• Bothias, L., 1987: Radiowave Propagation. McGraw-Hill, 330 pp.

  • Bullington, K., 1957: Radio propagation fundamentals. Bell Syst. Tech. J., 36, 593–626.

  • Cole, H. L., , and Hock T. F. , 2005: The driftsonde observing system development. Preprints, 13th Symp. on Meteorological Observations and Instrumentation, Savannah, GA, Amer. Meteor. Soc., CD-ROM, 3.4.

  • Grubišić, V., , Doyle J. D. , , Kuettner J. , , Poulos G. S. , , and Whiteman C. D. , 2004: Terrain-induced Rotor Experiment (T-REX). Scientific Overview Document and Experiment Design, 72 pp. [Available online at http://www.eol.ucar.edu/projects/trex.].

  • Hock, T. F., , and Franklin J. L. , 1999: The NCAR GPS dropwindsonde. Bull. Amer. Meteor. Soc., 80 , 407420.

  • Kottmeier, C., , Reetz T. , , Ruppert P. , , and Kalthoff N. , 2001: A new aerological sonde for dense meteorological soundings. J. Atmos. Oceanic Technol., 18 , 14951502.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mayr, G. J., and Coauthors, 2004: Gap flow measurements during the Mesoscale Alpine Programme. Meteor. Atmos. Phys., 86 , 99119.

  • Rappaport, T. S., 1996: Wireless Communications Principles and Practice. Prentice-Hall, 641 pp.

  • View in gallery

    (a) A plan view of the HAIPER flight track (yellow line), signal acquisition location along the flight track (black circle), and location of the ground-test dropsonde (green circle) during the T-REX IOP-1 test near Independence, CA. (b) The geometry of the T-REX HIAPER dropsonde ground test during IOP 1 (not to scale); Ha = 12.44 km, Uag = 200 m s−1. The test began at 1707:00 UTC and the signal was first reliably acquired at 1738:59 UTC, 65.75 km from the target or 3.75 km closer than the line-of-sight prediction (underlying maps courtesy of DeLorme, from Topo USA software).

  • View in gallery

    (b) (Continued)

  • View in gallery

    The geometry of platform-independent complex terrain signal loss for a valley target. We show an airborne platform dropping sondes with mean fall speed Ws while westbound in westerly flow at flight level Ha with ground speed Uag. The Owens Valley geometry is exaggerated for clarity (underlying maps courtesy of DeLorme, from Topo USA software).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 23 23 3
PDF Downloads 11 11 1

Targeted Dropwindsondes in Complex Terrain

View More View Less
  • 1 In-situ Sensing Facility, Earth Observing Laboratory, National Center for Atmospheric Research, * Boulder, Colorado
© Get Permissions
Full access

Abstract

The dropwindsonde (or dropsonde) is a frequently utilized tool in geophysical research and its use over ocean and flat terrain is a reliable and well-established practice. Its use in complex terrain, however, is complicated by signal acquisition challenges that can be directly related to the ground target location, local relief, and line of sight to flight tracks relevant to the observation sought. This note describes a straightforward technique to calculate the theoretical altitude above ground to which a ground-targeted dropsonde will provide data for a given airborne platform. It is found that this height HCq can be calculated from expected airborne platform horizontal velocity Uag, mean dropwindsonde vertical velocity Ws, the relevant barrier maximum HB, and the horizontal distance from the target area to the barrier maximum DB. Here, HCq is found to be weakly dependent on release altitude through Ws. An example from the Terrain-induced Rotor Experiment (T-REX) is used to show that for modern aircraft platforms and dropwindsondes signal loss can occur 1–2 km above ground if mitigation is not pursued. Practical mitigation techniques are described for those complex terrain cases where signal propagation problems would create a significant negative scientific impact.

Corresponding author address: Gregory S. Poulos, National Center for Atmospheric Research, P.O. Box 3000, Boulder, CO 80307-3000. Email: gsp@ucar.edu

This article included in the Terrain-Induced Rotor Experiment (T-Rex) special collection.

Abstract

The dropwindsonde (or dropsonde) is a frequently utilized tool in geophysical research and its use over ocean and flat terrain is a reliable and well-established practice. Its use in complex terrain, however, is complicated by signal acquisition challenges that can be directly related to the ground target location, local relief, and line of sight to flight tracks relevant to the observation sought. This note describes a straightforward technique to calculate the theoretical altitude above ground to which a ground-targeted dropsonde will provide data for a given airborne platform. It is found that this height HCq can be calculated from expected airborne platform horizontal velocity Uag, mean dropwindsonde vertical velocity Ws, the relevant barrier maximum HB, and the horizontal distance from the target area to the barrier maximum DB. Here, HCq is found to be weakly dependent on release altitude through Ws. An example from the Terrain-induced Rotor Experiment (T-REX) is used to show that for modern aircraft platforms and dropwindsondes signal loss can occur 1–2 km above ground if mitigation is not pursued. Practical mitigation techniques are described for those complex terrain cases where signal propagation problems would create a significant negative scientific impact.

Corresponding author address: Gregory S. Poulos, National Center for Atmospheric Research, P.O. Box 3000, Boulder, CO 80307-3000. Email: gsp@ucar.edu

This article included in the Terrain-Induced Rotor Experiment (T-Rex) special collection.

1. Introduction

Techniques for the use of dropsondes for field operations over large water bodies and relatively flat terrain are well established (Hock and Franklin 1999) and seldom encounter problems associated with signal loss. Similar dropsonde operations when executed in complex terrain are, however, more susceptible to negative impacts on data acquisition from signal termination. From well-known radio signal propagation literature, signal loss is generally understood as the result of interception by the terrain of the line of sight between the transmitting and receiving antennas for a given system (Bullington 1957; Bothias 1987). Diffraction (first Fresnel zone obstruction), reflection, and atmospheric refraction can also influence the accuracy of line-of-sight calculations of signal loss (Rappaport 1996). These effects are generally small or intermittent perturbations to line-of-sight signal loss calculations for upper-troposphere/lower-stratosphere dropsonde (400-MHz frequency/0.75-m wavelength) launches in complex terrain and they will not be considered in this note (T. F. Hock 2006, personal communication; Bullington 1957).

The loss of the dropsonde signal compromises the acquisition of the full atmospheric profile, depending on terrain relief characteristics, dropsonde characteristics, and the dropping platform. In the case of those dropsondes that retain data internally and broadcast their final position for retrieval via cell phone (Kottmeier et al. 2001), signal acquisition is not a concern where the cell signal grid is adequate. However, the retrieval of these sondes can be arduous and incomplete in complex terrain and does not allow for real-time access to data. The ability to access dropsonde data in real time is often a crucial component of scientific flight planning in modern-day field projects, particularly in the case of flight operations in hazardous or rapidly changing weather, such as during the recent Terrain-induced Rotor Experiment (T-REX; Grubišić et al. 2004).

If any portion of the complex terrain phenomenon of scientific or operational interest occurs in the lower levels of the terrain, then signal acquisition failure can compromise the optimal achievement of measurement goals. During the Mesoscale Alpine Program (MAP), for example, dropsonde data were found to be lost during attempted dropsonde flights in the Inn Valley and in Brenner Pass during studies of gap flows (Mayr et al. 2004). As a result, planned flight patterns for the study of gap flow were at times adjusted as the research aircraft completed circles over the target area within line of sight until the dropsonde reached the valley bottom. With the recent emergence of the High-Performance Instrumented Airborne Platform for Environmental Research (HIAPER, a Gulfstream V aircraft) and the present attempts to reduce the cost weight and form factor of dropsondes (Cole and Hock 2005), it is prudent to consider the constraints to their use for complex terrain atmospheric scientific research purposes.

2. The dropsonde ground test

The signal acquisition characteristics of a Vaisala RD93 dropsonde were tested during the T-REX field project. T-REX was held 1 March–30 April 2006 in the Owens Valley of California and focused its measurement efforts on cross-barrier flow-induced atmospheric rotors that occur in the lowest ∼3 km above ground near the town of Independence (see Fig. 1; Grubišić et al. 2004). Although only the ground-test dropsonde site relevant to this test is shown in Fig. 1, we note that up to two dropsonde aircrafts at a time focused their flights on understanding the mountain wave/rotor/boundary layer system in the vicinity of Independence.

Since dropsondes were a critical part of the measurement campaign and significant costs and planning were expended in the optimal use of hundreds of dropsondes for T-REX, this test would determine how important the possibility of lost data from dropsondes in Owens Valley might be for the scientific goals of the project. This single systematic test was executed during a flight of the HIAPER Gulfstream V research aircraft in the vicinity of Owens Valley. The purpose of this test was to determine how strictly the line-of-sight requirement holds for RD93 dropsondes and advanced signal acquisition systems (Hock and Franklin 1999).

The test simply required that the dropsonde be activated on the ground while HIAPER flew a cross-valley leg that included a significant leg length well beyond the line connecting the dropsonde, the relevant barrier maximum (HB), and HIAPER location along the flight track. Such a condition existed during T-REX intensive observing period (IOP) 1 when HIAPER flew an eastbound track at an altitude of Ha = 12.44 km mean sea level (MSL), as depicted in Figs. 1a and 1b. To execute the test, a previously agreed upon frequency, 405.98 MHz, was set for both the ground-based dropsonde and the HIAPER-based dropsonde receiver. As HIAPER reached its westernmost extent the dropsonde itself was activated. Then, as HAIPER entered the west end of the leg at 1707:00 UTC its dropsonde data system was activated, such that the time of signal acquisition could be tracked. The dropsonde signal was first acquired at 1738:59 UTC (Fig. 1a) and consistent signals were established 10 s later.

The relevant side view geometry for this specific test is shown in Fig. 1b. The horizontal distance between the Gulfstream V and the ground-test sonde at the time the signal would be acquired,
i1520-0426-24-8-1489-e1
was predicted to be 69.50 km based on the line of sight between the dropsonde and the aircraft flight track tangent to HB. Here, HS is the height of the airborne platform above the ground target and βT is the elevation angle of HB from the target, separated by a horizontal distance DB such that
i1520-0426-24-8-1489-e2
The actual point of first signal acquisition was 65.75 km from the sonde, or 3.75 km (5.4%) horizontally closer than predicted. Consistent signals were first acquired 2.5 km farther or a total of 9.2% difference from the line-of-sight prediction. At HAIPER ground speeds, Uag of ∼200 m s−1 signal acquisition was approximately 19 s later than would be expected from line-of-sight considerations. For this Owens Valley example, we consider the latter point to be that where useful dropsonde data would begin to be acquired on an eastbound leg, or equivalently, where dropsonde data would first be compromised by signal loss on a typical westbound flight leg. It should be noted that the signal loss condition described here may be somewhat worse than under typical drop conditions, where the signal is acquired and synchronized as the dropsonde exits the aircraft. This is a relatively minor contributing error, however (T. F. Hock 2006, personal communication), and is dependent on the electronics and synchronization software utilized. Note that in T-REX, due to externally imposed constraints on the westward extent of flight legs and the eastward extent of sonde launches, dropsonde signal loss was not noted as a major problem during operations.

3. Platform and terrain-independent geometry

Although the data from this ground-based test showed that sonde signal loss can occur at a distance 5%–9% less than the line-of-sight expectation, the application of the above test result to a given release platform and dropsonde is not as straightforward due to the relative movement of the two bodies. The simple geometry of the generalized system is shown in Fig. 2 using both the Owens Valley orography and an example of a westbound aircraft dropping sondes at a target point on the valley floor with two HC examples shown.

Let us assume, despite our test result, that the critical point where the dropsonde signal would be lost is the line-of-sight point or more specifically that point where the line between the dropsonde and airborne platform first becomes tangent to the topography. To calculate the critical signal loss elevation above ground in quiescent conditions, HCq, we first recognize that this line (thick line in Fig. 2) defines the critical signal loss angle, αC. We further define the horizontal distance from the target where the signal is lost as lC, the distance the sonde has dropped at the time the signal is lost as zC, and the period from the drop time to when the signal is lost as tC, such that
i1520-0426-24-8-1489-e3
For the signal to be lost αC must be smaller than the elevation angle βT. Physically, this confirms that for targets in valleys with greater relief the likelihood of signal loss is greater. Further, consider a mean aircraft ground speed along the flight leg Uag and a mean sonde fall speed Ws, where zC = WstC and lC = UagtC. We find
i1520-0426-24-8-1489-e4
Thus, the critical angle is defined only by the mean aircraft ground speed and mean sonde fall speed between the drop and the loss of the signal and is independent of the time it takes to reach that point. Thus, for all βTαC line of sight will be preserved.
Here, Ws can be calculated for the given dropsonde type. For a standard parachute-borne Vaisala RD93 dropsonde the relevant equation for instantaneous fall speed can be written as
i1520-0426-24-8-1489-e5
where ms is sonde mass, ρ is air density, Ap is the parachute area, the drag coefficient CD is 0.61, and g is the gravitational constant (Hock and Franklin 1999). Using the equation of state and integrating temperature T and pressure p from the aircraft flight altitude above mean sea level Ha to the critical line-of-sight intersection height HC we find
i1520-0426-24-8-1489-e6
where R is the dry gas constant for air. This equation does not consider the effects of atmospheric moisture on Ws and in practice could be calculated using the U.S. Standard Atmosphere, 1976 or radiosonde data. The value of Uag depends on Ha, the aeronautical characteristics of the particular airborne sonde platform, and environmental conditions. For example, at maximum cruising altitude, 15.27 km (51 000 ft), a dropsonde launched from HIAPER experiencing no net vertical motion during the drop will reach the ground in 876 s at Ws = 17.43 m s−1. HIAPER will have traveled approximately 232 km in this time (the current RD93 dropsonde is engineered such that the maximum signal range is 300 km).
Also note for the geometry shown in Fig. 2 that
i1520-0426-24-8-1489-e7
where HB is the terrain relief from the target to the relevant terrain maximum and DB is the horizontal distance from the vertical line intersecting the terrain maximum to the target. Here, HCq is the variable of interest and is the height above ground where line of sight would be lost in quiescent conditions. Equation (5) can be readily solved for HCq since DB, HB, Ws, and Uag are known such that
i1520-0426-24-8-1489-e8

In the limit where Uag approaches infinity or Ws approaches 0, we note from (8) that HCq approaches HB = HCmax (see Fig. 2). Physically, this means that very fast aircraft combined with very slow falling sondes create the worst-case scenario for signal loss in complex terrain, with the trivial limit being the loss of the dropsonde signal as soon as the dropsonde falls below the maximum relief. In typical atmospheric conditions, where a dropsonde follows a path dependent on environmental conditions at the time of the drop (thick dotted line in Fig. 2), the height above ground where the signal loss occurs will fall on the line from crest relief HB to H′, such as HC shown in Fig. 2. For cross-barrier flight tracks into the flow, HC > HCq, and signal loss will be worsened. For HCq = 0, αC = βT, and thus line of sight, will be maintained throughout the drop. The most desirable dropping platform and sonde characteristics or maximum flight leg distance from the target may be ascertained for a given terrain configuration, target, and scientific goal by evaluating this limit.

Table 1 shows HCq examples for the National Center for Atmospheric Research (NCAR) HIAPER, the Wyoming King Air, and the NCAR C-130 using typical Uag for westerly flight tracks at certain elevations in a quiescent flow using the Owens Valley terrain shown in Fig. 1b. It indicates that significant portions of the lower boundary layer would not be sampled due to signal loss for long westerly flight legs. Here, HC would be larger and therefore more of the lower atmosphere would not be sampled for typical midlatitude westerly flow and net easterly sonde drift (i.e., the T-REX science objectives; Grubišić et al. 2004).

4. Practical considerations

We have found that for any given dropsonde platform and flight leg orientation over well-characterized terrain, and a given dropsonde type and complex terrain target, the elevation above ground where the dropsonde signal will be lost can be estimated. Since HC is also one of the limits in the integral needed to calculate Ws, an iterative procedure must be used to solve the equations. For most applications, where Ha/HB > 2, a sufficiently accurate HCq can be estimated using the elevation of the target point (Hv in Figs. 1b and 2) as the lower bound on the integrand in (6) since the contribution of error to Ws will be small. Since our sonde ground test indicated that signal acquisition may be 5%–10% worse than line of sight would otherwise indicate, a more realistic and conservative estimate of HCq would use a value of HB in (8) that is this amount larger than the known value from a terrain cross section.

When an unacceptably large HCq has been calculated for a given scientific or operational application for dropsondes in complex terrain, there are four primary mitigation strategies that can be considered (for a given dropsonde/parachute combination and airborne platform):

  1. modify the flight plan such that βTαC;
  2. utilize a form of dropsonde technology, such as Kottmeier et al. (2001), if supported sufficiently by the cell grid and terrain accessibility;
  3. provide an additional dropsonde receiver on the ground near the target location; and
  4. install a high elevation signal repeater, in effect reducing HB to very small values.

Acknowledgments

We thank Errol Korn for the execution of the HIAPER onboard data acquisition and Terry Hock for his helpful comments (both of NCAR).

REFERENCES

  • Bothias, L., 1987: Radiowave Propagation. McGraw-Hill, 330 pp.

  • Bullington, K., 1957: Radio propagation fundamentals. Bell Syst. Tech. J., 36, 593–626.

  • Cole, H. L., , and Hock T. F. , 2005: The driftsonde observing system development. Preprints, 13th Symp. on Meteorological Observations and Instrumentation, Savannah, GA, Amer. Meteor. Soc., CD-ROM, 3.4.

  • Grubišić, V., , Doyle J. D. , , Kuettner J. , , Poulos G. S. , , and Whiteman C. D. , 2004: Terrain-induced Rotor Experiment (T-REX). Scientific Overview Document and Experiment Design, 72 pp. [Available online at http://www.eol.ucar.edu/projects/trex.].

  • Hock, T. F., , and Franklin J. L. , 1999: The NCAR GPS dropwindsonde. Bull. Amer. Meteor. Soc., 80 , 407420.

  • Kottmeier, C., , Reetz T. , , Ruppert P. , , and Kalthoff N. , 2001: A new aerological sonde for dense meteorological soundings. J. Atmos. Oceanic Technol., 18 , 14951502.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mayr, G. J., and Coauthors, 2004: Gap flow measurements during the Mesoscale Alpine Programme. Meteor. Atmos. Phys., 86 , 99119.

  • Rappaport, T. S., 1996: Wireless Communications Principles and Practice. Prentice-Hall, 641 pp.

Fig. 1.
Fig. 1.

(a) A plan view of the HAIPER flight track (yellow line), signal acquisition location along the flight track (black circle), and location of the ground-test dropsonde (green circle) during the T-REX IOP-1 test near Independence, CA. (b) The geometry of the T-REX HIAPER dropsonde ground test during IOP 1 (not to scale); Ha = 12.44 km, Uag = 200 m s−1. The test began at 1707:00 UTC and the signal was first reliably acquired at 1738:59 UTC, 65.75 km from the target or 3.75 km closer than the line-of-sight prediction (underlying maps courtesy of DeLorme, from Topo USA software).

Citation: Journal of Atmospheric and Oceanic Technology 24, 8; 10.1175/JTECH2065.1

Fig. 1.
Fig. 1.

(b) (Continued)

Citation: Journal of Atmospheric and Oceanic Technology 24, 8; 10.1175/JTECH2065.1

Fig. 2.
Fig. 2.

The geometry of platform-independent complex terrain signal loss for a valley target. We show an airborne platform dropping sondes with mean fall speed Ws while westbound in westerly flow at flight level Ha with ground speed Uag. The Owens Valley geometry is exaggerated for clarity (underlying maps courtesy of DeLorme, from Topo USA software).

Citation: Journal of Atmospheric and Oceanic Technology 24, 8; 10.1175/JTECH2065.1

Table 1.

Examples of HCq (quiescent atmosphere) for the Owens Valley terrain shown in Fig. 1b for DB = 15 800 m, HB = 2560 m. Note that Ws is [through (6)] dependent on release altitude, whereas HCq is not. These calculations are based on a 0.26-m-diameter dropsonde parachute.

Table 1.

* The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Save