Oxford MIPAS meeting#2
26 Mar 02

Present

Oxford: Anu Dudhia, Don Grainger, Dominik Kasprzyk, Francesco Lama, Mark Parrington, Chiara Piccolo, Clive Rodgers,
RAL: Victoria Jay

Future Meetings (AD) [Prev]

Given the number of people interested, and the distance some of you would have to travel to attend, the plan is to make these meetings slightly more formal, so

Put these notes on the web together with related plots/documents for the benefit of those who can't attend (although not linked from anywhere else, so not too "public")
Include one or two more formal talks (10-15min each) each meeting, advertised in advance

Instrument Status (AD) [Next]

Starting to take interferograms - all appears to be going well

Lv2 access software (VJ) [Prev] [Next]

RAL are responsible for supplying software to extract information from ESA's level 2 files and associated products

Got hold of Reinhold Spang's (Leicester) IDL code for reading Lv2 data from the Envisat CD
So far only reads profile values, but working with Reinhold to extend the code
Also installed new version of Enviview at RAL, used to verify results

Cloud Detection (DK) [Next]

Dominik Kasprzyk (physics undergrad) is 1 week into a month long vacation project looking at cloud detection in MIPAS spectra.

So far, he has run RFM simulations with cloudfree and thin/thick cloud at 15km using 8 different atmospheres to investigate the behaviour of the U.Leicester `Cloud Index' (defined as the ratio of radiances in two microwindows).
The ULe cloud-detection threshold of CI=1.8 looks like a good cut-off for complete cloud cover (<1.8) while 75% cloudy fields-of-view lie above this.
Future work to try and associate cloud optical thickness with CI, or some other function of radiances - more results next meeting.

Lv2 Pre-Processor (CP) [Prev] [Next]

Chiara is responsible for supplying software to extract information from ESA's level 1 files and associated products

Current plan is to perform the apodisation on the full Lv1 spectra in the interferogram domain.
However, this takes ~minutes for one spectrum, so need to investigate faster options
Fall back solution is to perform apodisation only for selected microwindows by a convolution in the spectral domain.
Presentation next meeting

Error Spectra (FL)

Summary of information presented at the Progress Meeting#18 (Firenze, 01 Feb 02)

Only recent changes (Dec'01) have been to increase the water vapour variability (v3.12) and non-LTE errors (AMIL2DA, from IAA, Spain).
Jan'02 occupation matrices selected to include these, but not the Jul'01 occupation matrices. However, the effects are small anyway.

Pointing Effects (FL)

Summary of information presented at the Progress Meeting#18 (Firenze, 01 Feb 02)

Microwindows are selected for "nominal" tangent altitudes, so what happens if the pointing introduces an altitude bias
Investigation of biases up to +/- 6km showed no unexplained/worrying behaviour (nominal margin is +/- 1.5km)

AMIL2DA Error Analysis (AD) [Next]

Part of the AMIL2DA project is to produce a uniform error analysis for each group's retrievals

This involves plotting averaging kernel matrices and microwindow systematic error analyses. Oxford CH4 results: [Averaging Kernels] [Delta-function Responses]
Vivienne Payne also using this as an opportunity to try out the `Residual and Error Correlation' (REC) analysis.
Example: for the RAL H2O/O3 retrieval, microwindow analysis shows ClONO2 as a significant absorber, and REC analysis shows a significant ClONO2 +ve signature in the residuals - consistent with the fact that ClONO2 is omitted as an absorber in this version of the retrieval. [RAL results]

Oxford MIPAS meeting#2 26 Mar 02

Oxford MIPAS meeting#2
26 Mar 02