Lidar observations of mixedphase clouds Robin Hogan Anthony

Lidar observations of mixed-phase clouds Robin Hogan, Anthony Illingworth, Ewan O’Connor & Mukunda Dev Behera University of Reading UK Overview • Enhanced algorithm for supercooled liquid water detection (Hogan et al. 2003, QJ in press) • Specular reflection from horizontally aligned plate crystals: A blessing or a curse? • Global distribution of stratiform supercooled water clouds from spaceborne lidar

Lidar and mixed-phase clouds Small supercooled liquid droplets Large falling ice particles • Typical concs: ice 20 l-1, liquid droplets 20 000 l-1 – The same mass of water is ~10 times more optically thick as liquid than as ice, so lidar return also 10 times greater – By contrast, D 6 dependence of radar makes the same mass of water ~1000 times more reflective as ice than as liquid! • Radiation calculations on 2 case studies suggest – When supercooled liquid present it is usually more radiatively important than the ice, even though tends to form thin layers – Crudely represented in current models

Integrated lidar backscatter • The integrated backscatter through a cloud of optical depth of is approximately (Platt 1973): – k = extinction/backscatter ratio (18. 75 sr for droplets) – = multiple scattering factor (~0. 7 for Chilbolton lidar ) • For large optical depth it reduces to = (2 k)-1 • If z 1 and z 2 encompass the 300 m around the strongest echo in a profile, we can identify thin liquid water layers with greater than, say, 0. 7

Lidar echo Integrated lidar echo Microwave radiometer LWP • Example of supercooled water detection at Chilbolton

Results for lidar 5° from zenith • Analysis of continuous Chilbolton CT 75 K lidar data from 2000 when looking off-zenith Frequency that cloud was observed Fraction of clouds containing supercooled water with >0. 7

Results for zenith pointing lidar • Analysis of Chilbolton lidar data from 1999 when pointing at zenith Enhanced occurrence between -10 and -20 °C: specular reflection from plates?

Supercooled water in models • A year of data from the Met Office and ECMWF – Easy to calculate occurrence of supercooled water with > 0. 7 Prognostic ice and liquid+vapour variables Prognostic cloud water: ice/liquid diagnosed from temperature

Specular reflection • Specular reflection from planar crystals can occur within 1° of zenith or nadir – Enhanced backscatter with no accompanying increase in extinction (very low k): radiative properties difficult to infer – Integrated backscatter in ice can exceed the asymptote corresponding to optically thick liquid cloud (recall ~(2 k)-1) – Is locating plate crystals useful? Currently nadir viewing is being considered for spaceborne lidars Calipso and Earth. CARE • To quantify, require lidar to be precisely at zenith: 20 days of data analysed so far at Chilbolton – Algorithm calculates integrated backscatter from 2 km up – Specular reflection deemed to occur if this integral is more than 1. 05 times the asymptote for liquid water – Excess above this value is attributed to pixels with highest

Probable plate crystals

~-23°C Pristine crystals are columns or needles Pristine crystals are plates Specular reflection: Results ~-9°C Fraction of clouds with specularly reflecting crystals • Around 20% of ice cloud profiles are strongly affected by specular reflection: enhancement > a factor of 2 – PDF of the maximum backscatter suggests that a further 30% of profiles are affected by specular reflection to a lesser extent – Big problem for interpreting backscatter measurements from space in terms of the radiative properties of ice clouds – Recommend operate spaceborne lidar a few degrees from nadir

Supercooled liquid water from the LITE lidar on the space shuttle in 1994 • LITE took 45 hours of data 9 -20 September 1994 • We use 532 nm channel: appeared most sensitive • Frequent changes in gain: only 10. 5 hrs of data for which saturation level high enough to detect supercooled layers unambiguously • Even then liquid water often saturated receiver, and multiple scattering more uncertain from space, so integral method not reliable • 280 000 km of ground covered: equivalent to 160 days of surface observation!

Cloud occurrence versus height, temperature and latitude

Supercooled water occurrence versus temperature and latitude

Comparison with Chilbolton • Cloud fraction: much better coverage from space – Lidar does not need to penetrate the cloudy boundary layer • Liquid detection: very similar, esp. below -10°C

Conclusions • Have shown that spaceborne lidar can identify supercooled liquid water clouds across the globe – Problems with LITE: saturation & severe multiple scattering • We will use long-term spaceborne lidar data: Ice. Sat: launched 12 Jan 2003: High polar orbit, 2 wavelengths Calipso: launch in Dec 2004: Includes depolarisation channel