WT's proposed update is based on the inclusion of the SLIT6
correction (wide slit burn-in) in the analysis of 1997 EGS data.
It increases the sensitivity (deduced physical units will
decrease) by ~15% at 584A.  Uncertainties will be (15+10)% @ 584
else (25+10) in quadrature.

GDZ confirmed that he did not apply the SLIT6 correction during
his analysis. Qualitatively at least it appears that the bumps he
finds in the calibration curves could match the regions where the
SLIT6 correction is largest.

JL reported that sample OIII ratios agree with the shape of the
proposed update. 

CDP reported that the MgX lines (609/624) also gave a ratio
closer to 2.0 for off-limb data with the new curve.

***Action***   Adopt WT update, but to be coordinated with
other updates, preferably by 01/12/01.


The effect of incorporating the SLIT6 correction for 368A is
still TBD, but thought to be much less eg ~5%.  EGS-97 and
SERTS-97 data will be re-analysed.

GDZ's calibration shape could be an 'anti-kink' contributed to by 
wide slit burn-in from the FeXVI lines at 335 and 361A.  Currently no 
correction is available for this effect.

***Action***   Repeat analyses to include SLIT6 and adjust
calibration accordingly.

NIS2 - second order

The GDZ and theoretical/lab curves agree pretty well in shape.
The absolute values will need to be revisited once the first
orders values are updated. Application of the second-order
calibration should be made available in the software.

JL is using the SiXI lines to check the NIS2 first-second order
ratio.  Also suggests using the mean of EGS and SERTS 303A
intensity to tie in calibration.

***Action*** Adjust absolute values and adopt theoretical/lab
shape and implement in software by 01/12/01.



WT reports no detectable change post-recovery, or perhaps a small 

***Action*** Keep same as pre-loss calibration.


Based on a limited dataset GDZ suggests no change post-recovery.
From SYNOP data, WT suggests a correction factor of 1.4 - ie a
senstivity loss of 30%. This should be checked though because of
possible blending effects in SYNOP 368A data.

***Action***  Check Mg 368 and other data.  All agree on factor to
be implemented in next update (01/12/01). 

NIS general

It should be made clear that the NIS pre-calibration corrections
are based upon ensuring that there is no longterm change in the
intensity of HeI 584. When SYNOP data are used for this, there is
a large distribution of daily intensities, particularly later in
the mission as the solar cycle progresses.  The lower intensity
boundary of this distribution is chosen as the fiducial.  The
fact that AP (and SUMER) still finds a small increase in HeI
intensity with time from the ICAL1 data could be explained on the
basis of data selection.  Maybe the lower boundary of the SYNOP
data is picking the quietest of the quiet sun while ICAL1 data
includes the network.  This could maybe be checked by segregating
the network from the SYNOP images.

I've started looking at the intensity distributions as a function
of time for the SYNOP data, which may shed some light on this


All updates to GIS calibration must await the results of the gain
depression analysis Carl is doing, but in the meantime...


GIS4 has been compared to SUMER data using the NeVIII line at
770A.  Note that the line is ghosted by 15% and that SUMER is
revisiting its own calibrations which may move the SUMER data by
approximately 8%. The current analysis however suggests that the
GIS4 sensitivity should be increased by

2.44 +/- 0.5  pre-loss
1.83 +/- 0.4  post-loss

From his analysis GZD suggests a pre-loss factor 1.9 +/- 0.5,
though this would need to be increased to about 2.1 to take
account of the new NIS2 calibration (ie including SLIT6

Suggestion was to take some weighted mean of the results from the
two analyses eg F=2.3 @ 750A and to keep it flat as a function of
wavelength.  To be consistent with other GIS channels, however, 
it would make more sense to keep the laboratory shape as a
function of wavelength.


GDZ has intercompared all the other GIS channels as well and
requires adjustment factors (from the current calibration) of 

Fg1 = 1.45 @ 206A
Fg2 = 2.8  @ 283A
Fg3 = 2.9  @ 406A

All of these factors will need adjusting for the NIS (SLIT6)
update. But since the wavelengths at which they are normalised
are the wavelengths at which the lab calibration was measured it
will be easy to use these factors to scale the calibration and to
using the shapes from the lab cal.


1) finish analysis of LTGD
2) decide on consequent GIS4 factor
3) assess impact of LTGD on GDZ factors for GIS1-3
4) keep (all?) shapes as they are as a function of wavelength
5) adopt post-loss factor for GIS4 from SUMER intercomparison
6) scale post-loss factors accordingly?

Comments from Giulio

Giulio wasn't able to attend the meeting and I've received the
following comments from him.  Suggested answers/responses would
be welcome.

1) There is no mention of GIS second order - what is the
situation with that?

2) If the suggestion for GIS is to take the theoretical/lab shape
with the scaled absolute values I found, then I do not understand
why my shapes as a function of wavelength should not be adopted.
Such a compromise will appear a fudge to users. 

3) I do not see how gain depression will affect my results since
most of the lines I used are not very bright and if there was a
significant longterm change in 1997 I would have seen it.  I
cannot exclude changes after 1997.


Bill's summary timeline - indentation indicates dependency.
Lack of indentation indicates it should be happening now! 

NIS-1 and NIS-2 2nd order:

EGS-97 reanalysis (WT)
SERTS-97 reanalysis (RT,WT)
  New calibration curves
      Use line ratios to check pre/post behavior (JL, others)
         New post-recovery calibration curves


Sumer A/B reanalysis (KW/US, AP)
GIS LTGD analysis (CF)
Effect of NIS-2 changes on GIS-4 (WT,GDZ)
  New GIS-4 calibration curve
    Effect of NIS changes on GIS 1-3 (WT,GDZ)
       New GIS 1-3 calibration curves
          New post-loss calibration curves (?)