5th GEWEX G-VAP

Tony Reale and Bomin Sun

NOAA STAR

Quality of instantaneous water vapor

profiles assessed with NPROVS+

5th GEWEX G-VAP – Workshop

04 - 05 November 2015

Lowell Center,

Univ. Wisconsin, Madison

Outline

About NPROVS

Long Term Monitoring (NARCS)

Shorter Term More Detailed Monitoring (PDISP)

Strength vs Weakness: NPROVS vs NPROVS +

GRUAN vs Conv RAOB

RAOB vs NWP

CALWATER

Summary 2

NPROVS

Collocation Archive

NPROVS NPROVS+

Conv Radiosondes

+GFS

GRUAN & DOE/ARM

Radiosondes

NUCAPS (S-NPP)

NUCAPS Parallel (STAR)

NOAA IASI MetOp-A (OSPO)

MetOp-B

(STAR FTP)

MIRS NPP (Op, v.11)

NOAA-18,19 MetOP-A,B DMSP F18 (STAR FTP,

OSAPO)

AIRS v.6 (NASA)

EUMETSAT IASI

MetOp-A,B (OSPO)

GOES (OSPO)

ATOVS NOAA-18,19 MetOp-A,B

(OSPO)

COSMIC (UCAR)

GRAS (EUMETSAT)

NPROVS+

Collocation Archive

3 day delay

FTP

VALAR

14 day delay

FTP

Algorithm Development

Visualization Tools: ODS

PDISP NARCS

INPUTS

PROCESSING

OUTPUTS

NPROVS/NPROVS+ Data Management Schematic

ECMWF

3

4

Collocation Criteria:

+/- 6-hour

250 km

Single Closest

5

Typical NPROVS Global Collocation Dataset

(1000 collocation records per day) 6

Maritime … Blue, Red

Land … Brown, Orange

(Drops … Purple)

N P R O V S

GRUAN and JPSS funded Dedicated (S-NPP) RAOB Sites

Over 10,000 RAOBS (1000 Dedicated) available since July 2013

N P R O V S +

7

8

Global Climate Observing System (GCOS)

Reference Upper Air Network (GRUAN)

JPSS Funded Dedicated RAOB

• DOE ARM (SGP, NSA, ENA)

CIMSS

(2) per week

GRUAN processed

dual vs single, etc

• AEROSE

• CALWATER

• PMRF

• Sterling Test Site …

Request coordination with

“other” intensive field

experiments particularly if

synchronized with

S-NPP

N P R O V S +

9

WEB Site

http://www.star.nesdis.noaa.gov/smcd/opdb/nprovs/index.php EDR

10 http://www.star.nesdis.noaa.gov/icvs/

SDR

WEB Site

Continental IR

GFS Forecast

GFS CFSR

ECMWF Analysis

AIRS Aqua IR+MW

NUCAPS IR+MW

NOAA IASI MetOp-A IR+MW

NOAA IASI MetOp-B IR+MW

EU IASI MetOp-A IR+MW

EU IASI MetOp-B IR+MW

GFS Forecast

GFS CFSR

ECMWF Analysis

AIRS Aqua IR+MW

NUCAPS IR+MW





7/1

3

8/1

3

9/1

3

11

/13

10

/13

12

/13

1/1

4

2/1

4

3/1

4

4/1

4

5/1

4

6/1

4

7/1

4

8/1

4

9/1

4

10

/14

11

/14

12

/14

1/1

5

2/1

5

3/1

5

4/1

5

5/1

5

6/1

5

7/1

5

8/1

5

7/1

3

8/1

3

9/1

3

11

/13

10

/13

12

/13

1/1

4

2/1

4

3/1

4

4/1

4

5/1

4

6/1

4

7/1

4

8/1

4

9/1

4

10

/14

11

/14

12

/14

1/1

5

2/1

5

3/1

5

4/1

5

5/1

5

6/1

5

7/1

5

8/1

5

Temperature 649.99 mb Layer Statistics

Water Vapor Percent Error 649.8 mb Layer Statistics

RM

S

RM

S

2.5

1.94

1.38

0.81

0.25

80

60

40

20

0

sum winter summer winter sum


11

1

NARCS NPROVS

Maritime IR

Temperature 649.99 mb Layer Statistics

Water Vapor Percent Error 649.8 mb Layer Statistics

7/1

3

8/1

3

9/1

3

11

/13

10

/13

12

/13

1/1

4

2/1

4

3/1

4

4/1

4

5/1

4

6/1

4

7/1

4

8/1

4

9/1

4

10

/14

11

/14

12

/14

1/1

5

2/1

5

3/1

5

4/1

5

5/1

5

6/1

5

7/1

5

8/1

5

7/1

3

8/1

3

9/1

3

11

/13

10

/13

12

/13

1/1

4

2/1

4

3/1

4

4/1

4

5/1

4

6/1

4

7/1

4

8/1

4

9/1

4

10

/14

11

/14

12

/14

1/1

5

2/1

5

3/1

5

4/1

5

5/1

5

6/1

5

7/1

5

8/1

5

RM

S

RM

S

2.5

1.94

1.38

0.81

0.25

80

60

40

20

0

GFS Forecast

GFS CFSR

ECMWF Analysis

AIRS Aqua IR+MW

NUCAPS IR+MW





GFS Forecast

GFS CFSR

ECMWF Analysis

AIRS Aqua IR+MW

NUCAPS IR+MW







12

0.3K improvement; less for AIRS 2

10% improvement

NARCS NPROVS

13

5/14 9/14 1/15 5/15 9/15

5/14 9/14 1/15 5/15 9/15

Maritime

Land

… sample 500 per week

… sample 4000 per week

weekly average SAT-minus-RAOB water vapor fraction

217 hPa

217 hPa

NARCS NPROVS

14

maritime land

PDISP NPROVS

15

NPROVS vs NPROVS+

16

Oct 14 Sept 15

Oct 14 Sept 15

NPROVS+

NPROVS

NARCS

monthly

weekly

(150)

(1500)

17

Oct 14

Oct 14

Sept 15

Sept 15

NPROVS+

NPROVS

NARCS

monthly

weekly

(150)

(1500)

18

H20 Weighting

3 Methods • W=1

– %RMS = 100*SQRT{ sum{((q_ret-q_tru)/q_tru)^2)} / N}

– %BIAS = 100*sum{((q_ret-q_tru)/q_tru)} / N

• W=q_tru – %RMS = 100*SQRT{sum{(q_ret-q_tru)^2/q_tru}} / SQRT{

sum{q_tru}}

– %BIAS = 100*sum{q_ret-q_tru} / sum{q_tru}

• W=q_tru^2 – %RMS = 100*SQRT{sum{(q_ret-q_tru)^2} / SQRT{sum{q_tru^2}}

– %BIAS = 100*sum{q_tru*(q_ret-q_tru)} / sum{q_tru^2}

• Equation pairs shown above can be used to compute standard deviation, %SDV = SQRT{%RMS^2 - %BIAS^2}

• Items shown in red are the original calculation done in simstat and would produce errors, due to the inconsistency, if standard deviation was computed

20

W W2

PDISP NPROVS+

21

W W2

PDISP NPROVS+

GRUAN vs GTS RAOB:

2011 vs 2012 vs 2013

(NPROVS+)

July 2014

Day 2011

VAISALA(digicora) - GRUAN Bias

Temp

H20 vapor

Fraction (%)

Conv RAOB cooler Conv RAOB dryer

PDISP NPROVS+


Temp

H20 vapor

Fraction (%)

Day 2012

PDISP NPROVS+



Temp H20 vapor

Fraction (%)

Day 2013

PDISP NPROVS+


Temp

H20 vapor

Fraction (%)

Nite 2013


PDISP NPROVS+

Radiosonde type relative humidity (RH) bias

27

RAOB-minus-SAT 300 hPa RH Bias

(Calculated RAOB BT)-minus-(observed SAT BT) @ 183+/- 1 gHz

RAOB aloft too dry

particularly during day

RS92

Sun et al. (2010, JGR)

Bomin Sun

Daytime samples … 317 Polar, 373 midLat, 75 Tropic

PDISP NPROVS+

Day

Day

Nite

Nite

Polar/mid-lat Polar/mid-lat

Nite

Tropic Tropic

H20 Vapor 2012 PDISP NPROVS+

30

IASI RH vs cloud fraction

Change of IASI (dotted) and RAOB (solid) relative humidity with

effective cloud amount for different atmospheric layers. IASI pass

qc retrievals and collocated RAOB are used.

Bomin Sun

NWP vs Conv RAOB:

2013 vs 2015

(NPROVS)

32 10-day PDISP Collocation Dataset

PDISP NPROVS

NWP

2015

PDISP NPROVS

34 NWP

2013

PDISP NPROVS

35

NPROVS+

2011

PDISP NPROVS+

150 hPa

36

PDISP NPROVS

150 hPa

Given two measurement (m1, m2), their uncertainty (u1, u2) and their

systematic variability (), then two observations are consistent if

“k” .le. 2

1)

2

2

2

1

2

21 uukmm

u2 is GRUAN uncertainty … given

u1 is SAT (or NWP) profile uncertainty … needed

σ is systematic variability between given platform pair… needed

GRUAN Reference Measurement Principles

37 NPROVS+

“k” based on GRUAN Uncertainty

Temperature Temperature

AIRS AQUA ECMWF NUCAPS NPP

Baseline: GRUAN Radiosonde

“Spec”

RMS (K) 2

for “k”=2, sum uncertainty under square root (1) in vicinity 500 hPa:

for SAT is 18(u1**2) … for ECMWF is 3(u1**2)

8.5 3.5

2)For ECMWF, assuming u1 ~ u2 (0.13K), then ~ 0.1K thru layer σ

0.8 1.6

11

PDISP NPROVS+

H20 Vapor Fraction H20 Mix Ratio

AIRS AQUA ECMWF NUCAPS NPP

Baseline: GRUAN Radiosonde

“Spec”

RMS ( % ) “k”

based on GRUAN Uncertainty 2

for “k”=2, sum uncertainty under square root (1) in vicinity 600 hPa

for SAT is 12(u1**2) … for ECMWF is 4(u1**2)

7 4 20 35

12

PDISP NPROVS+

CalWater 2/ACAPEX Field Campaign

Platform Range of Obs Duration Types of sensors

AR Observatories and Hydro-Met Testbed

ARO sites: CA(4), OR(2), WA(1)

Full campaign Snow level radar (S-band), 449 MHz wind profilers, soil moisture, 10 meter surface tower

NOAA WP-3D 1-22 kft, 4000 km range

80h over 4 weeks ~150 dropsondes, W-band radar (clouds), IWRAP Radar, Tail Dopper Radar, Cloud Probes, SFMR

NOAA G-IV 1-45 kft 90h over 6 weeks ~300 dropsondes, Tail Doppler Radar, NOAA O3, SFMR

DOE G-1 with ~40 instruments

1-23 kft 120h over 8 weeks

Cloud properties (Liq/water content, size), aerosol properties (concentration, size, CCN), trace gases (H2O, O3, CO)

NOAA R.H. Brown Can move 5 deg/day to stay within AR

30 days AMF2: Aerosol Observing System, Ka ,X, W-Band Cloud Radars, DOE, Micropulse LIDAR, Wind Speed, Rain Guages RS-92 Sondes: ~260 (~half dedicated overpass time)

• Interagency Campaign: • Scripps (Marty Ralph, Kim Prather) • NOAA (Allen White, Ryan Spackman) • DOE (PI: L. Ruby Leung) ACAPEX = ARM

Cloud Aerosol Precipitation Experiment • White paper at • http://esrl.noaa.gov/psd/calwater

11/13/2015 Gambacorta et al. 40

http://esrl.noaa.gov/psd/calwater

GRUAN and JPSS funded Dedicated (S-NPP) RAOB Sites

Over 10,000 RAOBS (1000 dedicated) available since July 2013

N P R O V S +

41

PDISP

CALWATER RAOB collocated with NUCAPS

All

(169)

Collocated with NUCAPS

(150)

42

PDISP


+/- 1 hour, 50km

(61)


+/- 1 hour, 50km

IR+MW pass QC

(24)

43 CALWATER RAOB collocated with NUCAPS

PDISP

H20 frac

+/- 2 to 6hr

Sample of NUCAPS IR which pass QC

ECMWF Analysis MIRS NPP Test NUCAPS NPP

Baseline: Reference Sonde GRUAN RAOB

44

PDISP

ECMWF Analysis MIRS NPP Test NUCAPS NPP

Baseline: Reference Sonde GRUAN RAOB

45 Sample of NUCAPS IR which pass QC

H20 frac

+/- 1hr

PDISP

ECMWF 12Z ( Feb 7th) to 6Z (Feb 8th)

12Z 18Z

00Z 06Z

X

X X

X

46

650 hPa ODS

SAT @ 1003Z … ECMWF @ 12Z … RAOB @ 1000Z

RAOB

ECMWF Analysis

MIRS NPP

NUCAPS (old)

47

xxxxx best

collocated

PDISP

SAT @ 2123Z … ECMWF @ 18Z … RAOB @ 2032Z 48

xxxxx

RAOB

ECMWF Analysis

MIRS NPP

NUCAPS (old)

PDISP


xxxxx

RAOB

ECMWF Analysis

MIRS NPP

NUCAPS (old)

PDISP


xxxxx

RAOB

ECMWF Analysis

MIRS NPP

NUCAPS (old)

best

collocated

PDISP

SAT @ … ECMWF @ 06Z … RAOB @ 0256Z 51

xxxxx

RAOB

ECMWF Analysis

PDISP

SAT @ … ECMWF @ 06Z … RAOB @ 0435Z 52

xxxxx

RAOB

ECMWF Analysis

PDISP

SAT @ 21Z Feb 7th … ECMWF @ 18Z

NUCAPS

IR+MW

NUCAPS

MW only

650 hPa H20 Vapor

ECMWF

ANL

53

NUCAPS

IR+MW

NUCAPS

MW-only

MiRS NPP

ECMWF

ANL

X X

X X

ODS


xxxxx

RAOB

ECMWF Analysis

MIRS NPP

NUCAPS (old)

PDISP


xxxxx

RAOB

ECMWF Analysis

MIRS NPP

NUCAPS (old)

best

collocated

PDISP

SUMMARY • Assessments of satellite, ground and nwp moisture profiles

available via NPROVS (conventional) and NPROVS+ (reference)

• LTM , shorter term ensemble and instantaneous assessments

demonstrated

• HIRS water vapor (Shi) to be appended to NPROVS+ (Sun)

• H20 vapor challenging and requires tight spatial and temporal

sampling … regional / seasonal

• Weighting important for RMS but not for “k” analysis

• Coordination with ongoing GAIA-CLIM (GAIA-CLIM WP3/WP5

Meeting Darmstadt, 29/Oct/2015 )

• Identify key areas of instantaneous assessment section for report

56

57

LIN

Winter

2014

PDISP NPROVS

Documents

5th GEWEX G-VAP