SMAP Cal/Val

SMAP Cal/Val

T. J. JacksonUSDA ARS Hydrology and Remote Sensing Lab

December 11, 2012

• Soil Moisture Satellite Missions: Past, Present and Future• SMAP

– Mission– Cal/Val

• Sparse Networks in Cal/Val• Challenges to using COSMOS

Outline

100 km 10 km 1 km

Day

Week

Month

ClimateClimate ApplicationsApplications

Weather ApplicationsWeather Applications

CarbonCarbon CycleCycle ApplicationsApplications

Resolved Spatial Scales

Res

olve

d T

empo

ral S

cale

sEvolution of Microwave Remote Sensing (Land)

Passive

Active

Historic Perspective on Remote Sensing

Scale ranges are based on the NRC Decadal Survey

100 km 10 km 1 km

Day

Week

Month





Res

olve

d T

empo

ral S

cale


AMSR-EASCAT

2002-2011AMSR-E X-band 40 km


ASCATGCOM-W

ALOS SARALOS-2

2012 GCOM-WSoil moisture product is the same as AMSR-E

• European Space Agency (ESA)• 1.4 GHz Microwave Radiometer• 40 km footprint, three day global coverage• Launch November 2009

Soil Moisture and Ocean Salinity Mission (SMOS)

100 km 10 km 1 km

Day

Week

Month





Res

olve

d T

empo

ral S

cale


AMSR-EASCATSMOS

1.4 GHz

6.0 GHz

10.0 GHz

• Same resolution but a better product

• Technology demo

Aquarius/SAC-D

390 km

Inner beam76×94 km

Outer beam96×156 km

Middle beam84×120 km

• Mission (NASA and CONAE)– Sun-synch orbit – 6 am (Des.)/6 pm (Asc.)– Night time look direction– 657 km Alt; 7 day revisit– Launch: June 2011

• Aquarius Instrument– L-band Polarimetric– Radiometer and Scatterometer– 3 Beam Pushbroom– Incidence angles of

29.36°, 38.49°, and 46.29°• SAC-D

– MWR – Other

100 km 10 km 1 km

Day

Week

Month





Res

olve

d Te

mpo

ral S

cale


AMSR-EASCATSMOS

Aquarius


Active and passive L-band but coarser spatial resolution and temporal repeat

SAOCOM: SAtélite Argentino de Observación COn Microondas

• Comisión Nacional de Actividades Espaciales (CONAE)-Argentina Space Agency

• Constellation of two identical satellites SAOCOM 1A and SAOCOM 1B carrying an L-band polarimetric SAR instrument

• SAOCOM will be in a sun-synchronous nearly circular frozen polar orbit (06:12 am LTAN/619.6 km)

• Repeat cycle of 16 days (8 days with full constellation of 2 satellites)

• Launch of SAOCOM 1A in 2015.• Challenge: Infer surface soil moisture

values from SAR measurements with varying incidence angles.

100 km 10 km 1 km

Day

Week

Month





Res

olve

d Te

mpo

ral S

cale


GCOM-WASCATSMOS

ALOS-2

Aquarius


Commitment to a soil moisture product

SAOCOM




Outline

SMAP Level 1 Science Requirements

TJJ–12

(a) North of 45N latitude, (b) Percent classification accuracy (binary freeze/thaw), (c) Volumetric water content, 1-σ in [cm3/cm3] units

RequirementRequirement Hydro-Hydro-MeteorologyMeteorology

Hydro-Hydro-ClimatologyClimatology

Carbon Carbon CycleCycle

Baseline MissionBaseline Mission Threshold MissionThreshold MissionSoil Soil

MoistureMoistureFreeze/Freeze/ThawThaw

Soil Soil MoistureMoisture

Freeze/Freeze/ThawThaw

Resolution 4–15 km 50–100 km 1–10 km 10 km 3 km 10 km 10 kmRefresh Rate 2–3 days 3–4 days 2–3 days(a) 3 days 2 days 3 days 3 daysAccuracy 0.04-0.06 (c) 0.04-0.06 (c) 80–70% (b) 0.04 (c) 80%(b) 0.06 (c) 70%(b) Mission Duration 36 months 18 months

• These are the L1 priority products and requirements. Other product accuracies derive from L2 requirements. Defines the baseline mission.

• The SMAP Project proposed the active-passive approach for meeting these requirements.

• The NRC Decadal Survey identified numerous potential applications for SM/FT observations.

• These were grouped into three categories with a spatial resolution, refresh rate, and accuracy.

• L-band unfocused SAR and radiometer system, offset-fed 6 m light-weight deployable mesh reflector. Shared feed for

1.26 GHz HH, VV, HV Radar at 1-3 km (30% nadir gap)

1.4 GHz H, V, 3rd and 4th Stokes Radiometer at 40 km

• Conical scan, fixed incidence angle (40o across swath

• Contiguous 1000 km swath with 2-3 days revisit (8 day repeat)

• Sun-synchronous 6am/6pm orbit (680 km)

• Launch October 31, 2014 (now in Phase C/D)

• Mission duration 3 years

SMAP Project Approach

TJJ–13

Product DescriptionGridding

(Resolution)

Latency**

L1A_Radiometer Radiometer Data in Time-Order - 12 hrs

Instrument Data

L1A_Radar Radar Data in Time-Order - 12 hrsL1B_TB Radiometer TB in Time-Order (36x47 km) 12 hrs

L1B_S0_LoRes Low Resolution Radar σo in Time-Order (5x30 km) 12 hrs

L1C_S0_HiRes High Resolution Radar σo in Half-Orbits 1 km (1-3 km) 12 hrs

L1C_TB Radiometer TB in Half-Orbits 36 km 12 hrsL2_SM_A Soil Moisture (Radar) 3 km 24 hrs

Science Data (Half-Orbit)L2_SM_P Soil Moisture (Radiometer) 36 km 24 hrs

L2_SM_AP Soil Moisture (Radar + Radiometer) 9 km 24 hrsL3_FT_A Freeze/Thaw State (Radar) 3 km 50 hrs

Science Data (Daily

Composite)

L3_SM_A Soil Moisture (Radar) 3 km 50 hrsL3_SM_P Soil Moisture (Radiometer) 36 km 50 hrsL3_SM_AP Soil Moisture (Radar + Radiometer) 9 km 50 hrs

L4_SM Soil Moisture (Surface and Root Zone ) 9 km 7 daysScience

Value-AddedL4_C Carbon Net Ecosystem Exchange (NEE) 9 km 14 days

* Over outer 70% of swath.** The SMAP project will make a best effort to reduce the data latencies beyond those shown in this table.

TJJ–14

SMAP Science Products

100 km 10 km 1 km

Day

Week

Month

SMAP Radar-Radiometer





Res

olve

d Te

mpo

ral S

cale


GCOM-WASCATSMOS

ALOS-2Scale ranges are based on the NRC Decadal Survey

SMAP will support established climate and carbon cycle applications and will open up new applications in weather and carbon cycle.

AquariusSAOCOM

SMAP L1 Requirements Impacting Cal/ValLevel 1 (Baseline) Science Requirements and Mission

Success Criteria

Provide estimates of soil moisture in the top 5 cm of soil with an error of no greater than 0.04 m3/m3 volumetric (one sigma) at 10 km spatial resolution and 3-day average intervals over non-excluded regions.

Conduct a calibration and validation program to verify data delivered meets the requirements.

Provide estimates of surface binary freeze/thaw state in the region north of 45N latitude, which includes the boreal forest zone, with a classification accuracy of 80% at 3 km spatial resolution and 2-day average intervals.

Threshold mission requirements are 0.06 m3/m3 and 70%

What the SMAP Project and NASA have agreed to do.

CEOS Validation Stages Adopted for SMAP

TJJ–17

Validation Stage Description

Stage 1 Product accuracy is assessed from a small (typically < 30) set of locations and time periods by comparison with in situ or other suitable reference data.

Stage 2

Product accuracy is estimated over a significant set of locations and time periods by comparison with reference in situ or other suitable reference data. Spatial and temporal consistency of the product and with similar products have been evaluated over globally representative locations and time periods. Results are published in the peer-reviewed literature.

Stage 3

Uncertainties in the product and its associated structure are well quantified from comparison with reference in situ or other suitable reference data. Uncertainties are characterized in a statistically robust way over multiple locations and time periods representing global conditions. Spatial and temporal consistency of the product and with similar products have been evaluated over globally representative locations and periods. Results are published in the peer-reviewed literature.

Stage 4 Validation results for stage 3 are systematically updated when new product versions are released and as the time-series expands.

Validation: The process of assessing, by independent means, the quality of the data products derived from the system outputs. The quality is determined with respect to the specified requirements.

SMAP Validation Methodologies

TJJ–18

Methodology Role Constraints Resolution

Core Validation Sites Accurate estimates of products at matching scales for a limited set of conditions

• In situ sensor calibration• Limited number of sites

• In Situ Testbed• Cal/Val Partners

Sparse Networks One point in the grid cell for a wide range of conditions

• In situ sensor calibration• Up-scaling• Limited number of sites

• In Situ Testbed• Scaling methods• Cal/Val Partners

Satellite Products Estimates over a very wide range of conditions at matching scales

• Validation• Comparability• Continuity

• Validation studies• Distribution

matching

Model Products Estimates over a very wide range of conditions at matching scales

• Validation• Comparability


matchingField Campaigns Detailed estimates for a very

limited set of conditions• Resources• Schedule conflicts

• Airborne simulators• Partnerships

SMAP Cal/Val ApproachPre-launch• Focus on insuring that there are means in place to fulfill the mission objectives

– Acquire and process data with which to calibrate, test, and improve models and algorithms used for retrieving SMAP science data products– Develop and test the infrastructure and protocols for post-launch validation

Post-launch• Focus on validating that the products meet their quantified requirements

– Calibrate, verify, and improve the performance of the science algorithms– Validate accuracies of the science data products as specified in L1 science requirements according to Cal/Val timeline

TJJ–19

Science Data Validation and Delivery Timeline

TJJ–20

In-Orbit Checkout (3 months)

Launch

L1 validation (6 months)

L2-L4 validation (12 months)

Formal start of SMAP Science Mission

Delivery of validated L1 products to Data Center

Delivery of validated L2-L4 products to Data Center

Pre-launchPreparation

Beta release of L1 products and start of routine delivery

Beta release of L2-L4 products and start of routine delivery




Outline

SMAP Validation Methodologies

TJJ–22

Methodology Role Constraints Resolution

Core Validation Sites Accurate estimates of products at matching scales for a limited set of conditions

• In situ sensor calibration• Limited number of sites

• In Situ Testbed• Cal/Val Partners

Sparse Networks One point in the grid cell for a wide range of conditions

• In situ sensor calibration• Up-scaling• Limited number of sites

• In Situ Testbed• Scaling methods• Cal/Val Partners

Satellite Products Estimates over a very wide range of conditions at matching scales

• Validation• Comparability• Continuity


matching

Model Products Estimates over a very wide range of conditions at matching scales

• Validation• Comparability


matchingField Campaigns Detailed estimates for a very

limited set of conditions• Resources• Schedule conflicts

• Airborne simulators• Partnerships

SMAP Cal/Val Partners Program• In situ observations are essential to SMAP Cal/Val• There were only a few high quality resources available• Increasing the number was constrained by

– The time and effort to establish a site– No $ to support these

• Action: ROSES DCL– No cost collaboration– Minimum standards– In situ data in exchange for early access to SMAP products

• Based on responses– Refined definitions– Missed some important resources

TJJ–23

SMAP Cal/Val Partners: Site Types• Core Validation Sites: In situ observing sites that provide well-

characterized estimates of a L2-L4 product at a matching spatial scale, a direct benchmark reference for the products. Additional minimum criteria are:– Provides calibration of the in situ sensors– Up-scaling strategy provided (implemented by Project)– Provides data in a timely manner– Long term commitment by the sponsor/host

• Contributing Validation Sites: In situ observing sites that provide estimates of a L2-L4 product but do not meet all of the minimum criteria for a Core Validation Site. (i.e. sparse networks)– Contributing Validation Sites are a supplemental resource (In assessing

meeting mission requirements but important in Stage 2 Validation).– The baseline approach to using sparse networks is the triple-collocation

technique. Efforts to improve this approach are desirable.

TJJ–24




– Up-scaling– Contributing depth– Integrating networks– Resolving “noise”

Outline

TJJ–26

Proposed best practices*:

First, apply temporal stability analysis (Cosh et al., 2006; 2008) to select sampling sites with temporal dynamics that best mimic footprint scale variability.

Second, use land surface modeling (Crow et al., 2005) and/or an intensive field campaign (De Rosnay et al. 2009) to refine understanding of the relationship between point- and footprint-scale variability (i.e., F↑ on left).

Third, apply triple collocation (Mirrales et al., 2010) to estimate impact of residual sampling errors on RMSE validation results.

Footprint-scale

θPOINTF↑ (θPOINT)

*Based on: Crow et al., “Upscaling sparse ground-based soil moisture observations for the validation of coarse-resolution satellite soil moisture products,” Reviews of Geophysics, 50, RG2002, doi:10.1029/2011RG000372, 2012.

Up-scaling Challenge: Using point-scale soil moisture observations to validate footprint-scale SMAP retrievals.

Challenge: Scaling Points to Footprints

Application of Triple Co-Location To Estimate Random Sampling Error in Sparse Ground Observations

Remote Sensing (RS)-SMAP

Land Surface Model (LSM)

Sparse Ground Observation (SPARSE)

TRUESPARSELSMSPARSERSSPARSE MSE ,

RS

LSM

SPARSE

1) Obtain three independent (and uncertain) estimates of footprint-scale soil moisture:

2) Assume independent errors and sample the following temporal average to estimate random sampling error in SPARSE:

3) Use this estimate to correct soil moisture RMSE estimates derived from RS versus SPARSE comparisons for sampling error in SPARSE.

TJJ–27

• Intensive sampling of a limited number of sites?• Exploit the Rover? How many conditions?

Challenge: Scaling Points to Footprints

Challenge: Matching Depth to a SMAP Product

• We all understand why the contributing depth varies with the wetness and shape of the profile.

• SMAP is only concerned with soil moisture of two layers; 0-5 cm and the root-zone (1 m) (for validation).

• Can COSMOS produce standard depth products?

Challenge: Integrating Networks

• Lots of points that are currently not compatible.• Still need to address the variable contributing depth issue but

there are more options for matching the depths of other networks.

• First step: co-location of instruments (i.e MOISST)

USDA-NRCS-Soil Climate Analysis Network – D. Harms

• Natural Resource Conservation Service monitoring

• SMAP Soil Moisture (surface and profile)

• CONUS - 181• Telemetry and FTP• Hourly• Status – Operating and Developing

Measurement Method Depths

Soil Moisture Hydra 5, 10, 20, 50, and 100 cm

Soil Temperature Hydra 5, 10, 20, 50, and 100 cm

Precipitation Tipping Bucket

-

TJJ–31

U.S. Climate Reference NetworkM. A. Palecki and J.E. Bell, NCDC

• USCRN observes climate change• Soil Moisture/Temperature Product Validation

with sparse network• 114 sites, 20 field calibrated in FY13• Satellite to NCDC, Internet to SMAP• 2-3 hours• Instruments in place, communication in place,

gravimetric sampling of subset planned for FY13

Measurement Type Method Depths (cm)

Soil Moisture Coaxial Impedance Dielectric

5,10,20,50,100

Soil Temperature Thermistor 5,10,20,50,100

MeteorologicalVariables (air T, prec, surface T,global solar

Platinum resistance thermometer, weighing bucket, IR, pyranometer

150Above Ground

TJJ–32

Challenge: Resolving “Noise”

• Vegetation, atmosphere,….

SMAP Cal/Val

Documents

Transcript of SMAP Cal/Val