Evaluation and Changes to CALIOP Cloud Ice and Water Phase DiscriminationMelody Avery1, Brian Getzewich2, David Winker1, Mark Vaughan1, Megan McKeown3, Robert Ryan2, Charles Trepte1 and Stuart Young2

1NASA Langley Research Center, Hampton, VA, USA, 2Science Systems and Applications, lnc., Hampton, VA, USA, 3NASA Langley NIFS Student Intern ; Contact: Dr. Melody Avery, Melody.A.Avery@nasa.gov

Tilted View: Case Study of Typhoon NidaTropical cyclones provide an ideal natural laboratory fortesting the CALIOP cloud extinction and phase algorithms,as they consist of cloud layers with a large range of opticalthickness within one storm system. Typhoon Nida was aCategory 4 Typhoon that formed in the Western Pacific inNovember of 2009. This case study is a nighttime A-Trainoverpass that occurred on November 28, 2009, when theTyphoon was near Guam.

Figure a) shows an IR ISCCP image of the Typhoon at15Z. The warm-colored line is the hurricane track, withwind speeds of > 100 kt plotted beneath the image. Figureb) shows the CALIOP backscatter measured during theoverpass, and Figure c) shows depolarization, which iscomplex and varies vertically. Figure d) is the GMAOGEOS temperature field, and Figure e) illustrates 5significant temperatures within a single cloud layer, andhow these change relative to one another as the opticalthickness and cloud particle sizes change.

One can observe areas of low depolarization at the bottomof layers in the storm core where vertical velocities arelikely to be very large. Is this water or oriented ice at thebottom of thermally thick layers with very cold, highlydepolarizing ice at the top of them?

Figures f) and g) show the retrieved cloud phase. Version4 (f) does a better job than Version 3. There are fewerstripes of HOI in thin Ci, and thanks to improved surfacedetection (Please see the poster by Vaughan et al. fordetails.) there are more cloud layers, although these tendto be thin and uncertain in phase.

b)

a)

c)

d)

e)

f)

g)

Nadir View: Case Study from TC4

In Figures e) - f) one cansee the difficulty that thephase algorithm has withusing layer-integratedquantities.

MC-HOI (dark blue) withhigh depolarization is morelikely to be mixed ice andwater, while that occurringwith depolarization close towater is predicted by theoryto be ROI mixed with HOI.

Figures a), b) and c) above show a monthly time series of the mean (dashed) and median (solid)integrated V4 backscatter, lidar ratio and optical depth for ice layers in 2007 and 2008. TheCALIOP viewing angle changed in November, 2007, indicated by the dotted line on these plots.HOI extinction is now treated appropriately, so that in V4 the statistics of cloud optical depth aresimilar between nadir and tilted views, an artifact in V3. This is an important result, as an error inthe phase type of opaque ice clouds does not appear to bias OD. Please see the poster byYoung, et. al. for an explanation of much improved extinction retrievals for opaque cloud layers.

a) b) c)

CALIOP Version 3 vs Version 4 Cloud Phase DeterminationCloud phase determination relies primarily on the relationship between layer-integrated 532 nm backscatter and depolarization. Secondary decisions are madeusing cloud temperature and other parameters. It is impacted by upstream changes incalibration, layer detection and cloud/aerosol discrimination. In the plots below,randomly-oriented ice occurs above the red line, and water between red and blue.

There are more cloudlayers in V4 than in V3,and the ice/waterclusters are generallywell separated. Thephase algorithm isconfident about ~80%+of clouds.

“Medium Confidence” Oriented Ice

Water

IceNadir (0.3o) vs Tilted(3o) viewing anglesshowing oriented-ice(“HOI”). Below the blueline is considered “highconfidence HOI”. Themixing line betweenROI and HOI is missingin the tilted view, as isthe high confidenceHOI.

(0.3o) (3o)

Lower confidence layersare dominated by thick,mainly opaque layersdesignated as “mediumconfidence HOI”, andvery thin layers with lowdepolarization that mightbe aerosols (low “CAD”).

Nadir vs Tilted Viewing Angle

V3 vs V4 high-confidence cloud layer phases. V3 V4

V3V3 vs V4 lower-confidence cloud layer phases.

V4In a 3o view these layers are likely to be ice with water or HOI at the bottom of them.

There are more layers in V4, so there are consequently also more “uncertain” layers.

More layers in V4 due to better surface detection; less HOI “striping” of thin Ci.

The IIR and cloud top temperature coincide at the eyewall, but IR does not “see” much of the anvil. The contrast between cloud top (dark blue) and base (light blue) temperature may be 60o, with ice at the top and water or HOI at the base.

Ice at the top, with HOI or water at the bottom.

Backscatter

Lidar Ratio

Optical Depth

In a 3o view these layers are likely to be ice with water or HOI at the bottom of them – or just misidentified ROI. (They may be eliminated in V4.)

In a 0.3o view there a significant number of HOI layers and mixing between ROI and HOI is evident.

e)

f)

~7N

Sampling at 4000 – 6000 seconds corresponds to ~5N – 8N in a) – c).

CPI cloud particle images from the DC-8

d)

a)

b)

c)

This case study of an overpass during the TC4mission is quite complex, and occurs during a nadir(0.3 degree) view by CALIOP. Figures a) - c) showthe backscatter, temperature (from GMAO) anddepolarization for a convective cloud near theITCZ, of the coast of Costa Rica. Figure d) showsaircraft observations of temperature and verticalvelocity, as well as phase estimated from the CPIimages (round water droplets vs shattered ice) andthe corresponding radar reflectivity measured byCPR on CloudSat. The depolarization field from 5-8N is quite complex, corresponding to the largeCPR reflectivity and observations of spherical waterdroplets. Profiles with near zero depolarizationoccur interlaced with profiles that look more likewater (increasing depolarization due to multiplescattering). Frequently the cloud tops aredepolarizing more like randomly oriented ice.

This case points out thatin a 5 km average, theremay well be layers withROI, water and HOIoccurring together, sointerpretation is tricky withthick cloud layers.

water HOI

ROI

V4 has less CiHOI “striping”

More water, less medium confidence ROI

V4 has more “unknown” layers, less HOI