Water vapor estimates using simultaneous S and Ka

Water vapor estimates using simultaneous S and Ka band radar measurements Scott Ellis, Jothiram Vivekanandan NCAR, Boulder CO, USA

Background • NCAR S-Pol radar upgraded with simultaneous S- and Ka-band measurement capability S-band – Matched beam widths – Matched range gates • For Rayleigh scatterers reflectivity differences can be related to attenuation by liquid and gas antenna Ka-band antenna

Objectives • Retrieve path-integrated humidity • Retrieve range-resolved liquid water content (LWC) and – Differential gaseous absorption – Compare reflectivity at nearest edge of median volume diameter cloud (MVD) through clouds – Differential absorption through clouds Path integrated water vapor profiles + + + Range resolved cloud liquid Range Height – Create profile by plotting mid-point of path integrated estimates

Method • Remove non-meteorological targets • Determine where Rayleigh scattering approximation is valid at Ka-band – Use S-band dual-pol measurements • < 1 mm drops – Estimate D 0 from S-band ZDR – D 0 must be < 0. 5 mm Example PPI plot of Rayleigh mask of cloud field during RICO Blue = Mie Brown = Rayleigh • Z at S-band < 20 d. BZ • ZDR < 0. 4 d. B – Produce Rayleigh mask • Estimate attenuations – Liquid – gaseous Distance from radar (km)

• Run radiation model many times varying T, P and specific humidity (SH, g m-3) • Compute polynomial fit of SH to attenuation Specific humidity (g m-3) Method: Humidity retrieval 1 -way atm attenuation (d. B km-1) SH = 201. 40 A 3 – 209. 60 A 2 + 120. 55 A – 2. 25 Where SH is specific humidity (g m-3) and A is gaseous attenuation (d. B km-1)

Method: Liquid retrieval • Liquid water attenuation at Ka-band (Aka) linearly related to LWC (g m-3) – No dependence on drop size distribution – Small temperature correction (CT) • MVD (mm) retrieved from LWC and reflectivity (Ze, mm 6 m-3) LWC = 0. 74*Aka*CT MVD 3 = 2. 16 x 10 -4*Ze/LWC

Results from RICO RMS difference: sounding (g m-3) + radar retrieval – primary ray + Radar retrieval – secondary ray - Sounding Without dry layer With dry layer 0. 85 1. 40

Results from RICO Average sounding computed at height H as the average specific humidity from 0 to 2*H + radar retrieval – primary ray + Radar retrieval – secondary ray - Sounding - Average sounding

Results from RICO Sounding average humidity for secondary ray + primary ray + secondary ray - Sounding - Average sounding

Results from RICO RMS difference: sounding (g m-3) + radar retrieval – primary ray + Radar retrieval – secondary ray - Sounding 0. 75

LWC and MVD retrievals S-band reflectivity (d. BZ) with C-130 track displaying LWC (g m-3) collected from RICO 12 January, 2005 C-130

LWC and MVD retrievals S-band reflectivity (d. BZ) LWC (g m-3) C-130 Distance from radar (km)

LWC and MVD retrievals MVD (mm) LWC (g m-3) Outliers due to attenuation underestimates Distance from radar (km)

Future work -- planned • Automate algorithms • Further verification – Use aircraft/radar matching technique for LWC • Algorithm refinements – Humidity • Compute humidity – attenuation relationships for several layers and use most appropriate one • Compute optimal humidity profiles using primary and secondary rays – LWC/MVD • Improve attenuation estimation to remove outliers • Account for increased attenuation in non-Rayleigh rain

Future work -- planned • Compare dual-wavelength humidity with refractivity, GPS, water vapor DIAL, radiometer… during REFRACTT 06 and COPS 07 • Combine humidity retrieval with near-surface refractive index humidity for radar based 3 -D moisture field • Detection/quantification of super-cooled liquid water – Microphysics – Aviation safety

Future work -- desired • Partner with international community to verify satellite derived and model microphysical products – Cloudsat • Cloud fraction • LWC • Ice content – NASA GPM • Partner with data assimilation community

Thank you. Questions?

Discussion • Possible to obtain accurate path-integrated water vapor estimates in boundary layer • Both horizontal and vertical distributions can be obtained • Depends on cloud distribution • Path integration limits resolution • Not automated yet • Can be automated

Sources of error • Failure to exclude contamination • Errors in humidity- and liquid- attenuation relationship • Radar calibration errors • Modification of humidity by environmental conditions by clouds; e. g. moist, cool outflows

Sources of error • Radar calibration – Errors in reflectivity differences Errors in attenuation and humidity resulting from errors in d. BZ differences (S – Ka)

Method: potential primary rays Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 1 Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 1 Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 1 Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 1 Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 2 (not implemented yet) Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 2 (not implemented yet) Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)

Method: creating secondary rays method 2 (not implemented yet) Ka-band reflectivity (d. BZ) S-band reflectivity (d. BZ)