Using GLAS to Characterize Errors in Passive Satellite Cloud Climatologies

1
Using GLAS to Characterize Errors in Passive
Satellite Cloud Climatologies

• Michael J Pavolonis and Andrew K Heidinger
• CIMSS/SSEC/UW-Madison
• NOAA/NESDIS

2
Motivation

• Passive sensors on operational satellites are
  limited in their ability to detect thin cloud and
  multi-layered cloud systems, but they provide a
  30-year data record, which makes them a critical
  component of the climate observing system.

Each passive climatology has its own error
characteristics and sensitivities, resulting in
large differences. Can satellite-based active
measurements help resolve these differences?
3
GLAS Overview

• The Geoscience Laser Altimeter System (GLAS) is a
  space-based two channel (532 nm and 1064 nm)
  laser altimeter located on board ICESat. Its
  data is primarily used for altimetry and
  cloud/aerosol studies.
• The GLAS near surface footprint is 70 m and the
  vertical resolution is 76.8 m.

GLAS measurements are more sensitive to thin
cloud/aerosol layers than passive measurements.
4
GLAS Overview

• Due to hardware failure, the GLAS dataset only
  covers a few short measurement periods.
• The October 16, 2003 November 18, 2003 time
  period is used in this study since the 532 nm
  laser, which is more sensitive to cloud/aerosol
  detection, was in operation during this time.
• Only data from the 532 nm laser is included.
• ICESat is in a near sun-synchronous orbit and had
  a 0600 0800 LST descending node equator
  crossing time during this period.

5
GLAS Data Analysis

• Coincident GLAS/passive observations are
  relatively rare.
• In this study, globally mapped GLAS observations
  from the descending node (morning) of ICESat are
  compared to various passive observations that
  occur within 4 hours of the ICESat overpass.
• All data are mapped to a 2.5o equal area grid.

6
Spatial resolution may cause some of these
differences, especially in the subtropics.
7
7 of clouds have an optical depth lt 0.1 23 of
clouds have an optical depth lt 0.5 Most thin
clouds are high clouds.
8
(No Transcript)
9
14 of clouds have an optical depth lt
2.0 Difference between passive and GLAS 40, so
many thicker clouds are also being missed. So
co-located active and passive data is needed to
better characterize differences.
10
High Cloud Detection Limitations
The GLAS cloud amounts were calculated as a
function of optical depth. Only daytime, open
ocean grid cells were considered, so this
represents the best case scenario for passive
measurements.
11
Consistency with Aircraft Measurements
Histogram of cloud optical depth from the Cloud
Physics Lidar (CPL) for MODIS Airborne Simulator
(MAS) clear FOVS (as determined by the
automated cloud mask.
From B. Holtz and S. Ackerman
Most undetected clouds have an optical depth of
less than 0.3.
12
Concluding Thoughts

• We are taking steps towards quantifying
  limitations and differences in global
  satellite-derived cloud climatologies, but much
  more work remains.
• A more complete error analysis can be performed
  when CloudSat/CALIPSO are launched. Aqua MODIS
  data can be used to simulate various cloud masks
  and cloud properties. This is the best way to
  tie together past and future data records (e.g.
  AVHRR/VIIRS) and to understand the differences
  between climatologies and help in the development
  of an optimal multi-sensor long-term cloud
  climatology (e.g. AVHRR/HIRS).
• The effects of spatial resolution on cloud
  detection also need to be addressed.
• Can active measurements from satellite (e.g. GLAS
  or CALIPSO/CloudSat) be used in conjunction with
  passive observations to improve estimates of
  cloud vertical structure and thereby effectively
  extend the active measurements in time?