Beam Attenuation Coefficient protocols

                    Data Reduction Scheme

    A 25 cm Sea Tech Transmissometer was interfaced with the
University of Washington's SeaBird CTD for all EQPAC cruises. 
For TT008, beam transmission values were extracted from raw CTD
files and passed through SeaBird's SeaSoft data processing
package with no data averaging and the spikes were removed.  The
data were compared with the output using the SeaSoft spike
removal and 1 db bin averaging routine.  There was no substantial
difference in the data using the two methods, so for TT007,
TT011, and TT012 the data were taken from the SeaSoft output
directly.  Beam transmission values were selected at 1db
intervals and de-spiked.  The data were corrected for factory and
field air calibrations and adjusted for changing Temperature,
Salinity, Index of Refraction, and Pressure.  Beam transmission
was converted to beam attenuation coefficients using
c=-(1/z)*ln(%Tr/100) where c=beam attenuation coefficient (m^-1),
z=beam path length (m), ln=the natural log, and Tr=% beam

    The two spring cruises and the two fall cruises were treated
as paired data sets and processed further as follows.  Most
profiles only went to 400 m, so for TT008 the minimum beam c was
determined over that depth range for each profile, the minimum
beam c for the cruise was determined and all profiles were
shifted to the cruise minimum beam c.  The average depth of the
minimum was 250 m (200 m for TT012), so stations shallower than
this amount were adjusted by the same amount as an adjacent station
rather than the actual minimum of the shallow station.  The above
procedure automatically corrects for any drift in the LED light
source and uses the reasonable assumption that there was no
change in the minimum beam c during the three weeks at the
equator.  It also eliminates errors in air calibrations.

     To calibrate beam c with particulate matter concentrations
(PMC), beam c was picked from the adjusted data at each depth
where samples were filtered on the equator.  Regressions of beam
c verses PMC for TT008 and TT012 yielded the equations:

     TT008      c=0.00221821*(PMC)+0.3103.   r^2=0.87
     TT012      c=0.00154626*(PMC)+0.3700.   r^2=0.90

The r^2 values for these two cruises are the best correlations
between beam c and PMC that we have seen for any data from
surface waters.

     Sea Tech transmissometers were factory calibrated to have a
beam c of 0.364 m^-1 in particle-free water, so the difference
between the regression intercept (e.g. 0.3103 for TT008) and the
factory calibrated intercept (0.364) was added to all profiles to
obtain calibrated profiles for analysis.  Beam cp, the
attenuation due to particles in the water, was obtained by
subtracting the attenuation due to water (0.364 m^-1).  Inverting
the above equations yielded the relationship between beam cp and
PMC (ug/Kg):
     PMC=451*cp         Spring Time Series
     PMC=647*cp         Fall Time Series
By comparison, the relationship for data in surface waters of the
North Atlantic Bloom Experiment was:
     PMC=1022*cp        North Atlantic Bloom Exp.

     For cruise TT007 beam c was calculated as above and
detrended for decay of the transmissometer LED.  The minimum beam
c was determined in the upper 400 m for each latitudinal Station
(1-11 profiles) and all profiles at a particular Station were
adjusted to the minimum beam c for the Station to preserve
latitudinal variations.  Cruises TT007 and TT008 used different
transmissometers, so data from the two cruises were tied together
with the assumption that the minimum beam c at the equator did
not change between the two cruises.  The difference between the
minimum beam c at the Equatorial stations was calculated for
TT007 and TT008 and all TT007 profiles were adjusted by that
amount.  Thus the filtration calibration made during TT008 could
be reasonably applied to TT007 data.  Values at 12 deg. N during
TT007 had incorrect air calibrations, so the profiles were
adjusted to have the same minimum in the upper 400 m as stations
at 9 deg. N (this required a decrease in beam c of only 
0.015 m^-1).

     Data from cruises TT011 and TT012 were processed and tied
together in the same way.  The spring and fall cruises are also
comparable because of the filtered water calibration.  The
minimum beam c in the upper 400 m at the equator for TT008 was
0.3828 m^-1 and 0.3830 m^-1 for TT012.