The Harvard total water instrument

The design of the newly developed total water instrument is based on the same principles as the water vapor instrument, and is intended to fly in conjunction with it. Conceptually, the total water instrument can be thought of as containing four subsystems:

1. An inlet through which liquid and/or solid water particles can be brought into an instrument duct without perturbing the ambient particle density.
2. A heater that efficiently evaporates the liquid/solid water before it reaches the detection axis.
3. Ducting through which the air flows to the detection axis without perturbing the water vapor mixing ratio.
4. A water vapor detection axis that accurately and precisely measures the total water content of the ambient air.

We will discuss each of these aspects of the instrument separately.

There are two basic design considerations for designing the total water inlet: (1) the position of the inlet relative to the aircraft skin and (2) the inlet shape. Theoretical and experimental studies of air flow and particle trajectories around aircraft fuselages [King, 1984; King et al., 1984] indicate that for quantitatively sampling particles up to 50 µm diameter, the inlet must be separated from the skin of the aircraft by a distance equal to the aircraft radius where the inlet is attached. The formula is derived for calculating the sampling efficiency of a modified Stokes # (S) where

and where D is particle diameter, v is flow velocity, r is fuselage radius, and n is viscosity of the air. We show in Figure 1a the particle enhancement factor calculated for a nose radius of 17.5" and d/b, the normalized distance from the skin, as a function of modified Stokes #. With d/b = 1, or an 18" distance from the aircraft skin, particles up to 50 µm diameter (S = 12) can be measured with an enhancement of 10% or less. Figure 1b illustrates the enhancement as a function of d/b.

The requirement of isokinetic flow means that the flow velocity through the inlet must be identical to the free stream velocity of the WB-57, or 200 m/sec. With a nozzle id of 1 cm, the velocity through the 2" square cross section detection axis will be about 6 m/sec when the flow is isokinetic. A Roots pump is used to both regulate this velocity as well as provide the ability to scan flow velocity for diagnostic purposes.

The NACA 1-55-100 inlet shape is used for the 1-cm id inlet [Soderman et al., 1991]. It has been proven by wind tunnel tests that to avoid flow separation or measurement enhancements it is necessary to attack the streamlines within an angle of a degree or two. As shown in the wind tunnel test of Murphy and Schein [1998], a shroud around the inlet allows flow at the inlet entrance to be significantly less sensitive to this angle. For our inlet, the shroud is 3" id, and is scaled directly from the inlet used by Murphy and Schein.

The heater design is predicated on the need to completely vaporize 50 µm diameter particles within the transit time from the heater to the detection axis, about 60 milliseconds. Based on typical ambient temperatures and pressures in the upper troposphere and lower stratosphere, a 600 watt heater is needed to raise the air temperature to about 30°C. The heater is constructed of strips of nichrome ribbon, mounted axially in the flow tube, providing extremely efficient heat transfer to the air stream. A thermistor just downstream of the heater monitors the air temperature and provides a signal used in a feedback loop by a pulse-modulated heater controller that maintains the air temperature stable to better than a degree.

The detection axis is functionally identical to that in the water vapor instrument that has been previously described in detail [Weinstock et al., 1994]. A simple diagram is shown in Figure 2. Briefly, 121.6 nm (L_a) radiation from an RF discharge lamp photodissociates water vapor in a 2-inch duct. A fraction of the resulting OH fragments are formed in their first excited electronic state (A²S⁺), and the OH fluorescence at ~ 315 nm is collected at right angles to the L_a beam through a filter and detected with a photomultiplier tube (PMT). Because the fluorescence is strongly quenched by collisions with O₂ and N₂ at a rate proportional to the air density, at altitudes of the upper troposphere and lower stratosphere the observed detector signal is proportional to the water vapor volume mixing ratio. Lamp scatter near 315 nm is measured by using a quartz window to periodically block the L_a beam. Changes in lamp intensity monitored with a vacuum photodiode opposite the lamp are used to normalize the fluorescence signal.

The pre- and post-flight laboratory calibrations are carried out simultaneously for the water vapor and total water instruments by flowing known amounts of water vapor in the air through the detection axes [Weinstock et al., 1994]. Saturated air-water mixtures are prepared by passing air through a two-stage bubbler apparatus. These are diluted with additional dry air before being sent to the detection axes, which are integrated into the laboratory flow system.

Figure 3 shows the total water instrument in flight configuration mounted in a WB57 pallet. The initial engineering flights for the total water instrument took place in May, 2001 from the NASA Johnson Space Flight Center at Ellington Field, Houston, TX. In August 2001 the instrument took part in CARRETT, a campaign based out of San Jose, Costa Rica.

The instrument performed very successfully, with the heater and pump subsystems and detection axis working well. Intercomparisons between water vapor measured by the two instruments typically showed agreement to better than 10%. During two of the test flights the WB57 flew through extensive cirrus layers. In figure 4 we illustrate the measurement of cirrus cloud ice content sampled near Houston, TX during the May 25 flight.

Because, as we have stated before, ice water content that can still have significant radiative impact could be as low as 5e¹² mol/cc, we illustrate in Figure 5 water vapor concentration measured by the total water instrument in the lower stratosphere over Houston, during the same May 25, 2001 flight.

We include in the figure water vapor concentration simultaneously measured by the water vapor instrument. We emphasize here that the concentrations are used here to illustrate that the sensitivity and precision of the two instruments is easily sufficient to measure by subtraction the ice water content of a thin cirrus cloud. The concentrations measured by the two instruments are not strictly comparable because the densities in the two instruments are different. The point is that the nominal precision of each instrument for a 10-second data point is 1 e¹¹ mol/cc. From model calculations this is roughly equivalent to 0.04 W/m² of radiative heating.

In August, 2001, the total water and water vapor instruments flew as apart of a focused mini-mission from San Jose, Costa Rica. In figure 6 we illustrate the agreement of the water and total water instruments in clear air near the tropopause.

Total Water Instrument

References