DIP_METTWR_2.txt

Preliminary meteorological data from the metgroup's (Andreas, Fairall, Guest,
Persson) tower at the SHEBA icecamp.  Note, data screening has been fairly strict in
an effort to reject all bad data; thus, some additional data may be recovered when
more intelligent editing/quality control methods are developed in the future.  Most
"bad" data was caused by rime-ice formation on the sensors.

Released January 18, 1998.  The file is approximately 70 kb consisting of 571 lines of
text with 20 tab-delimited columns.  The first two lines are column headers as per
JOSS instructions.

Contact Chris Fairall (cfairall@etl.noaa.gov) or Ola Persson (opersson@etl.noaa.gov)
for more information.

Data from the sonic anemometers and the slow sensors have all been used
to obtain hourly average values from Julian Day 304 (Oct. 31, 1997)
through Julian Day 327 (Nov. 23, 1997 ) in this one file.  Missing or excluded  data
are indicated by either 99.9,  99.99 or 999 depending on the variable. 

The parameters are:

Sheba date/time format: YYYYMMDDHHSS.SS

JD: Decimal Julian Day -the time is the start time for the hourly
averages (note, these time averages are not centered at the top of the hour)

ws-2.5 m; ws-10 m, wd-2.5 m, wd-10 m: mean wind speed (m/s) and wind direction
(deg clockwise from approximately true north - not corrected for
rotation of site since installation) linearly interpolated to 2.5 and 10 m
from the sonic anemometers.  Interpolation is only done if at least two
levels of good data exist bracketing the respective heights.
.
t-2.5 m, t-10 m, q-2.5 m, q-10 m: Hourly average temperature (deg C) and mixing
ratio(g/kg) linearly interpolated to 2.5 and 10 m from the Vaisala T/RH
sensors giving good data and immediately bracketing the height.  Accuracy is about
0.1 C for temperature and 2% for humidity.

Ustar, Hs:  Hourly mean friction velocity (m/s) and sensible heat flux
(W/m2) estimated from the median of the values given by the 5 levels of
sonic anemometers.  The raw 10-Hz data was processed using an FFT on 7
overlapping 14-minute periods and averaging the spectra.  A pressure of
1013 mb was used in the calculation of Hs from the w'T' covariance,
giving an uncertainty of about +-3%.  Ustar is the square root of -<wu>; in the case
where <wu> is positive, Ustar is given a negative sign.  Automated and some manual
quality control was done on the fluxes before the median was taken.  This
quality control included testing for reasonable wind speed and flux
values, and removing all sonic data for which the sigu/S or sigv /S >0.45
where u and v refer to the along and cross-stream wind components and S is the
wind speed.

t-0.12 m, t0, tice:  Hourly mean temperatures (deg C) at 12 cm above the
surface (from the General Eastern), at the surface (estimated from
radiative measurements of the Barnes radiometer), and at the ice/snow
interface (from a thermistor placed there at set-up - the snow depth
above it varied from about 25 cm at the beginning to about 45 cm at the end).  

lwd, lwu, swd, swu:  Hourly mean downward and upward longwave and
shortwave radiation (W/m2) from the Eppley radiometers.  Some editing
was done to remove values >300 w/m2 and < 100 W/m2 for the longwave and
values > 50 W/m2 and < 0 W/m2 for the shortwave (to be done for only the
dark time period which this is).

Prcp-or, prcp-wgh:  Hourly mean precipitation (mm/hr) from the optical raingauge. 
This sensor is not actually calibrated for snow, so data are inferential at best.  Note
also, that it does not successfully reject blowing snow so all readings in strong winds
are suspect.  An NCAR weighing bucket gauge was also used, but data are not given
in this report.