Some Recent Developments in Remote Sensing of Ice Sheets Kenneth Jezek The Ohio State University.

Some Recent Developments in Remote

Sensing of Ice SheetsKenneth Jezek

The Ohio State University

Motivation

Limited capability to capture 3-d detail of glacier bed and internal structure determine ice sheet internal temperatures

Science Drivers Internal temperature influences stiffness, which influences stress-strain

relationship and therefore ice deformation and motion Bed geometry is a strong control on ice flow

Research Goals Measure ice sheet internal temperatures remotely (radiometry) Image buried landscape as if the ice sheet were stripped away (radar) Investigate advantages of combined active and passive measurements

Physics of the Problem

Principle of Tomographic Radar Sounding

• Received signals

at each sensor:

xi: received signal of sensor i; k: 4/; di: distance of sensor i; i: arrival angle; p: number of sensors; si: signal; q: number of signals; ni: noise;

H

D

air

ice

base

Invert matrix to produce 3-d images of subsurface geometry

Emission Physics

• In absence of scattering, thermal emission from ice sheet can be treated as a 0th order radiative transfer process

• Similar to emission from the atmosphere: temperature profiling possible if strong variations in extinction with frequency (i.e. absorption line resonance)

• Ice sheet has no absorption line but extinction does vary with frequency

• Motivates investigating brightness temperatures as function of frequency

5

Progress in Radiometry

SMAP, 2015

Evidence from SMOS

1

SMOS data over Lake Vostok (East Antarctic Plateau)

55° 25°

The analysis of SMOS data point out a relationship between Tb and Ice Thickness

Brogioni, Macelloni, Montomoli, and Jezek, 2015

500 MHz Model results suggest Tb sensitivity at depth and dependence on presence of subglacial water

Jezek and others, 2015

Modeled UHF Behavior for Antarctic

Greenland Brightness Temperatures Cloud Model, SMOS, SMAP

• Cloud model Tb estimate based on temperature profiles derived from OIB thickness, CISM heat flux, RACMO SMB, MODIS surface temp. Parameter then corrected to match CC, NGRIP, GRIP temps.

• 1.4 GHz data forced to align with SMOS data (black) using a constant multiplier. Same multiplier applied to other frequencies.

• Variations are small at 1.4 GHz along flight path because temperature profiles are more uniform in depth. 500 MHz anomaly associated with region of assigned basal melt

240 250 260 270 280

0

500

1000

1500

2000

2500

3000

3500

Temperature (K)

Dep

th B

elow

Sur

face

(m)

0

0.5

1

1.5

2

x 109

0 0.5 1 1.5 2x 106

Range

Spectrum Waterfall

Rang

e

225

230

235

240

0.5 GHz B Model1.0 R “”1.4 C “”2.0 G “”SMOS Bla (thick) (Jan. 2014)SMAP B (thick) (April, 2015)

Oswald and Gogineni, Subsurface Water Map

Freq

uenc

y

Antarctica-Greenland Brightness Temp vs. Frequency

• Antarctic geophysical cases: low accumulation rates result in temp profiles that increase with depth

• Strong changes in TB vs. frequency

• Higher accumulation rates in Greenland (at least for GISP site) result in more uniform temp profile vs. depth

• Smaller changes in TB vs. frequency

• Need instrumentthat can capturethese variations

0 1000 2000 3000

150

200

250

Tb(

K)

frequency(Hz)

1585 Simulated Tb vs Freq profiles

Antarctica

Greenland(GISP) Greenland

(GISP)

Blue: With AntennaRed: Without Antenna

Blue: Simulated ProfilesRed: GISP Data

Additional Factors

• Layering is important. At present, include statistical model of density with depth (Gaussian variability with a defined correlation length)

• Model layering effects using coherent and partially coherent radiative transfer models

• Working on interface roughness

• Yet to include layer conductivity at depth

0 50 100200

210

220

230

2 GHz Tb: ice sheetwith upper low density strata

.1 m layer thickness

Total Strata Thickness (m)Tb

Figure 7. Brightness temperature for changing the total thickness of the near surface low density layer. The discrete layer thickness is constant at 0.1 m. Variability is a consequence of recomputing the density function for each calculation. The red curve shows only the effect so the subsurface layers. The blue cure includes the loss at the air snow interface.

0.2 0.4 0.6 0.8 1

0

20

40

60

80

100

Combined Drinkwater and random density

Density (gm/cc)

Dep

th (m

)

Forward Model Assessment• Used “Dome-C”-type physical parameters

• Including density fluctuations with correlation length parameter

• Results show:• Coherent effects can be significant if density correlation length << wavelength;

otherwise good agreement between models

0.5 1 1.5 2150

200

250l = 3cm

Brig

htne

ss T

empe

ratu

re (

K)

0.5 1 1.5 2150

200

250l = 5cm

0.5 1 1.5 2200

220

240

260l = 10cm

frequency (GHz)Brig

htne

ss T

empe

ratu

re (

K)

0.5 1 1.5 2200

220

240

260l = 40cm

frequency (GHz)

Cloud

DMRT/MEMLS

Coherent

(Tan and others, 2015)

Greenland Retrieval Studies

• Generated simulated 0.5-2 GHz observations of “GISP-like” ice sheets for varying physical properties (500 “truth” cases)

• Including averaging over density fluctuations

• For each truth case, generate 100 simulated retrievals with expected noise levels (i.e. ~ 1 K measurement noise per ~ 100 MHz bandwidth)

• Select profile “closest” to simulated data as the retrieved profile, and examine temperature retrieval error

• Errors in this simulation meet science requirements

• Additional simulations continuing over Greenland flight path

OSU Ulta-WideBand Software Controlled Radiometer

• Ice sheet temperature at 10 m depth, 1 K accuracy• 10 m temperatures approximate the mean annual temperature, an important climate

parameter

• Depth-averaged temperature from 200 m to 4 km (max) ice sheet thickness, 1 K accuracy

• Spatial variations in average temperature can be used as a proxy for improving temperature dependent ice-flow models

• Temperature profile at 100 m depth intervals, 1 K accuracy • Remote sensing measurements of temperature-depth profiles can substantially improve

ice flow models

• Measurements all at minimum 10 km resolution• Time stamped and geolocated by latitude and longitude

Presently building an instrument that can measure:

Progress in Radar Tomography

July 20, 2008, 17 km wide, 150 MHz radar tomography GISMO image (geocoded) of the upper surface of the ice sheet across Jacobshavn Glacier (right). 2000 Radarsat C-band image (center). Inset map from Radarsat mosaic (left). July 15, 2008, MERIS optical image (lower left). GISMO image located at about 69.3N, 48.3 W

Multi-frequency Images of Ice Sheet Surface

Lake

Ice stream

Crevasse Band

0 5000 10000 15000 200001260

1280

1300

1320

1340

1360

1380

1400

1420

ICEsat elevationtomographic eleva-tion

ICEsat along track (m)

ice

shee

t su

rfac

e el

evati

on

(m)

Surface Elevation Validation

Wu and others, 2011

Image: Ice thickness map of Jacobshavn, Greenland (2008) mosaiced from 2 GISMO swaths. Gray-scale indicates thickness. The lines locate Kansas University’s nadir ice sounder 2006 tracks.

Graphs: Ice thickness inter-comparisons have 18m and 14m rms errors.

Validation: basal topography accuracy

4.5 km

22 km

5x20 Km 3-d image of the base of the ice sheet. Scene is an orthorectified mosaic located just south of the main Jacobshavn Drainage Channel

GISMO Basal Imagery

Oblique downstream views of basal topography beneath the Greenland Ice Sheet compared with part of the now-exposed bed of the former Laurentide Ice Sheet near Norman Wells, Northwest Territories, Arctic Canada (60.3 N, 126.7 W; image ca 0.5 km in width).

Jezek and others, 2011

(a) Radarsat-1 image showing the location of the study area (red box) located about 14 km south of the main Jacobshavn Glacier drainage channel.

(b) Ice thickness in meters. Surface velocity vectors from radar interferometry.

(c) basal topography contours in m above the ellipsoid. Red (bright) and blue (weak) tones represent radar reflectivity.

(d) 5x20 km hill-shaded basal topography. Ice flow lines (red) are determined from

(e) Driving stress in Pascals Color tones correspond to radar reflectivity.

Basal Imaging: Southern Flank of Jacobshavn Glacier, Greenland

Jezek and others, 2011

Radar Sounding of Russel Glacier: Nadir Tracking and Tomography

Jezek, Wu, Paden, Leuschen, 2012

Basal Topography estimate of Isunguata Sermia Glacier computed by tomography (upper) and by interpolating nadir (lower).

Driving stress overlaid on Landsat-7 image. Lakes (white patches) generally correspond to locations of low driving stress.

Hill-shaded model of the tomography-derived basal topography (dark blue) overlaid on a hill-shaded model of the interpolated nadir-data topography (gray). In turn, these are overlaid on a lidar derived model of the ice-sheet, exposed-rock surface (light blue).

Example of shallow pockets at Umanaq, Greenland - 2011 IceBridge data

Example of shallow pockets at Umanaq: Intensity (top) and bed thickness (bottom)

dep

th (

3 km

in

air)

cros

s tr

ack

grou

nd r

ange

(3

km)

50 m

1160m

Along track (20 km)

Courtesy Wu, 2014

Anomalous Subsurface Object

Approximate Location of the feature

Details on Subsurface Structure

Ice thickness map in Polar stereo-graphic projection

0 100 200 300 400 500 600 700100

200

300

400

500

along track (5m/pixel)

thic

kn

es

s (

m)

Location of the data frame

dep

th (

3 km

in a

ir)cr

oss

trac

k gr

ound

ran

ge (

3 km

)

100 m 530 m

Intensity image

Ice thickness

Along track (3.75 km)

Courtesy Wu, 2014

Instrumentation Specifications

Hybrid couplers

GISMO Radar Parameters and Configuration

Parameters Year 2006 Year 2008

Radar carrier frequency 150 MHz

Signal bandwidth 20 MHz

Transmit pulse duration 3 s 3 s / 10 s

Duty cycle 10%

Peak transmit power 400 W 800 W

System loss -4 dB

Receiver noise figure 4.0 dBNumber of transmit antennae 2 4

Number of receive antennae 6 8

Antenna type dipole

A/D dynamic range 12-bits, 72 dB

Sampling frequency 120 MHz

PRF 15 kHz 10 kHz

Ultra-wideband software defined radiometer (UWBRAD)

• UWBRAD=a radiometer operating 0.5 – 2 GHz for internal ice sheet temperature sensing

• Requires operating in unprotected bands, so interference a major concern

• Address by sampling entire bandwidth ( in 100 MHz channels) and implement real-time detection/mitigation/use of unoccupied spectrum

• Supported under NASA 2013 Instrument Incubator ProgramFrequency Channels 0.5-2 GHz, 15 x 100 MHz channels

Polarization Single (Right-hand circular) Observation angle Nadir Spatial Resolution 1 km x 1 km (1 km platform altitude) Integration time 100 msec Ant Gain (dB) /Beamwidth

11 dB 30

Calibration (Internal) Reference load and Noise diode sources Calibration (External) Sky and Ocean Measurements

Noise equiv dT 0.4 K in 100 msec (each 100 MHz channel) Interference Management

Full sampling of 100 MHz bandwidth in 16 bits resolution in each channel; real time “software

defined” RFI detection and mitigation Initial Data Rate 700 Megabytes per second (10% duty cycle)

Data Rate to Disk <1 Megabyte per second

H = 37”

Diameter: 10 inches

Diameter: 1.1 inches

Cone Angle = 13.2°

56 Turns

Next Step: Radar and Radiometry

• Difficult to model fine scale structure necessary to accurately correct Tb data for near surface scattering

• Wideband radar may be suitable for characterizing scattering magnitudes

• Peake’s equation relates emissivity to backscatter coefficient based on conservation of incoming, scattered and emitted power

• Spaceborne systems already operate L-band radars and radiometers (Aquarius, SMAP)

• Potential for integrating UWBRAD with wideband radars developed by CReSIS. Will require additional development of tomographic and bistatic-SAR techniques. 2016 experiment will underfly CReSIS data.

Antarctic and Greenland Field Deployments

• April or October 2016 Greenland Airborne Campaign

• Continued discussions with Ken Borek Air, Ltd. for use of Bassler aircraft

• Budget for 5 days/ 40 flight hours consistent with project plan

• IFAC will deploy an L-band radiometer at DOME-C November 2015-January 2016 (30-45 day campaign)

• Plan to include UWBRAD tower deployment at DOME-C as part of the IFAC Project

• Would be desirable to include full 13 channel system, but a 4 channel system could provide valuable information

• Developing plan to deploy UWBRAD 4 channel system at DOME-C

Antenna

UWBRAD Enclosure

Summary• Radar Tomography:

• Technique proven in limited instances• Provides necessary measurement of 3-d ice sheet internal structure and basal

geometry• Requires more research to improve swath width and continuous coverage• Requires algorithm development for routine products useful for models

• Wideband Radiometry

• Operational L-band systems suggest brightness temperatures are sensitive to depth• Wideband simulations demonstrate that, within assumptions, physical temperature can

be measured at depth• UWBRAD will be tested in October 2015 and April 2016

• Radar and Radiometry

• Active and passive measurements may be required to correct radiometric data for scattering from the complex, near surface layers of the ice sheet