| Start Date | Start Time | End Date | End Time | Data Quality Metric |
|---|---|---|---|---|
| 04/01/1994 | 0000 | 04/09/1996 | 2359 | Suspect |
| Subject: | Reprocess: LBLRTM and AERI/LBL QME data |
| DataStreams: | sgpqmeaerimeansC1.c1, sgpaerilbldiffC1.c1, sgpqmeaerilblC1.c1, sgplblch2C1.c1, sgplblch1C1.c1 |
| Description: | DQR No:
Subject: LBLRTM and AERI/LBL QME data
Date Submitted: 7/2/96
Submitted By: Tim Shippert
___ Instrument Mentor
_X_ EST Member
___ Science Team Member
___ Other _____________________________
Organization: PNNL
Email address: tr_shippert@pnl
Telephone: 509-375-5997
Platform/Measurement:
LBLRTM model runs:
sgplblch1C1.c1
sgplblch2C1.c1
AERI/QME output:
sgpqmeaerilblC1.c1
sgpqmeaerimeansC1.c1
sgpaerilbldiffC1.c1
What level data: (raw,a0,a1,b1,c1 etc): c1
What location was the data collected at: SGP
Period of time in question
Begin Date 4/1/94
End Date current
Data should be labeled:
___ questionable
___ incorrect
___ wrong calibration
_X_ other: Usefulness of data is limited to instrument validation and
the evaluation of atmospheric state measurements, and
should not be used for the study of radiative transfer
modeling issues
Discussion of Problem:
The AERI/LBLRTM QME, which compares the output of the Line by
Line Radiative Transfer Model (LBLRTM) with the spectra measured by
the Atmospheric Emitted Radiance Interferometer (AERI) at the SGP CART
Central Facility, has been successful in identifying and assessing
problems with the measurement of downwelling spectral radiation and
the characterization of the atmosphere in the radiating column, but
there are serious concerns regarding the use of this data for
understanding and addressing radiative transfer modeling issues by the
scientific community. There are three main purposes of the QME:
(1) To improve modeling of radiative transfer;
(2) To assess the ability to measure high-resolution
downwelling radiation, in particular from the AERI
instrument;
(3) To assess the ability to characterize the atmosphere in
the radiating column, including water vapor.
The AERI/LBL QME has been instrumental in finding numerous
problems with (2) and (3), as seen in the attached timeline. Because
the AERI and sonde data are being released, it is important to
identify issues related to these datasets, which can be accomplished
using the AERI/LBLRTM QME. Therefore, the QME is a useful tool for
validating and establishing data quality for the instrument data
streams. As a result the QME process, difficulties associated with
ARM-style automated data collection are identified.
Because of the difficulties associated with the measurement of
spectral radiance and, most importantly, that characterization of
water vapor as part of the atmospheric state, the QME has been unable
to contribute to the improvement in radiative transfer modeling.
Therefore, the QME data should not be used as a basis for improvements
to radiative transfer models; the magnitude of the residuals due to
errors in the determination of the atmospheric path in the radiating
column are larger than those due to modeling error.
The following are periods of time which may be used for
general analyis of spectral radiance measurements:
(a) 5/94-12/94
This period was prior to the time period which has
been identified as containing radiosonde launches with
the Vaisala sonde calibration problem. The AERI00 data
has been reprocessed to correct for the field of view
obscuration. However, pseudo-Vaisala processing of
sonde data was used through 5/21/94. Therefore,
the model validations and the associated QME data
should be treated seperately from other time periods.
(b) 7/95-8/95
This period was prior to the time in which the dust on
the AERI01 mirror impacted the data. Some sondes
during this period have the Vaisala correction
problem, which are identifiable using serial
numbers. The MPL data is questionable due to issues of
dust and wear, which impacts the ability to determine
clear sky cases.
(c) 1/96-3/96
The mirror on the AERI01 was replaced, but the data quality
is currently unknown. The MPL00 data has the same
problems as in (b). Most importantly, the MWR was not
running during this period, so we have no comparisons
with sondes and our ability to detect clouds from
liquid water is limited.
(d) recent data from 4/96 on
A new "improved" mirror was put in the AERI01 in April '96
(data quality still unknown, however), and the MPL02 came
on line in May. Data from this period looks good;
however it has not been as throroughly analyzed as
older data, and potential problems have not had as
much time to surface.
-----
The following paragraphs summarize the problems and issues with the
instrument datasets related to the LBLRTM model runs.
AERI
----
The sporadic LN2 loss in the AERI00 dewar from April, 1994 through
July 25, 1995 is manifested in the gradual increase in noise into the
AERI00 measured spectrum, ultimately manifested by a Planck function
spectrum of the detector temperature as the detector warms upon running out
of the N2 coolant. These spectra were included in the data distribution of
the AERI00 spectra as well as the QME validations with LBLRTM, although
during subsequent reprocessing caused by the FOV problem (see next
paragraph) these samples have been flagged in the AERI data and skipped by
the QME. The current AERI01 instrument makes use of a stirling cooler and
no longer requires an LN2 dewar.
The field of view of the AERI was partially obscured from April,
1994 through July 6, 1995, which resulted in a several percent increase on
measured radiance [Knuteson, et al., 1995]. The effect of this obstruction
has been removed from the spectral measurements, and both the AERI and
AERI/LBLRTM QME data has been reprocessed and redistributed. Use of the
original, uncorrected data is not advised.
The accumulation of dust on the AERI01 collection mirror over the
period 9/1/95 through 12/22/95 resulted in an obscuration which may be
similar in nature to that of the AERI00 obstruction. A similar effect has
been identified over the period 1/18/96-4/9/96. The data has not been
corrected, although efforts are underway to remove the effect of the dust
from the radiance spectra. It is advised that the AERI01 data over these
time periods not be used for scientific study prior to reprocessing (see
AERI DQR #'s P960715.1 and P960715.2).
MWR
---
Other measurement datasets related to the AERI/LBLRTM QME studies
have been identified as having problems in certain time
periods. Temperature calibration issues in the MWR from 4/94 through 12/94
are being addressed, though the data which is currently released reflects
incorrect calibration coefficients.
The processing of MWR measured brightness temperatures in the
retrieval of total water vapor and total liquid water has changed starting
3/1/96, which has resulted in improved agreement between the column water
vapor product from the MWR platform and the retrieved total water vapor
which is done as part of the MWR/LBLRTM QME in conjunction with the
AERI/LBLRTM QME.
BBSS
----
The AERI/LBLRTM QME studies have identified several issues with the
radiosonde data. The nominal Vaisala processing of the raw radiosonde data
from 4/94 through 5/21/94 was not done, which resulted in the use of the
unprocessed data in the validations of the line by line model with the
radiance measurements and release of QME data. The radiosonde data was
subsequently processed using software written at PNL which intended to
mimick the Vaisala proprietary alogrithms, and the validations and QME
products were rerun. The model calculations and QME data for this time
period should be treated distinctly from those of subsequent time periods
as a result of the differing radiosonde processing software.
From 4/94 to 8/3/94, ground checks were routinely applied to
nominal BBSS PTU data. These ground checks were removed before using the
data as input to the LBLRTM. From 4/94 through 5/21/94, the period
described in the above paragraph, ground checks were removed as part of the
pseudo-Vaisala processing. From 5/21/94 to 8/3/94, the ground checks were
removed from the Vaisala nominal data by a seperate post-ingest routine.
After 8/3/94, ground checks were no longer applied to nominal BBSS data.
Over the time period of 1/20/94 through 9/22/95, some radiosondes
launched were incorrectly calibrated by Vaisala at the factory and
incorrectly measured water vapor through the atmosphere. This data,
including the model validations and QME products, cannot be corrected and
should not be used. For more information, including the serials numbers of
those sondes affected, refer to the DQR submitted 2/28/96 by Barry Lesht,
the instrument mentor.
In addition, some studies [Clough et al., 1996; Whitney et al.,
1996] have indicated that the radiosonde measurements contain systematic
errors in observing water vapor through the atmospheric path, and that
these errors are the largest source of disagreement between the line by
line model and the spectral radiance measurement.
----
The following is a timeline of most of the data streams related to
the AERI/LBL QME. It graphically illustrates just how many things can
affect the QME. The timeline table is 172 columns wide, so you will
probably need to resize your window or print it out in landscape mode to
see it completely:
ARM AERI/LBLRTM QME Timeline
MPL00 Era
MPL02 Era
_____________________________________________________________|____________________________________________________________________________________________ ____|____
|
| |
MWR in datastream (S/N 010)
MWR in datastream
_________________________________________________|_______________________________________________________________ ______________________|___________
|
| |
AERI01 Era
AERI00 Era _ _ _ _ _ _ _ _ _ _ _ _
_______________________________________|__________________________________________________________________
_________________________________|____|________________________|_______
| |
4/94 12/94 4/17/95 5/7/95 7/6/95 7/31/95
9/1/95 10/95 12/1/95 1/18/96 3/15/96 4/16/96 5/10/96
|-----|----/-----|------|-------|------|-------|-------|--------|-------|-----|------|-----|------|------|---|----|----|-------|-------|-------|------|----|----|----|--->
5/21/94 1/20/95 4/26/95 5/31/95 7/25/95
8/28/95 9/22/95 11/1/95 12/22/95 3/1/96 4/9/96 4/20/96
|______________| |___________|
|__________| |______________|
| |
| |
SCM IOP SCM IOP (7/17-8/6)
ARESE IOP (O3 sondes) IOP
|_______________________________________________________________________|
|
Sporadic AERI00 LN2 Dewar fill problem
|_______________________________________________________________|__________________________|___________________________|_______|______________________|___________________
| |
| | | |
AERI00 FOV obscuration (corrected) Mirror #1
Dust on mirror #1 Mirror #2 Residue on mirror #2 Mirror #3
|_____|
|_________________________________________________________________________|
| |
Pseudo-Vaisala Vaisala Calibration Problem (Central Facility)
Processing (no correction)
|________________|
|__________________________________
|
|
MWR Calibration Problem
Change in MWR data product processing
(to be reprocessed)
|_____________________________________________________________________________________________|
|
MPL00 Data Questionable (wear/dust issues)
References:
Knuteson, R., B. Whitney, H. Revercomb, & F. Best. (1995). The
history of the University of Wisconsin Atmospheric Radiance
Interferometer (AERI) prototype during the period April 1994 through
July 1995. (Technical Report with ARM PIF #'s P950525.4 and
P950224.1). University of Wisconsin-Madison.
Clough, S.A., P.D. Brown, J.C. Liljegren, T.R. Shippert, &
D.D. Turner. (1996) Implications for atmospheric state specification
from the AERI/LBLRTM QME and the MWR/LBLRTM QME. Proceedings of the
Sixth Atmospheric Measurement Program (ARM) Science Team Meeting.
Whitney, B.A., H.E. Revercomb, R.O. Knuteson, F.A. Best, and
W.L. Smith. (1996). Atmospheric Emitted Radiance Interferometer
(AERI), part II: Water vapor and atmospheric aerosols. Proceedings of
the Sixth Atmospheric Measurement Program (ARM) Science Team Meeting.
Suggested Corrections of the Problem:
AERI/LBL QME Data should be released only with this DQR attached.
Tim Shippert tr_shippert@pnl.gov
(509) 375-5997
FAX: (509) 375-3641 |
| Suggestions: | AERI/LBL QME Data should be released only with this DQR attached. |
| Measurements: | sgpqmeaerilblC1.c1: more
sgplblch2C1.c1: more sgplblch1C1.c1: more sgpaerilbldiffC1.c1: more sgpqmeaerimeansC1.c1: more |