Enhanced Satellite Cloud Products for Climate Studies
Thomas Greenwald1 and Ralf Bennartz2
1Cooperative Institute for Meteorological Satellite
2 Department of Atmospheric and Oceanic
Studies (CIMSS), University of Wisconsin-Madison
Sciences, University of Wisconsin-Madison
Contact: Tom Greenwald tomg@ssec.wisc.edu
Comparison of Cloud Properties
Comparison of AMSR-E LWP and
MODIS LWP for warm (cloud top
temperature > 273 K), overcast
nonprecipitating clouds reveals
significant differences. Most
differences appear to be
associated with regional systematic
errors in the AMSR-E products
LWP-effective radius
relationships
Annual distributions
Cloud geometric thickness
Time variation
Cloud drop number concentration (N)
MODIS LWP
LWP (kgm -2)
N (#/cm3)
LWP (kgm -2)
MODIS LWP
AMSR-E
Level 2A (V09)
Fast geolocation based
on inverse modeling
MYD35_05
MYD06_05
(Cloud mask)
(Cloud products)
AMSR-E LWP adjusted for systematic
errors*
Vertical mean liquid water content
“Polluted” clouds (defined as
N > 100 cm-3) exhibit a
different slope in the LWP-re
relationships than clean
clouds. Aside from having
smaller drop sizes than
clean clouds, polluted
clouds also tend to have
smaller LWPs
August 2002
Polluted
Clean
LWC (gm -3)
LWP (kgm -2)
(geolocation)
Polluted
AMSR-E LWP
Data Product Generation & Characteristics
MYD03_05
Clean
Thickness (m)
LWP (kgm -2)
AMSR-E LWP
LWP (kgm -2)
Motivation
Satellite-based climate studies of clouds rarely take advantage of the complimentary information provided by passive
microwave observations. These observations can supply, for example, cloud liquid water path (LWP) estimates in
conditions where solar/IR measurements cannot, such as overlaying ice clouds and mixed phase clouds, as well as
provide estimates of water vapor, surface winds, and sea surface temperature under overcast conditions. Solar/IR
observations, on the other hand, can complement microwave observations by characterizing the cloud conditions that
occur within the field-of-view (FOV) of the microwave instrument.
To take full advantage of the information content from these different observations for use in climate studies, new
merged cloud products are being developed that combine instantaneous, full-resolution Level 2 products from the
Advanced Microwave Scanning Radiometer-EOS (AMSR-E) and the MODerate resolution Imaging Spectroradiometer
(MODIS) on Aqua. This study examines a year’s worth of these new products (July 2002 – June 2003) with respect to
1) quantifying certain errors in AMSR-E LWP products, 2) comparing AMSR-E and MODIS cloud products (including
new MODIS cloud droplet number concentration and geometric thickness products; Bennartz 2007), and 3) assigning
cloud types to AMSR-E LWP products.
*AMSR-E LWP field above was subtracted by the
systematic error field below at left
Geolocated
MODIS scan indices
bin files
(per half-orbit)
AMSR-E
Level 2B (V06)
Wentz Ocean
Products
AMSR-E
Level 2A Tbs
Classification of AMSRAMSR-E LWP Products
•
•
•
•
•
•
Level 2C Merged
MODIS/AMSR-E
products
AMSR-E/MODIS L3 cloud products:
Liquid water path
Total precipitable water (TPW)
Sea surface temperature
Surface wind speed
Ocean product quality flags
Experimental precipitation flag (89 GHz)
A major limitation of current climatic LWP datasets derived from passive microwave measurements (e.g., O’Dell et al.
2008, Ferraro et al. 1996) is that they do not provide information on cloud type. With the new data products generated by
this study, MODIS-derived IR cloud phase and cloud top pressure (CTP) can be used to classify the clouds occurring
Liquid
within the AMSR-E FOV. These results are for nonprecipitating clouds only.
Annual distributions
AMSR-E cloud cleared L3 products:
Level 3 Merged
MODIS/AMSR-E
products
• Liquid water path
Low level (CTP > 680 hPa)
Mid-level (680-440 hPa)
High level (CTP < 440 hPa)
Mixture of levels
• Surface wind speed
• Surface wind speed local standard deviation
High thin clouds*
Estimation of AMSRAMSR-E LWP Errors
Systematic error
LWP (kgm -2)
Random error
The spatial pattern of
random errors (lower left)
resembles the mean cloudcleared TPW field (below),
suggesting an adverse effect
of water vapor on the LWP
retrievals
TPW (kgm -2)
LWP (kgm -2)
Random error
*CTP < 440 hPa and optical depth < 5
Summary and Further Work
Annual distributions
LWP (kgm -2)
Mixed phase
Greenwald et al. (2007) showed that cloud-cleared microwave-derived LWP products are useful for
estimating minimum errors in the LWP products themselves. These errors are attributed to uncertainties
in the ocean surface and the atmosphere. Cloud-cleared LWP values are a measure of systematic errors
(since the LWP should be zero), while the local spatial variation of the cloud-cleared LWP (represented
as standard deviations within 0.5o lat/lon grid boxes) is a measure of the random errors in the LWP
products. Here we investigate the time variation and geographic distribution of these errors.
LWP (kgm -2)
Ice
Ice
LWP (kgm -2)
•Total precipitable water local standard deviation
Systematic error
Mixed phase
LWP (kgm -2)
•Total precipitable water
Level 3 products are
monthly on 0.5o x
0.5o global grids
Time variation
Liquid
• Liquid water path local standard deviation
LWP (kgm -2)
AMSR-E LWP all cloud conditions
MODIS LWC all low liquid clouds
MODIS cloud geometric thickness all low liquid clouds
MODIS cloud drop number concentration all low liquid clouds
AMSR-E & MODIS LWP for warm overcast nonprec clouds
AMSR-E LWP for liquid, ice, mixed phase, low level, mid level, high
level, high thin, and mixed level nonprecipitating clouds
• Precipitation frequency
•
•
•
•
•
•
Time variation
LWP (kgm -2)
• Number of collocated pixels
• Cloud mask (number of confident
cloudy and probably cloudy pixels)
• Number of single/multi-layered
clouds (optical)
• Mean MODIS zenith angle
• Mean MODIS azimuth angle
All Level 2C
products are
AMSR-E single
FOV data
organized into
half-orbit files
LWP (kgm -2)
Cloud optical depth (mean/std)
Cloud effective radius (mean/std)
Cloud water path (mean/std)
Cloud phase (optical & IR)
Cloud top pressure (mean/min/max)
Cloud top temperature (mean/min/max)
LWP (kgm -2)
•
•
•
•
•
•
AMSR-E L2C products:
LWP (kgm -2)
MODIS sub-FOV L2C products and statistics:
ƒ Enhanced cloud products have been created that merge MODIS and
AMSR-E Level 2 products. These products include new cloud properties
currently not available in the MODIS cloud product suite, i.e., cloud
geometric thickness and cloud droplet number concentration
ƒ Detailed comparisons (at the FOV level) can now be done between
AMSR-E and MODIS LWP products to better understand their strengths
and weaknesses
ƒ These merged products provide for the first time a means of evaluating
certain AMSR-E LWP errors over large areas and long time periods and
provide a means of classifying the LWP according to cloud type
ƒ Plans are to produce these products over the entire Aqua mission and to
include MODIS aerosol products and CERES radiation budget products
References
Greenwald T., T. L’Ecuyer, and S. Christopher, 2007, Geophys. Res.
Lett., 34, L22807, doi:10.1029/2007GL031180.
Bennartz, R., 2007, J. Geophys. Res., 112, D02201,
doi:10.1029/2006JD007547.
O’Dell, C. W., F. J. Wentz, and R. Bennartz, 2008, J. Climate, 21,
1721-1739.
Ferraro, R. R., N.C. Grody, F. Weng, and A. Basist, 1996, Bull. Amer.
Meteor. Soc., 77, 891-905.
Acknowledgments
This work was supported by the NASA Atmosphere Product and
Evaluation and Test Element (PEATE) project (NNG05GN47A). We also
appreciate the work of Fred Nagle in developing the code for the AMSRE/MODIS collocation.