1	Using observations of clouds and radiation at Chilbolton UK to evaluate NWP models. Anthony Illingworth and Robin Hogan University of Reading, UK Chilbolton 24h/7d vertical profiles of clouds 94GHz radar and lidar – profiles 30sec/60m resolution. Infer cloud properties and compare with values held in operational models for Chilbolton grid box. 35GHz radar, 22/28/38GHz Radiometers, Raman lidar. 1275 clear air radar – boundary layer + refractivity 3GHz polarisation radar for precipitation.
2	COMPARE OBSERVATIONS AND REPRESENTATION IN MODELS. Typical day Is cloud fraction correct? Pdf OK? Is ice water content correct? Errors? Pdf OK? Errors when classified by weather regime? Example of one month data and model Cloud overlap not really maximum random? Cloud inhomogeneity? Supercooled layer clouds are common.
3	Standard Chilbolton observations on the web Radar Lidar, gauge, radiometers
4	Target categorization Combining radar, lidar and model allows the type of cloud (or other target) to be identified
5	Cloud fraction
6	Cloud fraction: one year of data Too much cloud at high levels, too little at mid-levels
7	Ice water content Cirrus in situ measurements suggest we can obtain IWC from Z to a factor of two Particles tend to be smaller at lower temperatures, so with additional use of temperature, error is reduced to -30%/+40% Less accurate between -10°C and 0°C because of strong aggregation
8	"Ice water content" Ice water content from Z and T Error in ice water content Retrieval flag
9	Ice water content: results First ever long-term evaluation of ice water content Underestimate of mean mid-level IWC in both models Seems to be due to factor-of-2 error in mean cloud fraction Mean in-cloud IWC appears to be reasonably good above 4 km
10	IWC distributions The Met Office Unified Model tends to simulate very high and very low ice water contents too infrequently
11	Clasification by regime Ascent at 700 hPa (>0.1 hPa/s) Ascent at 400 hPa (>0.1 hPa/s) Stability between 900 and 1000 hPa Descent and low level stable – UM can’t make 100% cloud cover
12	Eddy dissipation rate e 30-s standard deviation of 1-s radar velocities, plus wind speed, gives eddy dissipation rate (Bouniol et al. 2003)
13	PDF of e by cloud type Mean turbulence in different clouds: Stratocu: 10^-3m²s^-3 Mixed-phase: 10^-5 m²s^-3 Cirrus: 10^-6m²s^-3
14
15
16
17
18
19
20
21	Cloud overlap assumption in models Cloud fraction and mean ice water content alone not sufficient to constrain the rad iation budget Assumptions generate very different cloud covers Most models now use “maximum-random” overlap, but there has been very little validation of this assumption
22	Cloud overlap from radar: example Radar can observe the actual overlap of clouds We next quantify the overlap from 3 months of data
23	Cloud overlap: results Vertically isolated clouds are randomly overlapped Overlap of vertically continuous clouds becomes rapidly more random with increasing thickness ANALYTICAL EXPRESSION FOR VERTICAL DECORRELATION
24	5. Importance of ice-cloud inhomogeneity Non linear relation between optical depth and emissivity . For clouds which are inhomogeneous use of average optical depth gives wrong emissivity.
25	Cirrus fallstreaks and wind shear - inhomogeneities
26	Ice water content distributions In the near future, models will carry variables for the variance of water content, as well as the mean Derive variance of ice water content of cirrus from radar PDFs of IWC within a model gridbox can usually be fitted by a lognormal or gamma distribution
27	Analytic expression for effect of shear on pdf of iwc and vertical decorrelation as a function fo grid box size. Variance and decorrelation increase with gridbox size Shear makes overlap of inhomogeneities more random, thereby reducing the vertical decorrelation length Shear increases mixing, reducing variance of ice water content
28	Mixed-phase clouds – SUPERCOOLED LAYER CLOUDS SUPERCOOLED LAYER CLOUDS ARE COMMON SAME WATER CONTENT – BIG RADIATIVE EFFECT IF LIQUID DROPLETS – SMALL EFFECT IF ICE PARTICLES.
29	SUPERCOOLED CLOUD EXAMPLE
30	Supercooled water – model comparison Use ground-based lidar to estimate occurrence of supercooled water layers over a 1-year period Around 15% of mid-level ice clouds at Chilbolton contain liquid water with optical depth > 0.7
31	DUAL WAVELENGTH 35 AND 94GHZ RADAR LIQUID WATER CONTENT – THE 94GHZ RADAR IS ATTENUATED MORE THAN THE 35GHZ. ICE PARTICLE SIZE - Z AT 94GHz -MIE SCATTERING Z AT 35GHZ - RAYLEIGH SCATTERING RATIO OF Z GIVES PARTICLE SIZE. ONCE SIZE IS KNOWN CAN FIND N FROM Z, AND SO MORE ACCURATE IWC.
32	LIQUID WATER CONTENT PROFILE – FROM DIFFERENTIAL ATTENUATION OF Z AT 35 AND 94GHZ
33	ICE PARTICLE SIZE FROM 35 AND 94GHZ Z -35GHz Z–94GHz DELTA Z Do Better IWC Model IWC
34	25m dish: Scan on interesting days.
35	1275MHz Clear air ‘acrobat’ radar. Return is from changes in refractive index – turbulence on the scale of l/2 or 11.7cm. Changes in the summer dominated by humidity. Beamwidth 0.75degs – 660m at 50km range
36	CONVECTIVE ‘DONUTS’
37	RHI–6 AUG-03 TOP OF THE BOUNDARY LAYER
38	REFRACTIVITY in the boundary layer. Ground clutter targets Round trip time changes with refractive index. Detect as phase change in return. Refractivity, N, 1ppm change in refractive index. DN = 1 gives Df = 3 deg/km (round trip). DN = 1: »1% change in RH (summer) or 1K Technique developed by Fred Fabrey
39	Refractivity change over four hours in Nov 03. http://www.met.reading.ac.uk/radar/cloudnet/quicklooks/
40	3GHz/10cm BETTER RAINRATES - ZDR AND KDP IN RAIN
41	Cloud products on the web http://www.met.reading.ac.uk/radar/cloudnet/quicklooks/ Interested in data/collaboration? a.j.illingworth@reading.ac.uk r.j.hogan@reading.ac.uk