Observation operators for wind-profiler reflectivity, weather- radar dual-polarization observations, and weather-radar refractivity Olivier Caumont et.

Observation operators

for wind-profiler

reflectivity, weather-

radar dual-polarization

observations, and

weather-radar

refractivity

Olivier Caumont et al.

CNRM-GAME

IODA-MED meeting

31 January 2013

2

IODA-MED deliverables

3

Outline of talk

1. Reflectivity from wind profilerO. Caumont1, K. Y. Nawanti2, F. Saïd3, B. Campistron3, Y. Bezombes3, S. Derrien3, O. Bousquet1, J.-M. Donier1, T. Douffet1, O. Garrouste1, J. Van Baelen4, J.-L. Caccia5, H. Luce5

1. CNRM-GAME2. ENIT (Tunisia)3. LA/CRA4. LaMP5. LSEET

2. Weather-radar dual-polarization observations3. Weather-radar refractivity

4

Reflectivity from wind profiler

Observations = vertical profiles of: Doppler velocity (V, already assimilated) Reflectivity or Cn² (refractivity turbulence structure constant) Doppler spectrum width ( eddy dissipation rate ε)

Development of an observation operator (Stankov et al. 2003):

© S. DerrienCRA’s UHF wind profiler

23/4

23/22

3

41.2

dz

dV

L

LMC hw

n

dz

dq

Tg

N

T

q

T

pM

7800156001106.77

26

dz

dT

T

gN 2

Lw/L: ratio of the outer length scales for potential refractive index and shear (set to 4 here)

p: pressure (hPa)T: temperature (K)q: specific humidity (kg kg-1)g: acceleration of gravity (9.81 m s-2)Γ: adiabatic lapse rate (9.8 10-3 K m-1)

(Brunt-Väisälä frequency)

5

Reflectivity from wind profiler

Comparison of Cn2 from UHF profiler and radiosonde

5 to 10 % error, with large variability

Vertical profiles of mean relative error between Cn2 (dB) measured by LA’s UHF radar and simulated from radiosonde (solid, thick line), and their standard deviations (dashed lines).

Low mode

High mode

(F. Saïd)

6

a) b)

Reflectivity from wind profiler

Validation of observation operator: Boundary Layer Late Afternoon and Sunset Turbulence (BLLAST,

http://bllast.sedoo.fr/) field campaign in southwestern France (summer 2011)

CRA and CNRM-GAME UHF profilers

overall consistent; some discrepancies (black ellipses) but also some notable matching patterns (blue ellipses).

Times series of log10(Cn2) vertical profiles (a) measured by CNRM-GAME’s UHF radar in high mode and (b) simulated from Arome analyses between 15 and 28 June 2011

7

Reflectivity from wind profiler

Future work:

Study each component of the observation operator separately o Compare Arome, wind profiler, and independent measurements (RS)o For each component, determine which is the best estimate (Arome or wind

profiler)

Investigate the sensitivity of Cn2 to each of these components

If needed, use alternative formulations (other than Stankov’s)

8

Outline of talk

1. Reflectivity from wind profiler2. Weather-radar dual-polarization observations

O. Caumont1, C. Augros2, P. Tabary2, V. Ducrocq1

1. GAME2. DSO/CMR (Météo-France)

3. Weather-radar refractivity

9

Weather-radar dual-polarization data

Observations: Horizontal+vertical polarizations additional

information about hydrometeors (nature, size, shape, etc.)

Already 12 dual-pol radars (out of 26) in the French operational network (purple, blue, and green); more to come!

Widely-used technology on research radars

Complex observations need forsignificant work on observation operator

Work done: On-going development of a

versatile observation operator in Meso-NH’s post-processing

10

Weather-radar dual-polarization data

Status of observation operator in Meso-NH’s post-processing: Gridpoint observations of ZHH, ZDR, KDP (J.-P. Pinty):

o Rayleigh and Jameson scatteringo ZDR and KDP only sensitive to rain

11

Weather-radar dual-polarization data: Gridpoint observation operator

Aaaaa Bbbb

o Cccc Ddddd

Reflectivity (ZHH in dBZ)

Meso-NH simulation of 8 Oct. 2002, 21 UTCHorizontal cross sections at 2 km MSL

Differential reflectivity (ZDR in dB)

Specific differential phase shift (KDP in ° km-1)

12

Weather-radar dual-polarization data

Status of observation operator in Meso-NH’s post-processing: Gridpoint observations of ZHH, ZDR, KDP (J.-P. Pinty):

o Rayleigh and Jameson scatteringo ZDR and KDP only sensitive to rain

PPI (cones) of ZHH, ZDR, KDP, and more (C. Augros et al.):o Various scattering algorithms (Rayleigh(-for-spheroids), Mie, T-matrix)o Propagation ((differential) attenuation) and broadening effectso Move from T-matrix dynamic, burdensome scattering computations to lookup

tables to increase numerical efficiencyo Addition of ρHV

13

Rayleigh MieRayleigh spheroids

Rayleigh 6th order T-matrix

Meso-NH simulation of 8 Oct. 2002, 21 UTC

S-band radar (Nîmes)

Elevation: 2°

T=15°C everywhere

ZHH due to rain (dBZ)

Weather-radar dual-polarization data: PPI simulations (C. Augros)

14

ZDR (dB)

Rayleigh MieRayleigh spheroids

Rayleigh 6th order T-matrix

Weather-radar dual-polarization data: PPI simulations

Meso-NH simulation of 8 Oct. 2002, 21 UTC

S-band radar (Nîmes)

Elevation: 2°

T=15°C everywhere

(C. Augros)

15

KDP

Rayleigh MieRayleigh spheroids

Rayleigh 6th order T-matrix

KDP (° km-1)

Weather-radar dual-polarization data: PPI simulations

Meso-NH simulation of 8 Oct. 2002, 21 UTC

S-band radar (Nîmes)

Elevation: 2°

T=15°C everywhere

(C. Augros)

16

Weather-radar dual-polarization data

Status of observation operator in Meso-NH’s post-processing: Gridpoint observations of ZHH, ZDR, KDP (J.-P. Pinty):

o Rayleigh and Jameson scatteringo ZDR and KDP only sensitive to rain

PPI (cones) of ZHH, ZDR, KDP, and more (C. Augros et al.):o Various scattering algorithms (Rayleigh(-for-spheroids), Mie, T-matrix)o Propagation ((differential) attenuation) and broadening effectso Move from T-matrix dynamic, burdensome scattering computations to lookup

tables to increase numerical efficiencyo Addition of ρHVo Short-term plans:

Implement scattering models for icy hydrometeors Validation for a HyMeX case Add to official Meso-NH version & Write documentation

17

Outline of talk

1. Reflectivity from wind profiler2. Weather-radar dual-polarization observations3. Weather-radar refractivity

O. Caumont1, A. Foray2, L. Besson3, J. Parent du Châtelet3, C. Boudjabi4

1. GAME2. DIRIC (Météo-France)3. DSO/CMI (Météo-France)4. LATMOS

18

Weather-radar refractivity

radarr1 r2

target #1 target #2

radarbeam

Principle: Based on radar pulse’s propagation time through the atmosphere, which depends

on the index of refraction of air (n) or refractivity (N=(n-1)∙106) Phase change between radar and target and over time (δΔφ) depends on path-

averaged refractivity change over time (δN). Initially formulated for klystron transmitters by Fabry et al. (1997) Refractivity related to pressure (p in hPa), temperature (T in K), and water vapour

partial pressure (e in hPa) through N=77.6 p/T+3.73 105 e/T²

Work done: Formulation for magnetron transmitters (Parent du Châtelet et al. 2012) Link between refractivity and atmospheric phenomena (Besson et al. 2012) Observation operator for refractivity + sensitivity study Long-term comparisons of radar observations vs. Arome Real-time production of refractivity maps during HyMeX SOP1

19

Link between N and (T,q) (Besson et al. 2012)

Mediterranean

surface rain rate refractivity

temperature relative humidity

warm

coldmoist

moist

altitude

low level

20 Oct 2008

20

Monitoring: Time evolution for Nîmes radar

N integrated over time ( ΔN) Circle corresponds to 40-km range Filtering of observations

Radar observations(PPI at 0.6°-elevation)

Arome analyses(beam height at 2 m AGL)

21

Sensitivity to beam height

Different models LVLXX: model level (~terrain-following) RAYCT: 4/3-Earth-model path of beam centre RAYGE: 4/3-Earth-model path that hits the ground (most accurate operator) AWS: AWS- vs. model-derived refractivity

Best results when beam within ~60 m AGL No significant improvement when using more accurate models

(Caumont et al. in rev. BLM)

2 m AGL

~60 m AGL

~15 m AGL

~113 m AGL

22

Long-term observation-vs.-model statistics

Good consistency between refractivity observations and model Similar spatial variability

Time series of domain-averaged observation (red) and model (green) refractivity difference since the beginning of the period. Corresponding standard deviations in blue and purple, respectively.

(Caumont et al. in rev. BLM)

23

Real-time refractivity maps during HyMeX SOP1

Fabry’s algorithm adapted to magnetrons http://sop.hymex.org/, Home>Observations>Radars>Single operational

radars>refractivity

Humide

Sec

24

Weather-radar refractivity

Future/on-going work: Improvement of the raw data quality (use of dual-polarization and higher elevations,

use of quality index) Deeper understanding of measurement physics (including role of turbulence) Use in process studies

Further reading: Besson, L., C. Boudjabi, O. Caumont, J. Parent du Châtelet, 2012: Links between

refractivity characteristics and weather phenomena measured by precipitation radar. Bound.-Lay. Meteorol., 143(1), 77–95, DOI : 10.1007/s10546-011-9656-7.

Caumont, O., A. Foray, L. Besson, J. Parent du Châtelet: An observation operator for radar refractivity change: Comparison of observations and convective-scale simulations. Bound.-Lay. Meteorol., in revision.

Parent du Châtelet, J., C. Boudjabi, L. Besson, O. Caumont, 2012: Errors caused by long-term drifts of magnetron frequencies for refractivity measurement with a radar: Theoretical formulation and initial validation. J. Atmos. Oceanic Technol., 29(10), 1428–1434, DOI: 10.1175/JTECH-D-12-00070.1.

Thank you for your attention!