Oxford MIPAS meeting#63
12 Oct 04

Present

Oxford: Alastair Burgess, Anu Dudhia, Don Grainger Vivienne Payne, Chiara Piccolo, Gareth Thomas, Joanne Walker (new D.Phil student)
U.Bologna: Enzo Papandrea (visiting student)

Instrument Status [Prev] [Next]

Some details on the Envisat Web-Site including Instrument Availability web-site

Single-slide ("Backup Mode") tried out on 21st Sep - see Message from Rob Koopman
After a prolonged turn-off, the slides appear to function relatively normally when switched on again, for a few days at least.

Future Operations

Not yet decided whether future operations will be 2 slides @40% resolution or 1 slide @50% resolution, or some combination of slides. Note that the relative slide speed is constant so moving either one or both slides makes no difference to the time to acquire an interferogram at a particular resolution
Operations may temporarily resume in November to coincide with a validation campaign, probably in 2-slide, 40% resolution mode, but not guaranteed.
The MIPAS Science Team has (provisionally) defined a new nominal scan sequence consisting of 27 steps from 6-70km tangent altitude, with spacing varying from 1.5km in the lower atmosphere to 4km at the top - effectively keeping the same horizontal sampling as the previous nominal mode but trading reduced spectral resolution for increased vertical resolution [Next]

L1 Data (CP) [Prev] [Next]

ESA are sending reprocessed L1B data to the BADC (at RAL), but no files have yet appeared on the BADC site - see Message from BADC.

L2 OFL Data (CP) [Prev] [Next]

[Plots of profile locations]

Reprocessed v4.61 data continues to arrive
Data stored at Oxford gradually being sorted to separate the various versions we have now received.

L2 Anomalies (CP) [Prev] [Next]

Two further anomalies with the L2 NRT data have been noted at the recent MIPAS QWG-5 meeting

There are duplicate versions of L2 profiles, slightly different. (Possibly explained by use of analysis v forecast ECMWF data?)
There are format problems with the pT error propagation matrix (part of the L2 VMR retrievals) [Next]

H2O (VP) [Prev] [Next]

H2O as retrieved by OPTIMO (also ESA) is about 2ppmv higher than ODIN values and climatology in the upper stratosphere [Plot]
Using an alternative set of microwindows (actually 10 of the pre-launch, hand-selected microwindows which are mainly a mixture of C and D bands) produces more conventional results. [Plot]
Three possible explanations so far

Using more microwindows gives an improved retrieval
C band has spectral calibration/AILS characterisation problems (so including D band helps)
"Old" microwindows are less susceptible to the known problems of H2O lines at high altitudes in the current operational retrievals, either through use of weaker lines or more microwindows

MORSE (AD) [Prev] [Next]

MIPAS retrieval code being written at Oxford for public use, funded by NERC.

Problems with pT representation in VMR retrievals fixed
Major improvements in code optimisation - now takes approximately 4 hours on an Alpha to simulate the ESA offline retrieval for 1 orbit (=100 minutes of data) as opposed to 28 hours.
Currently working on improving flexibility (eg allowing microwindows to be specified by a simple list instead of the MW*DAT type databases, which are only really necessary if masks and irregular grids are required).

SF6 (AB) [Prev] [Next]

Zonal mean plots of retrievals of SF6 (using OPTIMO) show unrealistic equatorial features: minimum at 18km, maximum at 21km (similar to that observed in the ESA retrievals of CH4 and N2O)

Could be due to residual cloud effects, or
Problems modelling tropopause with 3km resolution

H2O2 (EP) [Next]

Using H2O2 as an example of working though the process of microwindow selection, LUT generation and retrieval using various bits of software available in Oxford. Started by using the RFM to calculate Jacobian spectra based on MIPAS mid-latitude day climatology profile.

CO (JW) [Prev] [Next]

As the first part of her thesis work the aim is to investigate the possibility of retrieving CO from MIPAS, especially the UTLS region,

Either night-time only, ignoring Non-LTE effects, or
day-time, jointly retrieving the vibrational temperature

(in contrast to the IAA approach which is to model the Non-LTE effects)

At the moment just learning to use the RFM to simulate simple cases such as nadir-viewing column retrievals.

Precision Estimates (CP) [Next]

By comparing MIPAS retrievals at orbit intersections - effectively where two observations are made of the same atmosphere a few hours apart - a check can be made of the actual scatter in the measurements compared to the estimated standard deviation given in the L2 data.

So far just looked at a single day (25 Nov 03), 14 orbits giving 12 intersections at approx 10degN.
Results: [CH4] [H2O] [HNO3] [N2O] [NO2] [O3] [PRE] [TEM] (shaded area shows observed SD about bias, red lines are predicted scatter using L2 error bars)
In most cases the observed scatter is quite well explained by the L2 error bars, although this may itself be surprising since the pT error propagation hasn't yet been included (see above).

Full Spectral Simulations (AD)

Using MIPAS OFL L2 data plus IG climatologies for minor species to construct atmospheric profiles around an orbit, then running the RFM to simulate the full MIPAS spectra for comparison with the actual L1B spectra.

Running the RFM in `OFM' mode without any special physics (LTE only, no line mixing or horizontal gradients) to simulate the spectra assumed within the retrieval

Also, continuum information is ignored so no estimate of cloud contamination effects (apparent at low altitudes), although note that these residuals only inclusde sweeps for which the L2 has performed a retrieval so obvious cloud-contamination should have been excluded

Currently only assuming idealised AILS

[Band A] [Band AB] [Band B] [Band C] [Band D]
Plots show L1B-RFM, residuals for 40 scans of orbit 10798 (24 Mar 04) with different tangent altitudes superimposed. Middle panel shows microwindows used in OFL retrievals. Lower plot shows spectral signatures of various gases (for 12km altitude).
[Zoom] on H2O MWs in C-band shows 2nd-derivative type of structure on all lines (consistent with REC analysis) although bear in mind that only the idealised AILS has been used here
[Zoom] on A-band shows O3 MWs 211 & 215 selected for 0.0625cm-1 retrievals (PT 215 lies on top of O3 215) which appear to have large chi-sq values in tests.

MIPAS QWG-5 Meeting (AD/CP) [Prev] [Next]

Held at ESRIN, 4/5 Oct 2004.

Minutes (from Simone Ceccherini, IFAC).

Oxford Presentations (PDF versions)

Email from Rob Koopman, ESRIN, 23rd Sep 2004

A very brief message with initial feedback from the first back-up mode tests. Unsurprisingly, data analysis is still in progress.
A sequence of very short tests has been performed on 21 September. The tests where only slide 1 has been moved showed no anomalies, although several speed deviation warnings were reported. The slide-2 tests showed no speed deviations at all, but the tests were interrupted due to an anomaly, and remaining tests were not performed. Further tests are being defined, and are to be executed this week (probably this afternoon) and next week.

Email from BADC (Anne De Rudder), 12th Oct 2004

Hi Chiara.
It turns out that you were looking at our MIPAS L1B near-real time archive. We do receive MIPAS L1B validated data from ESA (on tapes) and have started to copy them over but have currently storage limitations that do not allow us to make them available to the users yet. We expect to have solved this matter in a week or two. I'll let you know as soon as the data become available.

Oxford MIPAS meeting#63 12 Oct 04

Oxford MIPAS meeting#63
12 Oct 04