Notes on Sensors Deployed on M2 Mooring for EqPac Test

When you run the 'argos' program to decode the sensor data, you may
notice some anomolies.  This file tries to make sense of the data for you.
Following are the types of anomolies and other sensor information

All-zero data at startup
For the first 24 hours after startup, OASIS has no valid data to send,
so it will send messages of all zeros (other than packet type and checksum).
This may translate into odd looking data.  Ignore it.  This is one reason
that the program prints the raw data as well as the decoded data.


pCO2 data
This should come out directly as the difference between the concentration
of CO2 between the surface water and air, in ppm.  We seem to have some
plumbing problems right now, though, and the data from this sensor is always
either zero or one ppm.  It will be a more significant number when we fix
this problem.


NO3 data
The program currently yields just the A/D count for nitrate.  There is
a formula for converting to micromolar NO3, but it's dependant on some
scale factors, which in turn depend on the values returned from the
standard and blank insertion times.  Right now, the scientist (Hans Jannisch)
prefers to get the raw data and work the numbers himself.  Later we'll
probably be able to convert directly to micromolar readings.  This will
likely involve another calibration file for this instrument.

Note that this is the reason that the standard and blank values are encoded
into the data stream.  The convention is that the regular data stream is
rounded down to the next lowest even value.  Readings that result from
standard or blank insertion are odd numbers.  This occurs once every four
days, and results in a standard value followed by a blank value.  Consequently,
the precision of both regular data and standard/blank values is 2 A/D counts.

The temperature reported for the NO3 instruments has a precision of 0.05 C.


AC-9 Data
The AC-9 measures absorption (a beam) and attenuation (c beam) of light
at 9 different wavelengths.  The resulting 18 values are reported via
ARGOS once per day, for the sample at noon local time.  In addition,
the chlorophyll channel (a676, or absorption at 676 nm) is reported for
8 samples/day, at the same times as NO3 and CO2 (0000, 0300, 0600, etc).

The program currently reports the data as Vsignal/Vreference, which has
little scientific meaning.  There is a formula to convert to a scientifically
meaningful number (inverse meters?), which looks something like
inverseMeters = K1 - ln(Vs/Vr)/K2.  Currently, Francisco is doing this
conversion.  When I get the right formula and constants, I'll change the
software to yield this number.  It may also involve another calibration
file.


Spectroradiometer data

The spectroradiometers show incident, or downwelling light (Edxxx) and
reflected, or upwelling light (Luxxx) at 7 different wavelengths.  One
of the Ed channels is for PAR, or photosynthetically active radiation.
One of the upwelling channels is NatF, natural fluorescence.  The units
for Ed are microwatts per cm2-nm, and for Lu are microwatts per
cm2-nm-sterradian.

The calibration file has offsets for "dark voltages".  This is the value
the instrument reads when dark, and is subtracted from the measurement.
However, it's sensitive to temperature.  As a consequence, it's possible
for the decoded value to be negative, and is merely an indication that
the dark value is too high.

The spectroradiometer also has a fluorometer attached.  This channel
reads out as a raw voltage.


MCP (10 and 30 meter sensors)

The MCP's are single wavelength spectroradiometers (not sure of wavelength,
somewhere around chlorophyll active region?).  They were wired incorrectly
in this deployment.  Ignore this data.