1.0 Introduction - University of Oklahoma School of ...



5.5 TARGETED OBSERVATIONS OF THE BOUNDARY LAYER USING THE GLIDERSONDE

METOEROLOGICAL PACKAGE IN A RADIO CONTROLLED AIRCRAFT PART II: TURBULENCE STRUCTURE AND PLATFORM LIMITATIONS

Daniel B. Weber [pic], Frank W. Gallagher III[pic], Kenneth W. Howard[pic]

[pic]Center for Analysis and Prediction of Storms, University of Oklahoma, Norman, Oklahoma

[pic] University of Oklahoma, Cooperative Institute for Mesoscale Meteorological Studies, Norman, Oklahoma

[pic]National Severe Storms Laboratory, Norman OK

1. INTRODUCTION

As shown in Part I of this series (Weber et.al., 2001), accurate vertical profiles can be obtained for scalar quantities in the boundary layer using a remote controlled aircraft. This paper looks further into the data obtained by the Dataplane platform from a three-dimensional perspective. From the three dimensional data set, boundary layer turbulence characteristics are investigated by means of the horizontal average, variances, and the vertical flux of heat and moisture, further extending the platform versatility. As a consequence of the above research, the limitations of the observing system are discussed in terms of the flight characteristics and instrument accuracy and are extended to the observation resolution.

Data presented in this study were collected on a sunny, warm and humid day at the end of June 2000 in Central Oklahoma. The flights were conducted from the Central Oklahoma Radio Control Society (CORCS) Field in Norman, Oklahoma. The observing platform consisted of a radio control aircraft, Dataplane DP-0 and is described in more detail in Part I of this work (Weber et.al., 2001).

The paper is organized as follows: Section 2.0 presents the three-dimensional data set obtained from the ARPS Data Analysis System (ADAS) analysis and the details of the flight path. Section 3.0 presents results from first and second order derived turbulence quantities and Section 4.0 provides insight to the limitations of the data set resulting from the observing platform.

2. DATA ANALYSIS

As discussed in Part I, the majority of the flights were intended to produce a time series of consistent vertical profiles. But during a 25-minute period from ___________________________________________

Corresponding author address: Daniel B. Weber, CAPS, 100E. Boyd St., Room 1110, University of Oklahoma, Norman, Ok, 73019-1011; e-mail: dweber@ou.edu.

15:07 to 15:35Z the aircraft was flown in a box-like pattern in an attempt to portrait a three dimensional picture of the developing boundary layer. Figure 1 displays the flight pattern for the prescribed period. Two areas of high-density coverage exist, below 965mb and above 950mb. It is clear from the figure that the pilot executed an irregular flight pattern, not able to predict the path exactly due to a number of factors including wind and imprecise control input. But the aircraft was able to fill a volume, with respect to the ground, with observations during a 25-minute flight.

A three dimensional analysis was performed using ADAS. The background state for ADAS was the output from an Advanced Regional Prediction System (ARPS) simulation initialized with using the 6/29/00 12Z Norman, OK sounding. The Barnes option within ADAS was used to perform the three dimensional objective analysis. Following the analysis procedure, only points where observations exist were applied to the background field. The horizontal radius of influence was 90m on the first pass and reduced to 50m on the fourth and final pass. The vertical radius of influence was feathered from 60m to 30m. This allowed for a tight fit to the data within the data rich regions in each spatial dimension with a 20m spacing in each direction. The result is a three-dimensional data set suitable for further study. Figure 2 displays dew point contour plots in the north-south direction near the middle of the flight envelope and at a constant height. The data rich regions are clearly indicated on the plots by the strong gradients along the edge of the flight path, indicating a significant change from the background moisture profile. A drying of the airmass took place in the upper portion of the domain since the morning sounding as was also evident on the constant pressure analyses. The number of grid points covered by Dataplane varies according to the flight path. More points were collected in the north-south direction due to the field orientation and the majority of the data were collected at an elevation greater than 25 m above ground level. This is due in part to the fact that a cushion in terms of elevation is desirable at increasing distances from the control point, thus low-level data collection was

[pic]

Figure 1. Dataplane flight path from 15:07Z to 15:33Z June 29, 2000. Line indicates the path taken by the aircraft as recorded by the on-board computer and GPS.

(a)

[pic]

(b)

[pic]

Figure 2. Cross sections in the (a) North-South and (b) x-y directions taken from data collected with Dataplane from 15:07 to 15:33Z 6/29/00. Analysis was performed using ADAS and a background field from the 12Z Norman sounding. The analysis was performed with a resolution of 20m. Outside the flight envelope the data background field is unchanged. The contour interval is 0.2 g/kg.

avoided at extended ranges. As Figure 2 reveals, the moisture pattern is complex within the data rich regions. Regions outside the analysis are horizontally homogeneous and represent the background conditions from the Norman sounding. A series of eddies are apparent in the plots and indicate horizontal and vertical scales on the order of 100-250m. Entrainment of drier air from above the boundary layer is evident by the presence of air with mixing ratios less than 10g/kg, approximately 600m above ground. In addition, moist plumes are present at 800m with source regions near the surface. No consistent convective pattern is visible, since continuous volume filling observations were not collected.

3.0 TURBULENCE CHARACTERISTICS

Turbulence can be measured by a number of different quantities. Here we investigate the ability of the Dataplane platform to reproduce common boundary layer turbulence features as derived from the moisture and temperature observations. Specifically, the mean, variance, and horizontally averaged vertical moisture and heat fluxes are calculated and compared to theory and observations.

The horizontal mean was computed by summing the regions in which the analysis found data and divided by the total number of analyzed points for each level. The background field was unchanged in data sparse regions and one such layer existed just above the surface. Figure 3 displays the horizontally average, variance and vertical flux of potential temperature and moisture. The vertical fluxes were normalized by their near surface values. In general, the horizontally averaged mean follows boundary layer observations and displays the general well-mixed trend. But there is evidence of deviations from neutral conditions and may be linked to the small sampling volume carved out by the aircraft. Since no other measurements were taken within the volume the results are unverifiable. At the top of the sounding data, a large variation in moisture was measured, on the order of 6 C, in each of the vertical profiles (Weber et.al., 2001). The variance in temperature is small, except for an unexplained spike in the middle of the mixed layer, and in general reflects the near neutral conditions of the boundary layer. The thick line overlaid on the vertical moisture flux chart represents a typical boundary layer profile. The general trend of the observed profile follows that of a typical convective boundary layer and includes the entrainment region at the top of the mixed layer.

A few comments need to be made regarding the conditions in which the observations were taken. First, it is understood that the measurements were taken over a 25-minute period, approximately two convective time scales. The boundary layer was deepening during this time but at a reduced rate and was approximately 600m deep, an argument could be made for the existence of near steady state conditions. The near surface winds were light and out of the northeast, minimizing advection. The surface features varied from Bermuda grass to a large parking lot to the north and trees to the south and west. It is clear from Figure 2 that convective elements do exist with scales on the order of 100m. The horizontal sampling rate of 1 hertz and a ground speed of 10-15 m, generates 10-15 data points per observed cell, sufficient for computing gradients.

4.0 SYSTEM LIMITATIONS

Platform limitations consist of two main components, the aircraft flight characteristics and the instrument response.

The instrument used in this study collects data at a rate of 1 Hertz using the standard Vaisala RS-80 PTH sensor. The size of the aircraft plays an important role in the ability to obtain reliable measurements. A small temperature bias exists due to the dissipation of engine heat when the sensor is located on the fuselage and in the future, the sensors will be mounted near the wingtips. The larger Senior Telemaster allows the movement the sensors further from the engine, since the wingspan is approximately 2.5m. In addition, the distance between the two sensor pods is approximately 2 meters and can be used as a direct horizontal flux measurement from dual sensor pods.

The observation resolution, independent of the reference frame, is a function of the instrument sampling rate and the airspeed of the aircraft. The Telemaster is a slow flying aircraft with a top airspeed of approximately 20m/s. When directed into a breeze, the aircraft can be landed with little forward motion. But it is the airspeed, not the ground speed that affects the sampling rate. Lower airspeeds and higher sampling rates allow for increased observation resolution. The groundspeed is estimated by the onboard GPS while the airspeed and horizontal winds requires the use of a pitot tube. A pitot tube will be installed on the platform for future tests. The key variables that determine the airspeed required for level flight include the airfoil and wing loading. Light aircraft require less lift, which is a function of the airfoil and the air velocity over the wing. Lower airspeeds improve the observation resolution. For level flight, the lift is equal in magnitude to the weight of the aircraft and can be expressed by:

[pic] (1)

where [pic]is the airfoil lift coefficient, [pic] the air density, [pic] the air velocity over the wing, [pic] the wing area, and [pic] is the mass of the aircraft. Equation (1) can be rearranged and solved for the

A B

[pic]

E F

[pic]

Figure 3. Computed turbulence quantities (A) mean potential temperature, (B) mean mixing ratio, (C) variance of potential temperature, and (D) variance of mixing ratio from observed data collected 15:07 to 15:33Z 06/29/2000. Units are degrees K and grams per kilogram for (A), (C) and (B), (D) respectively. Plots E and F display the horizontally averaged vertical heat and moisture flux for the same period normalized by their near surface values. The line overlay in (F) represents vertical flux trend.

velocity required to offset the wing loading. Table 1 presents the estimated airspeed for the Telemaster aircraft for a given payload. The wing is a flat bottom airfoil type similar to the Clark Y-Profile presented by Misis (1959). A lift coefficient of 1.25 was chosen and represents a value near the maximum lift as a function of the angle of attack. The aircraft flight characteristics indicate that the Telemaster should fly with a minimum airspeed of 8.5m/s with no payload. A 2kg payload requires an airspeed of 10.3m/s. The sampling resolution is determined by dividing the required airspeed by the sampling rate. The result is an observation resolution on the order of 10m for the current platform. No airspeed computations were made but could be estimated using upwind and downwind pass ground speed data. The current instrument package weighs approximately 2kg.

|PayloadMass |Wing Loading|Required |Sampling Resolution|

|(kg) |(oz/ft) |Airspeed |(m) |

| | |(m/s) | |

|0 |15.5 |8.5 |8.5 |

|1 |19.0 |9.4 |9.4 |

|2 |22.86 |10.3 |10.3 |

|3 |26.67 |11.2 |11.2 |

|4 |30.48 |12.1 |12.1 |

Table 1. Aircraft flying weight without payload is 4kg. Wing loading is presented in English units for comparison purposes. A coefficient of lift of 1.25 was used for the lift calculations with a Clark Y-Profile wing shape as presented by Mises (1959). Sampling resolution was computed using the current instrument-sampling rate of 1 Hertz.

The data displayed in Figure 2 represents the volume sensed by the aircraft with respect to the ground. Under calm conditions, the flight path is representative of the source region. But when a mean flow is present and the observations require several minutes to complete, the volume observed by the aircraft does not represent the source region. The position of the sensor with resect to the atmosphere is given by:

[pic] (2)

where [pic] and [pic] represent the initial position and the current position, [pic] is the velocity of the aircraft, and [pic] the wind velocity. When the wind velocity is zero, the location of the observation is the same with respect to the ground and the atmosphere. For a non-zero ambient wind, the offset between the position of the observation with respect to the ground and the atmosphere (source region) increases with time. Figure 4 illustrates the difference between the path of the aircraft and the source region of the observations.

[pic]

Figure 4. Graphic depicting the difference between the ground flight track (left) and the atmospheric source region sensed by the instrument (right). Wind is from east and the aircraft starts near the origin and travels eastward on the initial leg.

5. CONCLUSION

The targeted observations were objectively processed and reveal a complicated boundary layer structure in terms of convective elements and vertical structure. The Dataplane platform, although limited in its current observational capabilities, was able to capture some of the key boundary layer turbulence signatures from data collected during a warm summer morning in Central Oklahoma. The vertical moisture flux as well as the mean and associated variances followed theoretical expectations for the most part, despite sampling problems. The observations were not verified, as no other instrument measurements were available. The results lead us to believe that detailed boundary layer motions could be characterized by a single aircraft but in a limited fashion due to the problem with the source region. A small fleet of aircraft flying in formation would reduce the source region problem under ambient flow conditions.

6.0 ACKNOWLEDGEMENTS

We would like to thank the REU grant for supporting part of this research and Dihema Longman for participating in the June field study. Also, we would like to thank the Central Oklahoma Radio control society for use of the field for testing and data collection on June 29, 2000.

7.0 REFERENCES

Mises, Richard Von, 1959: Theory of Flight, Dover Publishing, 620pp.

Weber, D.B., F. W. Gallagher III, and K. W. Howard, 2001: Boundary Layer Targeted Observations Using the Glidersonde Meteorological Package in a Radio Controlled Aircraft Part I: Results. Preprints AMS 11th Symposium on Meteorological Observations and Instrumentation (this volume).

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download