VERSION 2
	
AUTHOR:
	Stuart Beaton
	National Center for Atmospheric Research
	P.O. Box 3000
	Boulder, CO 80307-3000
	303-479-1038
	beaton@ucar.edu

DATA SET OVERVIEW:
	National Center for Atmospheric Research (NCAR) Lyman Alpha Hygrometer
	POST variable MRLA3, recorded and output at 100 samples/second
	Data covers research flights RF01 (2008-07-16) to RF06 (2008-07-28) 
	and RF08 (2008-08-01) to RF17 (2008-08-15).
	The instrument was located in the nose section of the CIRPAS Twin Otter.
	There was approximately 25 cm of 1/4" OD tubing between the inlet and the 
	instrument.

INSTRUMENT DESCRIPTION:
	The NCAR prototype lyman alpha hygrometer measures the water vapor
number density by absorption at the hydrogen lyman alpha line  (121 nm). The 
lyman alpha source is a Resonance Inc. RF powered lyman-alpha lamp. The detector
is a Hamamatsu R6800U-26 phototube. The absorption signal is affected by lamp 
and amplifier drift as well as atmospheric absorption due to oxygen and carbon 
dioxide. Since the instrument has no internal reference its output signal is 
referenced to the chilled mirror dew point hygrometer, constrained by lab 
calibrations as much as possible. 
	The absorption pathlength is chosen to maximize sensitivity at the 
anticipated scientifically important H2O mixing ratio. In this experiment, the 
pathlength was approximately 2.5 mm, appropriate for mixing ratios in the range
of 5-15 g/kg.

Specifications:
	Instrument output: Analog voltage for absorption, cell pressure, and 
		cell temperature.
	Instrument bandwidth: 100 Hz
	Data sample rate: 100 samples/second recorded by Twin Otter
	Accuracy: Mixing ratio accuracy is expected to be approximately 0.2 g/kg
		for mixing ratio near 10 g/kg.
	Precision: RMS noise is typically 0.05 g/kg for mixing ratio near 10 g/kg.

DATA COLLECTION AND PROCESSING:
	Three analog voltage were recorded by the Twin Otter data system at 100
samples per second with a resolution of 1 mV on 10 V full scale. The voltages 
correspond to the lyman alpha detector signal and the sample cell pressure and 
temperature. 

	Since the hygrometer does not contain an internal reference a 
multi-parameter least-squares fit to the chilled-mirror calculated mixing ratio 
is determined for each flight. The equation for the fit includes terms for H2O 
absorption, atmospheric absorption (largely O2 and CO2), amplifier gain and 
offset, and drift due to temperature and window contamination.
	
	The water vapor absorption ceofficient was determined by post-project 
calibration over the range of 0.08 to 15 g/kg at ambient pressure and 
temperature, and was consistent before and after cleaning the optical windows. 
Its value was fixed in the fits.

DATA FORMAT:
	The data file name e.g. RF01_MRLA3_100Hz.out indicates the research flight number,
the variable (MRLA3) and the data rate (100Hz). 

	Files consist of two columns of tab-delimited ASCII values. End of line character
is the unix LF. The first column is the mission time in seconds with 0.01 
second resolution as recorded by the Twin Otter data system and reported 
in the cabin data file. The second column is the derived mixing ratio in g/kg.
Note that the start time given in the file header is the UTC time of the first
record in the file.

DATA REMARKS:
	Discounting periods when the mixing ratio is below 1 g/kg, or the 
chilled-mirror dewpoint instrument was overshooting, or the lyman-alpha 
hygrometer was afflicted by offset shifts (see below), the calculated lyman 
alpha mixing ratio is usually within 0.2 g/kg of the chilled mirror mixing 
ratio. There are some periods, particularly at the beginning of flights when the
instrument is still warming up, that the error is larger. Also the error 
increases when the mixing ratio goes below 1 g/kg, at which point the 
lyman-alpha mixing ratio may even go negative. These erroneous values have not 
been removed from the data set since the variations in mixing ratio may still 
have scientific value even though the absolute value is in error.
	
	The RMS noise of the calculated mixing ratio is typically around 0.05 
g/kg, with peak-peak value of about 0.2 g/kg. During TO10, and to a lesser extent
during RF04 and flights after RF10, the output voltage showed offset jumps which 
translate to as much as 2 g/kg. The cause of these jumps has not been determined 
and they cannot be reproduced.

	Instrument response speed: This instrument was later deployed in the 
VOCALS project aboard the NCAR C-130 where it showed frequency response to 
near 100 Hz based on power spectral density plots. Similar plots for the POST 
flights could only confirm response to about 10 Hz, but they appear to be 
limited by the 1 mV voltage resolution of the Twin Otter data acquisition 
system. Examination of the data reveals regions where the mixing ratio changes 
by 0.5 g/kg in 0.04 seconds (4 samples) with little or no equilibration time 
detectable at the new value so the response speed in the Twin Otter installation
is probably faster than 10 ms. During some of the sawtooth manuevers in and out
of clouds there appears to be some asymmetry present. This may indicate some absorption
and desorption between the inlet tubing walls and the sample airstream when the mixing ratio
changes by a large amount.

There may be a delay in response due to the length of tubing (~ 25 cm) between the inlet 
and the sample cell. This delay has not been estimated at this time. 

CHANGES FROM VERSION 1
	The version 1 mixing ratio calculations were in error by as much as 20% since
they were fit to incorrect mixing ratios calculated from the chilled-mirror 
measured dew point.