2_3398_IGARS2011.pptx

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Evaluation of the SMAP Combined

Radar-Radiometer Soil Moisture Algorithm

1Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, USA2Massachusetts Institute of Technology, Cambridge, MA 02139, USA

3University of California, Santa Barbara, CA 93106, USA

IGARSS 2011Paper #3398

N. N. Das1

D. Entekhabi2

S. K. Chan1

R. S. Dunbar1

S. Kim1

E. G. Njoku1

J. C. Shi3

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

SMAP Measurements Approach

RadarFrequency: 1.26 GHz Polarizations: VV, HH, HV Resolution: 3 kmRelative Accuracy: 1.0 dB (HH ,VV), 1.5 dB (HV)

RadiometerFrequency: 1.41 GHz Polarizations: H, V, 3rd & 4th StokesResolution: 40 kmRelative Accuracy: 1.3 K

Shared AntennaConstant Incidence Angle: 40ºWide Swath: 1000 km

OrbitSun-Synchronous, 6 am/pm Orbit, 680 km

Overview of the SMAP Mission

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

L-band Active/Passive Assessment

· Soil Moisture Retrieval Algorithms Build on Heritage of Microwave Modeling and Field Experiments

MacHydro’90, Monsoon’91, Washita92, Washita94, SGP97, SGP99, SMEX02, SMEX03, SMEX04, SMEX05, CLASIC, SMAPVEX08, CanEx10

· Radiometer - High Accuracy (Less Influenced by Roughness and Vegetation) but Coarser Resolution (40 km)

· Radar - High Spatial Resolution (1-3 km) but More Sensitive to Surface Roughness and Vegetation

Combined Radar-Radiometer Product ProvidesBlend of Measurements for Intermediate Resolutionand Intermediate Accuracy

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Requirement Soil MoistureFreeze/Thaw

Resolution 10 km 3 km

Refresh Rate 3 days 2 days(1)

Accuracy 0.04 [cm3 cm-3] (2) 80% (2)

Duration 36 months

(1) North of 45°N Latitude(2) % volumetric water content, 1-sigma(3) % classification accuracy (binary: Freeze or Thaw)

SMAP Level 1 Science Requirements

ProductShort Name Description Data

Resolution

L2_SM_P Radiometer Soil Moisture 36 km

L2_SM_A Radar Soil Moisture 3 km

L2_SM_A/P Active-Passive Soil Moisture 9 km

L2_F/T_HiRes Daily Global Composite Freeze/Thaw State 1-3 km

L3_SM_P Daily Global Composite Radiometer Soil Moisture 36 km

L3_SM_A/P Daily Global Composite Active-Passive Soil Moisture 9 km

L4_SM Surface & Root Zone Soil Moisture 9 km

L4_C Carbon Net Ecosystem Exchange 1 km

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Definitions and Data Products Flow

0 3 6 9 12 15 18 21 24 27 30 33 360369

121518212427303336

km

km

L1_S0_HiRes σ

0 360

36

km

km

0 9 18 27 360

9

18

27

36

km

kmMerge Algorithms

C

Mnm

nc = 1

nf = 144

nm = 16

C = Coarse (~36 km Radiometer)Mnm = Medium (~9 km Merged Product)F nf= Fine (~3 km Radar)

L1C_TB TB

L2_SM_AP

TB disaggregation

(Das et al., Preliminary ATBD) (TGARS, submitted)

Fnf

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

L2_SM_AP Radar-Radiometer TB Disaggregation Algorithm

Same evaluated at scale Mj:

jppjj

ppjpp

B

jB

MCMCM

CMC

CT

MT

p

p

Temporal Changes in TB and σpp are

Related. Relationship Parameter β is Estimated Statistically at Radiometer C-Scale Using Successive Overpasses:

CCCCT ppBp

jppjjjB MMMMTp

Subtract Two Equations to Write:

DOY, 2002 TBh~4 km σvv ~800 m

176

178

182

183

186

187

188

189

R2 (Low: 0.65, High: 0.93) values between TBh and σvv

dBK

SMEX02

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

L2_SM_AP Radar-Radiometer Algorithm

Heterogeneity in Vegetation and Roughness Conditions Estimated by Sensitivities Γ in Radar HV Cross-Pol:

Cjpqjpp MM , Slope

TB( Mj ) is Used to Retrieve Soil Moisture at 9 km

TB-Disaggregation Algorithm is:

)]}()([

)]()([ {

)(

)(

CM

CM

CT

MT

pqjpq

ppjpp

B

jB

p

p

hvvvhh

hv

2

8RVI

hvvvhh

hv

2

8RVI

CppBT , Slope

Based on PALS Observations From: SGP99, SMEX02, CLASIC and SMAPVEX08

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

0 0.1 0.2 0.3 0.40

0.1

0.2

0.3

0.4

Average of Field Measurements [cm3/cm3]B

asel

ine

Alg

ori

thm

[cm

3/c

m3]

RMSE: 0.033 [cm3/cm3]

0 0.1 0.2 0.3 0.40

0.1

0.2

0.3

0.4

Average of Field Measurements [cm3/cm3]

Min

imu

m P

erf

orm

an

ce [

cm3/c

m3] RMSE: 0.055 [cm3/cm3]

Combined Airborne Data From: SGP99, SMEX02, CLASIC and SMAPVEX08

Active-Passive Algorithm Performance

Minimum Performance AlgorithmRMSE: 0.055 [cm3 cm-3]

Active-Passive AlgorithmRMSE: 0.033 [cm3 cm-3]

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

0 0.1 0.2 0.3 0.40

0.1

0.2

0.3

0.4

Average of Field Measurements [cm3/cm3]

Bas

elin

e w

ith

no

HV

ad

j. [c

m3/c

m3]

RMSE: 0.043 [cm3/cm3]

The Role of Cross-Pol in Capturing Heterogeneity

0 0.1 0.2 0.3 0.40

0.1

0.2

0.3

0.4

Average of Field Measurements [cm3/cm3]

Bas

elin

e A

lgo

rith

m [

cm3/c

m3 ]

RMSE: 0.033 [cm3/cm3]Active-Passive AlgorithmRMSE: 0.033 [cm3 cm-3]

Active-Passive Algorithm Without Cross-PolRMSE: 0.043 [cm3 cm-3]

0 0.1 0.2 0.3 0.40

0.1

0.2

0.3

0.4

Average of Field Measurements [cm3/cm3]

Min

imu

m P

erf

orm

an

ce [

cm3/c

m3] RMSE: 0.055 [cm3/cm3]

Minimum Performance AlgorithmRMSE: 0.055 [cm3 cm-3]

0

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

VWC [kg/m2]

RM

SE

[c

m3/c

m3]

Baseline L2_SM_A/PL2_SM_A/P with no cross-pol adjustmentMinimum performance

Parameter Added Uncertainty (1 std.)Brightness Temperature TB 1.5 [K]

Vegetation Opacity (τ) 10%Soil Temprature 2 [K]

Single-Scattering Albedo (ω) 5%Roughness (h) 10%

Sand fraction (sf) 10%Clay fraction (cf) 10%

Study region selected from the CONUS domain.

Assessment of L2_SM_AP Algorithm Using SMAP Algorithm Testbed

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Sample of L3_SM_APOutput from SMAP Algorithm Testbed

V/V

Global Composite Map of Soil Moisture for April 02

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Summary

PALS data verifies that the assumption (linear TB-log[σ] relationship) holds well as the basis for the L2_SM_A/P algorithm

With current baseline approach, the algorithm meets the SMAP Level-1 requirements

Algorithm relies on radar co-pols and cross-pols

L2_SM_AP processor developed in SMAP Testbed

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Optimize length of temporal window (balance between phenology and statistical robustness)

Develop and mature algorithm prior parameters database for Bayesian estimation

Develop and mature L2_SM_A/P error budget table

Work in Progress

National Aeronautics and Space Administration

Jet Propulsion LaboratoryCalifornia Institute of TechnologyPasadena, California

Acknowledgements

Andreas CollianderJet Propulsion Laboratory

Joel JohnsonOhio State University

NASA SMAP Project