water vapour raman lidar and microwave profiler ... · profilers – higher computational burden...

87th AMS Annual Meeting

Water Water vapourvapour RamanRaman lidarlidar and and microwavemicrowave profilerprofiler: :

comparisonscomparisons and and synergiessynergies

F. Madonna, A. Amodeo, C. Cornacchia, G. D’Amico, L. Mona, G. Pappalardo

Istituto di Metodologie per l’Analisi Ambientale CNR-IMAA, Potenza, Italy

WG1 Workshop WaVaCS – Lindenberg 21 -23 May 2008


BS

BS

DM

IF386 nm

PMTM

Nd:YAGLaser

M

M

M = mirror

P = pinhole

L = collimation lens

DM = dichroic mirror

BS = beam splitter

PMT = photomultiplier

IF = interferential filter

M = mirror

P = pinhole

L = collimation lens

DM = dichroic mirror

BS = beam splitter

PMT = photomultiplier

IF = interferential filter

M

L

P

355

nm

IF407 nm

PMT

DM

M

M

IF355 nm

PMT

IF407 nm

PMT

IF355 nm

PMT

BS

IF386 nm

PMT

DETECTORS: photomultiplierTHORN EMI 9202 QB355, 386, 407 nm

ACQUISITION: photon counting modeEG&G MCS - PCIMinimum dwell time 100 nsBandwidth 150 MHzPHILLIPS SCIENT. Fast discriminatorBandwidth 300 MHz

RECEIVING SYSTEM: Cassegrain telescope

Primary mirror diameter 0.5 mCombined focal length 5 mNight-time field of view 1 ÷ 2 mradDaytime field of view 0.2 mrad

LASER: Nd:YAG laser (Coherent – Infinity)

Max. Pulse Energy@ 355nm 170mJMaximum repetition rate 100HzBeam divergence <0.3 mradPulse duration 3 ÷ 4 ns

SPECTRAL SELECTION: interferential filter Wavelengths (nm) 386 407 355 Out of band rejection <10-10 <10-8 <10-6 Night-time bandwidth (nm) ~1.0 ~1.0 ~1.0 Daytime bandwidth (nm) ~ 0.5 ~ 0.5 ~ 0.5 Trasmission efficiency ~ 60% ~ 60% ~ 33%

CNR-IMAA Water vapour Raman Lidar


3


Measurements

Operative since May 2002

2 measurements per week are systematically performed

Since November 2006 systematic measurements performed with

PEARL (Potenza EARlinet lidar)

Involvements

Validation of ENVISAT/MIPAS water vapour product

NDACC (The Network for the Detection of Atmospheric Composition

Change) for UT/LS water vapour monitoring

Measurements Campaigns

EAQUATE campaign (6-10 September 2004)

LAUNCH campaign (September – October 2005)




Water Vapor profiles characteristics


Night time measurements

60 m – 12000 m a.l.s.

15-150 m vertical resolution

10 minutes temporal resolution

Daytime measurements

90 m – 4500 m a.l.s.

15 m vertical resolution

10 minutes temporal resolution

0

2

4

6

8

10

12

1E-3 0.01 0.1 1 10

Lidar 19:48-19:57 UT Sonda 19:52 UT

Alti

tude

a.s.

l. (k

m)

Water vapor mixing ratio (g/kg)

11 November 2002

0

1

2

3

4

5

6

-1 0 1 2 3 4 5 6 7 8 9 10Water Vapor Mixing Ratio (g/kg)

Alti

tude

a.s.

l. (k

m)

Lidar 12:07-12:17 UT Sonde start 12:12 UT

02 October 2005





● Routine comparisons with numerical models (CloudNET)

● Long time series of measurements (campaigns, special events)

● Improvement of model parameterization


CNR-IMAA Water vapour Microwave profiler

Courtesy of Radiometrics Corporation, Boulder, CO, USA

The microwave profiler measures the sky brightness temperature at 12 frequencies:

5 frequencies are in the K-band (22.235, 23.0335, 23.835, 26.235, 30 GHz), around 22 GHz water vapour resonance band;

7 frequencies are in the V-band (51.250, 52.280, 53.850, 54.940, 56.660, 57.290, 58.800 GHz), around 60 GHz spyn-rotation oxygen band.

Rate: > 12 sAccuracy: 0.5 KResolution: 0.25 KRange: 0 -700 KOperational range: -20° - 50° CScanning capabilities: 3D skyBeam width: 6.3° at 22.2 GHz, 4.9° at 30 GHz, 2.5° at 51.3 GHz and 2.4° at 58.8 GHz (full width half power)

Output products (Neural network retrieval)Temperature, water vapour, relative humidity and cloud liquid water profiles up to 10 km above the ground

Vertical step: 100 m from 0 to 1 km, 250 m above up to 10 km

Ancillary parameters.Cloud base temprature measured using an infrared thermometer (IRT).Surface meterological parameters (T, RH, p)

The microwave profiler measures the sky brightness temperature at 12 frequencies:

5 frequencies are in the K-band (22.235, 23.0335, 23.835, 26.235, 30 GHz), around 22 GHz water vapour resonance band;

7 frequencies are in the V-band (51.250, 52.280, 53.850, 54.940, 56.660, 57.290, 58.800 GHz), around 60 GHz spyn-rotation oxygen band.

Rate: > 12 sAccuracy: 0.5 KResolution: 0.25 KRange: 0 -700 KOperational range: -20° - 50° CScanning capabilities: 3D skyBeam width: 6.3° at 22.2 GHz, 4.9° at 30 GHz, 2.5° at 51.3 GHz and 2.4° at 58.8 GHz (full width half power)

Output products (Neural network retrieval)Temperature, water vapour, relative humidity and cloud liquid water profiles up to 10 km above the ground

Vertical step: 100 m from 0 to 1 km, 250 m above up to 10 km

Ancillary parameters.Cloud base temprature measured using an infrared thermometer (IRT).Surface meterological parameters (T, RH, p)



CNR-IMAA Water vapour Microwave profilerMeasurements

Operative since February 2004

Continuous measurements (24h, 7 days per week)

Involvements

CloudNET

Departement of the Italian Civil Protection

Measurements Campaigns

EAQUATE campaign (6-10 September 2004)

LAUNCH campaign (Ziegendorf, Germany, August – October 2005)

COPS (Hornisgrinde, Germany, June – August 2007)



• Comparisons with numerical prediction models (DWD- Lokall Modell) are routinely performed.

• Operational measurements (special scanning strategies)• Intercomparison campaign




Comparisons


● Differences in the resolution

● Microwave continuous measurements

● Possible synergies and integration LIDAR – Microwave profiler


The IPWV retieved using the neural network algorithm is used to caiibrate the water

vapour Raman lidar profile

Calibration

• Calibration factor results constant within 5%

Vaisala radiosonde RS80, RS90 and RS92 PTU measurements (also wind speed/direction for RS92)

Calibration campaigns with co-located radiosondes

More than 200 radiosonde launches since May 2002

Calibration is checked systematically with radiosondes

1100 1200 1300 1400 1500 1600 17000.0

0.5

1.0

1.5

2.0

2.5

3.0CNR-IMAA 09 September 2004

Wat

er V

apor

Col

umna

r C

onte

nt (c

m)

Minutes of the day

Radiometer Lidar

Continuous monitoring of Raman lidar calibration

constant stability

MICROWAVE PROFILER

RADIOSONDE



Case study: 30 Sep – 03 Oct 2005

๏ Presence of a dry intrusion around 5 km a.g.l.

๏ Minimum penetration level 2.5 km a.g.l. with a minimumvalue of the WVMR of 0.16 g/kg and a corresponding RH of 16 %.

๏ Strong uplift of moisture structures up to 7.5 km a.g.l. after the passage of the intrusion



• Intrusion of a dry layer coming from about 5.5 km a.g.l. and moving down to less than 2 km on 2 October 2005.

• At 00:00 UTC on 2 October, the lidar resolves three distinct dry layers. • The highest layer is characterized by a minimum WVMR of 0.18 g/kg and a

maximum penetration level of 1.75 km a.g.l. where a WVMR value of 0.46 g/kg is measured.




• Images from the VISSR (on board of Meteosat) water vapour channel at 5.7 -7.1 µm for 28 September 2005 (a), 29 September 2005 (b), 30 September 2005 (c) and 1 October 2005 (d) at 06:00 UTC (Courtesy of University of Dundee).




Integration techniques• Kalman filter integration MODULAR scheme

TIME UPDATE

RTM

SOEKF LIDAR SOEKF MWP

FINAL GUESS

RH (MWP)

FIRST GUESS


● Low time consuming

● Second-Ordereffects in the retrieval

● Expansion to other profilers


Integration techniques• Extended Kalman filter integration (1 Filter)

TIME UPDATE

RTM

FOEKF MEAS. UPDATE

FINAL GUESS

RH (MWP)

FIRST GUESS

F. MADONNA, C. CORNACCHIA, G. D'AMICO, L. MONA, A. AMODEO, and G. PAPPALARDO, Integration of Raman lidar and microwave profling measurements of water vapor using the Kalmanfiltering, in preparation.


● High time consuming

● Integration

● Expansion to other profilers – Higher computational burden

● Technique to reduce the burden


Work in progress…..

• COPS (Intensive scanning measurements)

• CloudNET (Raman lidar, microwave profiler, ceilometer)

• EuroClouds proposal



ThankThank youyou



• Potential vorticity at 300 hPa surface (in PVU, 1 PVU = 10−6 m2 s−1 K kg−1) on 28 September 2005 at 06:00 UTC from 31-level ECMWF analysis with 1.125° resolution.





Interpolated time series of the potential temperature retrieved using the pressure and temperature profiles measured with the radiosoundingsperformed at Lindenberg observatory on 30 September 2005.



Integration: preliminary resultsCASE I: 26-27/02/2006 (Night time)



Synergy

17.24.47 17.49.47 18.14.47 18.39.46 19.04.453

4

5

6

7

8

0.0

0.1

0.2

0.3

Clo

ud B

ase

(km

)

Time (UT)

Integrated Liquid Water (m

m)

18:02 18:12 18:22 18:32 18:42 18:52 19:020

10

20

30

40

50

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Optical D

ept @355 nm

Lida

r R

atio

(sr)

Time (UT)

LR OD

Optical depth in the last 10 minutes is 0.03

The time variation of the lidarratio can be correlated to the cloud phase next to its base.



18:00 18:20 18:40 19:000.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

LWC

(mm

) - O

D

Time UTC

NN+IRT NN+LIDAR OD

18:00 18:20 18:40 19:00

6000

6500

7000

7500

8000

8500

9000

9500

IRT PEARL

Clo

ud B

ase

Hei

ght (

m a

.s.l)

Time UTC

Synergy


Temperature – 20 June 2007

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS05:52

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS05:52

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS07:52

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS07:52

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS11:12

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS11:12

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS16:57

225 250 275 300 3250

1000

2000

3000

4000

5000

6000

7000

8000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS16:57

Temperature – 24 August 2007

250 275 3000

1000

2000

3000

4000

5000

6000

H

eigh

t (m

.a.s

.l)

Temperature (K)

MP3014 RS 05:32

250 275 3000

1000

2000

3000

4000

5000

6000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS 07:56

250 275 3000

1000

2000

3000

4000

5000

6000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS 10:58

250 275 3000

1000

2000

3000

4000

5000

6000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS 13:56

250 275 3000

1000

2000

3000

4000

5000

6000

Hei

ght (

m .a

.s.l)

Temperature (K)

MP3014 RS 17:01

Humidity – Models (15 July 2007)Met Office NAE (North Atlantic and European) model

MP3014 CNR-IMAA microwave profiler

Humidity – Models (20 July 2007)


Met Office NAE (North Atlantic and European) model

Humidity – Models (01 August 2007)Met Office NAE (North Atlantic and European) model


87th AMS Annual MeetingWG1 Workshop WaVaCS – Lindenberg 21 -23 May 2008


CLOUDNET MODELS

The EU CloudNET project offers an extended database of water

vapor profiles provided by five operational forecast models of

ECMWF, the MetOffice, MeteoFrance, KNMI and DWD.

Cloudnet is an European pilot network of stations for

clouds profiles observation.

Horizontal Resolution Temporal Resolution Level Number Forecast Time

DWD 7km 1h 35 t-12

ECMWF 40km 1h 60 t-12

MeteoFrance 25km 1h 41 t-12

MetOfficeGlobal 60km 3h 38 t

SMHI L24 44km 1h 24 t-12

SMHI L40 44km 1h 40 t-12


COMPARISON APPROACH

1. Lidar high resolution profiles are reduced into a large grid boxes: vertical and

temporal resolution are reduced to those of the model

2. The new time grid is calculated on the base of wind speed to take into account

the advection time.

3. Only lidar data with a total error less than 20% are considered.

Long record of measurements (about 30 hours) has been collected at CNR-

IMAA on 1-3 October 2005, during the LAUNCH 2005 international campaign,.


COMPARISON APPROACH - AN EXAMPLELidar at ECMWF model gridLidar at ECMWF model grid

ECMWF model dataECMWF model data

Lidar at MeteoFrance model gridLidar at MeteoFrance model grid

MeteoFrance model dataMeteoFrance model data


COMPARISON APPROACH - AN EXAMPLE

More quantitative comparison can be carried out in terms of the probability

density function (pdf) calculated for both models and lidar data reduced at

models resolutions in different altitude ranges.

0 1 2 3 4 5 6 7 8 9 100.0

0.1

0.2

0 1 2 3 4 5 6 7 8 9 100.0

0.1

0.2

Freq

uenc

y of

Occ

uren

ces

Freq

uenc

y of

Occ

uren

ces

Water Vapor Mixing Ratio (g/kg)

0-2 km of Altitude

Meteofrance model

Lidar Measurements

0 1 2 3 4 5 6 7 8 9 100.0

0.1

0.2

0 1 2 3 4 5 6 7 8 9 100.0

0.1

0.2

Freq

uenc

y of

Occ

uren

ces

Freq

uenc

y of

Occ

uren

ces


0-2 km of Altitude

ECMWF model

Lidar Measurements

ECMWF modelECMWF model MeteoFrance modelMeteoFrance model


LONG TERM COMPARISON

Sistematic comparison has been perfomed between lidar data

and models.

1 comparison per month is carried for September 2002 –

February 2006 period.

The longest measurements run is chosen for each month.

Models and lidar data reduced at models grid are collected for

different altitude ranges.

Lidar and models Pdf are compared for 0-2 km, 2-4 km, 4-6

km, 6-8 km altitude ranges.


LONG TERM COMPARISON – ECMWF

0 2 4 6 8 100.0

0.1

0.2

0.3

0.4

0.5 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25

0.30 Lidar Measurements

Freq

uenc

y of

occ

uren

ces

ECMWF model

Freq

uenc

y of

occ

uren

ces

Water Vapor Mixing Ratio (g/kg)0 1 2 3 4 5 6

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7 0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4


Freq

uenc

y of

occ

uren

ces

ECMWF model

Freq

uenc

y of

occ

uren

ces


4-6 km4-6 km 6-8 km6-8 km

0 2 4 6 8 100.00

0.05

0.10

0.15

0.20 0 2 4 6 8 100.00

0.05

0.10

0.15


Freq

uenc

y of

occ

uren

ces

ECMWF model

Freq

uenc

y of

occ

uren

ces


0-2 km0-2 km

3 modes are sligthly shifted in

ECMWF

0 2 4 6 8 100.000.050.100.150.200.250.300.350.40 0 2 4 6 8 100.00

0.05

0.10

0.15


Freq

uenc

y of

occ

uren

ces

ECMWF model

Freq

uenc

y of

occ

uren

ces


2-4 km2-4 km

Higher values occurrences

underestimated by ECMWF

Pdf shape is generally well reconstructed in each interval

range


LONG TERM COMPARISON – 6-8 km Altitude Range

0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4

0.5 0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4


Freq

uenc

y of

occ

uren

ces

DWD model

Freq

uenc

y of

occ

uren

ces


0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4

0.5

0.6 0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4

0.5


Freq

uenc

y of

occ

uren

ces

smhi L40 model

Freq

uenc

y of

occ

uren

ces


0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7 0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4


Freq

uenc

y of

occ

uren

ces

MeteoFrance model

Freq

uenc

y of

occ

uren

ces


0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4

0.5

0.6 0 1 2 3 4 5 60.0

0.1

0.2

0.3

0.4

0.5


Freq

uenc

y of

occ

uren

ces

MetOfficeGlobal model

Freq

uenc

y of

occ

uren

ces


MetOfficeMetOffice SMHI L40SMHI L40

DWDDWD MeteoFranceMeteoFrance

Pdf well reconstructed, but larger pdf for observated wv high

values



0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25

0.30 0 2 4 6 8 100.000.050.100.150.200.250.300.350.40 Lidar Measurements

Freq

uenc

y of

occ

uren

ces

DWD model

Freq

uenc

y of

occ

uren

ces


DWDDWD

0 2 4 6 8 100.0

0.1

0.2

0.3

0.4

0.5 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25


Freq

uenc

y of

occ

uren

ces

MeteoFrance model

Freq

uenc

y of

occ

uren

ces


0 2 4 6 8 100.0

0.1

0.2

0.3

0.4

0.5

0.6 0 2 4 6 8 100.0

0.1

0.2

0.3

0.4

0.5


Freq

uenc

y of

occ

uren

ces


Freq

uenc

y of

occ

uren

ces

Water Vapor Mixing Ratio (g/kg)0 2 4 6 8 10

0.0

0.1

0.2

0.3

0.4

0.5

0.6 0 2 4 6 8 100.0

0.1

0.2

0.3

0.4

0.5


Freq

uenc

y of

occ

uren

ces

smhi L40 model

Freq

uenc

y of

occ

uren

ces



MeteoFranceMeteoFrance

Pdf well reconstructed, but larger pdf for observated wv high

values



0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25

0.30 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25


Freq

uenc

y of

occ

uren

ces

DWD model

Freq

uenc

y of

occ

uren

ces


0.000.050.100.150.200.250.300.350.40 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25


Freq

uenc

y of

occ

uren

ces

MeteoFrance model

Freq

uenc

y of

occ

uren

ces


0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25

0.30 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25


Freq

uenc

y of

occ

uren

ces


Freq

uenc

y of

occ

uren

ces


0.00

0.05

0.10

0.15

0.20

0.25

0.30 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25


Freq

uenc

y of

occ

uren

ces

smhi L40 model

Freq

uenc

y of

occ

uren

ces


DWDDWD



Pdf shape well reconstructed

2 modes observed

DWD overestimates the center of 2nd mode

Observed values in 0-8 g/kg range

Model underestimatesoccurrences of high values



0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25

0.30 0 2 4 6 8 100.00

0.05

0.10

0.15

0.20

0.25


Freq

uenc

y of

occ

uren

ces

DWD model

Freq

uenc

y of

occ

uren

ces


0.00

0.05

0.10

0.15

0.20 0 2 4 6 8 100.00

0.05

0.10

0.15


Freq

uenc

y of

occ

uren

ces

MeteoFrance model

Freq

uenc

y of

occ

uren

ces


0 2 4 6 8 100.00

0.05

0.10

0.15

0.20 0 2 4 6 8 100.00

0.05

0.10

0.15


Freq

uenc

y of

occ

uren

ces


Freq

uenc

y of

occ

uren

ces


0.00

0.05

0.10

0.15

0.20 0 2 4 6 8 100.00

0.05

0.10

0.15


Freq

uenc

y of

occ

uren

ces

smhi L40 model

Freq

uenc

y of

occ

uren

ces


DWDDWD



Fair agreement, but large differences for each model exist.


PRELIMINARY RESULTS

Pdf is well reconstructed even if data show also

higher wv values that models do not forecast

Fair agreement in terms of pdf, but

large differences for each model exist.

0-2 km altitude range

4-6 km and 6-8 km altitude ranges

2-4 km altitude range

Pdf shape is well reconstructed.

2 modes are observed on DWD grid, but DWD

overestimates the 2nd mode center.

MeteoFrance and ECMWF underestimate

occurrences of values > 2 g/kg


FUTURE PLANS

Investigation of possible seasonal behavior of

differences in model/observed data

Classification of data in cases with and without Free

Troposphere humid layers

Extending comparison dataset adding further data

water vapour raman lidar and microwave profiler ... · profilers – higher computational burden...

Documents