Oxford MIPAS meeting#115
3 Apr 07

Present

Oxford: Anu Dudhia, Jane Hurley, Luis Millan-Valle, Chiara Piccolo, Joanne Walker, Claire Waymark

Envisat currently performing an orbit control manoeuvre, so MIPAS switched off 1-3 April. Should resume today with 3 days of nominal mode measurements.

Mixture of NOM and MA mode observations scheduled for the next 2 weeks, in line with the increased duty cycle (now 60%). In the absence of other requirements, the nominal operational sequence is now a 10 day cycle consisting of:

Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7	Day 8	Day 9	Day 10
MA	NOM	NOM	NOM	off	off	NOM	NOM	off	off

Mission Planning Document for 2007 [XLS] [HTML].
Currently defined measurement modes
Envisat Web-Site

L1B Data (JH) [Prev] [Next]

Coverage maps updated (after a couple of problems - see next item) to show data acquired up to the end of February 2007.

[L1B Data Coverage]

BEAT problems (JH/AD) [Next]

http://www.science-and-technology.nl/beat/

Problems have been encountered running BEAT v4.2.0 to read L1B data (this is the original format L1B data, software v4.67 or earlier). The following are examples of IDL code that worked with the previous version of BEAT but no longer work (assuming that a L1B file has been opened successfully using PF = BEAT_OPEN ( filename ))

Problem#1 (loading the Geolocation data structure)

GEOADS = BEAT_FETCH ( PF, 'GEOLOCATION ADS' )

results in error message % BEAT-IDL ERROR -102: "invalid fieldname argument (GEOLOCATION ADS) (libbeat/beat-type.c:814)"

Problem#2 (extracting the nominal number of sweeps from the specific product header)

SPH = BEAT_FETCH ( PF, 'SPH' )
NOMSWP = SPH.NUM_SWEEPS_PER_SCAN

the first line results in error message % BEAT-IDL ERROR -300: "product error detected (value for ascii integer too large for int16)" but only for data from orbits 25126 (20Dec06) and later, earlier orbits OK.

A query was sent to STCORP and rapidly resolved - see below for solutions

Precision Validation (CP) [Prev] [Next]

ACPD Paper discussion now closed.

Final draft now completed and sent to ACP

NOx (JW) [Prev] [Next]

Constructing monthly mean maps of NOx to examine the stratospheric variability at middle and low latitudes

Pressure v Time series for 90S:60S zonal average [N2O] [NO2]
Comparing MIPAS NOy (=NOx+HNO3) with N2O shows expected anticorrelations in lower stratosphere (ie 'new' air has relatively more N2O than NOy, whereas in 'old' air more N2O has been converted to NOy),
However, in the upper stratosphere both appear to be correlated - could be related to assumption that nighttime NO2 accounts for all NOx, less valid at high altitudes where nighttime NO could be significant.
Clear denitrification signals in the lower stratosphere polar winters
Next: see if there is a quantitative relationship between decrease in N2O and increase in NOy, eg is (2N2O + NOx + HNO3) conserved?

Cloud Parameters (JH) [Prev] [Next]

Scheme to retrieve Cloud Top Height, Cloud Top Temperature and Cloud optical thickness.

Now works with simulated data. [Plot] showing 'true' (diamonds) and retrieved (asterisks plus red error bars) values for the three parameters (left hand axis) as a function of position of cloud top relative to centre of the FOV (lower axis) and different log10(extinction coefficients) (different rows of plots - thick cloud in top row, thin cloud bottom row)
Trying it out on an orbit real data currently produces plausible-looking results but difficult to validate.
Also looking into SVD patterns from simulated data to see if any obvious cloud-signatures emerge.

Mesospheric Retrievals (LMV) [Prev] [Next]

Trying a linear least-squares fit retrieval for pT for altitudes 52, 60 and 68km using entire A band (screening out lines affected by non-LTE effects and gases other than CO2). If this works, intend to apply it to MA mode data.

Solved problem with application of hydrostatic constraint
Linearised the problem as far as possible by retrieving Planck function and pressure (instead of temperature and log(pressure))
Now investigating whether it is possible to improve on the single iteration method using knowledge on how we expect the temperature and pressure Jacobians to change with pT, ie we not only know the precomputed Jacobians at the starting point but, for a given pT retrieval we can also estimate what they would look like at the retrieved point.

O3 Isotopes (CP) [Prev] [Next]

Trying to run a Kalman filter (within each latitude band for 1 month) to improve S/N but encountering problems with passing retrieved profile from one scan as a priori for the next

Reinterpolating for slightly different tangent point grid smooths profile, causing gradual reduction in ozone peak
Passing the un-interpolated profile on the set of tangent points results in a gradual build-up of oscillation features

Solution is probably to apply some sort of intelligent smoothing between scans to reduce oscillations but not the peak value

MIPAS/TES Retrievals (CW) [Prev] [Next]

Aim is to combine MIPAS limb measurements of stratospheric O3 with TES nadir-viewing measurements to obtain more information on tropospheric O3 profile (basically how TES was supposed to work before their problems with limb-views)

Currently trying to get a joint p,T,O3,H2O,surface temperature nadir retrieval working with simulated data.

Stratospheric Trends (AD) [Prev] [Next]

Running our local MIPAS processor on selected days throughout the MIPAS mission to try and estimate stratospheric trends.

Retrieve the ESA key species plus ClONO2, N2O5, F11, F12 (the additional 'HIRDLS' species) and CO (LTE assumption). Days have been selected on the basis of being approx 3 months apart (6 months for full-res data) and MIPAS operating for almost the whole day in nominal mode.

The following days are being processed (more may be added later)

Mode	Dates
FR17	09JUL02	23SEP02	23MAR03	22SEP03	21MAR04
RR17	17SEP04
RR27	28JAN05	21MAR05	16MAY05	10AUG05	12JAN06	13MAR06	25SEP06	14JUN06	04JAN07

ACPD Papers [Prev] [Next]

Papers selected from the ACPD website for discussion

"Global statistics of liquid water content and effective number density of water clouds over ocean derived from combined CALIPSO and MODIS measurements": Y. Hu et al. (published 28.03.2007, open until 23.05.2007); Introduced by JH.
"Retrieval of temperature profiles from CHAMP for climate monitoring: intercomparison with Envisat MIPAS and GOMOS and different atmospheric analyses": A. Gobiet et al. (published 27.02.2007, open until 24.04.2007); Comments being collected by CP.
"Validation of MIPAS-ENVISAT NO2 operational data": G. Wetzel et al. (published 02.03.2007, open until 27.04.2007); Comments being collected by CW.
"Global peroxyacetyl nitrate (PAN) retrieval in the upper troposphere from limb emission spectra of the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS)": N. Glatthor et al. (published 29.01.2007, discussion now closed); Comments submitted by LMV.

Other papers with 'MIPAS' in the title in the Open Discussion phase, which may be adopted at a future meeting

"Bias determination and precision validation of ozone profiles from MIPAS-Envisat retrieved with the IMK-IAA processor": T. Steck et al. (published 30.03.2007, open until 25.05.2007)

Following discussions with Sander Niemeijer from STCORP,

Problem#1: The following works (insert '_' in the field_name)

GEOADS = BEAT_FETCH ( PF, 'GEOLOCATION_ADS' )

Full explanation from SN:

This is a change that we introduced from BEAT v3 to v4. Previously it was possible to traverse into data sets by using either the data set name (which could contain spaces) or the corresponding field_name of the data set (which contained underscores). In order to streamline the software and make it better compatible with other data formats we dropped the special treatment of data set names (which was an ENVISAT format specific feature). I would advise you to have a look at the several CHANGES documents that are include with the BEAT package. Although we tried to maintain backward compatibility, a lot has changed between v3.x and the current v4.2.

Problem#2: The following works (add extra argument specifying required field directly)

NOMSWP = BEAT_FETCH ( PF, 'SPH', 'NUM_SWEEPS_PER_SCAN' )

This is actually a more subtle problem due to an overflow in the SWEEP_ID field in the MIPAS Specific Product Header which happened between orbit 25125 (20Dec06) (SWEEP_ID=+31118) and 25126 (SWEEP_ID=+34020) - there is some ambiguity as to whether this should be a signed or unsigned 16-bit field. BEAT 4.2.0 actually checks this - the earlier version didn't - so reading the entire SPH is now flagged as causing an error.

Assuming you don't specifically want the SWEEP_ID value, the solution - as above - is to use a second argument to the BEAT_FETCH procedure that directly targets the field that you do want. If you do want the SWEEP_ID field (or the entire SPH), the following is suggested (until the next BEAT release)

If you have downloaded the source package of BEAT (i.e. Linux/Mac OS X/Unix platform) you can perform a small patch yourself by editing the file src/beat-dd-dictionary.c, searching for '"sweep_id"', changing the 'native_type_int16' in the first line that you find to 'native_type_uint16', and rebuilding and installing BEAT.

Oxford MIPAS meeting#115 3 Apr 07

Oxford MIPAS meeting#115
3 Apr 07