The PATMOSx Cloud Climatology Recent Activities

1
The PATMOS-x Cloud Climatology-Recent Activities

• Michael Pavolonis and Andrew Heidinger
• Cooperative Institute for Meteorological
  Satellite Studies (CIMSS)
• NOAA/NESDIS Office of Research and Applications
• Madison, Wisconsin

2
Recent Improvements/Activities

• Characterizing sub-pixel cloudiness
• Comparison of PATMOS-x cloud amounts and cloud
  top properties to those derived from a
  space-borne lidar
• Identifying volcanic ash plumes and ash plume
  height estimation
• Cloud property trend analysis and comparisons to
  MODIS

3
Recent Improvements/Activities

• Characterizing sub-pixel cloudiness
• Comparison of PATMOS-x cloud amounts and cloud
  top properties to those derived from a
  space-borne lidar
• Identifying volcanic ash plumes and ash plume
  height estimation
• Cloud property trend analysis and comparisons to
  MODIS

4
Identifying and Characterizing Sub-Pixel
Cloudiness in the CLAVR-x Cloud Mask

• The CLAVR-x cloud mask is relatively unique in
  that spatial uniformity tests are used to help
  identify partially cloudy pixels.
• How should partially cloudy pixels be weighted
  when calculating the PATMOS-x cloud fraction?
• Partially cloudy pixels should be weighted such
  that the weight, f, is 0.0 lt f lt 1.0.

5
How is the cloud weight, f, determined?

• In the CLAVR-x cloud mask there are two partially
  cloudy categories, Probably Clear and Probably
  Cloudy.
• Probably cloudy pixels were determined to be
  cloudy by one or more cloud tests but are
  spatially non-uniform.
• Probably clear pixels were determined to be clear
  by all cloud tests but are spatially non-uniform.

6
How is the cloud weight, f, determined
(continued)?

• Cloud fraction for a given PATMOS-x grid cell is
  given by
• cldfrac (N0f0 N1f1 N2f2
  N3f3)/N0123
• (clear probably clear
  probably cloudy cloudy)/total
• Naturally, f0 0.0 and f3 1.0.
• Initially, a simple radiometric balance approach
  was used to determine the weight applied to N1
  and N2, where f1 0.13 and f2 0.88 for 4 km
  AVHRR GAC data.
• MODIS 250 m resolution data was used to test the
  the accuracy of these weights.

7
MODIS 250 m Data
Rayleigh-corrected 0.65 and 0.86 um reflectances
are used to create an optimal 250 m cloud mask
over clean non-sunglint open ocean.
100 km
CLEAR CLOUDY
100 km
The true 4 km pixel cloud fraction is
calculated from 250 m cloud mask (binary).
8
Simulate GAC cloud mask results using MODIS and
compare to true cloud fraction derived from 250
m MODIS.
9
Analyzing the current weights
Based on the analysis of this scene and others,
the probably clear weight was changed to
0.35. The probably cloudy weight was assumed to
be reasonable as is. Any significant changes to
the cloud mask algorithm will lead to a change in
the probably clear weight.
MOD35 0.80
Based on all non-high clouds outside of glint for
a single MODIS granule.
10
Recent Improvements/Activities

• Characterizing sub-pixel cloudiness
• Comparison of PATMOS-x cloud amounts and cloud
  top properties to those derived from a
  space-borne lidar
• Identifying volcanic ash plumes and ash plume
  height estimation
• Cloud property trend analysis and comparisons to
  MODIS

11
Comparisons to GLAS
12
Recent Improvements/Activities

• Characterizing sub-pixel cloudiness
• Comparison of PATMOS-x cloud amounts and cloud
  top properties to those derived from a
  space-borne lidar
• Identifying volcanic ash plumes and ash plume
  height estimation
• Cloud property trend analysis and comparisons to
  MODIS

13
Volcanic Ash Detection and Height Assignment in
CLAVR-x
A volcanic ash mask and ash plume height
algorithm will soon be used in all CLAVR-x
processing and ash products will be included in
the PATMOS-x data set. PATMOS-x can be used to
create a satellite-based volcanic eruption
climatology.
14
Mount St. Helens Example
VAAC Height up to 11000 m VAAC Height up to 6000 m
Retrieved heights agree well with VAAC analysis
in the thickest regions of the plume.
15
Recent Improvements/Activities

• Characterizing sub-pixel cloudiness
• Comparison of PATMOS-x cloud amounts and cloud
  top properties to those derived from a
  space-borne lidar
• Identifying volcanic ash plumes and ash plume
  height estimation
• Cloud property trend analysis and comparisons to
  MODIS

16
Future work

• Improve nighttime cloud detection over land.
• Explore multilayered clouds
• Look at the big picture. Explore the relationship
  between changes in cloud properties and the
  regional/global dynamics.