Passive microwave and visible/near infrared
retrievals of boundary layer cloud physical
properties
Ralf Bennartz1,2, John Rausch1,Tom Greenwald2,
Greg Elsaesser3,4, Chris O’Dell4, Matt Lebsock5
1: EES, Vanderbilt University
2: SSEC, University of Wisconsin
3: NASA-GISS/Columbia University
4: CIRA, Colorado State University
5. JPL
Outline
1. Cloud Droplet Number Concentration from
13 years of MODIS data
• Long-term climatology
• Error analysis
2. Microwave-based LWP climatology from
28 years of passive microwave observations
• Long-term climatology
• Error analysis using combined MODIS/CloudSat/AMSR-E
3. Outlook
Cloud Droplet Number Concentration from
MODIS
Cloud Droplet Number Concentration from
MODIS
LWC
Cloud Droplet Number Concentration from
MODIS
• Updated climatology based on 13 years of full
Level 2 data (Bennartz and Rausch 2017, ACP)
• Fixed errors in condensation rate parameterization
• Used full Collection 6 data
• Limited evaluation to cases consistent with adiabatic
assumption (r16 < r21 < r37)
• Investigated differences C5 versus C6. Paper
submitted with Kerry Meyer/Steve Platnick
• Initial results using updated climatology,
DOI: 10.15695/vudata.ees.1
Climatology results
CDNC over global oceans (13 yrs average)
Validation against aircraft observations
Our retrieval
Aircraft observation reported by
Painemal and Zuidema (2011)
Climatology results
CDNC over global oceans (13 yrs average)
Climatology results
Climatology results
See ‘plumes’ of high retrieved CDNC around
various volcanic islands
Climatology results
See ‘plumes’ of high retrieved CDNC around
various volcanic islands
Climatology results
CDNC over global oceans (13 yrs average)
Climatology results
East China Sea
Climatology results
Consistent annual cycle with earlier studies
using AVHRR (PATMOS-x)
(Bennartz et al. 2011; Rausch et al., 2010)
Uncertainty estimates
• How to calculate uncertainty estimates:
• Propagate errors in tau/reff forward and
average?
• Use standard deviation over 1x1 degree box?
Uncertainty estimates
MW cloud liquid water path climatology
• Based on RSS V7 retrievals, SSM/I since 1987,
AMSR-E, and TMI
• Monthly diurnal mean liquid water path
• Climatological diurnal cycle
• Elsaesser et al. JCLIM, submitted.
(O’Dell, Wentz, and Bennartz, JCLIM, 2008)
• Ongoing NASA Measures project (2013-2018)
MW: Principle of retrieval
• Use 2 channels and
polarization
difference to
estimate WVP, LWP
• Also affected by
rain water
• Separation of
RWP/LWP critical.
Data Record
• SSM/I, SSMIS
Morning/Evening
Coverage since 1987
• TRMM/GPM
crisscrossing in LEXT
since 1997 resp 2014
• AMSR-E/AMSR-2
13:30 LEXT
• MWI on EUMETSAT/
EPS-SG early
afternoon orbit
The diurnal cycle of LWP
Long-term satellite studies of LWP must account for
the diurnal cycle. Otherwise, satellite drifts will lead to
an aliasing of the diurnal cycle onto trends of LWP.
Time Series: California
• MAC is higher than
MODIS by about 20 gm-2
• Terra/Aqua bias
21
Uncertainty analysis – Clear-sky bias
• Clear-sky bias a function of wind speed and TPW
• V7 retrievals are improved over older versions.
• Data used for correction of climatology
Uncertainty analysis – Cloud/Rain
partitioning
• Developed new cloud/rain partitioning based on
AMSR-E/Cloudsat collocation
• Significantly improved agreement with Cloudsat and
MODIS.
Conclusions
• CDNC Climatology published:
• Bennartz and Rausch (2017, ACP). Dataset published: DOI:
10.15695/vudata.ees.1
• MAC-LWP Climatology:
• Elsaesser et al. JCLIM under review. Supported by the NASA
MEaSUREs program
• To be hosted at GES-DISC (Goddard Earth Science Data
and Information Services Center)
http://disc.sci.gsfc.nasa.gov/measures
• Particularly important for global climate modeling
validation as well as aerosol-cloud interaction process
studies.
• Would be highly beneficial to tie MODIS dataset
backward in time to include Patmos (AVHRR) and
forward to VIIRS.