An Intercalibrated Microwave Radiance Product for Use in Rainfall Estimation Level 1C

1
An Intercalibrated Microwave Radiance Product
for Use in Rainfall Estimation Level 1C

• Christian Kummerow, Wes Berg, G. Elsaesser
• Dept. of Atmospheric Science
• Colorado State University

3rd IPWG Workshop Melbourne, Australia Oct.
23-27, 2006
2
Original Motivation
GPM will fly a constellation of radiometers to
provide 3-hourly sampling. If the radiometers
are not inter-calibrated, neither will the rain
rates. Except for the dedicated GPM
radiometers (GMI), the remaining set belongs to
US and International agencies that have differing
experience and requirements for calibration.
These agencies may or may not support calibration
activities once their own requirements are met.
Level 1C is an attempt to overcome political
and programmatic issues and return calibration
and data stewardship to the rainfall
community Level 1C aims to simplify formats and
toolkits. If a graduate student knows how to
read data from one sensor, they can read it from
all sensors
3
Todays GPM constellation
Inc Ang
Sensor
4
AMSR-E Calibration Differences
5
Effect of SSM/I Calibration on rainfall
6
Inter-calibration Procedure

• Select coincident overpasses (30 min.) of TMI
  and SSM/I (other sensors will be done in same
  fashion but waiting for next version of algorithm
  below)
• Use current retrieval method (OE in non-raining
  PR adjustment in rain) to derive geophysical
  parameter field that is consistent with each TMI
  Tb
• Use geophysical parameters to compute Tb for each
  SSM/I sensor using appropriate channels, view
  angles and spatial resolution
• Compute biases in computed SSM/I Tb versus
  observed Tb.
• Use as much data as possible to reduce random
  errors.

7
SSM/I TMI Crossover
8
(No Transcript)
9
SSM/I Calibration
10
SSM/I Calibration Offsets
11
Level 1C Calibration Differences
TB Bias (Simulated Minus Observed) in Kelvins
3 2 1 0 -1 -2
TMI
SSM/I F13
SSM/I F14
SSM/I F15
AMSR-E
19 H 19 V 22 V
37 H 37 V 86 H
86 V
Approximate Channel Frequency
12
Level 1C Projecthttp//mrain.atmos.colostate.edu/
LEVEL1C
13
Level 1C development Quality Control Procedures

• Pixel check for nonphysical values Checks that
  individual antenna temperatures values are within
  physical limits (currently 50 to 325K), that
  lat/lon values are within range (-90 to 90 and
  -180 to 360), and that the distance between
  pixels along a scan are reasonable (10 to 30km).
• Scan check for nonphysical seperation Checks
  that the distance between adjacent scans (nadir
  point) is between 20 and 30 km. Scans outside of
  this range are flagged and are currently
  eliminated (not written out).
• Scan check for deviations from climatology
  Compares the antenna temperatures for each
  channel along the scan to climatological values.
  If more than 30 of the pixels along a given scan
  are greater outside the range of the mean
  climatological TA /- 3 sigma, and this
  represents a change of 30 or more from the prior
  scan, the scan is flagged and eliminated (not
  written out).

14
Level 1C Data Availability
15
Naming Convention
F11.1C.CALT1.R1.19990303.37470.HDF
T TMI was used as reference standard 1
Maturity level 0 Nothing has been done to
calibrate the Tb to the reference sensor 1 Data
has been re-calibrated to match the reference
instrument 2 The intercalibration has been
verified by at least one independent effort 3
The intercalibration has been accepted by the
Precipitation Science Team
R Revision number 1 Initial Version
R2, R3 etc.) indicate substantive changes to
the data, data format, and/or file contents,
which imply reprocessing of the entire data
record. Minor revisions to the files, such as
might occur with a change to the metadata, are
indicated by a lower case letter following the
revision number (i.e. 1a, 1b, etc.).
16
Available Rainfall Maps