herman g.j. smit

Mitg

lied

de

r H

elm

holtz-G

em

ein

sch

aft

3. Februar 2011 SPARC/IOC/WMO-IGACO-Meeting, WMO, Geneva, 25-27 Jan.2011 Slide 1

Herman G.J. Smit Institute of Chemistry and Dynamics of the Geosphere: Troposphere (ICG-2),

Research Centre Juelich (FZJ) , D-52425 Juelich, Germany (e-mail: [email protected])

SPARC/IOC/WMO-IGACO Workshop on

Past Changes in the Vertical Distribution of Ozone

Geneva, January 25-27 2011

Overview of the performance of ozone sondesand their uncertainties

3. Februar 2011 Slide 2SPARC/IOC/WMO-IGACO-Meeting, WMO, Geneva, 25-27 Jan.2011

Performance ECC-Ozone

Sounding: Intercomparison

0

5

10

15

20

25

30

35

Alt

itu

de

[

km

]

180 200 220 240 260 280 300

Temperature [Kelvin]

0 5 10 15 20

Ozone , Pressure [ mPa ]

Ozone Pressure

Temperature

Grey Line = ECC-Sonde Kasshidoo

ECC-Ozone sensor converts sampled ozone flow into electrical current:

�In aqueous KI-solution Ozone is converted into Iodine molecules

�In electrochemical cell Iodine is converted at Pt-Cathode into Iodide-ions


JOSIE-1996: Juelich Ozone Sonde Intercomparison Experiment

-30 -20 -10 0 10 20 30Relative Deviation [ % ]

0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-A-MOHp [1]BM-orig.

File : JSRMD1C.GRF


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-B-CMDL [2]ECC/ENSCI-1Z

File : JSRMD2C.GRF

-30 -20 -10 0 10 20 30

Relative Deviation [ % ]

0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-D-IMD [4]Indian

File : JSRMD4C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-E-ASP [5]BM-Hybrid

File : JSRMD5C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-G-CNRS [7]ECC/ SPC-5A

File : JSRMD7C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-H-JMA [7]KC79

File : JSRMD8C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-F-AES [6]ECC/SPC-5A &ECC/ENSCI-1Z

File : JSRMD6C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-C-FZJ [3]ECC/SPC-6A

File : JSRMD3C.GRF

INDIANECCBrewer-Mast

ECC

ECC

ECC JapaneseKC79

BM-Hybrid


Conversion Efficiency ECC-Sonde: Influence of Sensing Solutions (KI & Buffer)

Johnson, et al, J. Geophys. Res., 2002

Stoichiometry (O3 + KI) Reactionyielding I2 at 1% KI, full Buffer:

At launch stoichiometry is unity

During flight evaporation of H2O

Increase concentrations of KI & Buffer

Increase of stoichiometry

Majorly due to increase of buffer

Ozone (mPa)


Manufacturer Model Type Years Manufactured.

Science Pump SPC-6A 1995- present

EN-SCI ENSCI-Z 1997 –present

PH-Buffer

Sensing Solution Type (SST)

KI

[g/L] NaH2PO4.H2O

[g/L]

Na2HPO4.12H2O

[g/L]

KBr

[g/L]

SST1.0: 1.0% KI & full buffer (a) 10 1.250 5.0 25

SST0.5: 0.5% KI & half buffer (b) 5 0.625 2.5 12.5

SST2.0: 2.0% KI & no buffer (c) 20 0 0 0

Controversy ECC-sondes since Mid 1990‘s:Two Manufacturers and 3 Different Sensing Solution Types


JOSIEJülich

Ozone SondeIntercomparison

Experiment

ASOPOSAssessment forStandard

OperatingProcedures forOzone

Sondes

BESOSBalloon

Experiment onStandards forOzone

Sondes

Since 1996 April 2004, atLaramie, USA

JOSIEJülich

Ozone SondeIntercomparison

Experiment

ASOPOSAssessment forStandard

OperatingProcedures forOzone

Sondes

BESOSBalloon

Experiment onStandards forOzone

Sondes

Since 1996 April 2004, atLaramie, USA

Sources: JOSIE [Smit et al., J.Geophys.Res., 2007] BESOS [Deshler et al., J.Geophys.Res., 2008]

WCCOSWorld

Calibration

Centre for

Ozone

Sondes

At Juelich,

Germany

JOSIE – ASOPOS - BESOS


Sondes versus OPM : BESOS versus JOSIE

1. Systematic difference of 3-6%

2. Compared to sondes: OPM in

BESOS about 3-6 % lower than

in JOSIE.

3. Limited performance OPM during

BESOS

4. Behaviour of sondes in

stratosphere very consistent

during JOSIE 1996, 1998, 2000

and BESOS

-15 -10 -5 0 5 10 15 20 25Relative Differences [%]

(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]

JOSIE-2000: SST1.0 (N=8sondes)

BESOS-2004: SST1.0 (N=3sondes)

ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



SPC-6A


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40A

ltit

ude [k

m]



SPC-6A

SST1.0 SST0.5

SST1.0 SST0.5

A B

C D


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



SPC-6A


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40A

ltit

ude [k

m]



SPC-6A

SST1.0 SST0.5

SST1.0 SST0.5


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



SPC-6A


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40A

ltit

ude [k

m]



SPC-6A


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



ENSCI-Z


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40

Alt

itude [k

m]



SPC-6A


(Sonde-OPM)/OPM

0

5

10

15

20

25

30

35

40A

ltit

ude [k

m]



SPC-6A

SST1.0 SST0.5

SST1.0 SST0.5

A B

C D

SPC-6ASST1.0

SPC-6ASST0.5

ENSCI-ZSST1.0

ENSCI-ZSST0.5

JOSIE JOSIE

JOSIE JOSIE

BESOSBESOS

BESOS BESOS


Sondes versus Sondes : BESOS versus JOSIE

-10 -5 0 5 10 15 20Relative Differences [%]

(ENSCI-Z minus SPC-6A)/OPM

0

5

10

15

20

25

30

35

40A

ltit

ude

[km

]

ENSCIZ: SST1.0 & SPC6A: SST1.0






0

5

10

15

20

25

30

35

40

Alt

itude

[km

]





JOSIE2000

BESOS

2004

A B



0

5

10

15

20

25

30

35

40A

ltit

ude

[km

]







0

5

10

15

20

25

30

35

40

Alt

itude

[km

]





JOSIE2000

BESOS

2004

A B

JOSIE2000

BESOS2004


For each Sensing Solution Type

(SST1.0 , SST0.5, and SST2.0) :� ENSCI 5-10 % higher than SPC� Precision about 3-6 %

Each sonde type (ENSCI or SPC):

�SST1.0 ≈ 5% larger than SST0.5�SST0.5 ≈ 5% larger than SST2.0�SST1 ≈ 10% larger than SST2.0

JOSIE 2000 & BESOS 2004: Comparison SPC-6A&ENSCI-Z @ Different Sensing Solutions

SST1.0 (1.0%KI,Full Buffer), SST0.5 (0.5%KI,Half Buffer), SST2.0 (2.0%KI, No Buffer)

[Data processed after Komhyr 1986, IB0 (PO2), No Total O3 Normalization]


1.) Small changes of

a. ozone sonde instrument (e.g. manufacturers) b. operating procedures (e.g. sensing solutions)

can have large impact on sonde data quality

2.) Non-uniformity in data processing

3.) Establishment of Standard of Operating Procedures

(SOP) for ozone sondes by ASOPOS recomendations

4.) After standardization and homogenization

improvement of precision and accuracyby about factor 2 can be expected

ECC-Ozone Sonde Performance:What we learned from JOSIE and BESOS


Established with following plan:

1) Construct/derive empirical transfer functions from existing data:

Relative Diffference PO3 (Z)

Absolute Difference PO3 (Z)

Absolute Difference PO3 (PO3)

2) Evaluate transfer functions: testing a set of homogenized O3S data on consistency (e.g. use MATCH, satellites, MOZAIC).

3) Recommendations on the methodology to homogenize data

4) Review by independent panel of scientists (O3S data users)

5) Approval of recommendations by WMO/GAW-O3-SAG and NDACC-SSC

Task force group established as WMO/GAW-IGACO activity in

collaboration with the NDACC and SHADOZ ozone sounding networks.

Task force group:

Transfer functions to homogenize ozone sonde records


JOSIE-1996: Overview Comparison Ozone Sondes


0

5

10

15

20

25

30

35A

ltitude

[ km

]

Fig.-A-MOHp [1]BM-orig.

File : JSRMD1C.GRF


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-B-CMDL [2]ECC/ENSCI-1Z

File : JSRMD2C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-D-IMD [4]Indian

File : JSRMD4C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-E-ASP [5]BM-Hybrid

File : JSRMD5C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-G-CNRS [7]ECC/ SPC-5A

File : JSRMD7C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-H-JMA [7]KC79

File : JSRMD8C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-F-AES [6]ECC/SPC-5A &ECC/ENSCI-1Z

File : JSRMD6C.GRF

-30 -20 -10 0 10 20 30


0

5

10

15

20

25

30

35

Altitude

[ km

]

Fig.-C-FZJ [3]ECC/SPC-6A

File : JSRMD3C.GRF

INDIANECCBrewer-Mast

ECC

ECC

ECC JapaneseKC79

BM-Hybrid


A B

JOSIE-BESOS: Improvement KC96-Ozone Sonde

Before After


Electrochemical Ozone Sonde: Introduction

In an electrochemical cell:

A small pump forces ambient air through

a KI-solution

Ozone is converted into iodine by the reaction:

At a Platinium cathode the Iodine is converted

to Iodide:

In external electrical circuit a current is generated

directly related to the uptake rate of ozone in the

sensing solution

KOH2OIOHOKI2 2223 ++++++++→→→→++++++++

PtI2e2I2 →→→→++++ −−−−PtI2e2I2 →→→→++++ I2e2I2 →→→→++++ −−−−

(((( ))))BMP

PC3O II

T04307.0P −−−−⋅⋅⋅⋅

ΦΦΦΦ⋅⋅⋅⋅ηηηη⋅⋅⋅⋅====

Pump Flow [cm3/s]

Ozone Pressure[mPa]

Conversion Efficiency

BackgroundCurrent [µA]

MeasuredCurrent [µA]

Pump Temperature [K]


0.0 5.0 10.0 15.0 20.0Ozone Partial Pressure [mPa]

0

5

10

15

20

25

30

35

Altitude [k

m]

0 2.5 5 7.5 10Relative Uncertainty to Ozone [%]


0

5

10

15

20

25

30

35

Altitude [k

m]

0 5 10 15 20Relative Uncertainty to Ozone [%]

A: Mid Latitude B: Tropics

Pressure O3 (PO3)

Rel.Unc.PO3

Rel.Unc.IBack & IM

Rel.Unc.Conversion O3 to I2

Rel.Unc.Pump Flowrate

Rel.Unc.Pump Temperature


0

5

10

15

20

25

30

35

Altitude [k

m]



0

5

10

15

20

25

30

35

Altitude [k

m]




0

5

10

15

20

25

30

35

Altitude [k

m]



0

5

10

15

20

25

30

35

Altitude [k

m]



Pressure O3 (PO3)

Rel.Unc.PO3

Rel.Unc.IBack & IM




Pressure O3 (PO3)

Rel.Unc.PO3

Rel.Unc.IBack & IM




Pressure O3 (PO3)

Rel.Unc.PO3

Rel.Unc.IBack & IM




Instrumental Uncertainties of ECC-Ozone Sonde


Two rate constants:

• 1/10th buffer: τ = 19 s, τ’ = 24 min

• full buffer : τ = 19.7 s, τ’ = 28 min

(similar to Davies et al., QOS 2000)

�� Can be used for time lag corrections!

ECC-Sonde: Downward Step Response[Source: Voemel & Diaz, AMT 2010]

'

00

'

00)( ττ

tttt

eIeItI

−−

−−

+=


ECC-Background Current Correction @JOSIE 1998

BG1 = Before O3 Exposure (IB0) , No F(PO2)BG2 = Before O3 Exposure (IB0) , F(PO2)

BG3 = After O3 Exposure (IB1) , No F(PO2)BG4 = After O3 Exposure (IB1) , F(PO2 )

ENSCI-Z

BG1 BG2

BG3 BG4

SPC-6A

BG1 BG2

BG3 BG4

No F(PO2) = Constant F(PO2) = O2-pres. dependent


0.52±±±±0.100.10±±±±0.0528.6±±±±3.227.7±±±±2.9AverageENSCI-Z

0.52±±±±0.150.16±±±±0.0721.5±±±±2.924.0±±±±1.7AverageSPC-6A

StratosphereTroposphereStratosphereTroposphere

Offset (mPa)

(@t=5min)

Response Time (s)Sonde

Nr.Sonde Type

JOSIE 1998: Downward Response Tests ECC-Sondes (13xSPC6A and 13xENSCI-Z-)


Icell - ITEI = α·Icell + β

Offset parameter β:

Modified background

Slope parameter α:

Correction for stoichiometry

IO3 = (1-α) • ICell – β

full buffer :α = 0.090, β = 0.014µA

1/10th buffer : α = 0.031, β = 0.007µA

0.5% : α = 0.024, β = 0.009µA

ECC-Sonde: Background measurement with O3[Source: Voemel & Diaz, AMT 2010]


0

10

20

30

40

Alt

itude

[km

]

1000

100

10

Pre

ssure

[h

Pa]

270 280 290 300 310Temperature [K]

0 1 2 3 4 Rel. Deviation (TInt-TExt)/TInt [x100%]

Pump Temp. Extern

Pump Temp. Intern

Rel.Deviation (TINt-TExt)/TInt

TIntern

TExtern

RelativeDeviation

Pump Temperature:Internally or Externally Measured

Internal pump temperature:

inside the Teflon block of the pump

External pump temperature:

outside the Teflon block of the pump

corresponds to box temperature


Ozone Sonde:Total Ozone Column and Normalization Factor (NF)

TOC = TOS + ROC

TOC = Total Ozone Column between surface and top of atmosphere (TOA)

TOS = Total Ozone Column between surface and burst point (PB, ZB)

ROC = Residual Ozone Clumn above burst point (PB, ZB)

TOS : Dependent on Accuracy of Z, P

ROC : Dependent on Accuracy of ZB, PB

CMR-Method OR SBUV-Method

TO3 = Total Ozone Column measured by Dobson or Brewer Or Satellite

NF = TO3 / TOC


1.00 1.10 1.20 1.30

Pump Correction Factor

1000

100

10

Pre

ssu

re

[hP

a]

Komhyr 1986

Komhyr 1995

CMDL 1996

CMDL 2000

WFF 2000

0

10

20

30

40A

ltit

ud

e [

km

]

Pump Flowrate Correction at Low Pressures

JOSIE :[ CMDL(NOAA) and

WFF(NASA) are

averages from about 6 individually calibrated

pumps ]


1) Review different laboratory methods by “experts”

2) Intercomparison of existing methods by ring-experiment

of a set of SPC6A and ENSCI pumps by testing laboratories (NOAA, UWY, WFF, KMI & JMA) [Blind &

Referee].

3) Evaluation

4) Implications

5) Recommendations

6) Approval by WMO/SAG-O3 to include as

recommendation into the SOP’s

Task force group “O3S-Pump Flow Correction”

Time schedule: 2010-2012IGACO-O3 activity


In how far O3-Sondes usable for trend analysis?

Altitude:ECC : 0- 30 km (uncert. 5-10%)

BM: 0- 25 km (uncert. 10%) ; KC79/96: 0-25 km (uncert.10%)

Long term stability:

Only for limited number of stations after selection on „objective“criteria to be defined by panel of experts.

Level of confidence:

Limited with respect to poor coverage in space and time

Yes, but depends on stations, sonde type , and length of record

Internal consistency of single station

External consistency: Intercomparison excercises like JOSIE, BESOS and other field activities (e.g. dual-soundings)

herman g.j. smit

Documents