EXPERIMENT

1
Remote Sensing of Water Vapor, Cloud Liquid, and
Temperature during RHUBC by Microwave and
Millimeter Wave Radiometers E. R. Westwater1,4,
D. Cimini2, V. Mattioli3, M. Klein4, V. Leuski4,
A. J. Gasiewski1,4 and D. Turner5 1Department
of Electrical and Computer Engineering,
University of Colorado at Boulder, USA 2CETEMPS,
University of LAquila, LAquila,
Italy 3Dipartimento di Ingegneria Elettronica e
dell'Informazione, Università di Perugia,
Perugia, Italy 4Center for Environmental
Technology, University of Colorado,
USA 5University of Wisconsin, Madison, Wisconsin,
USA
MOTIVATION

• EXPERIMENT
• Radiative Heating in UnderExplored Bands Campaign
  (RHUBC)
• Period February 15 March 15, 2007
• Location ARM NSA, Barrow, Alaska
• GOALS
• Comparison of Microwave and Millimeter
  wavelength radiometers in deriving PWV and LWP
• Retrieval of Temperature and Humidity Profiles

• As a continuation of our 1999 and 2004 NSA
  experiments, we participated In RHUBC, where the
  principal focus was on low amounts of
  Precipitable Water Vapor (PWV lt 2 mm), and a
  secondary goal of small amounts of cloud liquid.
  Several improvements were made in the GSR, both
  in number of frequencies, and performance of
  selected channels. The GSR was successfully
  operated from February 20 to July 1, 2007.

GSR
MWRP
MWR
Fig.1 List of instruments and experimental set-up
deployed during RHUBC.
Retrievals of PWV

• GSR RETRIEVAL METHOD
• A Priori Linear Statistical Retrieval using Tbs
  as predictors
• Historical data set of Vaisala RS90/RS92 taken at
  NSA
• (Substantial QC was necessary)
• RTE model based on Liljegren et al. (2005)
• Tbs adjusted to Liljegren et al. 2005 model
  based
• on WVIOP_2004 data comparisons
• Covariance matrix of experimental errors
  determined from WVIOP_2004 data
• MWR LOS RETRIEVAL METHOD
• A priori Linear Statistical Retrieval using t
  (optical depth) as predictors.
• Monthly values of Mean Radiating Temperature
• RTE model based on Liljegren et al. (2005)

Fig.2 Real time display during RHUBC of GSR Tb
data. Note the menu of available choices of
available data.
Fig.3 Comparison of PWV retrievals from the MWR
LOS, the GSR, and RS92 Sondes on Julian day 65,
2007
Fig.4 Scatter plot comparisons between the GSR
(blue) and the MWR LOS retreivals of PWV compared
with RS92 Radiosondes
RETRIEVALS OF LWP

LWP LESS THAN 60 g/m2
CLEAR CONDITIONS FROM VCEIL AND IRT
Fig. 5. Left. Scatter Plots of LWP retrievals
from the MWR-LOS (red), MWRRET(blue) and the
MIXCRA (BLACK) vs. the GSR. Right MIXCRA VS.
GSR. Note the excellent agreement between MIXCRA
and the GSR
Fig. 6. LeftGSR vs. MIXCRA. Right MIXCRA vs.
MWRRET. Note the outliers associated with the
MIXCRA in both figures . This could be due to
either an algorithm or an AERI hatch problem.
RETRIEVALS OF LWP AND T AND Q PROFILES

We have recently developed One Dimensional
Variational (1D-VAR) retrievals of GSR data. This
method is fully physical and is based on the
Liljegren et al. (2005) RTE model. It initial
guess is the NCEP forecast model, and the
algorithm iterates until the residuals are less
that the estimated GSR noise levels. These
retrievals use zenith GSR data up to 183.31 GHz.
In Fig. 6, we showed LWP retrieval for contents lt
60 g/m2. However, on Julian day 73, there was a
significant amount of liquid shown by the MWR,
the GSR, and the AERI. Unfortunately, the MWRRET
data did not pass quality control and are not
available for this day. We show in Figs. 7
retrieval comparisons for the higher liquid
amounts.
Fig. 7. Left Retrievals of GSR-MIXCRA. Right
LOS-GSR
Fig. 8. 1DVAR retrievals profiles of Temperature
and Specific Humidity.
CONCLUSIONS

• PWV retrievals of the GSR compared well with
  RS92 Radiosondes and yield a relative accuracy of
  about 5.
• Comparisons of LWP from the GSR, as well as
  MIXCRA and MWRRET are quite encouraging for LWP gt
  20 g/m2. The use of the 340 GHz channels is
  promising for the smaller LWP amount.
• Also encouraging are the preliminary 1DVAR
  retrievals of temperature and specific humidity.
  Further development will focus on cloud liquid
  and the complete set of 27 GSR channels.

Corresponding author Ed.Westwater_at_colorado.edu
ARM Science Team Meeting, March 10-14, 2008,
Norfolk, VA