The SGP97 Tethersonde Program Christa Peters-Lidard and Luke Davis Georgia Institute of Technology Atlanta, GA Les Showell NOAA/National Severe Storms Laboratory Norman, OK

 

Overview

The Data

Instrument Description

Collection Procedures

Quality Control/Quality Assurance

The Files

Size and Format

Name and Directory Information

The Science

Plan

Results

Data Access and Contacts

FTP Site

Points of Contact

References

 

Tethersonde Data

Surface Flux data

GA Tech Civil Engineering

NOAA Severe Storms Lab

Soundings Page

Overview

Georgia Institute of Technology (Georgia Tech) and NOAA's National Severe Storms Laboratory (NSSL) jointly operated a tethersonde at the DOE ARM/CART Central Facility during SGP97 to acquire high-temporal and vertical resolution profiles of temperature, humidity, pressure, and wind speed and direction in the lower atmospheric boundary layer.

The Data

Instrument Description

The tethersonde system deployed during SGP97 is an Atmospheric Instrumentation Research, Inc. (AIR) Atmospheric Data Acquisition System (ADAS) model AIR-3A consisting of a meteorological sensor package (Figure 1) powered by a sealed 9V alkaline battery and suspended below a gas-filled tethered balloon raised and lowered using a heavy-duty winch (Figure 2).  The tether line is 1000 meters in length and the speed of ascent or descent is controlled via a manual dial on the winch.

Meteorological sensors include dry and wet bulb thermistors, an aneroid capacitance barometer, a three-cup anemometer with tachometer and a magnetic compass. Humidity is obtained using dry and wet bulb measurements and the psychometric equation.   The sensor specifications are given in Table 1.  Two sensor packages, or AIRsondes, were available for the experiment.

The sensor data is transmited by a 10 milliwatt transmitter at a frequency of 403.5 MHZ to the AIR ground station, which post-processes and logs the data. The data is transferred to a laptop computer for analysis and distribution.

Collection Procedures

The tethersonde launch site was located at the DOE ARM/CART Central Facility near Billings, OK.  The elevation of the launch site was 313 m above mean sea level.

The tethersonde data collection consisted of ascent-descent sequences.  Ascents were initiated hourly from 0700 local time (CDT) through 1100 CDT.  Each ascent/descent sequence lasted approximately 45 minutes, yielding two profiles per hour in the morning hours.  In the afternoon, ascents were initated every one and one-half hours from 1230 CDT to
1700 CDT, winds permitting.

Samples were obtained at a rate of one every 10 seconds, which yielded a 2-5 meter vertical resolution in the atmosphere. The rise rate typically varied from 0.2-0.5 m/s.

The tethersonde could not be deployed in winds greater than 15 knots.  Due to the windy conditions in Oklahoma, the tethersonde could be operated on only 16 days of the experiment.  The dates are shown in Table 2 along with the maximum altitude achieved on each date.   The maximum altitude attained during the experiment was 993 m AGL on July 6, 1997, although typical maximum altitudes for each ascent/descent sequence were significantly less due to wind.

The ADAS system worked well throughout the experiment with the exception of the 0700 flight on June 25, 1997, when the first AIRsonde began to quickly drain batteries and lose data.  This instrument was replaced by the second AIRsonde for the remainder of SGP97.

On the basis of weather conditions (particularly winds) as well as the total number of flights per day and maximum altitude attained, we have divided the data into "Gold", "Silver", "Bronze" and "Other" days.  The dates are given in Table 3 for each category.  
 

Table 1.  AIR-3A tethersonde instrument specifications.
Sensor Type	Range	Precision	Resolution
Wet and Dry Bulb Thermistors	+50 to -70 C	0.5 C (-40,40) 1.0 C (-70,50)	0.01 C
Aneroid Barometer	1050 to 600 mb	3 mb	0.1 mb
Wind Speed	0 to 20 m/s	0.25 m/s	0.1 m/s
Wind Direction	2-358 deg.	5 deg.	1 deg.

 
 

Table 2.  Tethersonde flight dates and maximum altitudes.
Date	Maximum Altitude (m AGL)	Pressure at Maximum Altitude (mb)
6/18	600	908.8
6/19	121	959.3
6/22	619	909.3
6/25	738	895.4
6/26	486	925.7
6/27	700	901.3
6/28	212	948.4
7/02	658	901.0
7/03	566	910.5
7/04	675	904.4
7/05	778	893.4
7/06	993	869.6
7/14	790	890.6
7/15	330	940.7
7/16	697	900.6

 
 

Table 3.  Ranking of dates on which tethersonde data was collected.
Rank	Characteristics	Dates
Gold	> 7 flights/day  Calm to light winds  Altitude > 700 m AGL	6/25, 7/05, 7/06, 7/14, 7/16
Silver	5-6 flights/day  Light to moderate winds  Altitude 500-700 m AGL	6/18, 6/27, 7/03
Bronze	3-4 flights/day  Moderate to high winds  Altitude < 500 m AGL	6/22, 7/04
Other	< 3 flights/day  High winds  Misc. problems/weather conditions	6/19, 6/26, 6/28, 6/29, 7/02, 7/15

 

Quality Control/Quality Assurance

Initial in-field comparisons between the tethersonde data and collocated ARM/CART radiosonde data suggested a high degree of correlation between measured temperature and humidity profiles.  After the field experiment concluded and more detailed data analysis began, this observation was confirmed for tethersonde flights prior to June 25, 1997. Tethersonde data after this date were collected by a second AIRsonde (as discussed above) and significant and variable differences were found between the tethersonde and radiosonde profiles of temperature and humidity.  Examples of these differences on July 5, 1997 are shown in Figures 3, 4 and 5, which indicate a cool and dry bias in the AIRsonde data.  The temperature data are shown as dry bulb temperature, since that variable is measured directly by both the tethersonde and the radiosondes.  The humidity data is shown as specific humidity since the tethersonde measures wet bulb temperature, and the radiosonde measures relative humidity.  For data collected after June 25, 1997, the average difference in dry bulb temperature was found to be 2.5 K, and the average difference in specific humidity was 5.1 g/kg.

    In an attempt to derive a bias correction procedure for the temperature and humidity data, an experiment was performed in the metro Atlanta area from July 9-12, 1998.  Pressure, temperature and humidity data was collected using both AIRsondes as well as a surface meteorological station.  Again, the first AIRsonde reproduced the surface meteorological observations quite well while the second AIRsonde displayed differences of the same sign but different magnitude as compared to the differences between the tethersonde and radiosonde during SGP97.  Since the differences were not of the same magnitude as those recorded during SGP97, this approach was abandoned.

    The final bias correction procedure was derived using the results from a regression analysis in which the tethersonde dry bulb temperature and specific humidity profiles were corrected to match radiosonde profiles from the launch closest in time to the tethersonde launch. This regression analysis was justified on the basis that features such as inversion height were quite similar for the tethersonde and radiosonde profiles, as shown in Figure 4. Because the radiosonde data has a vertical resolution of approximately 10-15 meters, it was linearly interpolated to the pressure levels of the tethersonde data to provide a collocated data set.  Possible predictors were tethersonde dry bulb temperature, potential temperature, relative humidity, specific humidity and pressure.  The software used in the regression analysis was Minitab 10.5 for Windows.

    Temperature Correction

    Both the tethersonde instrument and radiosonde measure dry bulb temperature directly.  Therefore, we chose the radiosonde dry bulb temperature as the dependent variable, and found the best fit to be a simple linear regression model using only tethersonde dry bulb temperature as a predictor:
  TDTC = 1.82 + 0.994 TDT

where TDTC is the corrected tethersonde dry bulb temperature and TDT is the measured tethersonde dry bulb temperature.  This equation has an adjusted correlation coefficient of 86.8, a Mallows (cp) statistic of 931.8, and a root mean square error of 1.82 K.  Other predictors were also used in a multiple linear regression which yielded no significant improvement over the above correction formula.

    We applied the above correction to the all tethersonde data collected after June 25, 1997, and Figure 6 shows the comparison of corrected tethersonde data versus radiosonde data for those times when both are available.  An example of the corrected dry bulb temperature profiles is shown in Figure 7, and an example of corrected potential temperature profiles is shown in Figure 8.

   Humidity Correction

    The tethersonde instrument measures wet bulb temperature and then solves for relative humidity using both wet and dry bulb temperature.  The radiosonde measures relative humidity directly.  Rather than use a relative measure of atmospheric moisture, we chose specific humidity as our dependent variable.
    We attempted simple and multiple linear regression, nonlinear regression, and artificial neural network techniques on the specific humidity data.  The nonlinear regression approach provided the best results, and the correction equation is:
  TSHC =  4.14 + 45.7*(1/TWT) - 164*(1/TRH) + 0.155*(TWT^2) - 0.232*(TSH^2) - 0.00498*(TWT^3) + 0.0123*(TSH^3)

where TSHC is the corrected tethersonde specific humidity, TWT is the measured tethersonde wet bulb temperature, TRH is the measured tethersonde relative humidity, and TSH is the measured tethersonde specific humidity.  This equation has an adjusted correlation coefficient of 74.1, a Mallows (cp) statistic of 362.5, and a root mean square error of 1.25 g/kg.  We applied this equation to all data collected after June 25, 1997, and the results are plotted in Figure 9.  An example of the corrected specific humidity profiles is shown in Figure 10

The Files

Size and Format

The tethersonde data set consists of 150 ascii text files and eight plot files in postScript format (which may also be downloaded in .gif format from this document).

Average file sizes are about 11.5 KB for the data files and 38 KB for the plot files. The data set requires approximately 2 MB of disk storage.

Name and Directory Information

The Tethersonde data reside on disc.gsfc.nasa.gov in directory

/ftphttp://disc.sci.gsfc.nasa.gov/data/sgp97/atmos_sounding/tethersonde

The tethersonde file names are of the form

sgpmddhhmm.x.txt

where

sgp = "Southern Great Plains", campaign identifier;
mddhhmm = month, day, hour,minute;
x = "a" for ascending or "d" for descending pass;
txt = ascii text file

The Science

The Plan

In support of the eventual goal to integrate remotely sensed observations with coupled land­atmosphere models, Georgia Institute of Technology and the National Severe Storms Laboratory provided vertical profiles of atmospheric pressure, temperature, humidity, wind speed and wind direction during the Southern Great Plains 1997 field experiment (June 18­July 17). Our sounding design was based on three science needs directly related to the existing objectives of the experiment:

1. Provide boundary and initial conditions for coupled atmospheric­hydrologic modeling;

2. Provide data necessary for atmospheric correction of thermal remote sensing; and

3. Support water vapor and heat budget computations over the SGP97 domain.

In addition to these science needs, surface and boundary layer profiles provide data to support the estimation of surface layer similarity parameters and to study boundary layer top entrainment processes particularly during the morning transition.

Results

Much of the work to this point has focused on the use of  1-D boundary layer heat and vapor budgets to estimate regional latent and sensible heat fluxes.  Early results from these analyses were presented at the Spring 1998 AGU meeting (Davis et al., 1998) and will be presented in more detail at the 1999 AMS Conference on Hydrology (Davis et al., 1999).

Other ongoing work includes estimating roughness lengths and analysis of the morning transition.  

Data Access and Contacts

FTP Site The SGP97 Tethersonde data is in the following GES DISC ftp site:

SGP97 Tethersonde Data

Points of Contact

The principal investigator for the Tethersonde data is

 

Christa Peters-Lidard

Environmental Hydraulics and Water Resources

School of Civil and Environmental Engineering

Georgia Institute of Technology, Atlanta, GA 30332-0355

E-mail: cpeters@ce.gatech.edu

Voice: 404-894-5190

Fax: 404-894-2677

For more information regarding GES DISC data, contact:

Hydrology Data Support Team

Goddard Earth Sciences
Data and Information Services Center (GES DISC)

Code 610.2

NASA Goddard Space Flight Center

Greenbelt, Maryland 20771

E-mail: hydrology-disc@listserv.gsfc.nasa.gov

Voice: 301-614-5165

Fax: 301-614-5268

References

Davis, L. H., C. D. Peters-Lidard, and L. C. Showell, 1999:  Estimating Regional Evapotranspiration using an Atmospheric Boundary Layer Conservation Approach.  Preprint volume:  14th Conference on Hydrology, American Meteorological Society, Dallas, Texas.

Davis, L. H., C. D. Peters-Lidard, and L. C. Showell, 1998:  Estimating Regional Evapotranspiration using an Atmospheric Boundary Layer Conservation Approach.  Poster presented at:  American Geophysical Union Spring Meeting, 26-29 May 1998, Boston, Mass.

---

Last update:Thu Jan 28 09:34:12 EST 1999 Page Author: Hydrology Data Support Team -- hydrology-disc@listserv.gsfc.nasa.gov Web Curator: -- Website Curator: Anthony Drake NASA official: Steve Kempler, DAAC Manager -- Steven.J.Kempler@nasa.gov