Preliminary results from the new AVHRR Pathfinder Atmospheres Extended PATMOSx Data Set

1
P2.6
Preliminary results from the new AVHRR Pathfinder
Atmospheres Extended (PATMOS-x) Data Set Andrew
Heidingera, Michael Pavolonisb and Mitch
Goldberga aNOAA/NESDIS Office of Research and
Applications bUW/Cooperative Institute for
Meteorological Satellite Studies (CIMSS),
Madison, WI
Emails Andrew.Heidinger_at_noaa.gov,
mpav_at_ssec.wisc.edu, Mitch.Goldberg_at_noaa.gov
Other PATMOS-x Products
Introduction
PATMOS-x has over 100 products. Here is a sample
of some of the more common ones. These are
monthly averages from July 1987.
Validation of the Cloud Detection/ Clear Radiance
Quality
Validation of the Cloud Products (Cloud Type)

• PATMOS-x is
• an extension of the original AVHRR Pathfinder
  Atmospheres (PATMOS)
• extends PATMOS by processing the NOAA-klm data
  and the data from AVHRRs in morning orbits
• includes many algorithms for new cloud and
  surface products
• Is part of a larger NESDIS Data Stewardship
  Initiative which also include activities aimed at
  improving the AVHRR calibration and navigation
• GOALS of PATMOS-x
• Use the improved AVHRR observations to make a
  data-set useful for satellite climatology work
  within NESDIS and others based on accepted
  procedures.
• Contribute to bringing consensus to satellite
  cloud climatologies (where there is little now)
• Work with NPOESS and EOS to develop AVHRR
  climatologies that are consistent with the future
  climate records.

We have compared the SSTs from PATMOS-x computed
daily(a) and monthly averaged (b) to the
Reynolds Optimally Interpolated SST climatology
(d). The histogram of the SST OISST for the 4
cloud mask values show the desired behavior with
little indication of cloud contamination in the
clear radiances (no cold tail).
We are in the process of publishing and
validating all cloud algorithms used in
CLAVR-x/PATMOS-x. One of the algorithms already
published is the cloud type algorithm. We
derived 6 cloud types for each pixel (fog, water,
supercooled water, opaque ice, cirrus,
multilayer). The validation shown below was
based on MODIS and RADAR overpasses compiled by
Jay Mace of University of Utah.
Global Precipitation Index
(a)
(b)
(c)
(d)
Normalized Vegetation Index
Histogram of multilayer detection results
RADAR data showing a multilayer cloud during a
MODIS overpass
P2.6

• PATMOS-x Products
• Radiance Mean and Standard Deviations of all
  channels for all cloud mask values (clear,
  probably clear, probably cloudy, cloudy
  all-sky)
• Cloud Amounts (total, high, mid, low, ice and
  water), 6 Types (including multilayer), cloud
  temperature, emissivity, optical depth, particle
  size and liquid/ice water path
• Surface Sea, Land and Ice Surface Temperature,
  NDVI
• Aerosol Optical depths using NOAAs operational
  algorithm
• Precipitation Global Precipitation Index (GPI)
• Other Fire, Dust and Volcanic Ash
• developed but not yet implemented

Multilayer Cloud Fraction
Comparison with other Satellite Climatologies
(ISCCP)
Improvements over PATMOS
We are comparing our cloud climatologies to those
from other satellite derived climatologies
(ISCCP, UW/HIRS). While philosophical
differences often prevent close agreement in the
absolute values, we do see agreement in annual
cycles (here July January) and other relative
measures of cloudiness
One of the problems apparent in the PATMOS data
were the large jumps in some cloud product time
series during transitions from one satellite to
the next (vertical lines in figure to the right).
CLAVR-x (and therefore PATMOS-x) has reduced
this problem by improving the physical basis of
the cloud mask. This also allowed for processing
morning satellite data in a consistent way.
Cloud Top Temperature
AVHRR Data Improvement Activities

• A large part of ORAs effort is focused on
  improving the radiometric and geolocation
  accuracy. Some of these activities are
• using simultaneous nadir observations between
  AVHRR and MODIS to transfer MODISs on-board
  reflectance calibration to AVHRR (see below)
• Using advanced hyperspectral sensors such as
  Hyperion on NASAs EO1 satellite to improve our
  spectral knowledge of radiometric targets (i.e.
  desert sites) used for reflectance calibration
• Using AVIRIS data for characterizing and removing
  artifacts in climate records from the spectral
  differences between AVHRRs

• Conclusions
• ORA is developing an improved AVHRR data-set
  (1982-200?)
• PATMOS-x will use this improved data to develop a
  new climate data-set
• PATMOS-x data will made available as orbital,
  daily and monthly averages in a self describing
  format (HDF4)
• Work is ongoing to finish publication of all
  algorithms but initial results and comparison are
  encouraging and show PATMOS-x adds new
  information to the existing satellite
  climatologies
• We actively seek collaboration with others on the
  use of this data

Using Hyperion to improve our knowledge of AVHRR
and its relation to other sensors (i.e. MODIS)
Comparison of MODIS versus AVHRR (0.63 micron)
Continuity in the PATMOS-x and EOS/MODIS Climate
Records
While developing the PATMOS-x algorithms we have
tried to ensure physical continuity with the
comparable climate records from EOS/MODIS For
example, we use a split-window algorithm to
estimate cloud temperature and cloud emissivity
while MODIS uses a better CO2 slicing approach.
While AVHRR is spectrally limited, we feel we can
produce comparable climatologies of cloud
temperature and emissivity in many regions.
Comparing to MODIS (see right) helps us
characterize the weakness and strengths of the
PATMOS-x products.
AVHRR Cloud Temperature
MODIS Cloud Temperature (MOD06)