© Crown copyright Met Office CAVIAR field campaigns meeting Stuart Newman Exeter, 29 April 2010.

© Crown copyright Met Office

CAVIAR field campaigns meetingStuart Newman

Exeter, 29 April 2010


Contents

This presentation covers the following areas

• NPL laboratory calibrations

• Field campaign profile selection

• Initial continuum results


NPL lab calibrations


CAVIAR work package 3.2Low temperature blackbody calibration

• ARIES calibrated against NPL low temperature blackbody

• Range of temperatures relevant to atmospheric remote sensing – we achieved a range of -75 to +30 ºC

• ARIES blackbody target temperatures also varied as a test of target emissivity


ARIES target emissivity tests

Positive bias with “hot” target at 81ºC, “cold” target at 20ºC

Negative bias with “hot” target at 12ºC, “cold” target at 41ºC

(c.f. NPL target with 0.9975 emissivity at -74.8ºC)


Finding best fit target emissivities


ARIES target emissivity results

• “hot” target at 81ºC, “cold” target at 20ºC• “hot” target at 12ºC, “cold” target at 41ºC


Merge results

Experiment NPL target temperature

Fitted (CBB) Fitted (HBB)

“N004”

22 May 09

-45ºC 0.993 0.996

“N005”

26 May 09

-75ºC 0.980 0.992

“N006”

27 May 09

-75ºC 0.976 0.992


Account for calibration uncertainty?


Best estimate for uncertainty?

• Standard deviation gives very large uncertainty

• Standard deviation of mean (standard error) reduces this by √N but for 500-1000 measurements this is an over-constraint

• May be appropriate to use number of independent calibrations (blackbody views) as N, approx. 20-30 per set of conditions

Mean & std. dev.


Field campaign profile selection


Summary of flightsFlight Date Weather conditions Comments

B466 16.07.09 Cloud at times over Jungfraujoch Partial NPL data

B467 19.07.09 Initially thin cirrus which cleared Good NPL data

B468 20.7.09 Good clear sky conditions FLASH sonde* + MetOp overpass

B469 25.07.09 Cloud at times over Jungfraujoch Partial NPL data

B470 26.07.09 Some thin cirrus encroaching Good NPL data

B471 27.07.09 Excellent clear sky conditions MetOp overpass

B472 29.07.09 Excellent clear sky conditions ARIES failure

B473 01.08.09 Excellent clear sky conditions No TAFTS

B474 04.08.09 Partial cloud over Jungfraujoch Cancelled am flight, pm only

* Institute of Applied Physics in Bern launch radiosondes from Payerneequipped with RS92, Snow White and FLASH-B (Lyman-alpha) hygrometers


Meeting at IC in December

• Identify sources of profile information (dropsondes, aircraft probes, model fields)

• Agree a method for sub-selecting data (three areas: NW, over JFJ & SE, different profile for each aircraft run)

• Agree a method for combining profile information (as a first attempt use the nearest co-located dropsonde and perturb using model spatial-temporal evolution… however, see next slides)


Model biases

Statistics of model fields versus 55 dropsondes

ECMWF mean biases are small (though with large scatter)

Cosmo mean biases are larger, particularly for humidity


Model biases cont.

Really bad example for Cosmo


Model biases cont.

Better example for Cosmo

Cosmo particularly bad on 27/7/09 and 1/8/09 (for some reason)

So far have used ECMWF exclusively


Profile issues

• Using model to evolve dropsonde profile only works if model is close to true atmosphere

• If model biases are present, and profile is extrapolated too far in time and space, errors can be introduced

• It seems best to concentrate first on situations where the spectrometer data are closely located with a dropsonde (or aircraft profile)


Can we use FWVS data?

Dropsondes are widely recognised as most accurate source of humidity data from the aircraft

However, for runs immediately after a profile descent (e.g. 35000 ft down to 15000 ft) FWVS may be more representative

Attempt to use FWVS humidity data in LBLRTM simulations of ARIES at 15000 ft


CAVIAR exampleB467 19-Jul-2009

• Initial run at high level for radiance measurements (here looking up)

• Spiral descent over Jungfraujoch observatory measuring in situ water vapour (rapid response FWVS probe used here)

• Subsequent run at lower level for radiance measurements (here looking up)

• Determine change in radiance due to water vapour in atmospheric path

• Derive continuum strength, compare to MT_CKD model in LBLRTM



• Consistent discrepancy in centre of water vapour band for channels at 1586 and 1598 cm-1 (MT_CKD continuum is relatively strong at 1586 cm-1 compared with measurements)


Humidity information from MARSS

• Helpful to constrain humidity above the aircraft using microwave observations at 183 GHz

Microwave Airborne Radiometer Scanning System (MARSS)


MARSS humidity channels

B467 (19/7/09)Good agreement for channel centred closest to 183 GHz line

B471 (27/7/09)Worse agreement for channel 183±1 GHz, better for 183±3 GHz



• Inconsistent results here, all retrieved continuum strengths < MT_CKD

• Is this because LBLRTM profile is too moist?

• Possible to use MARSS data for profile constraint?


Initial continuum results from Jungfraujoch

flights and satellite data


• 8 MetOp cal/val flights• Sea – Gulf of Mexico• Land – Oklahoma• 2 night, 6 day flights• All MetOp collocated

Joint Airborne IASI Validation Experiment (JAIVEx) – April-May 2007


IASI water vapour bandFlight B290, 30-Apr-2007


IASI water vapour bandFlight B285, 19/20-Apr-2007


IASI water vapour bandFlight B290, 30-Apr-2007

• Coudert et al. water vapour spectroscopy updates since HITRAN2004

• MT_CKD_2.5 in LBLRTM_v11.7

• Continuum channels relatively insensitive to spectroscopic database


IASI water vapour band

• Implied continuum strength is less than MT_CKD < 1500 cm-1, but greater than MT_CKD > 1900 cm-1

• Retrieved continuum is sensitive to uncertainties in atmospheric profile


Preliminary resultsSelected data from flights B467-B474


Questions and answers


IASI atmospheric retrieval


Spectral retrieval

© Crown copyright Met Office CAVIAR field campaigns meeting Stuart Newman Exeter, 29 April 2010.

Documents

c slide

humidity slide

cosmo slide

ft slide

aircraft profile slide

lblrtm slide

slides slide

fit target emissivities