nadir angles Attenuation 9 dB/km Ice thickness 2.0 to 2.5 km Process: create reflectivity map of surface and subsurface Construct phase history data (PHD) using inverse chirp scaling Process PHD with COTS SAR processor and interfeometric processor to 80 looks

42. Key instrument and geometry parameters Platform Height: 600 km Center Frequency: 430 MHz Chirp Bandwidth: 6 MHz Pulse Length: 20 us PRF: 2 kHz Antenna Length: 12.5 m Antenna Boresight Angle: 1.5 o Baseline: 45 m

43. DEMS in slant range geometry echo delay caused by the ice thickness at nadir (a) surface DEM (b) basal DEM 148.8 km (ground range) 148.8 km 137.5 km

44. Amplitude images of the phase history data

45. Amplitude images of the SLC data

46. +2500 m +2137 m +2850 m +1714 m surface and basal interferogram Band-pass filtered interferogram original ice thickness Derived ice thickness (expected performance reduction pass 50 km) Images are 130 km vertical (azimuth and 70 km ground range (11.9 km slant range) Simulation Results for GISMO Error is 0 to 20 m out to 50 Km

47. Interferogram range spectrum The peak near 0 frequency represents the surface contribution and the peaks at the right side are from base contribution.

48. Airborne SAR processor Modify Vexcel’s current fast-back- projection space-borne spotlight SAR processor to be able to process the simulated airborne stripMap SAR data

49. Modifications for airborne simulation Airborne platform with air turbulence Varying PRF Interferometric mode with 2+ receiving antennas Inverse chirp-scaling modifications for varying PRF and sensor velocity Airborne SAR processor

50. GISIR Airborne Radar Experiment

51. Scaling GISMO to Aircraft Altitudes Preserve the fringe rate separations: Find that basal distance for comparable fringe separations scale as the square root of the platform hieghts: 600km elevation satellite imaging a swath from 10 km to 60 km corresponds to aircraft at 6 km imaging 1 to 6 km. Preserve number of fringes: imposes a restriction on the baseline. 45 m spaceborne baseline means a 4.5 m baseline on aircraft. 10 to 20 m better. Maintain geometric correlation: Limiting phase change over a resolution cell imposes a restriction on bandwidth. Find 45-90 MHz depending on baseline. Larger bandwidth also preserves number of samples per swath Achievable with P-3; on the edge with Twin Otter

52. Airborne Radar Depth Sounder: Rely on Design Heritage

53. Radars for operation on P-3 or Twin Otter aircraft Three more antenna elements are being added on each wing of a Twin Otter to eliminate grating lobes

54. GISIR System Block Diagram Return loss of a single antenna element. Sufficient bandwidth at 150 and 450 MHz

55. Preliminary Conclusions (1) The interferometric sounding concept has the potential to make the desired ice sheet depth measurement for depths > 1km For depths < 1km and for ice layering, conventional sounding approach may be sufficient –Results depend on P-band clutter roll-off

56. Preliminary Conclusions (2) At smaller depths (~1km) surface clutter contamination dominates: must be able to show clutter rejection for 10dB clutter to signal ratios for highest attenuation rates For other depths, clutter to signal ratios are much better and can be removed using IFSAR filtering At ~4 km depths, SNR is poor but sufficient to meet height accuracies

57. There are no major technological hurdles in the design, although the power requirements are likely to be high (for deep ice) Airborne mission is planned for 2006 with multiple antennas Preliminary Conclusions (3)

58. Project Management

59. GISIR Management Chart

60. GISIR Schedule – January, 2006

62. GISMO Project Linkages

63. Infusion Opportunities International Polar Year IPY Umbrella Proposal: Global Inter-agency IPY Polar Snapshot Year (GIIPSY) The goal of this proposal is to develop the most effective mechanism by which to plan and synchronise IPY satellite acquisition data requests Represents scientists in ESA, US, Canada, China, Korea, Australia, Japan, Norway, Denmark, UK, Germany, Itaty IPY Investigator Proposal: A proposed international campaign to map the interior and base of the polar ice sheets Campaign for Airborne Sounding of Ice Sheets (merged with ASAID) Airborne instrumentation development and data acquistion Glaciodyn Project (Devon Island overflight in 2007) NRC Decadal Survey RFI Response: Glaciers and Ice Sheets Mapping Orbiter (GISMO) Captain Ashley McKinley holding the first aerial surveying camera used in Antarctica. It was mounted in the aircraft ‘Floyd Bennett’ during Byrd’s historic flight to the pole in 1929. (Photo from The Ohio State University Archives)

64. CARISMA: P-band Nadir Sounding Potential for collaboration with GISMO – Flight of opportunity?? Infusion Opportunities: Carbon Cycle and Ice Sheet Mapping Mission ESA earth explorer proposal Flight Direction

65. IIP Potential Collaborations Keith Raney, Applied Physics Lab Pathfinder Advance Radar Ice Sounder (PARIS) IIP project. Delwyn Moller, JPL – Ka-band Radar IIP Project

66. GISMO Web Connection Web site –www-bprc.mps.ohio-state.edu/rsl/gismo/ FTP Site –ftp-bprc.mps.ohio-state.edu –GISMO transfers under /rsl

67. GISIR Publications Presentations Jezek, K.C., E. Rodriguez, P. Gogineni, A. Freeman, J. Curlander, X. Wu, J. Paden and C. Allen, in press. Glaciers and Ice Sheets Mapping Orbiter Concept. In press Journal of Geophysical Research - Planets. Rodriguez, E., and Wu, X. Interferometric ice sounding: a novel method for clutter rejection, in JPL review, to be submitted to IEEE Trans. Geoscience and Remote Sensing, 2005. Jezek, K.C., The Future of Cryospheric Research. Byrd Polar Research Center Colloquy, The Ohio State University, Columbus, Ohio, October, 2005. Jezek, K.C., Spaceborne Observations of Antartica: The RAMP and GISMO Project. USRA Lecture Series, University of Alaska, Fairbanks, October 2005. Curlander, J., K. Jezek, E. Rodriguez, A. Freeman, P. Gogineni, J. Paden, C. Allen, X.Wu. Glaciers and Ice Sheets Mapping Orbiter. Advanced SAR Workshop, Montreal, Quebec, 2005. Forster, R., Jezek, K., E. Rodriguez, S. Gogineni, A. Freeman, J. Curlander, X. Wu, C. Allen, P. Kanagaratnam, J. Sonntag, W. Krabill, "Global Ice Sheet Mapping Orbiter", FRINGE 2005 Workshop, Advances in SAR Interferometry from ENVISAT and ERS missions”, European Space Agency, ESRIN Frascati, Italy 28 November - 2 December 2005 Jezek, K., E Rodriguez, P. Gogineni, A. Freeman, J. Curlander, X. Wu, C. Allen, P. Kanagaratnam, J. Sonntag, and W. Krabill. Glaciers and Ice Sheet Mapping Orbieter, EOS Trans., AGU, 86(52, Fall Meet. Suppl., Abstract IN13B-1092.

68. Key objectives for next 6 months Complete Vexcel airborne simulations Complete KU RF/digital design and system testing Complete first airborne campaign planning document Confirm aircraft availability Complete JPL motion compensation programs and back-propagation processor

69. Backup Slides

70. What is the Effect of Surface Slopes? Spectral broadening due to surface slopes Relative spectral broadening is much less than the spectral peak, as long as the angle of incidence is small: changes in the angle of incidence are dominated by changes in cross- track distance rather than changes in slope. This argument does not hold for higher incidence angles.

71. SAR Tomography Potentials of GISMO SAR tomography background SAR tomography simulation Results of E-SAR tomography tests GISMO potentials for SAR tomography applications

72. SAR Tomography Background Conventional SAR Imaging Ground Reference Point Idealized Straight Flight Paths

73. Multi-Pass SAR Imaging Synthetic Elevation Aperture Ground Reference Point Synthetic Aperture