• Alamaro, M., , Emanuel K. , , Colton J. , , McGillis W. , , and Edson J. , 2002: Experimental investigation of air–sea transfer of momentum and enthalpy at high wind speed. Extended Abstract, 25th Conf. on Hurricanes and Tropical Meteorology, San Diego, CA, Amer. Meteor. Soc., CD-ROM, 17C.6.

  • Brown, E. N., , Friehe C. A. , , and Lenschow D. H. , 1983: The use of pressure fluctuations on the nose of an aircraft for measuring air motion. J. Climate Appl. Meteor., 22 , 171180.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brunke, M. A., , Fairall C. W. , , Zeng X. , , Eymard L. , , and Curry J. A. , 2003: Which bulk aerodynamic algorithms are least problematic in computing ocean surface turbulent fluxes? J. Climate, 16 , 619635.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crawford, T. L., , and Dobosy R. J. , 1992: A sensitive fast-response probe to measure turbulence and heat flux from any airplane. Bound.-Layer Meteor., 59 , 257278.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dobosy, R. J., , Crawford T. L. , , MacPherson J. I. , , Desjardins R. L. , , Kelly R. D. , , Oncley S. P. , , and Lenschow D. H. , 1997: Intercomparison among the four flux aircraft at BOREAS in 1994. J. Geophys. Res., 102 , D24. 2910129111.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eckman, R. M., , Crawford T. L. , , Dumas E. J. , , and Birdwell K. R. , 1999: Airborne meteorological measurements collected during the Model Validation Program (MVP) field experiments at Cape Canaveral, Florida. NOAA Tech. Memo. ARL/ATDD-233, NOAA Air Resources Laboratory, Silver Spring, MD, 54 pp.

  • Eckman, R. M., , Dobosy R. J. , , Strong T. , , and Auble D. L. , 2004: Development and initial deployment of an omnidirectional pressure-sphere anemometer for observing winds and turbulence in tropical cyclones. Extended Abstract, 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 368–369.

  • Emanuel, K. A., 1986: An air–sea interaction theory for tropical cyclones. Part I: Steady-state maintenance. J. Atmos. Sci., 43 , 585604.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 1995: Sensitivity of tropical cyclones to surface exchange coefficients and a revised steady-state model incorporating eye dynamics. J. Atmos. Sci., 52 , 39693976.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 2003: A similarity hypothesis for air–sea exchange at extreme wind speeds. J. Atmos. Sci., 60 , 14201428.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fairall, C. W., , Kepert J. D. , , and Holland G. J. , 1994: The effect of sea spray on surface energy transports over the ocean. Global Atmos. Ocean Syst., 2 , 121142.

    • Search Google Scholar
    • Export Citation
  • French, J. R., , Johnson R. , , Beard S. , , and Crawford T. , 2004: Modification of an airborne gust probe for hurricane boundary layer research. Extended Abstract, 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 332–333.

  • Green, S. I., , Ostafichuk P. , , and Rogak S. N. , 2001: Measurement of three-dimensional unsteady flows using an inexpensive multiple disk probe. J. Atmos. Oceanic Technol., 18 , 883891.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hacker, J. M., , and Crawford T. L. , 1999: The BAT probe: The ultimate tool to measure turbulence from any kind of aircraft (or sailplane). J. Tech. Soaring, 23 , 4346.

    • Search Google Scholar
    • Export Citation
  • Katsaros, K. B., and Coauthors, 1994: Measurements of humidity and temperature in the marine environment during the HEXOS main experiment. J. Atmos. Oceanic Technol., 11 , 964981.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Krayer, W. R., , and Marshall R. D. , 1992: Gust factors applied to hurricane winds. Bull. Amer. Meteor. Soc., 73 , 613617.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunkel, K. E., , and Bruce C. W. , 1983: A sensitive fast-response manometric wind sensor. J. Climate Appl. Meteor., 22 , 19421947.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Large, W. G., , and Pond S. , 1981: Open ocean momentum flux measurements in moderate to strong winds. J. Phys. Oceanogr., 11 , 324336.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lenschow, D. H., 1986: Aircraft measurements in the boundary layer. Probing the Atmospheric Boundary Layer, D. H. Lenschow, Ed., Amer. Meteor. Soc., 39–55.

    • Search Google Scholar
    • Export Citation
  • Miller, B. I., 1964: A study of the filling of Hurricane Donna (1960) over land. Mon. Wea. Rev., 92 , 389406.

  • Oost, W. A., , Worrell E. W. H. , , Schaap J. W. , , van Oort C. , , and Kraan C. , 1991: An improved version of the pressure anemometer. J. Atmos. Oceanic Technol., 8 , 575584.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ooyama, K., 1969: Numerical simulation of the life cycle of tropical cyclones. J. Atmos. Sci., 26 , 340.

  • Paulsen, B. M., , and Schroeder J. L. , 2005: An examination of tropical and extratropical gust factors and the associated wind speed histograms. J. Appl. Meteor., 44 , 270280.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peters, G., , and Fischer B. , 1998: One-year operational measurements with a sonic anemometer-thermometer and a Doppler sodar. J. Atmos. Oceanic Technol., 15 , 1828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Powell, M. D., , Houston S. H. , , and Reinhold T. A. , 1996: Hurricane Andrew’s landfall in South Florida: Standardizing measurements for documentation of surface wind fields. Wea. Forecasting, 11 , 304328.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Powell, M. D., , Vickery P. J. , , and Reinhold T. A. , 2003: Reduced drag coefficient for high wind speeds in tropical cyclones. Nature, 422 , 279283.

  • Rediniotis, O. K., , and Kisner R. E. , 1998: Development of a nearly omnidirectional velocity measurement pressure probe. AIAA J., 36 , 18541860.

  • Rosenthal, S. L., 1971: The response of a tropical cyclone model to variations in boundary-layer parameters, initial conditions, lateral boundary conditions, and domain size. Mon. Wea. Rev., 99 , 767777.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Smith, S. D., 1988: Coefficients for sea surface wind stress, heat flux, and wind profiles and a function of wind speed and temperature. J. Geophys. Res., 93 , 1546715472.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thurtell, G. W., , Tanner C. B. , , and Wesely M. , 1970: Three-dimensional pressure-sphere anemometer system. J. Appl. Meteor., 9 , 379385.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tjernström, M., , and Friehe C. A. , 1991: Analysis of a radome air-motion system on a twin-jet aircraft for boundary-layer research. J. Atmos. Oceanic Technol., 8 , 1940.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tuleya, R. E., 1994: Tropical storm development and decay: Sensitivity to surface boundary conditions. Mon. Wea. Rev., 122 , 291304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Vickery, P. J., , Skerlj P. F. , , Steckley A. C. , , and Twisdale L. A. , 2000a: Hurricane wind field model for use in hurricane simulations. J. Struct. Eng., 126 , 12031221.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Vickery, P. J., , Skerlj P. F. , , and Twisdale L. A. , 2000b: Simulation of hurricane risk in the U.S. using empirical track model. J. Struct. Eng., 126 , 12221237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Y., , Kepert J. D. , , and Holland G. J. , 2001: The effect of sea spray evaporation on tropical cyclone boundary layer structure and intensity. Mon. Wea. Rev., 129 , 24812500.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wesely, M., , Thurtell G. W. , , and Tanner C. B. , 1970: Eddy correlation measurements of sensible heat flux near the earth’s surface. J. Appl. Meteor., 9 , 4550.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Conceptual diagram showing the general operation of the ET probe.

  • View in gallery

    Photograph of the ET probe on a laboratory stand. A pen has been placed next to the stand as a scale reference.

  • View in gallery

    (a) The interior components of one probe hemisphere. The A/D card is a data acquisition module that digitizes the analog sensor outputs. (b) A close-up of the circuit board, showing the location of the absolute and differential pressure sensors.

  • View in gallery

    (a) The pressure port geometry used to compute the wind vector. Port 0 is closest to the flow stagnation point, and ports 1–4 form a cruciform pattern about the center port. (b) The angle of attack α in the vertical xz plane. (c) The angle of sideslip β in the xy plane.

  • View in gallery

    The 3-m tripod used for deploying the ET probes into hurricanes. The computer and batteries used to run the probe are located in an enclosure at the base of the tripod.

  • View in gallery

    Scatter diagram of sonic and ET probe airspeeds during a 46-min road test in dry conditions starting at 1932 UTC 26 May 2004.

  • View in gallery

    Velocity spectra computed from a 500-s subset of road test data beginning 1940 UTC 26 May 2004. The sonic anemometer spectra (gray) have been displaced downward one decade to provide separation from the ET spectra (black). (a)–(c) The u, υ, and w spectra. On the vertical axes, the spectral values are multiplied by the frequency.

  • View in gallery

    Composite velocity statistics for the sonic anemometer and ET probe based on eight road tests conducted between 8 Apr and 4 Aug 2004. (a) The mean airspeed U; (b)–(d) the velocity standard deviations σu, συ, and σw. In (d), the circles are based on the original 50-Hz ET data, whereas the triangles are based on ET data after block averaging to 10 Hz.

  • View in gallery

    Airspeeds observed during a road test in the rain on 10 Jun 2004. The ET speeds (gray) do not show the large number of rain spikes observed with the sonic anemometer (black). The periods near 1540 and 1600 UTC when the speed drops to zero are times when the vehicle was either turning around or stopped at an intersection.

  • View in gallery

    Spectrum of ET probe w velocity component for a 400-s period starting 1527 UTC 10 Jun 2004. The vehicle was driving through rain at the time.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 17 17 9
PDF Downloads 9 9 3

A Pressure-Sphere Anemometer for Measuring Turbulence and Fluxes in Hurricanes

View More View Less
  • 1 Field Research Division, NOAA/Air Resources Laboratory, Idaho Falls, Idaho
  • 2 Atmospheric Turbulence and Diffusion Division, NOAA/Air Resources Laboratory, Oak Ridge, Tennessee
  • 3 Field Research Division, NOAA/Air Resources Laboratory, Idaho Falls, Idaho
© Get Permissions
Full access

Abstract

Turbulence and air-surface exchange are important factors throughout the life cycle of a tropical cyclone. Conventional turbulence instruments are not designed to function in the extreme environment encountered in such storms. A new instrument called the Extreme Turbulence (ET) probe has been developed specifically for measuring turbulence on a fixed tower in hurricane conditions. Although the probe is designed for surface deployment, it is based on the same pressure-sphere technology used for aircraft gust probes.

The ET probe is designed around a 43-cm-diameter sphere with 30 pressure ports distributed over its surface. A major obstacle during development was finding a method to prevent water from fouling the pressure ports. Two approaches were investigated: a passive approach using gravity drainage and an active approach using an air pump to flush water from the ports. The probes were tested in both dry and wet conditions by mounting them on a vehicle side by side with more conventional instruments. In dry conditions, test data from the ET probes were in good agreement with the conventional instruments. In rain, probes using the passive rain defense performed about as well as in dry conditions, with the exception of some water intrusion into the temperature sensors. The active rain defense has received only limited attention so far, mainly because of the success and simplicity of the passive defense.

* Deceased

Corresponding author address: Dr. Richard M. Eckman, NOAA/ARL, Field Research Division, 1750 Foote Dr., Idaho Falls, ID 83402. Email: Richard.Eckman@noaa.gov

Abstract

Turbulence and air-surface exchange are important factors throughout the life cycle of a tropical cyclone. Conventional turbulence instruments are not designed to function in the extreme environment encountered in such storms. A new instrument called the Extreme Turbulence (ET) probe has been developed specifically for measuring turbulence on a fixed tower in hurricane conditions. Although the probe is designed for surface deployment, it is based on the same pressure-sphere technology used for aircraft gust probes.

The ET probe is designed around a 43-cm-diameter sphere with 30 pressure ports distributed over its surface. A major obstacle during development was finding a method to prevent water from fouling the pressure ports. Two approaches were investigated: a passive approach using gravity drainage and an active approach using an air pump to flush water from the ports. The probes were tested in both dry and wet conditions by mounting them on a vehicle side by side with more conventional instruments. In dry conditions, test data from the ET probes were in good agreement with the conventional instruments. In rain, probes using the passive rain defense performed about as well as in dry conditions, with the exception of some water intrusion into the temperature sensors. The active rain defense has received only limited attention so far, mainly because of the success and simplicity of the passive defense.

* Deceased

Corresponding author address: Dr. Richard M. Eckman, NOAA/ARL, Field Research Division, 1750 Foote Dr., Idaho Falls, ID 83402. Email: Richard.Eckman@noaa.gov

1. Introduction

Turbulent air-surface exchange is a controlling factor throughout the life cycle of tropical cyclones. While these storms are over the ocean, their intensity depends strongly on the turbulent transfer of momentum and enthalpy between the ocean and the atmosphere (Ooyama 1969; Rosenthal 1971; Emanuel 1986, 1995). The rapid decay of the cyclones once they make landfall is also determined by air-surface exchange, particularly a steep reduction in surface enthalpy flux (Miller 1964; Tuleya 1994).

Atmospheric turbulence also plays an important role in property losses associated with tropical cyclones. Structural engineering design is often based on gust factors (Krayer and Marshall 1992; Powell et al. 1996; Paulsen and Schroeder 2005), which are ratios of the peak wind speed based on a short averaging time (e.g., 2 s) to the mean wind speed based on a longer averaging time (e.g., 10 min). These gust factors are directly related to the velocity power spectra. Gust factor algorithms are commonly used in hurricane models designed for risk assessment (Vickery et al. 2000a, b).

Given the importance of boundary layer turbulence to tropical cyclones, accurate estimates of the turbulence and surface fluxes are critical to the success of any cyclone modeling effort. Direct observations of turbulence fluctuations have been rare under the extreme conditions encountered in hurricanes, particularly observations of vertical fluxes. Hence, the commonly used parameterizations for air–sea fluxes in hurricanes are based on extrapolations from observations at lower wind speeds (Large and Pond 1981; Smith 1988; Brunke et al. 2003). Accumulating evidence from theoretical studies (Emanuel 1995), hurricane wind speed profiles (Powell et al. 2003), and laboratory wind-wave tanks (Alamaro et al. 2002) indicates that these extrapolations are inaccurate in hurricane-force winds. The momentum drag coefficient CD, for example, appears to level off at high wind speeds rather than steadily increasing with wind speed as predicted by standard parameterizations. This change in behavior of CD may be related to the generation of sea spray in high winds (Fairall et al. 1994; Wang et al. 2001; Emanuel 2003).

Significant improvements in the parameterization of turbulence and fluxes would be possible if direct observations of turbulence could be made in hurricane conditions. The National Oceanic and Atmospheric Administration (NOAA) Air Resources Laboratory (ARL) has been developing a turbulence probe capable of withstanding the extreme conditions encountered in hurricanes. This probe, called the Extreme Turbulence (ET) probe, is designed for tower deployments at the surface but is based on the same technology used in the design of aircraft gust probes. The ET probe operates well in high winds and is also designed to mitigate the effects of heavy rain and spray. This paper describes the initial development and testing of the probe.

2. Development and design

Conventional turbulence instruments are not designed to function under hurricane conditions. Modern 3D sonic anemometers often contain built-in algorithms to detect and exclude spurious signals produced by particles in the sound path, but these function best in light precipitation (Peters and Fischer 1998). When the sonic transducers become coated with water or are struck by large rain drops, the output usually is severely contaminated (Katsaros et al. 1994). A hurricane intercept team at Texas Tech University has been deploying 3D sonic anemometers as part of their instrument package, but the output has generally been too contaminated to be of use (J. L. Schroeder 2004, personal communication).

The ET probe concept was motivated by several years of development work on aircraft gust probes. ARL has been steadily improving a nine-hole aircraft gust probe called the Best Aircraft Turbulence (BAT) probe (Crawford and Dobosy 1992; Hacker and Crawford 1999; French et al. 2004). This probe design was originally used on a light single-engine aircraft (Dobosy et al. 1997; Eckman et al. 1999) but has since been adapted to a variety of aircraft, including the four-engine NOAA WP-3D Orion research aircraft (French et al. 2004). Since the aircraft probes are routinely operated at airspeeds exceeding 50 m s−1, they are a logical starting point for developing a surface-based hurricane turbulence probe.

a. Operating principles

The ET probe is based on pressure-sphere anemometry, which has been used for many years in the design of aircraft gust probes (Brown et al. 1983; Lenschow 1986; Crawford and Dobosy 1992). In high Reynolds number flow past a sphere, the pressure distribution near the stagnation point, where the oncoming flow is perpendicular to the sphere’s surface, closely follows the potential flow solution
i1520-0426-24-6-994-e1
Here, q is the dynamic pressure, ps is the static pressure, s is a unit normal vector at the stagnation point on the sphere, and p(n) is the total pressure at the point on the sphere with unit normal vector n. This equation is applicable within about 60° of the stagnation point. All pressure-sphere anemometers measure the pressure at several different locations on a spherical surface and then use Eq. (1) to derive the relative flow velocity (Brown et al. 1983; Lenschow 1986).

The details of the pressure and velocity computations depend on the probe design and intended use. Aircraft probes can assume that s falls within a narrow cone pointing ahead of the aircraft and therefore are usually hemispheres with less than 10 pressure measurement ports (Brown et al. 1983; Lenschow 1986; Crawford and Dobosy 1992). A similar hemispherical design was used by Thurtell et al. (1970) and Wesely et al. (1970) as a surface-based instrument; it was rotated into the mean wind direction either manually or with a motor assembly. To circumvent the directional limitations of hemispherical probes, Rediniotis and Kisner (1998) developed a nearly omnidirectional pressure sphere intended for wind tunnel applications. Their pressure ports are distributed almost evenly over the surface of a sphere, so the probe can measure velocities from nearly any direction except those where the unit normal s is within about 10° of the mount that holds the probe in place.

The ET probe was developed independently of the Rediniotis and Kisner probe, but it shares many similarities. Figure 1 is a conceptual diagram of the probe’s operation. The instrument is designed around a rigid sphere with no moving parts. Pressure ports are distributed over the sphere’s surface. Hence, the raw outputs are the pressure distribution together with a temperature measurement T. Using the pressures, the system must first locate the pressure port closest to the velocity stagnation point. The pressure measurements near the stagnation point are then used to compute the 3D velocity vector (u, υ, w).

Although pressure spheres like the ET probe are the most common type of manometric wind sensor, alternate designs have been successfully developed. Kunkel and Bruce (1983) describe a wind sensor that converts the dynamic pressure into an acoustic signal, which is then measured with a microphone. Oost et al. (1991) developed a pressure anemometer based on pairs of metal tubes set at angles to one another. The wind sensor developed by Green et al. (2001) uses three orthogonal disks with pressure sensors on each side of the disks.

b. Hardware

In high winds near the surface, the magnitude of the vertical velocity component is small compared to the horizontal components, so the stagnation point on the ET probe will fall within a limited band around the great circle of the sphere that is in the horizontal plane. The pressure ports can therefore be concentrated near this great circle. Adding additional ports near the top and bottom of the sphere would actually create practical problems, in that the top ports would quickly fill with water in hurricane conditions and become unusable.

Given these considerations, the final exterior design of the ET probe is shown in Fig. 2. The sphere is made of fiberglass and is 43 cm in diameter. Three rows of pressure ports run horizontally around the sphere. The center row is on the great circle of the sphere that lies in the horizontal plane, and the other two are on small circles 18° above and below the center row (in spherical coordinates). Within each row, the ports are situated 36° apart. A total of 30 ports is therefore used. At the top of the sphere is a “mushroom” housing that contains the temperature sensors.

For ease in construction, the ET probe splits along a vertical seam into hemispheres that use the same internal hardware. Figure 3a shows the internal construction within one hemisphere. Each external pressure port is attached to plastic tubes that run to a series of pressure sensors on a circuit board. The reasons for the small and large tubes are described in the following section on rain defense. One of the hemispheres on each probe also contains the wires that run from the temperature sensors in the mushroom housing and connect to the circuit board.

The circuit boards (Fig. 3b) were specially designed at ARL for use in the ET probes. Their primary function is to amplify the output voltages from the sensors, which are on the order of 10 mV, to the ±10-V range expected by the data acquisition module located beneath the circuit board. Along the outer edge of the circuit board are 10 differential pressure sensors. Five of these sensors are attached to individual top and bottom ports on the sphere (Fig. 2) and therefore measure vertical pressure differences. The other five are attached to adjacent ports on the great circle and thus measure horizontal pressure differences. At least one absolute pressure measurement is also required for the wind computations, so the circuit board contains three absolute sensors connected to alternate ports on the great circle. All the analog sensor outputs are gathered into a 26-pin socket that is attached to a data acquisition module by a ribbon cable. The entire circuit board is powered by a 12-V dc source, typically a bank of batteries.

The differential pressure sensors must be chosen so that they will not overrange in hurricane winds. Currently, the probes use Honeywell model DCXL10DN differential sensors with an operating pressure range of ±25 hPa. The dynamic pressure q associated with an airspeed U is given by (Brown et al. 1983; Lenschow 1986)
i1520-0426-24-6-994-e2
where ρ is the density. Thus, the DCXL10DN sensors are capable of directly measuring q in speeds up to about 65 m s−1 at sea level densities. However, the port arrangement on the ET probe means that no differential sensor ever observes the full value of q. From Eq. (1), the maximum differential pressure at 36° port separation is about 0.78q. The current sensors should therefore allow the probe to function in airspeeds up to about 73 m s−1, which cover all but the most intense hurricane winds. The absolute pressure sensors are Honeywell model XCX15ANH, with a full-scale pressure of 1034 hPa.

Two temperature sensors are contained within the mushroom housing on top of the sphere. The first is a YSI Temperature microbead thermistor model E33A401C. It has a diameter of 0.127 mm and a time constant in still air of 0.12 s. This sensor is capable of providing high-frequency temperature fluctuations suitable for computing heat fluxes, but its output is linear over only a relatively narrow temperature range. The circuit board contains an adjustable bridge/amplifier circuit that allows the sensor’s range to be adjusted for expected operating temperatures. The second temperature sensor is a YSI 44212 thermilinear network, which is a combination of thermistors and resistors designed to provide accurate temperature measurements over a wide range. This sensor has a linear range of −50 to +50°C and an accuracy of ±0.1°C but a much larger time constant of 10 s in still air. The 44212 is primarily used as a temperature reference to accurately calibrate the fast-response sensor.

Each hemisphere contains a Data Translation model DT9800 data acquisition module positioned below the circuit board (Fig. 3a). It can digitize up to 16 single-ended analog channels at 16-bit resolution. The module receives analog input from the circuit board via a 26-pin ribbon cable and outputs the digitized data through a USB cable connected to a computer.

c. Rain and spray defense

One of the major obstacles in the ET probe development was designing a method to keep rainwater and spray from fouling the pressure ports. The aircraft probes developed by ARL (Crawford and Dobosy 1992; Hacker and Crawford 1999) were not designed to function in rain and therefore do not include any defense against water intrusion. The special BAT probe developed for use on the NOAA WP-3D hurricane hunters (French et al. 2004) does include an air pump to backflush water from the ports. However, this pump is not designed to operate while measurements are being collected. Rather, it is temporarily activated just prior to collecting data in dry sectors of tropical cyclones.

To increase the overall chances for success, two different approaches to the water-fouling problem were pursued during development. One of these is based on simple gravity drainage and is therefore a passive defense. The other uses an air pump to flush the ports and is therefore an active defense.

1) Passive defense

The first ET probe prototypes were built with pressure ports that were the same as on the ARL aircraft probes. These ports were 1 mm in diameter and were connected to plastic tubes with an inside diameter of 1.6 mm. Such “pinhole” ports have certain advantages, in that there is minimal disruption of the spherical surface, and the small tubes take up less space inside the probe. Tests with these first-generation probes demonstrated that the overall design performed well in dry conditions but had serious drawbacks in rain.

When a raindrop strikes a pinhole port, the pressure sensors connected to this port register a large transient spike. This immediately causes problems in the data acquisition, because the rain spikes must be filtered from the pressure signals before the wind vector can be computed. Rain spikes can also cause problems in localizing the flow stagnation point. During each data scan, the acquisition software searches for the great circle port with the highest pressure to localize the position of the stagnation point. A transient rain spike can cause the software to misidentify the location of this point. As a result, the computed stagnation point appears to jump around the sphere in response to the random pattern of raindrop strikes.

Rather than developing complex spike removal algorithms, a far more preferable solution is to avoid the spikes in the first place. The Pitot tubes used on aircraft provided one possible strategy for avoiding rain spikes. A Pitot tube is a differential pressure sensor similar to those used on the ET probe. Aircraft routinely fly through rain, but there is no evidence that they lose their airspeed measurements as a result of water intrusion into the Pitot tubes. The main reasons for this are that the Pitot tubes tend to be fairly large in diameter, and they are sloped upward so that gravity assists in keeping the tubes clear. (Some Pitot tubes are also heated, but this is mainly to avoid ice buildup.) Large radome gust probes used on aircraft have sometimes employed large pressure ports similar in size to Pitot tubes (e.g., Brown et al. 1983).

Some of the ET probes were redesigned to match the rain defense used with Pitot tubes. First, the pressure ports were enlarged to 6.4 mm in diameter. Second, sections of the small-diameter tubes were replaced with larger plastic tubes having the same inside diameter as the ports. This modified design is called the “big-hole” ET probe. Figures 2 and 3 both show the big-hole design. As shown in Fig. 3a, the large tubes slope upward from the ports to the top of the sphere. This helps any water entering a port to drain back out before it can clog the tubes. The large tubes terminate at the sphere’s top, and the original 1.6-mm tubes are used for the remainder of the paths to the pressure sensors. Other than the changes described, the big-hole probes use the same sensors and data acquisition system as the original design.

2) Active defense

An active defense ET probe was actually conceived prior to the passive design, partly because of the success reported by Oost et al. (1991) in repelling rain and sea spray through continuous pneumatic backflushing of their pressure anemometer. One ET probe was modified for an active defense by adding an air pump that flushes air through each of the pressure ports. Early tests had shown that a 2 mL s−1 flow rate was sufficient to flush each port, so the pump was sized to supply this flow rate to all the ports.

During testing the active defense successfully kept water from entering the ports. However, this design still suffers from spikes in the pressure data due to direct impacts of rain drops on the ports. Because of the success of the passive rain defense described above, the focus shifted away from the active method. All of the test results described later in this paper were taken with probes using the passive defense.

d. Data acquisition

As mentioned previously, each probe hemisphere contains 10 differential pressure sensors, half measuring vertical differences and half measuring horizontal differences. In addition, three absolute pressure sensors are present in each hemisphere. One of the hemispheres also collects the data from the two temperature sensors in the mushroom housing. Hence, the DT9800 data acquisition module in one hemisphere receives 15 channels of analog input, whereas the module in the other receives 13 channels. Proper operation of the probe requires that both DT9800 modules synchronize their analog-to-digital conversions. This is accomplished by using the counter/timer subsystem on one module (the master) to pace the digitization on both modules.

The DT9800 modules are configured to digitize the analog inputs at 1 kHz. These digitized data are then transferred via USB cable to a data acquisition program on a nearby computer. As an antialiasing filter, the raw 1-kHz data are block averaged to 50 Hz. The original motivation for transferring 1-kHz data to the computer was the need to account for possible rain spikes in the data. In the early stages of development, the focus was on the active rain defense described in section 2c. Because of the small-diameter pressure ports, data from an active defense probe are significantly affected by rain spikes. The original intention was therefore to develop a despiking algorithm that could be applied to the raw 1-kHz data prior to the block averaging to 50 Hz. Despiking at 1 kHz is not necessary with the passive defense, since the raw data show little evidence of rain spikes.

After the block averaging to 50 Hz, the first task of the data acquisition program is to locate the port along the great circle that is closest to the flow stagnation point. By combining the absolute and differential pressure measurements, it is possible to compute the total pressure at each of the 10 ports along the great circle. For each 50-Hz data scan, the great circle port with the highest total pressure is assumed to be closest to the stagnation point. This port becomes the center point for the wind computations.

In Fig. 4a, the great circle port closest to the stagnation point is identified as port 0. Four other ports labeled as 1–4 in the figure create a cruciform pattern about the center port. The pressure differences observed within this cruciform pattern are the basis for the wind computations. A right-handed coordinate system is also defined in the figure, with the x axis pointing along the outward normal to the sphere at port 0, the y axis tangent to the sphere in the horizontal plane, and the z axis tangent to the sphere in the vertical plane.

If pi is the total pressure at port i = 0, . . . 4, then the ET probe sensors provide three raw differential pressure measurements within the cruciform pattern: p2p0, p0p3, and p4p1. However, for the wind computations it is more convenient to form combinations of these raw values:
i1520-0426-24-6-994-e3
The aircraft probes developed at ARL use a similar triad of pressure variables, with the main difference being in the definition of δpx (Crawford and Dobosy 1992; Eckman et al. 1999).

The oncoming wind vector makes two angles relative to the coordinate system x axis. (Note that port 0 is the closest great circle port to the stagnation point but not necessarily right at the stagnation point.) One angle α is in the vertical xz plane and is defined such that the angle is positive when the z velocity component is positive (Fig. 4b). Because of past linkages to aircraft probes, we still call α the angle of attack for the ET probe. Likewise, the angle β in the horizontal xy plane is still called the angle of sideslip (Fig. 4c), and β is positive when the y component of the wind is positive.

Using the potential flow solution given by Eq. (1), it is straightforward to derive equations linking the differential pressures to α and β:
i1520-0426-24-6-994-e4
i1520-0426-24-6-994-e5
i1520-0426-24-6-994-e6
Here ϕy is the horizontal angle between the ports (36°), ϕz is the vertical angle (18°), and
i1520-0426-24-6-994-e7
Since α and β are the unknowns, we want to invert Eqs. (4)(6) to solve for these variables. This is done by defining the ratios Hy = δpy/δpx and Hz = δpz/δpx and then expanding according to Eqs. (4)(6). The resulting equation for β is quadratic, but one of the solutions applies only for |β| > 45°, so the only solution retained is
i1520-0426-24-6-994-e8
Once tanβ is known, α is given by
i1520-0426-24-6-994-e9
Here α and β describe the orientation of the wind vector, but additional computations are required for the magnitude of the wind. The dynamic pressure q is easily obtained from any of Eqs. (4)(6). An estimate of the static pressure ps is also required for the wind computations. By solving Eq. (1) for port 0, ps is given by
i1520-0426-24-6-994-e10
The total pressure p0 at the center port is therefore required for computing ps and is obtained from the absolute sensors.
Once α, β, q, ps, and the temperature T are known, the wind computation is basically the same as with aircraft probes (Brown et al. 1983; Lenschow 1986; Eckman et al. 1999). The Mach number M is
i1520-0426-24-6-994-e11
where γ is the ratio of the specific heat at constant pressure for moist air to the corresponding specific heat at constant volume. The wind speed U is obtained from
i1520-0426-24-6-994-e12
with Rm being the gas constant for moist air and T ′ being the absolute temperature corrected for adiabatic heating (Lenschow 1986). Equations (11) and (12) are used rather than Eq. (2) mainly because they are slightly more convenient for dealing with the temperature correction. The speed computed from Eq. (12) is equivalent to that from Eq. (2).
The wind vector (u, υ, w) in the coordinate system of Fig. 4 is computed from
i1520-0426-24-6-994-e13
i1520-0426-24-6-994-e14
i1520-0426-24-6-994-e15
Since port 0 can be any of the great circle ports, a final horizontal rotation is required to convert (u, υ, w) to standard earth coordinates (ue, υe, we), with ue being the easterly component, υe being the northerly component, and we = w being the unchanged vertical component. During setup the ET probe is always oriented with a specific great circle port facing east, and the data acquisition software assumes this when rotating to earth coordinates.

e. Deployment tower

Since the ET probes are designed for hurricane deployments, ARL was faced with designing a deployment kit that can withstand hurricane winds and keep the electronics dry. This kit must also be portable and quick to install. Figure 5 shows the final design that was developed. The tower is a 3-m tripod made from aluminum tubing. Each leg of the tripod is secured by a combination of 76-cm screw anchors and heavy ballast currently consisting of cinder blocks. The ballast also helps the tower resist high-frequency buffeting by the wind. Unlike guy wires, the tripod design will continue to provide structural support even if the screw anchors fail due to saturation of the soil or other factors.

Both the data acquisition computer and 12-V batteries are located within a watertight enclosure at the base of the tower. This enclosure is also secured with screw anchors, and the batteries provide considerable ballast. To provide power for several days of continuous operation, four 12-V deep-cycle batteries are normally used with the passive design ET probe. The active design requires more batteries because of the additional power draw from the air pump.

Load testing of the tower design indicated that it is capable of withstanding wind loads exceeding 100 m s−1 when properly deployed. Of course, neither the probe itself nor the tower is designed to withstand the impact of large flying debris in major hurricanes. Care must be exercised in siting the probes to minimize the risk of such encounters with large debris.

3. Calibration and testing

Extensive field tests were performed with the ET probe to determine its overall accuracy and ability to function in rain. The probe’s ability to function properly in high winds was not a major issue of concern, given ARL’s long experience with aircraft gust probes. Moreover, the dynamic pressure fluctuations increase with the square of the oncoming wind speed [Eq. (2)], so the probe’s response actually improves as the speed increases.

a. Sensor calibrations and errors

Accurate calibration of the pressure sensors is critical to the proper operation of the ET probe. Just one badly calibrated differential sensor can potentially cause the data acquisition software to incorrectly identify the location of the wind stagnation point. Before deployments, all the pressure sensors are calibrated using a benchtop high-accuracy pressure gauge or transducer as a reference. Usually, the reference has been a Mensor Corporation pressure transducer with a pressure differential range of ±48 hPa and 0.01% full-scale accuracy. Experience using the Honeywell sensors has shown that they are highly linear over their specified range, and they exhibit little calibration drift over time. The two temperature sensors in each probe are calibrated using a mercury bulb thermometer or other stable reference.

Another issue with the probe design is possible errors due to real-world departures from the theoretical equations in section 2d, which are derived from Eq. (1). Such errors are magnified for aircraft-based probes, because these systems must compute the ambient wind as a small residual resulting from the addition of two large vectors having nearly opposite directions (Tjernström and Friehe 1991). Data from the ET probe are expected to be less sensitive to such errors, since the probe is directly measuring the wind in earth coordinates. The ET probe software does include empirical corrections to α, β, and q to account for possible wind errors. However, the field tests have so far not indicated any significant deviations from the theoretical equations.

b. Road tests in dry weather

Like other pressure spheres, the ET probe does not function well at low airspeeds, because the dynamic pressure fluctuations given by Eq. (2) rapidly diminish with decreasing airspeed. Field experience with the probes using the current Honeywell sensors indicate that the velocity measurements increasingly drop out once airspeeds fall below roughly 8 m s−1, when q is about 40 Pa. There is not a sudden failure of the system but rather a gradual increase in data dropouts at lower speeds. The turbulence levels also have an influence on the dropout rate at these low speeds.

Testing the probes requires a reliable method of attaining airspeeds within the range encountered in tropical cyclones. The simplest approach that was accessible during development was to mount the probe on a vehicle and drive the vehicle at highway speeds. The probe was mounted on a post that extended about 1.5 m above the top of the vehicle. It is important to recognize that the quantity being measured during a road test is the air velocity relative to the moving vehicle, not the ambient wind velocity. This relative velocity is affected by both ambient wind fluctuations and variations in the vehicle velocity. The aim of these tests is to compare the ET probe with a reference instrument and not to study ambient turbulence levels along a roadway.

In the early tests, a cup anemometer was mounted beside the ET probe so that the basic functionality of the system could be evaluated. These early tests demonstrated that the overall design of the ET probe appeared to be sound, providing accurate airspeed measurements in line with the performance of aircraft gust probes. Later, the cup anemometer was replaced on the vehicle by a Gill Windmaster Pro sonic anemometer. This switch allowed direct comparisons of 3D velocity fluctuations between the ET probe and sonic anemometer.

Initially, the sonic–ET comparisons were performed in dry weather to test the overall system performance in the absence of rain. Figure 6 is a scatterplot of airspeed from one such test on 26 May 2004. For this comparison, the 50-Hz ET probe measurements were block averaged to the 10-Hz sampling frequency of the sonic anemometer. The airspeeds measured by the ET probe are in close agreement with the sonic. For all the data during the test, the median ratio of the ET speed to the sonic speed is 1.01, with an interquartile range from 0.96 to 1.08. A standard linear regression produces a slope of 0.98, intercept of 0.82 m s−1, and a squared correlation coefficient of 0.97.

The road tests with the sonic also allow velocity spectra to be evaluated. This is most appropriate for periods when the mean airspeed is nearly constant. Spectra for one such period on 26 May 2004 are shown in Fig. 7. The mean airspeed was approximately 32 m s−1 during this period. Before computing the spectra, the data from both instruments were first rotated into an orientation resembling conventional micrometeorological coordinates, with the x axis pointing along the mean direction of the relative airflow and the z axis pointing vertical. Overall, the sonic and ET spectra show excellent agreement. At airspeeds of 32 m s−1, the sonic’s sampling frequency of 10 Hz is clearly too low to observe the rolloff of the spectra at higher frequencies.

Care must be exercised in the interpretation of Fig. 7, since the spectra are for the relative velocity vector on the vehicle, not for the ambient velocity. Nonetheless, at high frequencies both the u and w ET spectra have −5/3 slopes. It is unlikely that changes in the vehicle’s velocity will generate a −5/3 spectral slope, so the most plausible explanation is that turbulence is the primary source of the velocity fluctuations at these frequencies. The υ spectrum clearly falls off faster than the other two. This pattern is not a fluke, since it has appeared in ET spectra computed from other road tests. The most likely explanation is related to the asymmetry of the port configuration in Fig. 4. Ports 1 and 4 used for the δpz measurement are 36° apart, whereas ports 2 and 3 for δpy are 72° apart. These separations will of course have an effect on the smallest observable flow features, including turbulent eddies. The arc length between ports 2 and 3 is about 27 cm. Only flow features with horizontal scales much larger than 27 cm will therefore be fully resolved by the probe. If we consider “much larger” to be 10 times larger, then the probe’s δpy measurements may start to show some degradation in resolution for features having horizontal scales less than about 2.7 m. At 32 m s−1, such features correspond to frequencies above 12 Hz, and this is consistent with the enhanced rolloff of the υ spectrum in Fig. 7. Because of the smaller port separation associated with the δpx and δpz measurements, a similar degradation is not expected until frequencies exceed about 24 Hz, which is near the Nyquist frequency of the instrument.

Figure 8 shows composite velocity statistics from eight different road tests spanning the period from 8 April to 4 August 2004. In each case the mean airspeed was nearly constant, and the sonic and ET data were rotated so that the x axis points along the mean relative velocity and z is vertical. The three velocity standard deviations σu, συ, and σw are computed in these rotated coordinates. The sonic and ET statistics are generally in good agreement with the exception of σw. As shown in Fig. 7, the sonic 10-Hz sampling rate caused this instrument to miss a significant fraction of the w velocity fluctuations at higher frequencies, leading to an underestimate of σw. The triangles in Fig. 8d were obtained by first block averaging the ET probe w data to 10 Hz, and these 10-Hz ET statistics show much better correspondence to the sonic estimates.

Most of the ET probe testing was based on the road tests, since they provided the high airspeeds that the instrument was designed for. However, a series of static tests in dry weather was performed on the afternoon of 15 May 2003 when wind speeds were marginal for the ET probe at 8–15 m s−1. The sonic–ET comparisons from these static tests produced results similar to the road tests (Eckman et al. 2004), so they are mentioned only in passing here. Additional static tests were planned, but most days that looked promising had marginal wind speeds similar to 15 May 2003, so the emphasis shifted to the road tests.

c. Road tests in rain

After evaluating the performance of the ET probe in dry conditions, the focus shifted to its performance in rain. Most of the rain tests have been conducted on the big-hole probes that use the passive rain defense. These tests were performed with the same vehicle mounting used in the dry tests. The results of one such test are shown in Fig. 9. The sonic measurements contain a large number of rain spikes that make the data of questionable utility. Similar problems have been encountered in hurricane deployments of sonic anemometers (J. L. Schroeder 2004, personal communication). In contrast, the ET measurements show little evidence of rainwater contamination. In fact, the ET velocity observations from the rain tests are very similar to those from the dry tests.

Figure 10 shows the w velocity spectrum for a 400-s portion of the 10 June 2004 rain test (again remembering that this is the w component of the relative velocity on the vehicle). The overall shape of the spectrum is similar to the dry test in Fig. 7c and indicates that there is little or no degradation of the ET probe’s velocity measurements in the rain. Subsequent experience during actual hurricane deployments in 2004 has supported this result, but there is still a possibility that the passive rain defense may be overwhelmed at some point in a strong hurricane.

The rain tests did indicate a weakness in the design of the mushroom housing for the temperature sensors. In heavy rain events some water was able to intrude into the base of the housing and cause voltage spikes in the temperature outputs. Attempts to correct this problem by applying a watertight silicone grease to the housing base were not entirely successful. The wind measurements are only weakly sensitive to temperature errors [Eq. (12)], so the problems with the temperature data have little effect on the wind statistics and momentum flux measurements. However, they directly affect the probe’s ability to measure the sensible heat flux. Since the ET probes will initially be deployed into landfalling hurricanes, the sensible heat flux is not of critical importance. It appears that correcting the problem will require a future redesign of the mushroom housing.

4. Summary and conclusions

The understanding of both tropical cyclone intensification and wind damage patterns has been hampered by an inability to directly measure surface fluxes and 3D turbulence in hurricane conditions. As a result, many tropical cyclone models use turbulence and flux parameterizations that are based on observations in low wind speeds. There is increasing evidence that extrapolating these parameterizations to hurricane-force winds leads to significant errors. The NOAA Air Resources Laboratory has been developing a surface-deployed turbulence probe capable of functioning in hurricane conditions. This Extreme Turbulence (ET) probe uses the same pressure-sphere technology as aircraft gust probes, which are routinely subjected to airspeeds similar to those encountered in strong tropical cyclones.

The ET probe is based on a 43-cm-diameter pressure sphere containing 30 pressure ports distributed over its surface. A major problem in deploying this type of sensor in tropical cyclones is to prevent water from fouling the ports and contaminating the pressure measurements. Two different rain defense approaches were pursued. In the passive defense, the pressure ports are enlarged and the internal tubing is sloped upward to allow gravity drainage. The active defense uses an air pump to flush water from smaller pressure ports.

Testing of the probe’s performance has largely focused on a series of road tests, where the probe was mounted beside a conventional instrument (cup anemometer or 3D sonic anemometer) on a vehicle and then driven at highway speeds. In dry conditions, the data from the ET probes have compared well with the observations from the conventional instruments, indicating that the ET probe has a performance comparable to the aircraft gust probes from which it was derived. Road tests in the rain have largely focused on the passive defense probes, and they indicated that these probes continue to function well in the rain. Of course, the precipitation and spray in a tropical cyclone may at some point become intense enough to overwhelm the passive defense.

Analysis of data from the active defense probe has received little attention, primarily because of the success of the simpler passive design. However, the active defense (possibly in combination with the passive defense) may still prove useful in more demanding situations. For example, an ET probe on a buoy may sometimes be immersed in water and then require some form of active defense to flush out the ports.

To reduce development costs, the ET probes built so far use many “office” hardware components, such as the Data Translation DT9800 data acquisition module and the standard notebook computer used for processing and archiving. Further development of the probes will likely switch to more rugged components. One promising approach would be to use embedded PC/104 modules to replace the existing computer components. PC/104 modules are small, self stacking, and often designed for a wide range of environmental conditions. A PC/104 computer would easily fit inside the ET sphere and would use significantly less power than the current notebook computer. Another future development would be to add remote satellite communications to the probes. Currently, satellite connections are too slow to transmit the full 50-Hz data, but they are sufficient for transmitting summary statistics and probe diagnostics.

Acknowledgments

The initial idea and guidance for the ET probe were provided by Dr. Timothy Crawford, who died unexpectedly in 2002. Development of the ET probe was greatly assisted by many staff members of the Air Resources Laboratory, with notable contributions from Dr. Jeffrey French and Messrs. Shane Beard, Randall Johnson, Roger Carter, and J. Randall White. The ET probe development was a joint collaboration between the Office of Naval Research CBLAST program and NOAA through the USWRP program. This work supports the Research and Development for Hurricane Observing and Prediction capability under the NOAA Science, Technology, and Infusion program.

REFERENCES

  • Alamaro, M., , Emanuel K. , , Colton J. , , McGillis W. , , and Edson J. , 2002: Experimental investigation of air–sea transfer of momentum and enthalpy at high wind speed. Extended Abstract, 25th Conf. on Hurricanes and Tropical Meteorology, San Diego, CA, Amer. Meteor. Soc., CD-ROM, 17C.6.

  • Brown, E. N., , Friehe C. A. , , and Lenschow D. H. , 1983: The use of pressure fluctuations on the nose of an aircraft for measuring air motion. J. Climate Appl. Meteor., 22 , 171180.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brunke, M. A., , Fairall C. W. , , Zeng X. , , Eymard L. , , and Curry J. A. , 2003: Which bulk aerodynamic algorithms are least problematic in computing ocean surface turbulent fluxes? J. Climate, 16 , 619635.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Crawford, T. L., , and Dobosy R. J. , 1992: A sensitive fast-response probe to measure turbulence and heat flux from any airplane. Bound.-Layer Meteor., 59 , 257278.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dobosy, R. J., , Crawford T. L. , , MacPherson J. I. , , Desjardins R. L. , , Kelly R. D. , , Oncley S. P. , , and Lenschow D. H. , 1997: Intercomparison among the four flux aircraft at BOREAS in 1994. J. Geophys. Res., 102 , D24. 2910129111.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eckman, R. M., , Crawford T. L. , , Dumas E. J. , , and Birdwell K. R. , 1999: Airborne meteorological measurements collected during the Model Validation Program (MVP) field experiments at Cape Canaveral, Florida. NOAA Tech. Memo. ARL/ATDD-233, NOAA Air Resources Laboratory, Silver Spring, MD, 54 pp.

  • Eckman, R. M., , Dobosy R. J. , , Strong T. , , and Auble D. L. , 2004: Development and initial deployment of an omnidirectional pressure-sphere anemometer for observing winds and turbulence in tropical cyclones. Extended Abstract, 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 368–369.

  • Emanuel, K. A., 1986: An air–sea interaction theory for tropical cyclones. Part I: Steady-state maintenance. J. Atmos. Sci., 43 , 585604.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 1995: Sensitivity of tropical cyclones to surface exchange coefficients and a revised steady-state model incorporating eye dynamics. J. Atmos. Sci., 52 , 39693976.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 2003: A similarity hypothesis for air–sea exchange at extreme wind speeds. J. Atmos. Sci., 60 , 14201428.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fairall, C. W., , Kepert J. D. , , and Holland G. J. , 1994: The effect of sea spray on surface energy transports over the ocean. Global Atmos. Ocean Syst., 2 , 121142.

    • Search Google Scholar
    • Export Citation
  • French, J. R., , Johnson R. , , Beard S. , , and Crawford T. , 2004: Modification of an airborne gust probe for hurricane boundary layer research. Extended Abstract, 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 332–333.

  • Green, S. I., , Ostafichuk P. , , and Rogak S. N. , 2001: Measurement of three-dimensional unsteady flows using an inexpensive multiple disk probe. J. Atmos. Oceanic Technol., 18 , 883891.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hacker, J. M., , and Crawford T. L. , 1999: The BAT probe: The ultimate tool to measure turbulence from any kind of aircraft (or sailplane). J. Tech. Soaring, 23 , 4346.

    • Search Google Scholar
    • Export Citation
  • Katsaros, K. B., and Coauthors, 1994: Measurements of humidity and temperature in the marine environment during the HEXOS main experiment. J. Atmos. Oceanic Technol., 11 , 964981.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Krayer, W. R., , and Marshall R. D. , 1992: Gust factors applied to hurricane winds. Bull. Amer. Meteor. Soc., 73 , 613617.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kunkel, K. E., , and Bruce C. W. , 1983: A sensitive fast-response manometric wind sensor. J. Climate Appl. Meteor., 22 , 19421947.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Large, W. G., , and Pond S. , 1981: Open ocean momentum flux measurements in moderate to strong winds. J. Phys. Oceanogr., 11 , 324336.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lenschow, D. H., 1986: Aircraft measurements in the boundary layer. Probing the Atmospheric Boundary Layer, D. H. Lenschow, Ed., Amer. Meteor. Soc., 39–55.

    • Search Google Scholar
    • Export Citation
  • Miller, B. I., 1964: A study of the filling of Hurricane Donna (1960) over land. Mon. Wea. Rev., 92 , 389406.

  • Oost, W. A., , Worrell E. W. H. , , Schaap J. W. , , van Oort C. , , and Kraan C. , 1991: An improved version of the pressure anemometer. J. Atmos. Oceanic Technol., 8 , 575584.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ooyama, K., 1969: Numerical simulation of the life cycle of tropical cyclones. J. Atmos. Sci., 26 , 340.

  • Paulsen, B. M., , and Schroeder J. L. , 2005: An examination of tropical and extratropical gust factors and the associated wind speed histograms. J. Appl. Meteor., 44 , 270280.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peters, G., , and Fischer B. , 1998: One-year operational measurements with a sonic anemometer-thermometer and a Doppler sodar. J. Atmos. Oceanic Technol., 15 , 1828.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Powell, M. D., , Houston S. H. , , and Reinhold T. A. , 1996: Hurricane Andrew’s landfall in South Florida: Standardizing measurements for documentation of surface wind fields. Wea. Forecasting, 11 , 304328.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Powell, M. D., , Vickery P. J. , , and Reinhold T. A. , 2003: Reduced drag coefficient for high wind speeds in tropical cyclones. Nature, 422 , 279283.

  • Rediniotis, O. K., , and Kisner R. E. , 1998: Development of a nearly omnidirectional velocity measurement pressure probe. AIAA J., 36 , 18541860.

  • Rosenthal, S. L., 1971: The response of a tropical cyclone model to variations in boundary-layer parameters, initial conditions, lateral boundary conditions, and domain size. Mon. Wea. Rev., 99 , 767777.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Smith, S. D., 1988: Coefficients for sea surface wind stress, heat flux, and wind profiles and a function of wind speed and temperature. J. Geophys. Res., 93 , 1546715472.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thurtell, G. W., , Tanner C. B. , , and Wesely M. , 1970: Three-dimensional pressure-sphere anemometer system. J. Appl. Meteor., 9 , 379385.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tjernström, M., , and Friehe C. A. , 1991: Analysis of a radome air-motion system on a twin-jet aircraft for boundary-layer research. J. Atmos. Oceanic Technol., 8 , 1940.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tuleya, R. E., 1994: Tropical storm development and decay: Sensitivity to surface boundary conditions. Mon. Wea. Rev., 122 , 291304.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Vickery, P. J., , Skerlj P. F. , , Steckley A. C. , , and Twisdale L. A. , 2000a: Hurricane wind field model for use in hurricane simulations. J. Struct. Eng., 126 , 12031221.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Vickery, P. J., , Skerlj P. F. , , and Twisdale L. A. , 2000b: Simulation of hurricane risk in the U.S. using empirical track model. J. Struct. Eng., 126 , 12221237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, Y., , Kepert J. D. , , and Holland G. J. , 2001: The effect of sea spray evaporation on tropical cyclone boundary layer structure and intensity. Mon. Wea. Rev., 129 , 24812500.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wesely, M., , Thurtell G. W. , , and Tanner C. B. , 1970: Eddy correlation measurements of sensible heat flux near the earth’s surface. J. Appl. Meteor., 9 , 4550.

    • Crossref
    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

Conceptual diagram showing the general operation of the ET probe.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 2.
Fig. 2.

Photograph of the ET probe on a laboratory stand. A pen has been placed next to the stand as a scale reference.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 3.
Fig. 3.

(a) The interior components of one probe hemisphere. The A/D card is a data acquisition module that digitizes the analog sensor outputs. (b) A close-up of the circuit board, showing the location of the absolute and differential pressure sensors.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 4.
Fig. 4.

(a) The pressure port geometry used to compute the wind vector. Port 0 is closest to the flow stagnation point, and ports 1–4 form a cruciform pattern about the center port. (b) The angle of attack α in the vertical xz plane. (c) The angle of sideslip β in the xy plane.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 5.
Fig. 5.

The 3-m tripod used for deploying the ET probes into hurricanes. The computer and batteries used to run the probe are located in an enclosure at the base of the tripod.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 6.
Fig. 6.

Scatter diagram of sonic and ET probe airspeeds during a 46-min road test in dry conditions starting at 1932 UTC 26 May 2004.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 7.
Fig. 7.

Velocity spectra computed from a 500-s subset of road test data beginning 1940 UTC 26 May 2004. The sonic anemometer spectra (gray) have been displaced downward one decade to provide separation from the ET spectra (black). (a)–(c) The u, υ, and w spectra. On the vertical axes, the spectral values are multiplied by the frequency.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 8.
Fig. 8.

Composite velocity statistics for the sonic anemometer and ET probe based on eight road tests conducted between 8 Apr and 4 Aug 2004. (a) The mean airspeed U; (b)–(d) the velocity standard deviations σu, συ, and σw. In (d), the circles are based on the original 50-Hz ET data, whereas the triangles are based on ET data after block averaging to 10 Hz.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 9.
Fig. 9.

Airspeeds observed during a road test in the rain on 10 Jun 2004. The ET speeds (gray) do not show the large number of rain spikes observed with the sonic anemometer (black). The periods near 1540 and 1600 UTC when the speed drops to zero are times when the vehicle was either turning around or stopped at an intersection.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Fig. 10.
Fig. 10.

Spectrum of ET probe w velocity component for a 400-s period starting 1527 UTC 10 Jun 2004. The vehicle was driving through rain at the time.

Citation: Journal of Atmospheric and Oceanic Technology 24, 6; 10.1175/JTECH2025.1

Save