Data Set Overview
=================

The balloon data set comprises time-ordered telemetry readings (internally 
referred to as 'level 2', entailing some interpolation and reconstruction - 
see CRISP1990; KREMNEV1986) from the balloons released into the Venus 
atmosphere by VEGA 1 and VEGA 2.  A 'level 3' product for each balloon is also
developed, integrating these measurements with a trajectory estimated from 
Very Long Baseline Interferometry (VLBI - see PRESTON1986; SAGDEEV1992) 
measurements of the balloon position on Venus' disk measured by radio 
telescopes, and from a model of the balloon altitude/pressure history - 
see CRISP1990 and DORRINGTON2013.

The lander data set is a pressure and temperature profile obtained from the 
METEO instrument (see LINKIN1986B) on the VEGA-2 lander during its descent in 
the Venus atmosphere.  It represents the most complete profile obtained by 
any Venus probe to date  (the precision and altitude resolution of the Venera
landers is not well-documented, and the results have not been available in
tabular form;  the Pioneer Venus data are reasonably well-documented and of
fairly high quality, but no direct temperature measurements below 12km were
obtained because of sensor failures on all four probes at that altitude).

The START-TIME and STOP-TIME are parachute opening/telemetry start, and
landing times, Moscow time of reception at the ground station.


Parameters
==========

The 'level 2' dataset (archived as VG1BL_EDR.DAT, VG2BL_EDR.DAT) comprises 
time, temperature, pressure, vertical wind (relative to the balloon) and 
cloud backscatter (nephelometer) data, as physical units together with data 
quality flags.

The 'level 3' dataset (archived as VG1BL_RDR.DAT, VG2BL_RDR.DAT) adds the 
latitude, longitude from radio tracking, and the inferred zonal and 
meridional wind speeds.  The vertical speed of the balloon is derived from 
the balloon measurements (W_b), as distinct from the motion of the air 
relative to the balloon (W_a). 

The lander dataset comprises time, temperature, pressure and derived altitude.
 
Processing and Data Quality
===========================

The implementation of the Vega balloon experiments was challenged by severe 
resource demands, and data from the Vega balloons was sent directly to Earth 
via a radio link whose power was limited by the lithium battery carried in 
the balloon gondola.  The transmitter was operated for short periods (75s) 
with intervals in between (see KREMNEV1986), and the data were heavily 
compressed - specifically only the least significant 6 bits of each 12-bit 
measurement word were transmitted typically, with the most significant 6 
bits (not expected to change rapidly) telemetered much less often. Since the 
physical quantities changed more rapidly than expected, the MSBs had to be 
sometimes estimated.

Linear interpolation was used to fill the data drop-outs (e.g. data with 
quality set to 0). The points with data quality = 0 are not used in 
subsequent processing, but made the files easier to plot on the primitive 
computers at that time. 

As an example of the ambiguity introduced by the compression scheme, 
consider the sequence of units and tens data

4, 7, 2, 6, 7, 9, 3, 30, 6, 9, 2, 4, 6, 7, 1, 70, 8, 2, 1, 7, 9, 1, 3, 20,...

Are those numbers:

14, 17, 22, 26, 27, 29, 33, 30, 36, 39, 44, 56, 67, 71, 70 .... ?

Or was it some other permutation?  

24, 27, 32, 26, 27, 29, 33, 30, 36, 49, 54, 66, 67, 71, 70 ....

Additionally, there is a small glitch in the pointer used for the wrap-around 
output data buffer that lost track of where (which variable) was being 
actually readout on each downlink.  This was determined by Lee Elson by 
matching data patterns.  Dave Crisp then reconstructed the rest of the  
dataset. Fortunately, the gain of the temperature sensor was set such that 
it rarely changed by more than 1 MSB.  The quantity, Tn, is the 'calibration' 
temperature recorded on the nephelometer board. This variable, which was 
critical for interpreting the nephelometer data, was not sent back often, but
it was one of the only variables that was sent back as a full 12-bit word.  
Although not very useful scientifically, this variable served primarily as a 
sanity check for the balloon model. The outputs from the anemometers also 
suffered some challenges.  There was some of static friction (stiction) near 
zero RPM and a finite response time. 

Examples of results from the Balloon model are included in the files 
Crisp_BalloonModel_VEGA1.dat and Crisp_BalloonModel_VEGA2.dat in the /extras
directory. Those model data were used in the level 3 products.

The relevant pressure, temperature quantities for the lander were selected 
from one of four temperature sensors and one of three pressure sensors, with 
various mounting characteristics. The data were calibrated for dynamic 
effects.

The lander altitude is a derived quantity.

Confidence Level Overview
=========================

The balloon dataset is believed to be complete and of high quality, subject 
to the reconstruction caveats above.

The source of the lander data is a tabulation in [LINKIN1987]. Data were
transcribed from a PDF of that paper via Optical Character Recognition and
errors identified by discontinuities etc. and corrected by hand by R. Lorenz.

The last few datapoints (from #424 onwards) correspond to measurements on the
surface - variations of about 0.5K and 0.04bar are seen.  These may be ambient
meteorological variations ; they may be actual atmospheric effects but
artifacts of the landing (e.g. the impingment of the probe wake onto the
surface) or they may reflect the intrinsic precision of the measurement due
to instrument noise.