EPA-600/4-77-045
November 1977
Environmental Monitoring Series
              VERTICAL  FLUXES  AND EXCHANGE
   COEFFICIENTS  IN  THE  AIR  OVER ST.  LOUIS
                               Field  Program  1975
                                Environmental Sciences Research Laboratory
                                     Office of Research and Development
                                    U.S. Environmental Protection Agency
                               Research Triangle Park, North Carolina 27711

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology.  Elimination of traditional grouping  was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.   Environmental Health Effects Research
      2.   Environmental Protection Technology
      3.   Ecological Research
      4.   Environmental Monitoring
      5.   Socioeconomic Environmental Studies
      6.   Scientific  and Technical Assessment Reports (STAR)
      7.   Interagency Energy-Environment Research and Development
      8.   "Special"  Reports
      9.   Miscellaneous Reports

This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and  instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations.  It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia  22161.

-------
   VERTICAL FLUXES AND EXCHANGE COEFFICIENTS
           IN THE AIR OVER ST.  LOUIS
              Field Program 1975
                      by
               Bernice Ackerman
          Illinois State Water Survey
            Urbana, Illinois  61801
            Grant Number:  R803682
                Project Officer

               James L. McElroy
Environmental Monitoring and Support Laboratory
           Las Vegas, Nevada  89114
  ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
      OFFICE OF RESEARCH AND DEVELOPMENT
     U.S. ENVIRONMENTAL PROTECTION AGENCY
      RESEARCH TRIANGLE PARK, N.C.  27711

-------
                                 DISCLAIMER
     This report has been reviewed by the Environmental  Sciences  Research
Laboratory, U.S. Environmental Protection Agency,  and approved for publi-
cation.  Approval does not signify that the contents necessarily  reflect
the views and policies of the U.S. Environmental  Protection Agency, nor
does mention of trade names or commercial products constitute endorsement
or recommendation for use.
                                    ii

-------
                                 ABSTRACT
     A field program was carried out in the greater metropolitan area of
St. Louis, Missouri during February and July of 1975 as part of the
Regional Air Pollution Study  (RAPS).  The purpose of the program was to
collect atmospheric measurements needed for future studies of the planetary
boundary layer  (PEL) over urban and industrial areas and surrounding rural
areas.  The overall goals of  the PEL study are to (1) describe the thermo-
dynamic, wind and turbulence  fields over the region; (2) determine the
magnitude and vertical variation of the vertical fluxes of heat, moisture
and momentum as a function of land use; (3) obtain estimates of the
exchange coefficients of these variables; and (4) determine the dependence
of turbulence intensity on land use.

     Three measurement systems were used:  1) a network of double-theodolite
pilot-balloon stations; 2) two tethered-balloon sounding stations; and
3) an aircraft  instrumented for air motion measurements.  The pilot-balloon
stations provided simultaneous measurements of the wind profile with
vertical resolution of about  50 m up to 2 km from five or six locations
in the area.  The tethered-balloon sounding systems yielded thermodynamic
and wind profiles, with vertical resolution of 20 m from surface to about
500 meters.  The instrumented airplane provided measurements of the three
components of wind velocity and of high frequency fluctuations in velocity,
temperature and humidity.

     The observational periods were scheduled for 3 or 4 hour durations
during field experiments, or  missions, in which all available measurement
systems were operated in modes to best attack particular experimental
objectives.  The mission objectives served the overall goals listed above.
They were (a) mapping missions to delineate the thermodynamic, wind and
turbulent fields over the region, (b) flux missions to provide estimates
of the true vertical fluxes of momentum, heat and moisture simultaneously
with vertical profiles of these variables, and (c) nocturnal missions to
provide information on the strength of the nocturnal heat island circu-
lation.

     About 1000 wind profiles were obtained from the pibal wind measuring
network, and the tethered-balloon systems yielded over 200 good thermo-
dynamic and wind profiles.  One hundred hours of scientific data were
collected during 25 flights with the instrumented airplane.  All of these
data have been processed and  are stored on magnetic tape for further
processing and  computer analysis.
                                     111

-------
     This report was submitted in fulfillment of Grant Number EPA 803682
by the Illinois State Water Survey under the sponsorship of the U.S.
Environmental Protection Agency.   This report covers the period February 1,
1975 to November 15, 1976 and work was completed under this grant as  of
December 20, 1976.

-------
                               CONTENTS

Abstract	iii
Figures	vii
Tables	  ix
Acknowledgements	x

   \.  Introduction	1
          Background	1
          Objectives	2
   2.  Facilities	4
          Aircraft	4
          Pilot balloon measurements	8
          Tethered balloons:  Boundary layer profilers	11
          Operations base	14
   3.  Operations summary	15
          General field procedures	15
          February program	18
          July program	18
   4.  Data summary	21
          Aircraft data	21
          Pilot balloon wind measurements	21
          Boundary layer profiler measurements	25
          Status of data reduction	26
   5.  Data Sample and Research Projections	27
          Nocturnal Observations July 26-27, 1975	27
          Nocturnal Observations July 25-26, 1975	36
          Summary	38

References	41
Appendices

   A.  Availability of data:  boundary layer profiles	42
   B.  An investigation of the errors in estimating wind
          velocity by double-theodolite pilot-balloon observations....47

       Introduction	47
       Test procedures	49
       Error analysis	49
          Error in wind direction	51
          Error in wind speed	51
          Vector error	56
                                    v

-------
                             CONTENTS
         Error in calculated balloon ascent rate	56
         Error in balloon height	58
    Discussion	58
    References	60
C.   Illinois State Water Survey Data Policy 	,	61
                                vi

-------
                                FIGURES

Number                                                             Page

1.  Photographs of the airplane and instruments	6

2.  Map showing location of observing sites	,	9

3.  Photographs of boundary layer profiler and sensors	13

4.  Typical flight patterns	16

5.  Root-mean-square errors in the pibal measurements	24

6.  Average wind speed and direction during special tests of the
    pibal systems	24

7.  Time-height analyses of temperature at the two profiler stations
    and of the urban-rural temperature difference on the night of
    July 26-27	29

8.  Simultaneous temperature profiles at downtown city and rural
    stations	30

9.  Time-height analyses of mixing ratio at the two profiler stations
    and of the urban-rural difference on the night of July 26-27	31

10. Time-height analyses of the wind speeds and direction, from
    double theodolite observations on the night of July 26-27	32

11. Time-height analyses of wind speed at the two profiler stations
    on the night of July 26-27	35

12. Time-height analyses of temperature at the two profiler stations
    and of the urban-rural temperature differences on the night of
    July 25-26	37

13. Time-height analyses of mixing ratio at the two profiler stations
    and of the urban-rural difference on the night of July 25-26	39

14. Time-height analyses of wind speed at the two profiler stations on
    the night of July 25-26	40
                                  v

-------
                                   FIGURES


A-l.   Locations of observing stations	46

B-l.   Mini-networks used in tests of accuracy of the double theodolite
      pibal systems	50

B-2.   Profiles of average wind direction on three test days	52

B-3.   Profiles of average wind speed on three test days	53

B-4.   Profiles of estimated error in wind direction, for each test
      day	54

B-5.   Profiles of estimated error in wind speed and the vector error in
      wind velocity for each test day	55

B-6.   Profiles of average balloon ascent rate and of estimated error in
      ascent rate for each test day	57

B-7.   Profiled of estimated error in balloon height estimated for each
      test day	59
                                     viii

-------
                                    TABLES







Number                                                                  Page




   1.  Pilot balloon stations 	 10




   2.  Boundary layer profiler sensor characteristics 	 12




   3.  February operations 	 18




   4.  July operations 	 20




 A-l.  Pilot balloon measurements available for February 1975 	 43




 A-2.  Pilot balloon measurements available for July 1975 	 44




 A-3.  Availability of data from boundary layer profilers 	 45

-------
                              ACKNOWLEDGMENTS

     A field program  the size  of   that   carried   out   during   1975   can be
 successful  only  through  the  dedicated  efforts of many  people.   The
 contributions of the  many who assisted in  this  program and in the massive
 processing  task are gratefully acknowledged.

     Mr.  J.  William Mansell supervised  the field operations of the  pilot
 balloon  component and, with  the able assistance of Mrs. Yueh Liu,  the data
 reduction.  The  field  observations  were  made by the  5th Weather Squadron
 the  Air Weather Service, U. S. Air Force under the supervision of M.  Sgt.
 Lewis  Jones  in February and of M. Sgt.  Joe Markham in July.   There would,
 of course,  be no data without the  efforts  of the  20 enlisted observers who
 spent so many hours driving  through  St. Louis traffic and peering through
 theodolite  telescopes.

     Mr. Gregory Fetter  and Mr.   Gregory  Dzurisin  operated  the Boundary
 Layer  Profilers during the  July   program  and  Dr.  Peter  H.  Hildebrand
 supervised  the development  of the  computer  processor and  of  the  data
 reduction of these measurements.  Mr. R.  B.  McBeth  and his colleagues at
 the  Field  Observing  Facility of NCAR provided  invaluable  cooperation  in
 providing   immediate  assistance  in  the frequent repair of  the  profiler
 sensor packages.

     NCAR pilots, Mr. Tom McQuade and Mr. Clay Orum, masterfully guided the
 airplane through  the  tall towers, aircraft and  deep-blue haze of the St.
 Louis boundary layer, while  maintaining good relations  and cooperation with
 the  FAA  controllers  at the St. Louis control tower, who were  most helpful
 in working  the air  operation  into their heavy work schedule.  Mr. Richard
 Friesen,  NCAR technician, kept the complicated scientific package  running
 with a minimum of down time.

     The personnel of the   Creve   Coeur  offices  of   RAPS  were extremely
 helpful  during both the  pre-program installations and  the data collection.
 It would be difficult to name all those who provided assistance at one time
 or another, but the cooperation of  Frank   Schiermeier  was an important key
 to the success of this effort.

     Credit for the success of this program must go  to all of  these people
 and to many others back at the various home facilities  who  pitched  in to
 respond  to emergency calls.  Most  of all the professionalism and the  most
willing  and good-natured  cooperation  of  all  under  the  fatiguing  and
stressful  conditions  of  a  concentrated   field  program    was  greatly
appreciated.

-------
                                 SECTION 1

                               INTRODUCTION
BACKGROUND

     The dispersion of surface-generated materials and gases in  the   lower
atmosphere is determined by transport with the mean wind  and  by spreading
due  to small-scale turbulence.  The direction of the mean wind  determines
the general  orientation  of_ the pollutant transport, whereas the pollutant
concentration is a function of the variability of the wind, the small  scale
turbulence, and the mechanically and convectively driven vertical currents.

     Urban pollutant dispersion models have proliferated  over the past few
years.  A few are relatively  simple  but  many  have very elaborate source
distributions in space and time.  Virtually  all  have  a major weakness —
the simplicity  of  the  atmospheric module, which frequently consists of a
single wind measurement  and  simple diffusion parameters.  Applications of
these models usually use  estimates  of  the  air  motion  from the surface
boundary layer and from a  single  urban  or  rural site.  However both the
mean and turbulent components of the  wind  vary  with  height and with the
character of the underlying surface and pollutant concentration is strongly
affected  by the wind structure throughout the atmospheric  boundary   layer
over-riding the area of interest.

     The  simplicity  of  the atmospheric modules in most models stems from
lack  of  information,  not  from  lack  of  importance.  Both the mean and
turbulent components of  the  regional  wind  field in the lower atmosphere
must be considered in  models  predicting  regional air quality or regional
climate.  This is particularly true in the case of metropolitan areas  since
the physical structure of cities and  the concentration of human activities
may cause systematic differences in the structure of the planetary boundary
layer, and  thus in the wind distributions over urban and surrounding rural
terrains.

     The mean winds are expected  to  change  in  the  vicinity  of a  city,
first, because differences in the surface  roughnesses  of  urban and rural
terrains modify the frictional force and, secondly,  because  distortion in
pressure  surfaces arising from differences in the thermal  characteristics
of  urban  and  rural  surfaces causes local modification of  the  pressure
gradient  force.    The  magnitude  and  characteristic  of  the  turbulent
component  of  the  air  motion  is  also  expected to vary over  a  region
including an urban  complex.  The mechanical turbulence should vary because
of the  differences  in  the  nature,  size  and  number  of  the roughness
elements, and the convective turbulence because  of urban-rural differences


                                -1-

-------
in  the  low-level  static   stability   stemming  from  surface  temperature
differences.

     There have  been  reports  of  observations  which indicate that the mean
winds do indeed  change as the air  approaches  and passes over a metropolitan
area, both  at   the  surface  (Chandler,  1960;  Findlay and Hirt, 1969), and
some distance into the planetary boundary layer (Angell et al., 1971, 1973;
Acis3rman, 1972,  1974a, b;  Hass,   et al.,   1967).   There have been very few
observations of  the turbulence  in   air   passing  from rural to urban sites,
but the work by  Bowne and  Ball  (1970)  indicates that there are significant
urban-rural  differences  in  turbulence   intensity.    Beyond some  "plume"
studies,  virtually  nothing  is known  of the magnitudes  of  the  vertical
fluxes  and  flux  divergences  —  yet  these are  critically  important  in
determining the  depth of mixing, the dilution,  and large-scale transport of
pollutants.

     All in all, knowledge is fragmentary and there continues to be a great
need for detailed documentation and description  of  the  three-dimensional
airflow in the   planetary boundary layer  over a metropolitan area, if it is
to  be  adequately  treated  in  urban  dispersion   models.   The  following
sections of this report describe a field  program  which  was  designed to
collect the  data  needed  to   develop  such a description.  The program was
carried out in   the St. Louis metropolitan area in 1975 in conjunction with
the Regional Air Pollution Study (RAPS)  and the Metropolitan Meteorological
Experiment  (METROMEX).


OBJECTIVES OF THE STUDY

     The research carried out under this  grant  is  part of an urban boundary
layer program which seeks to describe in  detail the three-dimensional  wind
field  in the planetary boundary layer  over-riding a mesoscale region which
includes an urbanized area.  This  program was initiated in 1971 as part  of
METROMEX and  has  continued  since then,  becoming part of the RAPS program
also in 1975.  It has been supported over the years by the National Science
Foundation  (NSF), the Atomic Energy Commission   (AEC),  the Energy Research
and Development  Administration  (ERDA),  the National Center for  Atmospheric
Research (NCAR), and the United States  Air Force (USAF).   Under this grant,
the Environmental Protection Agency has provided support for the collection
of boundary layer measurements  during the 1975  RAPS winter and summer field
expeditions and  for the basic reduction of the  data collected.

     The overall urban boundary layer program  at   the Illinois State Water
Survey has the following objectives:

     1.     Description of  the  wind, temperature, humidity, and turbulence
           fields  over   the   city    and surrounding  country  side  and
           delineation of  the  perturbations  induced  by  the city on the
           ambient fields.

     2.     Determination   of   the vertical  fluxes  of  momentum,   heat,
           moisture, and aerosols  in the Ekman   layer, and their variations
           with  height and land use.


                                 -2-

-------
     3.     Estimation of  the  exchange coefficients of the four parameters
           given in (2) and their variations with height and land use.

     4.     Determination of the variations in  turbulence  intensity and in
           the input scales with land use.

     The  field  efforts  of  this  1975  project addressed  all  of  these
objectives,  with particular emphasis on the second and third.   They  were
implemented during the periods of 15 through 28 February and  1  through 30
July so as to study conditions in both winter and summer.
                                  -3-

-------
                                 SECTION  2

                                FACILITIES

     Observations were made with three  measurement   systems:   a network of
double  theodolite  pilot balloon stations, an   aircraft   instrumented  for
measurements of air motion, and (during July  only)   two   tethered  balloon
sounding  stations.   Supplementary routine measurements  are  available from
the EPA/RAPS radiosonde and tower networks, from the  Illinois   State  Water
Survey  METROMEX  surface  networks  and  from the St.  Louis  Air  Pollution
Control  and Illinois  EPA  networks.   During   some   experimental  periods
additional  data  are   available   from  other   RAPS  and METROMEX  field
experiments,

AIRCRAFT

     An instrumented aircraft, pilot, and technician  were provided  by  the
Research Aviation  Facility of NCAR.  In  February the airplane  was based at
Lambert Field; in  July  it  was  based   at  Alton  Civic Memorial Airport,
Illinois.

     The airplane was the Queenair 306  which  was  instrumented to provide
measurements  of  three  components  of wind velocity  and high  frequency
fluctuations  in  velocity, temperature and humidity,  as  well   as  standard
state parameters and aircraft position.   Equipment to measure concentration
of  total  condensation  nuclei,  obtained on loan from EPA-Las Vegas,'  was
added during the July program.

     Measurements  of  the  turbulent  or fluctuating  components  of  the
atmospheric  structure were obtained using rapidly responding   instruments,
most  of which were mounted on a boom extending  about 3 meters  out in  front
of  the  aircraft.   The  remaining  instruments  were    either   wing-   or
fusilage-mounted.

     The airplane instrumentation  permitted  measurement  of the following
atmospheric, surface, and aircraft parameters:

     1.   Air Temperature
           a.     Slow  response  (a  few  seconds)  —  NCAR    reverse-flow
                 thermometer.
           b.     Medium  response  (a few tenths  of a second)  —  Rosemount
                 platinum resistance thermometer.
           c.     Fast response (a few hundredths  of a second),  fluctuations
                 only — NCAR "K-probe",  platinum resistance  thermometer.


                                -4-

-------
     2.    Dew-point  temperature  (slow  response   —  about  2  seconds),
           Cambridge dew-point hygrometer.

     3.    Atmospheric  refractive  index (response, a few hundredths  of  a
           second).  NCAR microwave refractometer.

     4.    Cloud  liquid water  content  —  Johnson-Williams  liquid  water
           content meter.

     5.    Concentration of condensation nuclei (July only) -- Environment I
           nuclei counter.

     6.    Surface Characteristics
           a.    Radiation-temperature (analog IR), Barnes PRT-5.
           b.    Visual, aerial  color photography — downward-viewing time
                 lapse movies, 4 sec/frame.

     7.    Altitude
           a.    Static pressure probe (transducer).
           b.    Geometric  altitude  above  ground (for altitudes of  less
                 than one kilometer only) -- Radio altimeter.

     8.   Airspeed (dynamic pressure)
           a.    Wing-mounted Rosemount pitot.
           b.    Boom-mounted pitot and transducer.

     9.    Orientation to airstream (boom-mounted)
           a.    Angle of attack, fixed vane.
           b.    Angle of attack, rotating vane.
           c.    Sideslip angle, fixed vane.

    10.    Boom  accelerations —  vertical   and   lateral  accelerometers,
           boom-mounted.

    11.    Orientation  to  fixed system, aircraft attidude  angles,  pitch,
           roll, and yaw.

    12.    Aircraft velocity  relative  to  ground  — Inertial Navigational
           System  (INS)
           a.    Heading
           b.    Ground speed
           c.    Vertical velocity
           d.    Vertical acceleration

    13.    Aircraft position
           a.    INS-computed latitude and longitude.
           b.    Distance to several DME stations (July only).

In Figure 1 are  photographs of the airplane and externally-mounted sensors
or sensor housings.

     All  measurements  (except  the  aerial  photography of  course)  were
recorded on magnetic tape.  All analog outputs were  recorded  at  16 times

                                  -5-

-------
i
!~
I
                                                    GUST
                                                    jPROBE
                      MICROWAVE
                    .REFRACTOMETER
WING-MOUNTED
 INSTRUMENTS
                                                                  FIXED VANE:
                                                                  angle of attack
                                                                                                         ROTATING VANE:
                                                                                                          angle of attack
                                                                                DYNAMIC
                                                                                PRESSURE
                                                                                               IFAST RESPONSE
                                                                                               THERMOMETER
                         MICROWAVE
                       REFRACTOMETER
                                                                    PITOT
                                                                   (dynamic
                                                                   pressure)
                                                         THERMOMETER
                                                                     LIQUID WATER
                                                                     CONTENT METER
                                                                                                REVERSE FLOW
                                                                                                THERMOMETER
                                c                                                        d
     Figure 1.   (a)  The NCAR  Queenair 306  instrumented  to collect  air motion measurements.   (b) Closeup of
     the  gust probe.   (c) Closeup  of the micro-wave refractometer.   (d) Closeup of wing-mounted instruments.

-------
per second; digital  outputs  (e,g,  from the INS) were  recorded  at  8 or 10
times per sec.   Some of the  basic  instruments were fitted  with  electronic
filters  having  4   Hz   bandwidth.    During the basic processing, which was
provided  by  NCAR,   all  measurements  were  brought  to a common rate and
frequency response through a series of cubic interpolations and application
of  a  "matching"  digital filter.    The  "standard  rate"  processed  data
provided by NCAR on magnetic  tape  were recorded 8 times per second, and
1-second (8 point) averages   at   a one second rate.  However,  all data have
been  filtered   so   that  the measurements  contain information  only  for
frequencies of 4 Hz  or  lower.

     The fast response  boom-mounted  dynamic  pressure and temperature, and
refractometer measurements were  also recorded,  before electronic filtering
at high rates of 256, 128,  and   64  times per second, respectively.   These
data  were  scaled   and recorded,  without any  digital  filtering,   at  the
original  rates.   This  permits  study of the fluctuations  in  total  air
motion, temperature  and humidity in the high frequency domain.

     Some  of  the parameters to be analyzed were measured directly;  others
were calculated  from the measurements from several sensors.  The parameters
required for  the study are  listed below.  The numbers in parentheses refer
to the basic measurements listed above which enter into their calculation.

           a.    Aircraft position,  in  the  horizontal, by integration of
                 measurements from  (12),  with update from (6b) and (13b)
                 when  available;    height   from  (7b)  below  900  m  and
                 calculated  pressure altitude from (7a) at all altitudes.
           b.    Surface characteristics and land use — qualitative  (6b).
           c.    Surface temperature — computed from  (6a)  assuming black
                 body radiation.
           d.    *Absolute air temperature (Ib), with (la) as backup.
           e.    *Absolute air humidity (2).
           f.    Pressure changes   (spatial)   below  900  m  AGL:  D-values
                 computed from (7a) and (7b).
           g.    Turbulence  intensity: (8b).
           h.    Temperature fluctuations: (Ic) and (Ib).
           i.    Humidity fluctuations — computed from (3),   (Ic),   (7a),
                 (8b).
           j.    Wind components,  u, v, w, computed  from (11),  (12),  (8a),
                 or  (8b), (9a or b), (9c),  with corrections based on (7a),
                 (10).
           k.    Turbulent fluxes  of momentum, heat, and moisture,  based on
                 auto-correlations  of derived  parameters (h), (i), and (j)
                 above.
           1,    Spectra  and co-spectra of fluctuations  in  total   wind,
                 temperature, humidity, and vertical wind component,   based
                 on  derived  parameters (h), (i), (j) listed  above and from
                 (8b).
   * Absolute  used here in sense of "mean" value, as opposed to  fluctuating
component.

                                  -7-

-------
PILOT-BALLOON MEASUREMENTS

     Wind  measurements  were  obtained   up    to    about    2000    m  using
double-theodolite  techniques.   These   involve tracking   a  pilot  balloon
(pibal) with theodolites from two well-separated locations.

     The personnel required to make these observations were provided by the
Air  Weather Service (AWS) of the U. S.  Air  Force   from   the  6th  Weather
Squadron  (Mobile).   During  February the AWS  unit  consisted of  6  two-man
double  theodolite teams plus two supervisors.   During July,  only 5 two-man
teams and one  supervisor  were  available  for this assignment.   Standard
equipment required for the  observations  plus   the  vehicles  necessary to
transport personnel and equipment to  observation  sites were also provided
by the AWS.

     Personnel  and the pibal operations center were based  at  Scott  Air
Force Base,  Illinois,  about 32 km ESE  of downtown  St. Louis.  During each
operational period (see Section 3) the observers were deployed to specified
sites  scattered   in  and   around  St.  Louis.   The  sites  varied  from
operation-to-operation depending on the  mission for the  day.    These were
selected from 12 previously surveyed locations  (Figure 2).

     The  12  pibal sites were established on the  basis   of  a  number  of
criteria, some scientific, some logistic.

           a.    Sites should be arrayed such that experiments  pertinent to
                 the objectives could be carried out.
           b.    The  area  around  each  site   should  have,  as  much  as
                 possible, homogeneous surface  conditions.
           c.    The  baseline  (distance  between the observers)  should be
                 600 m or longer.
           d.    The two observers should be able to see each other through
                 their theodolites,  and it  should be possible to string
                 telephone line between  them   without  crossing   walks  or
                 heavily traveled roads.
           e.    There should be no obstructions to  the view of the balloon
                 in any direction.
           f.    Personal safety of the  observers must not be endangered in
                 any  way, and phone and toilet facilities must  be  within
                 reasonable walking distance.
           g.    Equipment could be set  up in a reasonable length of time.
           h.    The site must be within 1-hour driving time from Scott AFB
                 (an AWS regulation).
           i.    Permission to use the   property from  the  owner  or  the
                 responsible public official must be granted.

     It was not always possible to meet  all of  these criteria.  Those which
were essential and had  to  be  met  were   (a),  (d),   (f),   (h),  and (i).
Criterion (c)  was  met  in  seven of the sites. The shortest baseline was
about 378 m  long,  and  the longest was nearly 800  m in length.   The other
criteria were met with  varying degrees  of success.  The most difficult one
to satisfy was (b), particularly  at  sites in  the city.   In fact, even for


                                   -8-

-------
Figure 2.  Map of St. Louis showing the location of the pilot balloon sites
(numbered) and the boundary-layer profiler stations (BP1 and BP2).
                                   -9-

-------
criterion (a), the locations,  although   suitable, were not  always  optimal.
The locations and baseline lengths of the 12  sites are given in Table  1.
          TABLE 1.  DOUBLE-THEODOLITE PILOT-BALLOON STATIONS USED
                               DURING 1975.
     Station                                               Baseline
     Number	Location	length  (m)
         33      East side of main runway, Scott Air Force
                 Base, IL.	657

         35      Troy Road, just south of 1-270, 4 km east
                 of Glen Carbon, IL.	763

         44      South side of parking lot, Korvette Shopping
                 Center, Rott Road and Hwy. 61, Sunset Hills,
                 MO.			611

         70      Triad High School, Hwy 40, 3 km west of
                 Hwy. 4, IL.		-			449

         71      Poag Road, southwest of Poag, IL.	790
         72      Brown Road, opposite northeast corner of
                 Lambert Field, Berkeley, MO.	 457

         74      Granite City Army Depot, Granite City, IL.	 795
         75      Army Publication Center, Woodson Road
                 and Page Ave., Overland, MO.	552

         76      County Road, 6 km north of O'Fallon, IL.	640

         77      Peabody Road, just west of Hwy. 159, about
                 8 km south of Belleville, IL.	663

         78      Along taxiway north side of Bi-States
                 Airport, Cahokia, IL.	 541
         79      Forest Park, along Hwy. 40, east of Hampton
                 Avenue, St. Louis, MO.	378
     When establishing  the  site,  the  survey  team selected two suitable
spots for the  theodolite  "pads", requiring that the two be  line-of-sight.
They then measured off  the  straight  line  distance  between the two with
transit and tape and the  difference  in  elevation with a stadia rod.  The
orientation of the baseline relative to  true  north  was  determined later
from sightings on Polaris on clear nights.


     Almost all of the pibal sites were on unprotected  public  and private
property,  so  that only small wooden ground stakes marking the  theodolite
locations  were  located  permanently  at  the sites.  The observers had  to

                                   -10-

-------
transport, set  up,  and  then   dismantle  all   equipment  for  each  day's
observations.   This took roughly  an hour at  both the start and the end  of
the day, in addition to travel  time.

     Two observers, one at  each theodolite,   composed a pibal  "team".   They
were linked by a  land  phone   line  for  communication.   Into this  line were
hooked a  tape recorder and a  tone generator  that was activated by a timer.
These, along with  pertinent   electrical  connectors   were  packaged into a
small attache case for  convenience  of   transport and storage.  The timing
tone was set for  a  timing  interval of 20 seconds and sounded  for  roughly 3
seconds.  At the  start of the   tone  the  observers centered the balloon  on
the cross-hairs of the theodolite  and kept it   there until the tone ended.
Then, before changing the dial  settings, each observer,   in turn,  read the
azimuth  and elevation angles  of the balloon  (to hundredths of a  degree)
into their head phones for  recording on  a cassette tape.    Azimuth  angles,
as  given,  were  in  the   coordinate system  with y-axis parallel to the
baseline.  Rotation  of the coordinate system to true north is accomplished
in the data reduction.

     Pilot balloons were inflated  (to within  the accuracy of the  gas-flow
meters) for estimated rise  rates   of about  2.2  to 2.5 meters per second
(mps).  Flow meters, rather than weights  were  used in inflation because  of
the short time between the  termination of one run and the start of the next
and the difficulty  of  inflating  balloons in the open.   Although  inflation
rates were assigned, and  were   adhered  to as closely as  possible,  this was
not an essential  factor since   the  computation  of  the   wind  velocity  is
independent  of   the ascent rate of  the  balloon  when  double theodolite
measurements are  available.  In fact,  the rise   rates  are  calculated and
provide  valuable  estimates of the vertical  air motions   in  the   boundary
layer.

TETHERED BALLOONS:  BOUNDARY LAYER PROFILERS

     Two tethered-balloon systems  for measuring  detailed profiles up to a
nominal operational altitude of 750 m were loaned to  the  State Water Survey
for the July  field  effort by  the Field Observing Facility of the  National
Center for Atmospheric Research.   One profiler   (BP1)  was  located  at the
RAPS radiosonde station 141 in   downtown  St. Louis and the second (BP2)  at
Triad High School on Highway 40 about 2.5 km  west of  Highway 4 (Figure 2).

     Each balloon system required   a senior operator  and  an assistant.  The
operators were State Water  Survey  staff  who were quartered at  Collinsville,
Illinois, about midway between  the two  sites.    The   assistants were  local
area college students.

     Details of the system  may  be  found  in a  paper by Morris  et al.,  1975.
Briefly,  the  system  consisted  of  a   relatively  small  plastic balloon
(inflated  volume  about 3.25 m3)  with an aerodynamic shape from which was
suspended  the sensor package.   The balloon was tethered  to the surface and
was raised  and   lowered  using an  electric  winch   in   order to  obtain
measurements in the vertical.

                                    -11-

-------
            TABLE 2,  NCAR  BOUNDARY  LAYER PROFILER SENSORS AND
                            THEIR  CHARACTERISTICS

Variable
Pressure
Temperature
Wet-bulb
temperature
Sensor
Sealed aneroid capsule
Ventilated bead ther-
mister in radiation
shield
Wick-covered thermis-
ter, mounted 2-cm
behind dry temper-
ature sensor
Precision
+1 mb
+0.5°C
+0.5°C
Resolution
0,5 mb
0.1°C
0.1°C
     Wind speed     3-cup anemometer,
                    small, light-weight      +.25m/s        0.1 m/s
                    cups  (min. speed:
                    0.5 m/s)

     Wind           Magnetic compass,
     direction      based on assumption      +5°             2°
                    that balloon is effec-
                    tive vane
     The sensor package provided measurements  of pressure,  dry and wet  bulb
temperature,  wind  speed  and  wind  direction.   The   sensors  and their
precision and resolution are given  in  Table   2.    However,   this is a new
system, and these estimates of precision may prove optimistic,  particularly
those for the wind measurements.  The airborne package   also  carried a 403
MHz transmitter and circuitry to condition  sensor output  and   modulate the
transmitter.    The   ground   station   consisted    of    a   receiver   and
Esterline-Angus chart  recorder.   Photographs of   the   balloon and sensor
package are shown in Figure 3.

     The recording was in a time-multiplex  data format with cycling through
8 channels  in  20 seconds.  As operated during the  July field program,  dry
and  wet bulb temperatures  and  wind  speed   were   each  recorded  on   two
channels.   Thus, these variables were sampled twice  every cycle  whereas
pressure and wind direction were sampled once.   Each channel recorded for
approximately 1.5 second.

     The maximum height actually  reached   by   the profiler depended on the
stability and wind speed, decreasing  with  increasing   stability  and  wind
speed.    The balloon was let out at 150-feet per minute  and brought back to
the surface at nearly twice that rate.   This   equipment cannot be operated
under strongly turbulent conditions or in wind speeds of  greater  than 10
m/sec.

                                   -12-

-------

               c                                    d
Figure 3.  (a) Boundary layer profiler tethered in parked position.  (b) Close-
up of sensor package showing harness by which it was attached to balloon.
(c)  Sensor package.  The rectangular base box contains electronics and radio
transmitter.  (d) End-on view of thermister housing.
                                   -13-

-------
OPERATIONS BASE

     The Principal Investigator under this  grant  directed  the  day-by-day
operations during both field efforts.  She  also  served  as  observer   and
technician on the airplane.

     During the February program, the RAPS offices in Creve Coeur, Missouri
served  as  a  base  of  operations.  Following the RAPS weather  briefing,
decisions as to mission operations were relayed to the pibal unit  at Scott
Air Force Base and to the pilot at Lambert Field by telephone.

     During the July program, the State Water Survey METROMEX base at Alton
Civic Memorial Airport served as base  of  operations.  Facilities included
facsimile  weather  maps  and  Service  A  teletype.   Additional   weather
information was obtained by phone.  Decisions as  to the planned operations
were relayed  to the RAPS headquarters and to FAA (necessary because of the
profilers) and the three observing components were alerted by telephone.
                                 -14-

-------
                                  SECTION 3

                            OPERATIONS SUMMARY

GENERAL FIELD PROCEDURES

     The operations were usually  scheduled as  3-  to   4-hour  field  experi-
ments, or missions, in which  all  available systems were  operated in  manners
to best  attack particular  experimental  objectives.   The mission objectives
served the overall goals listed in  Section 1.

     Missions fell into the following general  categories:

1.   Mapping Mission:

        Objective:   To  map   the  fields   of  temperature,  moisture,  wind,
     vertical  velocity, turbulence intensity,  and vertical  fluxes over the
     region.

        Flight Pattern:  "Checkerboard"  pattern  at   one or two  levels over
     and around the greater  St.  Louis  area,  with tracks oriented parallel
     and perpendicular to the wind  (Figure 4a).

        Pilot Balloon Array:  Five or six sites  activated in  as uniform an
     areal   distribution   as feasible.    Balloons   launched  at    20-min
     intervals,

2.  Flux Cross-Sections:

        Objective:    To  estimate   vertical   exchange    coefficients   for
     momentum, heat and moisture  and their variation  with height over urban
     and rural surfaces.
        Flight Pattern:  Two  "cross" patterns   with   arms about  15 to 20 km
     long, one parallel and  the  other perpendicular  to  the mean wind,  One
     cross pattern was over a rural  area  and the other over metropolitan
     St. Louis (Figure 4b).   During July,  the   two  "crosses"  were centered
     near the two boundary-layer  profiler sites;  in February they were near
     an urban and a rural pibal site.  Traverses  were made  at  three  levels,
     and the entire pattern repeated at  least  once.
        Pilot Balloon Array:   Three sites  were  activated,   two  located  in
     "homogeneous"  urban   and rural locations  (in  July at  or near  the
     boundary-layer  profiler sites), and the  third  at   a   downwind  urban
     site.  Balloons were launched  every 10 minutes at   the two key sites,
     and at 20-minute intervals at  the third site.

3.  General Cross Section:
        Objective:  To map, in vertical  planes  across the city parallel and
     perpindicular to the mean wind,  the vertical  fields of temperature,
     moisture, wind, vertical  velocity,  turbulence intensity,  and vertical
     fluxes.
        Flight Patterns  Long traverses  extending across the city and well
     into surrounding areas,  one  parallel  to the  mean   wind direction, and
                                   -15-

-------
Figure 4.  Typical flight tracks on (a) mapping missions,  (b)  flux cross
sections, (c) and (d) general cross sections, and  (e) and  (f)  nocturnal
heat island circulation missions.
                                  -16-

-------
     one or more perpendicular to the wind  (Figures 4c and 4d).  All tracks
     were repeated at three  levels,  and   in   some cases the whole pattern
     repeated.

        Pilot Balloon Array:   Five or six  sites  activated,  in  a  "cross"
     array to the extent possible, with one arm  parallel  and  the  other
     perpendicular to the wind.

4.    Nocturnal Circulation:

        Objectives^   (a)  To check on the  existence of the nocturnal urban
     heat  island  circulation and to measure  its strength and (b) to check
     on the   linearity  of the wind field which is assumed when calculating
     divergence over an area  from  point measurements, and to evaluate the
     error in these divergence estimates.
        Flight Patterns:    Circuits  around   the   inner   city,   greater
     metropolitan area, and an urban area equal in size to the metropolitan
     area, made  at  two  levels  and repeated at least once (Figure 4e) or
     straight tracks between  activated pibal  sites which formed triangular
     areas (Figure 4f).

        Pilot Balloon Array:  Five sites activated to form a box around the
     metropolitanarea and a triangle over a predominately  rural  area.
     (Optimal array not possible because of limited personnel.)

5.  Miscellaneous:
        A  few  special  missions  were  carried  out to address particular
     objectives which  did  not  necessarily require corrdinated use of all
     facilities.   These  were  usually carried out when one or more of the
     facility components was not  available due to malfunction of equipment
     or to unavailability of personnel.

        These missions were of two types:
           a.    Tests and/or checkout of systems.
           b.    Study  of   the  diurnal   evolution  of  the  temperature,
                 moisture, and humidity  profiles  or  evolution during the
                 morning and evening transition periods.   These  utilized
                 only pibals at two sites and  the boundary layer profilers.

     On all missions in July the  boundary  layer profilers were operated in
40-minute sequences, from  their permanent sites.

     Decision as to mission was made by the Program Director  (the Principal
Investigator  under this grant) by 9 a.m. each  morning on  the  basis of the
general weather conditions.   Following  decisions  as  to  timing,  flight
pattern  and  balloon  sites  to  be  used  in the mission, all units  were
notified by   telephone  as  to their particular operations.  This had to be
done  two  to three hours prior to the start of the mission because of  the
travel  and   preparation   time  required  by   the  various    observational
components.   Because of this lead time  and  lack of   field communication,
missions could not be changed when  late changes in weather  forecast or  some
equipment failure indicate a modification might be desirable,


                                   -17-

-------
February Program

     The winter program was carried  out  during   the  last half of February.
The AWS personnel and equipment  arrived  in   the   St.  Louis area on the 12th
and the NCAR airplane and personnel  on the   evening of the 15th,   Rain fell
nearly  continuously  between  the 12th  and  15th,  making it  impossible  to
carry out any operation.  However, this  time was  needed to  modify  some of
the  AWS  equipment  and  to relocate some of the previously located  pibal
sites.  The pibal sites had been  located and surveyed in January and  early
February.  However, much-above normal precipitation in January and February
resulted in very  soft muddy surfaces at some of  these locations,  rendering
them unworkable.

     The weather continued to be  considerably less  than  optimal   for  the
program during the last half of the month, with predominately low  ceilings,
frequent  heavy rains  and  snow,  and   unusually   strong  winds   on  about
one-fourth of the days.  Nevertheless nine missions were carried out (Table
3).  In one instance last  minute  failure of a key piece of the scientific
package on the airplane caused cancellation  of that part of the mission.
                    TABLE 3.  FEBRUARY 1975 OPERATIONS

Date
2/17
2/19
2/20
Mission
Mapping
Mapping
General cross
Airplane
Time*
(Local)
1345—1745
1135—1440
i i on i A A r
Pibal
Time*
(Local)
1220—1520
1 1 cr> i mn
No. of
Stations
6
£.
               section
      2/20    Mapping
      2/21    Mapping; general
               cross section
      2/25    General cross
               section
      2/26    Mapping
      2/27*** Mapping
      2/28    Flux cross
               section
1915—2130

1140—1540  1140—1450

1210—1600  1320—1520

1230—1615  1220—1550**
2045—2358  1210—1510

1325—1650  1230—1520**
3
6
       * CST prior to 23 February, CDT 23-28 February
      ** Pibals on 10-min launch schedule
     *** Flight delayed because of equipment malfunction
July Program

     The NCAR equipment (airplane and  profilers)  and personnel arrived in
St.  Louis on 30 June.  The  AWS equipment and personnel arrived late on the
                                   -18-

-------
evening of 2 July.  The KCAR units   departed  on  31   July and the AWS on 1
August.   The  effective period for full-scale missions  was  5  through  30
July.  Although a slip of  10 days would have  been preferable  in order to
fit  the  dates  of  the   main  RAPS expeditions, NCAR scheduling  made  the
aircraft available to this program  only during the month of July,

     The approach  to  the operations  was  to work the weather,  rather than
any set schedule.   However, in practice, AWS and NCAR guidelines  for  their
personnel necessitated scheduling some  days-off  a day   or  so  in advance.
Although this resulted  in the  loss   of   a few good days,  the  weather was
generally  favorable  and   did   not  seriously  affect  the   overall   data
collection.

     There were few failures  in the   pibal systems  and a relatively  small
fraction of the scheduled  launches  were lost because  of  equipment  problems.
Except  for  an  early  failure in   the inertial  navigation system,   the
scientific package on the  airplane  also presented  few  problems detectable
in  the  field.   However, failures of aircraft equipment   did   force
cancellation of missions on at  least three  occasions.

     Major  equipment  failures did occur however in  the sensor  packages of
the boundary layer  profiler  systems   and   occassionally in the recorders.
NCAR  had  not  provided   substitute parts,  wiring   diagrams,  or repair
instructions, largely because these  were   new  systems  and such things did
not exist.  Consequently   the sensor packages had to  be  returned to NCAR in
Boulder each time there was a   malfunction.    Although  repairs  were   made
immediately  and the malfunctioning instrument was back   within  36 to  48
hours, there was still nearly a 30% loss in desired profiles.

     A total of 25 missions were carried out between  the 1 July  and 31  July
 (Table 4).  Three of these were tests or calibration  missions.
                                   -19-

-------
TABLE 4.  JULY 1975 OPERATIONS


Date
7/01
7/02
7/03
7/05
7/07
7/08

7/09-10
7/10-11
7/12
7/14
7/15
7/16
7/18

7/21
7/22
7/24-25
7/25-26


7/26-27
7/27
7/28
7/29
7/29
7/30



7/31

Mission
Test
Test
Mapping
Mapping
Mapping
Flux cross
section
Nocturnal
Nocturnal
Mapping
Flux cross
section
Flux cross
section
Mapping
P re- rain
mapping
Flux cross
section
Mapping
Nocturnal
Nocturnal


Nocturnal
Mapping
Profile
evolution
Mapping
Flux cross
section &
Profile
evolution
Test
Airplane
Time
(Local)
1510-1627

1420-1806

1255-1700
1218-1550

2133-0010
2132-0025
1300-1700
1250-1627
1319-1705
1142-1435
1106-1457

1230-1620
1247-1702
2133-0020
2100-2230
0010-0130
0300-0425
2116-2342
1300-1450


1350-1525*

1155-1430



* Note airplane flight period
Pibal
Time
(Local)



1100-1400
1320-1640
1230-1530

2140-0020
2140-0020
1340-1620
1310-1610
1310-1610
1200-1500
1120-1420

1320-1620
1320-1620
2130-0010
2120-2400


2140-0020


0600-1500


0600-1500


0920-1145
included
Profilers
No, of Time No. of
Stns. (Local) Stns,



5
5
3

5
5
5
3
3
5
5

3
5
5
5


5


2


2


1

1210-1405
1100-1600
1100-1400
1300-1730
1200-1700

2120-0040


1300-1700
1220-1700
1240-1440
1040-1320

1200-1700
1200-1720
1940-0100
2040-0420


2020-0020
1350-1500
1700-2150
0030-1500


0600-1500




1
2
2
2
2

2


2
2
2
2

2
1
2
2


2
1
1
1


2



in period of pibal and
profiler observations for profile evolution
mission on 7/28
and 2!
            -20-

-------
                                  SECTION 4

                                DATA SUMMARY
AIRCRAFT  DATA

     Basic  processing  of   the  digitally-recorded measurements was done by
the  Research  Aviation   Facility   at NCAR.   Because of a huge  backlog  of
flight  programs  to  be  processed at NCAR,  reduced aircraft data did  not
become available to ISWS  until late Spring and Summer of 1976.

     The  reduced data was recorded  on magnetic tape as described in Section
2, in a convenient form   for further  computer  processing  and  analysis.
These tapes contain scaled values (in  appropriate  units)  of parameters 1
through 13 listed in the  aircraft section of  Section 2.  In some instances,
the measurements  have been  corrected for known performance characteristics
of the instruments  (e.g., temperature has been corrected for compressional
heating of the air).   In addition position, based on integration of track
information from the navigation  equipment,  and the three components of the
air velocity were also provided  at  a rate of  once per second.

     The  voice recordings of observer comments have all  been  transcribed.
Flight  track  information   has  been extracted from the time-lapsed  aerial
photographs for three flights of particular interest for use in  correcting
the computed tracks.  Programs to read the NCAR tapes (generated on the CDC
6600-7600)  on  the  University  of Illinois IBM 360 have been written and
others for initial  "massaging"  of the  data  are  under  development.   A
program for calculating divergence  around  closed  tracks  has been written
and debugged.

     Because of the delay in receiving  the  data  and the mass of numbers
involved, it  is  not  possible  to estimate  the  overall  quality or the
quantity  of useable  data at  the   time  of  this writing.  Review by NCAR
personnel of the processed   data and  a  spot  check  at ISWS after it was
received, have not indicated any unexpected  problems with the basic data.
However during initial exploratory  analyses of two flights, timing problems
have been discovered.   These arose  during generation of computer-compatible
tapes from the air tapes.  The problem is easily corrected but does require
reprocessing  of  parts   of  the flights  by  NCAR.   This  discovery  has
precipitated  review  of  all  25 processed tapes, now underway, to  detect
errors of this type on other flights.

PILOT BALLOON WIND MEASUREMENTS

     The pilot balloon measurements  consisted  of  two  sets of azimuth and
elevation angles to the balloon,  one  from   each  end  of  the "baseline".


                                  -21-

-------
These  angles,  read  from the theodolite scales  and  orally  recorded  on
cassette recorders,  plus the site baseline statistics (length, orientation
relative to true  north,  and  elevation difference in elevation at the two
ends) permit calculation of  the  average  wind for the layer through which
the balloon passes between successive sightings.

     The theodolite readings have been computer-processed to  obtain winds.
The computations were based on the methods described by Thyer (1962).  This
technique  solves  for location of the shortest distance between  the  rays
from  the two theodolites and estimates the most probable position  of  the
balloon  along  the  line  joining the rays at this location.  The  minimum
distance  between  the  two  rays provides an estimate of the error in  the
calculated  position  of  the  balloon  due  to errors in the angles to the
balloon.  Thus, the  technique provides a means for detecting errors in the
theodolite readings and for correcting them.

     The average wind velocity for the layer between two successive reading
times is  provided  by  the  difference  between  the two balloon positions
divided by the  difference  in  the  two  times.  The rate of ascent of the
balloon  is  determined by  the  height  difference  divided  by  the  time
interval.

     During the 1975 operations sightings to the  balloons were recorded at
20-second  intervals,  which nominally permits calculation of average  wind
velocity  for  layers  of about 50 meter depths.  In  actuality  there  was
considerable variation in the depth of the layer covered in  the 20 seconds
both because of inaccuracies in the flow meters used to gauge the amount of
gas  entering  the  balloon  and  because  of  the  frequent  existence  of
atmospheric vertical currents which  provided increments to the ascent rate
of the balloon due to the free lift.

     Because of the 20-minute launch interval the  balloons  were scheduled
for abandonment after 13 minutes of tracking, providing, in still air, wind
data up  to about 2 km.  Although a fair number of the soundings did go the
full distance, many  did not, either because the balloon entered a cloud or
because it passed behind a cloud or surface obstruction.

     In  Tables  A-l and A-2 of  Appendix  A  are  tabulated  the  profiles
obtained during the winter and summer programs.  Of the 1068 wind soundings
scheduled,  1000  (94%)  were  actually  made.  About half  of  the  missed
soundings were cancelled because of rain.  Of the 1000  balloons  launched,
739 (74%) were tracked to heights in excess of 1 km by both theodolites.

     Computer processing of  the  1000  soundings has been completed.   They
have been stored  on  magnetic  disk  for  editing, and further processing.
After editing, the data  are  archived  on  magnetic  tape,  in unformatted
binary code.  The tapes contain the following: (a) site information; (b) at
each reading time, the azimuth and elevation angles, height and position of
the  balloon relative to the release point,  and  the  "shortest"  distance
between the  two  theodolite  rays;  (c)  calculated  wind  velocity (u, v,
components and speed,  direction),  position mid-point in layer (height and
position relative to the  Arch),  and  the rate of rise of the balloon, all
                                   -22-

-------
for the interval between two  successive   readings.   To  date  30%  of the
soundings have been edited and archived.

     The personnel making the pilot balloon observations were  trained  Air
Force  upper-air  observers.  In addition  they were  given several  days  of
special training and practice in the methods  used  in this  program prior to
start of  the field expedition.  Therefore there is  considerable confidence
in the general quality of the data.

     Nevertheless  the  overall  accuracy  of  winds   determined  from  pilot
balloons,  even  using double theodolite techniques  is  generally  unknown.
The Thyer method permits an  estimate  of  the  possible uncertainty in the
angular reading and therefore in balloon position.   An  analysis  of these
estimated uncertainties for a similar wind program in 1971 (Ackerman et al,
1975)  indicates that  at  times  they  can   be  sizeable.   However,  they
represent position errors and  are  usually   systematic  so that the errors
largely  cancel  out  when  time   differentials  are  taken  in  the  wind
calculations.

     In  order  to try to obtain estimates of the   accuracy  of  the  wind
calculations, special experiments were carried out in July 1975,  and during
an  earlier  similar program in 1972.  In  these experiments several   double
theodolite  teams  set up with parallel baselines, tracked the same  balloon
for  6  or  8 launches.  Lacking a standard,  the average of the  individual
measurements may be considered as the best estimate  of the true value.   The
square root  of the pooled variances, which is referred to hereafter as the
root mean square   (RMS)  deviation,  gives  an estimate of the error in the
wind measurements.  A summary of the results  of these experiments are given
below.  A detailed description of the experiments  and  description   of the
results are given in Appendix B.

     In Figure 5 are  shown  the  profiles  of  the  RMS deviations for wind
speed, wind direction, wind vector,  and ascent rate of the balloon, and in
Figure 6 the wind profiles for the three experiments.  The RMS deviation in
wind direction was very small — of the order of 2°  to 3° up to 1500 m, and
very similar in magnitude in all 3 tests.   The RMS  deviation in wind speed
was  more  significant, and unfortunately  also varied significantly  between
the three experiments, ranging from about  0.2 mps in the  most  stable case
in 1972 to 0.5 to 2 nps for the 1975 test.    The  RMS  deviations in ascent
rate were far smaller, from 0.1 to 0.2 mps on the most stable day to 0.2 to
about 0.8 mps on July 31, 1975.

     A number of factors can  contribute   to  these differences in indicated
errors.  Perhaps the over-riding one is  the  velocity  structure itself --
the ease and accuracy with  which  the  balloon  can be tracked is directly
related to the turbulence, the wind speed  and the variability in direction.
Other factors include the individual measurement "system," in this instance
the personnel and equipment, the angle of  the wind to the baseline,  and the
distance (range) to the balloon.

     All factors were present in this set  of tests.   The  personnel  and
equipment were different in 1975 than in 1972 and   were not identical even


                                   -23-

-------
            2    3    4
          RMS DEVIATION
        WIND SPEED (mps)
502    4    6
      RMS DEVIATION
    WIND DIRECTION (deg)
                                             8 0
                   2    3    4
                   RMS VECTOR
                 DEVIATION (mps)
5 0  0.4  0.8  1.2
    RMS DEVIATION
   RISE RATE (mps)
   Figure 5.   Root mean square (RMS) errors for wind  speed  and direction,
   balloon ascent rate and  RMS vector error for three field tests.
1.6
1.4
1.2
1.0
0.6
0.4
0.2
         1

        Aug 9 1972 /
                          f July 31  1975
                          y
        4     6     8     10    12
            AVERAGE WIND SPEED (mps)
         -  \
                                                   July 31 1975
                                                     \
                                                            Aug 1 1972
                                                                               Aug 9
14    16 -40   -20    0     20     40    60    80   100
       AVERAGE WIND DIRECTION, BASELINE COORDINATES  (degrees)
Figure 6.   Mean  profiles of wind speed and wind  direction (in base-
line coordinates)  for  the durations  of the three tests  of the double
theodolite  system.
                                        -24-

-------
in  the  two  1972  experiments.  The  angle of the  wind  to  the  baseline
averaged 50° on August 1972, 42° on August 1,  1972,   and  18°  on  31  July
1975.   The  computational technique developed by Thyer is not as sensitive
to  "small  angle"  problems  as other methods.  However, the fact that the
largest RMS deviation occurred where the  crossing angle to the baseline was
largest suggests that even with this method the double theodolite technique
loses some of its advantage when the triangle  becomes very obtuse.

     For this group of experiments, the   crossing  angle and the wind speed
appear to be the important elements in determining accuracy.  The RMS error
was largest on  the  day  that  the  crossing  angle was smallest.  The RMS
deviation was also larger the stronger the wind, increasing from 0.2 or 0.3
raps  (less than 5% of  the  speed)  in  the lower kilometer on 9 August 1972
when the wind speed was 3  or 4 mps to 0.5 to  2 mps (10 to 20%)  on July 31,
1975  when  the wind speed was 5  to   9   mps.   In  addition  to  increased
difficulty in tracking, strong winds also result in  greater  range  to the
balloon at a given level so that small angular errors,  either  because of
reading error or because the balloon was  not exactly on  the  crosshairs at
reading time, represent larger (non-systematic) position errors.

BOUNDARY LAYER PROFILER (BLP) MEASUREMENTS

     The sensors on the BLP  package   were  interrogated  and  recorded  in
sequence over a 20-second interval.    The system  used a pen recorder that
provided a chart similar to the standard  radiosonde chart.   Ten values were
recorded sequentially  in each record  covering 20 seconds,  in the following
order:  high reference,  low  reference,  temperature,  wet bulb, pressure,
wind speed, temperature, wet bulb, wind speed, wind direction.

     The BLP records have been digitized  using  a digitizer (the Autotrol)
which produces punched IBM cards.  All of the useable  profiler data have
been  processed  through  a  program   which  scales  the  input  parameters
according to the calibration provided  by  NCAR  and generates,  in  addition,
the  height (using the hypsometric equation),  vapor pressure, mixing ratio,
dewpoint, relative humidity, saturation vapor  pressures at both dry and wet
bulb   temperatures,   virtual  temperature,   potential  temperature,   and
potential  wet bulb temperatures.  The vertical  "gradient"  of  potential
temperature, vertical shears in speed  and direction, and Richardson number
are calculated for successive readings also   but  are  not very meaningful
because  small  variations in basic values introduce  large  variations  in
these terms due to the  shallowness  of   the   layer  over  which  the  dif-
ferentials are taken.

     Although the design  limit  of  the  balloon and package is 750 m, the
maximum height attained varied as a function of wind speed, turbulence, and
stability.  All but 11 exceeded 300 m  and nearly half exceeded 500 m.

     Both the ascending and descending sections of  the soundings have been
processed.   The rate of ascent was approximately 45  m/min,  the  rate  of
descent  was  at  least  twice that rate.  Comparison of  the  profiles  of
temperature and humidity for the two parts of  the run  indicated  that  the
descending  sections  were  useable  despite the faster rate except for the


                                    -25-

-------
wind  speed  and direction, and particularly the former.  The rapid rate of
descent introduced large error in wind speed and a smaller less significant
error in wind direction.

     The performance of the BLP systems was very disappointing.  Twenty-six
percent of the scheduled soundings could  not be made because of instrument
malfunction or unavailability of sensor backage.  Of the 228 soundings that
were made, nearly 20% are unuseable because of instrument malfunction.  The
remaining 208 soundings appear  to  be  good,  although  they have not been
subjected to an in-depth review.

     As of this date, all of  the  profiler  data  have been processed.  In
Table A-3 of Appendix A are tabulated the dates and times for which useable
profiler data are available.  These data have been  edited for large errors
but  require  some  additional  examination for less obvious errors  before
incorporation  into  analyses.   Scaled  and  derived values, and the basic
chart readings are archived on magnetic tape.

Status of Data Reduction

     As has been indicated,  the  basic  reduction of the data collected by
each of the facility components  has  been  completed,   but  only  a  small
fraction  of  the data has been  edited.    The  pilot  balloon  data,   when
completely edited,  will  be  available  through  the  RAPS data bank.   The
airplane and profiler  data will reside at the Illinois State Water Survey.
Data are available to RAPS researchers, subject to the Illinois State  Water
Survey data policy (Appendix C).
                                   -26-

-------
                                 SECTION 5

                   DATA SAMPLE AND RESEARCH  PROJECTIONS


     The analysis of the data collected under  this  grant is being funded by
the National Science Foundation*.  Because of  unfortunate delays and lapses
in  funding  analysis  of  the data is barely  underway.   A  study  of  the
nocturnal  heat island circulation, based in part on data collected in  the
field  program  described  above, is nearing completion and a report of the
results will be available  by  midsummer  1977.   Currently funded research
covers a  study of the turbulent fluxes of heat, moisture and momentum over
urban and rural surfaces.  Initial  results  of this research should become
available by the end of 1977.

     One of  the  case  studies  carried out in the research on heat island
circulation was based  on data obtained on July 26-27, 1975.  In the course
of this study, the  profiler  data  were analyzed to determine the temporal
evolution of the nocturnal boundary  layer at  both the downtown and country
locations (BP1 and BP2, resp, Figure  2).    This analysis and a similar one
for the night of July 25-26 are  presented below, to illustrate the type of
information contained in this portion of the data bank.

NOCTURNAL OBSERVATIONS, July 26-27, 1975

     The weather conditions during the day and evening of July 26 were very
favorable for the development of a significant nocturnal heat island.   With
only thin high  cloudiness  and  light winds,  the midafternoon temperatures
reached around 30C.   The  heat  island  started to increase rapidly around
1400 CDT**, reached  its  maximum  around 2200 and started to decrease very
slowly shortly after midnight.

     On the night of July 26-27  observations  were  made  between 2020 and
2320 at  the  rural  station  (BP2)  and 2100  to 0030 in downtown St.  Louis
(BP1).  (Instrument problems  forced  cancellation of some of the scheduled
soundings.)  The boundary layer  winds  were light and from the SSE so that
the  two  stations were located  crosswind   with  respect  to  each  other.
Moreover the upwind fetches were homogeneous  at  both locations, urban for
the  downtown  station  and  rural for the   country  station.    Thus,  the
measurements  should  be  representative  of  the  general  locale  of  the
observing station.
* Grants DES 74-13931 and ATM 76-15870.
**In the following discussion all times  are  given  in Central Daylight Time
 (CDT).

                                   -27-

-------
     Time-height  analyses  of  temperature at both  the  urban   and   rural
profiler stations are shown in Figure 7. At the country station most of the
low-level  cooling  had  already  occurred  by  the time  the  observations
started, and a strong surface-based inversion existed  (Figure  7b).  Hie top
of the inversion was quite low, only 100 m above the surface,  and increased
only slightly in height   (25 m) during the three hours of observation.   The
temperature at the top of the inversion also was nearly constant  during the
three hours, varying between 24.3 and  24.7C,  but  with  no   evidence  of a
trend with time.  Most of the vertical temperature change  (80-90%) occurred
within the first 25 to 50 m above  the  surface.   The  region of  maximum
temperature at the top of the inversion extended through 50 m  at  least,  and
above that the temperature lapse was roughly neutral.

     The  situation  in  the  city was quite different, as one would expect
(Figure  7a).  The temperature was still high at 2200 and although the  rate
of cooling  at  the  surface  was  not  very  different than over the rural
surface  during  the common period of  observation,  the  cooling  extended
through a much deeper layer.   A  surface  based  isothermal,  or very weak
inversion, layer was very shallow (less  than  50  m).    The main nocturnal
inversion was  elevated,  and delayed, barely forming by 2130.  The base of
the inversion lowered during the observation period from about 175 m to  125
m, while the top remained almost constant in height at 200 to 220 m.

     In the city the whole layer  from  surface  to  200  m  was stable  and
apparently shared in the nocturnal cooling.  In  the  country, on the other
hand, the cooling was concentrated below 100 m.    From 2200 on the base of
the elevated inversion in the city lay just about  at the height of the  top
of  the stable layer in the country.  The temperature at  the  top  of   the
inversion was nearly the same as that at the top of  the country inversion,
varying  from  23.6  to 24C, but with an indication of an increasing  trend
with time.

     The  urban  heat island was  very  shallow  (Figure  7c).   The  urban
temperature excess dissappeared above 60  m  during  the  part of the night
covered by the observations.  A 'negative' heat  island  (i.e.  city cooler
than country)  occurred between 75 and 175 m from 2220 on.  A  comparison of
the temperature profiles (Figure 8) indicates that the reversal in the sign
of the urban-rural temperature  difference was related to the  fact that the
low-level cooling in the city  extended  well beyond the top of the country
inversion.  Urban-rural equivalency in temperature occurred  only above  the
urban stable layer.

     The  nocturnal moisture distributions over country and city sites   are
shown in Figure 9.  Well mixed conditions existed downtown,.but there was a
significant  vertical  gradient  in the moisture in the lower  50 m  of   the
rural  PEL.   Initially this gradient was a lapse, but shortly  after 2100 a
moisture inversion developed in the lowest 25 m.  A pocket of  dry air about
200 m  deep  appeared late in the evening at both locations, slightly  later
and about 100 m higher downtown than in the country.  At both  locations the
minimum  mixing  ratio  observed was abtrat  9.5  g/kg;  however,  since it
occurred at the last observing  time  at  both locations, this may not  have
been the  minimum reached at either one.  At both locations this dry pocket
occurred at the top of the nocturnal inversion.

                                   -28-

-------
           600
        o
        o:
        CD
        o
        CO
NJ
<^>
I
           500-
           400-
           300
           200
            100
                                                    JULY 26-27, 1975
                                                     TEMPERATURE (°C)
i      i      i
a. BP 1, DOWNTOWN
                   22
    I      i     I      I
b. BP 2, RURAL (TRIAD H.S.)
                                               24-
                   2100  2200  2300  0000
1	1	1	\	
 c.  URBAN-RURAL
    DIFFERENCE BP 1-BP 2

                                                                                                 -°-5
                       2100  2200  2300  0000

                              TIME, CDT
                                2100   2200  2300   0000
                Figure  7.  Time-height distributions of a) temperature in downtown St. Louis,
                b) temperature at Triad High School which is located in a rural setting surrounded
                by farmland, and c) temperature difference between the two stations, on the night
                July  26-27, 1975.   In a) and b) shading indicates inversion or isothermal layers,
                and heavy  lines the tops of inversions.

-------
w
o
                     500
                     400
                     300
                     200
                     100
                       16
                                 JULY 26 1975
                                   2300 CDT
                                                            I
                                                                     I
                                17
18
19
20        21

TEMPERATURE, "C
                                                                              22
                                             23
24
25
                  Figure 8.  Temperature profiles at a  station in downtown St.  Louis (solid  line) and

                  at  a rural station 20 miles  to the ENE  at 2300 CDT on July 26,  1975.

-------
   600
   500
S  400
o
oi
CD
o
co

-------
      WIND DIRECTION (10 degrees)
                                                                      WIND SPEED  (mps)
    o\- SITE  70, July  26-27  1975
             26
  a.
2120

Rural location
                 2220         2320
                      TIME, CDT
                                     0020  2120
2220         2320
      TIME, CDT
                                                                                    0020
      WIND DIRECTION (10 degrees)
                                                                      WIND SPEED (mps)
        SITE  78, July  26-27  1975
        SITE  44, July  26-27  1975
                2220         2320
                     TIME, CDT
                                     0020 2120
2220          2320
     TIME,  CDT
0020
b. Upwind metropolitan  stations
                                           -32-

-------
   WIND DIRECTION  (10 degrees)
2.2
2.Op SITE 72, July  26-27  1975
1.8 -
     SITE 74, July  26-27'1975

  2120        2220        2320
                  TIME, CDT
 Downwind metropolitan stations
0020 2120
2220        2320
     TIME, CDT
                                                                         0020
   Figure  10.   Time-height distributions of wind direction and wind speed
   on the  night of July 26-27, 1975 as obtained from double theodolite
   pibal observations a)  at a rural location 25 miles east of St.  Louis,
   b) at two  stations on the south  (upwind) side of  the  metropolitan area,
   and c)  at  two stations on the north (downwind)  side of the metropolitan
   area.   Solid heavy lines indicate maxima in the wind  speed profiles
   and dashed  heavy lines indicate minima.
                                    -33-

-------
      The  urban-rural differences in mixing  ratio  (Figure  9c)  were  negative
 (city dry) only  in the lowest 50 m.  The magnitude of  the  urban  deficit  de-
 creased  with  time  and by 2300 the  near-surface  moisture   was   slightly
 greater downtown (by 0.6 g/kg) than in the  country.   This  change was no
 doubt due to the loss of moisture from the  lowest air layers  to the ground
 in the country as the surface cooled down and the air  became saturated.  In
 the city, on the other hand, the temperature remained  so high  that even at
 midnight  the surface relative humidity was above 60%.

      Above the   lowest  50  m,  the  air  downtown  was more moist  than  the
 country air up to about 200 m (slightly higher later in the evening).  This
 more  moist layer coincided with  the  layer  of  relatively  cool   city  air
 (Figure 7c), and was probably  due  to  the  same  cause,  namely the deeper
 mixed layer over the city.

      On clear summer nights  with low wind speeds, the wind profiles in  the
 St.  Louis area typically show speed maxima forming in  the  lowest 200 to  400
 m around  sunset  or shortly thereafter,  increasing  in magnitude and height
 as  the evening progresses and frequently splitting  into two or  three peaks
 late  in the evening.  The evolution of the  wind structure, as indicated by
 the pibal observations (Figure 10) followed this sequence  at   all stations.
 At  the rural location (Figure 10a) the 'jetlet' reached 6  mps while at  the
 stations  on the  periphery of the metropolitan area (Figures lOb  and c)  the
 low level speed  maxima reached 8 to 10 mps, 2 to 4  mps  greater than   the
 speed minima  just  above.   The  levels  of  these  low level jetlets were
 generally higher,  by  several   tens  of  meters,  at  the   two   downwind
 metropolitan stations (72 and 74) than at the country  stations.

      The  wind was southerly through the lowest  several  hundred meters on
 this  night, but  at about 800 m  it  veered sharply from south  to west.    The
 depth  of  the  shear layer was 200 or 300 m early in the night but increased
 (and  consequently the magnitude  of  the  shear  decreased)  as  the  night
 progressed.   This shear could represent  concentration  of  the  shift  to
 gradient  flow to a shallow layer  during  the  afternoon   when   the  mixing
 distributed the  frictional dissipation of gradient momentum  vertically  and
 the flow  to the  left of the gradient  wind  was  found  through  the deeper
 layer.  This is  indicated also by the veering of  the  wind with time below
 the shear layer.    This veering occured nearly simultaneously  through depth
 at  the  rural and near rural sites (Stations 70 and 78)  so that a  layer of
 near  constant  wind direction was maintained.   At the downwind urban sites,
 the isogonal analysis indicates a gradual turning of the wind with height.

     The  profiler data provides  a  more  detailed look at the structure of
 the   low-level  maximum  in  wind  speed.   In  Figure  11  are  shown   the
height-time distributions of wind speed at the  two  profiler  sites,  with
 greater height  exaggeration  than in Figure 10 since the  profilers provide
much greater height  resolution.    There was a very strong gradient in wind
speed off the surface  in the country (Figure lib), with most  of the change
occurring in the lowest 50  to  75m.    The low level jetlet was  at  200 m at
2050 but dropped rapidly to 125  m  and  remained  between 125  and 175 m.
Shortly after 2200 a secondary maximum formed around 300 m.  Both the level


                                    -34-

-------
                               JULY 26-27,  1975
                               WIND SPEED (MRS)
                                             I      I      1
                                                b. BP 2
                                                  RURAL  (TRIAD H.S.)
            2100  2200  2300  0000
2100  2200  2300  0000
                                  TIME, CDT
Figure 11.  Time-height distributions of wind speed on the night
of July 26-27, as measured by boundary layer profilers aj  in
downtown St. Louis and b) at Triad High School located in  a rural setting
surrounded by farmland.  The shaded areas indicate regions of maxima and
the heavy lines the peak values in the vertical profiles.
                                  -35-

-------
and the magnitude of the maximum in the speed profile are in good agreement
with those in the profiles calculated from pibal measurements  at  the same
location  (Figure lOa).

     The  low level vertical gradient in wind speed was much smaller in  the
city than in the country.  The surface wind (which was calm in the country)
was about 2  mps  downtown.   Although  there  appears  to  have  been some
intermittent formation of maxima below  100  m, the primary maxima occurred
between 200 and 350  m.   There  was  definitely  a double structure to the
jetlet from 2130 on,  with  the  lower maximum around 200 m, and the second
about 75  m  higher.   Both  increased  in height and speed with time and a
third maximum formed  around  225  m  again at about midnight.   The maximum
speed increased from under 5 mps to a little over 8 mps.

     Blakadar (1957) has shown that nocturnal  low level jets and nocturnal
inversions  are related.   This relationship existed on  this  evening  even
though the speed maxima were relatively small.  In both the urban and rural
boundary  layers, the layers of speed maxima (shaded in  Figure 11) occurred
just at the top of the nocturnal inversion (highlighted in Figure 7).

NOCTURNAL EXPERIMENT, July 25-26, 1975

     The  profiler data have also been examined for the night of July 25-26,
1975.   On this  night  the  period  of overlapping observations at the two
profiler  stations was considerably longer,  extending from 2100 on the 25th
to  0400  on the 26th.   The low level winds were from the east so that the
upwind fetch at BP2  was  rural and air arriving at the downtown site (BP1)
had passed  over  several miles  of urban- industrial surface.  Therefore the
measurements were again representative of the locale of the observations.

     As   will be seen in  the  following  discussion,  the  boundary  layer
structures of all three variables,  temperature,  moisture  and  wind,   were
very similar to that found on the night of July 26.

     Again,  a strong surface-based inversion had developed in the  country
by 2100 and continued throughout the night (Figure 12b).  The  stable layer
deepened  from about 125  m early in the evening to about  250  m  at  0400.
Downtown, the main low-level stable layer was a rather shallow inversion at
around  200 m, lifting about 50 m after midnight (Figure 12a).  The base  of
the elevated  inversion was just above the top of the rural inversion.  The
temperatures at the  tops  of  the urban and rural inversion were about the
same.

     The  urban heat island reached  a  maximum intensity .of about 8C at the
surface by  2300  and  remained within a degree of that for the rest of the
night.   The heat  island  was  less  than  100 m in depth  (Figure 12c).  A
reversal  in the sign  of  the  urban-rural  temperature difference occurred
above the low level heat  island,  but  not until midnight and the negative
heat island did not reach any significant magnitude until nearly 0300.  The
layer of  the 'negative heat island1  was   based at about 100 m, or a little
higher,    and  increased  in depth and magnitude during  the  early  morning
hours.    The  magnitude of the urban temperature deficit reached  a  little
                                   -36-

-------
1ULV 25-26, 1975
TEMPERATURE (°C)
                        /I///I//I/ ////I//
                        b. BP 2, RURAL (TRIAD H.S.)
                          I  /X//V/A//V/ h / /I///
                        c. URBAN-RURAL DIFFERENCE, BP 1 - BP 2 '
                                         S-0.2 < At < +0.2
                    2000
2200
                                                    0400
                                        0000      0200
                                        TIME, COT
Figure  12.   Time-height distributions of  a) temperature in downtown  St. Louis,
b) temperature at Triad High School  which is located  in a rural setting
surrounded  by farmland,  and c)  temperature difference between the  two
stations, on the night  of July 25-26,  1975.  In a) and b)  shading  indicates
inversion or isothermal layers, and  heavy lines the tops of inversions.
                                   -37-

-------
over 1C by the time the observations ended at 0400.   As on July 26-27 this
reversal of temperature difference was  a  consequence of the mixing in the
city and thus a sharing of  the surface heat loss through a deeper layer in
the city than in the country.

     The urban and rural moisture distributions on the  night of July 25-26
(Figure 13) were also similar to those observed on the night of July 26-27.
The lowest 500 m were very well mixed downtown and  little  change occurred
during the night except for the development of a weak moisture inversion at
around  25 m above ground after 0100.  In the country, there was  a  strong
moisture gradient  near  the  surface and a low level moisture inversion at
about 25 m from 2120 on, which increased in strength as the night wore  on.
Whereas the surface  mixing ratio remained between 10 and 10.7 gm/kg in the
city it decreased  from  11.9  to  9.1  in  the country.  Thus in the early
morning hours the surface air  was more moist downtown than in the country.
Above 50 m, the boundary layer was fairly well mixed at both locations.

     A dry pocket  appeared  about  200 m at both locations on this evening
(slightly later in the  city).  The minimum mixing ratio was 0.5 g/kg lower
in the city than in  the  country.   This dry pocket could have been due to
advection at both locations although it  is  unlikely that the air over the
rural site (BP1) could have been the  same  that  appeared  downtown in St.
Louis within the hour.   Moreover the occurrence of these dry pockets around
midnight on both nights would suggest that they are  due  to  local effects
rather than due to advection.

     Between 100 and 200 m the urban moisture deficit decreased with  time,
and the  urban-rural  difference  in  moisture became positive in the early
morning (Figure 13c).  However on this evening the city remained drier than
the country  until  well  after  midnight,  both at the surface and in this
elevated layer, and the urban moisture excess aloft was quite small.

     Low level wind speed maxima also  occurred  on this evening, extending
into the early morning hours  (Figure  14).  In the country a single jetlet
was well  established  by shortly after sunset and continued at a height of
about 125 to  150  m  throughout the night.  The maximum wind speed however
occurred in the middle  of  the  night  and  had started to decrease by the
morning  hours.   Multiple low level  maxima  in  wind  speed  occurred  in
downtown St.  Louis.  The main layer  of maximum speed was based between 200
and 250 m with considerable variation in  height  with  time.   The maximum
speed was observed early in the night rather  than  in  the middle, and was
slightly higher than the maximum value in the country.   However,  it  then
decreased so that by the morning hours the maximum speeds  in  the  country
and city were about the same.

     The base of the layer of high winds occurred just above the top of the
inversion  during  most  of  the night in the city.   However, this was the
case only  until  shortly  after  0200  in  the  country, at which time the
inversion  started to deepen significantly  but  the  height  of  the  main
maximum in wind speed remained fairly constant.

SUMMARY
     The material  above has been presented to illustrate some of the kinds
of information contained  in  the  data  collected  with the boundary layer
                                    -38-

-------
 JULY 25-26, 1975
MIXING RATIO (g/kg)
 I //I/7777
a. BP 1, DOWNTOWN
                        ///I/////I///I/////I
                       b.  BP 2, RURAL  TRIAD H
                  500
                  400
                  300
                S 200
                  100 -
                   2000
                        V///////'///i////
                      c. URBAN-RURAL DIFFERENCE, BP 1 - BP 2
  2200
                                                  0400
                                     0000     0200
                                     TIME, CDT
Figure 13.  Time-height distributions  of  a)  mixing ratio in downtown St.
Louis, b) mixing  ratio at Triad High School  which is located in a rural
setting surrounded by farmland, and c) difference in mixing ratio at the
two stations, on  the night of July 25-26,  1975.   In a)  and b) heavy lines
indicate the tops of moisture inversions.
                                 -39-

-------
O
        e
         «<
       Q
       O
       co
          600
          500 -
          400 -
          300 -
5  200 -
(—t
LU
rc

   100 -


     0
     2000
                      2200      0000      0200
                                 TIME, CDT
                                                    JULY 25-26, 1975
                                                    WIND SPEED (mps)
                                                                b. BP 2, RURAL (TRIAD H.S.)
0400   2000
2200      0000      0200
            TIME,  CDT
0400
                               Figure 14.   Time-height distributions of wind speed on the
                               night of July 25-26,  1975,  as measured by boundary layer
                               profilers a)  in downtown St.  Louis and b) at Triad High
                               School located in a rural setting surrrounded by farmland.
                               The shaded areas indicate regions of maxima and the heavy
                               lines the peak values in the  vertical profiles.

-------
profiler.  These  examples  indicate that a great deal can be  learned  about
the evolution of  the  structure in the lower PEL from these  data.   In  the
currently  funded studies  of the turbulent fluxes the profiler data will be
used  in  quite   a different  manner.   However it is the intention to seek
funding  at  a  later  date to do intensive studies of the evolution of PEL
structure over the inner city and country surfaces.
                                 REFERENCES

1.   Ackerman,  B.   Winds  in  the  Ekman  layer  over St.  Louis.   Preprints,
     Conf.  on Urban Environment and  Second  Conference  on  Biometerology,
     Philadelphia,  Pennsylvania,  October 31-November 2,  1972.   pp.  22-27.

2.   Ackerman,   B.    METROMEX:   Wind Fields Over St.  Louis   in   Undisturbed
     Weather.   Bull.  American Meteorological Society, 55(2):93-94,  1974a.

3.   Ackerman,  B.   Wind Fields  Over the St. Louis Metropolitan Area.   J. of
     Air Pollution  Control  Assoc., 24(3):232-236, 1974b.

4.   Angell,  J.  K.,   D. H.  Pack, C.  R. Dickson, and W.  H.  Hoecker.   Urban
     Influence   on   Nighttime Airflow Estimated from Tetroon Flights.   J.
     Applied Meteorology,  10(2):194-204, 1971.

5.   Angell, J.   K.,   W.  H.  Hoecker,  C. R. Dickson,  and  D.  H. Pack.   Urban
     Influence  on Strong  Daytime  Flow as Determined from  Tetroon  Flights.
     J. Applied Meteorology,  12(6):924-936, 1973.

6.   Blakadar,  A. K.   Boundary Layer Wind Maxima and Their  Significance  for
     the   Growth of  Nocturnal Inversions.  Bull.  American  Meteorological
     Society,  38(5):283-290.   1957.

7.   Bowne, N.  E.,  and J. T.  Ball.  Observational Comparison of   Rural   and
     Urban  Boundary   Layer   Turbulence.    J.    Applied  Meteorology,
     9(6):862-873,  1970.

8.   Chandler,  T.   J.    Wind as a Factor of Urban Temperatures,  a Survey in
     Northeast  London.  Weather, 15(6):204-213, 1960.

9.   Findlay,   B. F.,  and M.  S. Hirt.   An  Urban-Induced  Meso-Circulation.
     Atmospheric Environment, 3(5):537-542,  1969.

10.  Mass,  W. A., W.  H.  Hoecker, D.  H. Pack, and J. K.  Angell.   Analysis of
     Low-Level,  Constant Volume Balloon (Tetroon)  Flights   over New  York
     City.  Quarterly J.  of Royal Meteorological  Society,  93(398):483-493,
     1967.

11.  Morris, A.  L.,  D. B.  Call, and R. B. McBeth.  A Small  Tethered Balloon
     Sounding     System.      Bull.    American   Meteorological     Society,
     56(9):964-969, 1975.

12.  Thyer, N.   Double Theodolite Pibal Evaluation by Computer.   J.  Applied
     Meteorology,  l(l):66-68,  1962.


                                   -41-

-------
                                APPENDIX A

               AVAILABILITY OF DATA:BOUNDARY LAYER PROFILES

     The  availability  of  useable profile  data  collected  in  the  ISWS
Boundary Layer Program during the field  efforts  in  1975 are shown in the
following three tables.

     Tables A-l and A-2 show the times and stations for which wind profiles
obtained using pilot balloons are available.  Table A-3 shows the times and
stations for  which  temperature,  humidity and wind profiles obtained with
the boundary layer  profiler  (tethered balloon) system are available.  The
locations of all stations are shown in Figure Al.

     The equipment and techniques used in both systems are described in the
main body of this  report.    All  the data are archived on 9-track-1600 BPI
magnetic tapes in IBM unformatted binary code.
                                   -42-

-------
              Table Al.  Pilot  balloon  measurements available for
                                  February 1975.
 Symbols:  0 = no data, / = data  to 200 m of ni,jner,   = d.na _,;„.  ;....,s  >.nan  ±<)0  ">.
SITE     DATE      	RELEASE Tim,  (local)**	

      19 Feb.      1220    1240    1300     1320     1400     1420    1U40    1500
33                    v'///////
<*<»                    •///////
71                    0       /       /        /        Q        /       /
75                    /O       //////
78                    /       0       /        /        0        0        /       o
79                    ////////

      20 Feb.      1150    1210    1230     1250     1340     1400     1420    1440
33                    0       ///////
44                    ////////
71                    ////O///

75                    ////////
79                    ////////

      21 Feb.      1140    1200    1220     1240     1330     1350     1410    1430

33                    ////////
44                    ////////
71                    ///O////

75                    ///////       0
79                    ////////

      25 Feb.      1320    1340    1400     1420     1440     1500

33                    0       0       0        0        /        /
44                    //////
71                    0       /       /        /        /        /
75                    0       0       /                /        /
78                    //////
79                    0       /       /        /        /        /

     * 26  Feb.     1220    1240    1300     1320     1400     1420     1440    1500   1520

74                    /V////*/>'/v/
75                    /////////
79                    //////*^^^
74
75
79
                   1230    1250    1310    1330     1410     1430     1450     1510    1530
       27 Feb.     1210    1230    1250     1310     1350     1410    1430    1450
 33                    ///////
 1*4                    ///////
 •jl                    00/OQ03
 74                    //////''
 75                    0       0       0       0        /       J       *'
 79                    Q       J       •!       '•'''
 33
 78
                   1240    1300    1320     1340    1400    1420    1440    1500
  * = W nin staggered releases  at situs S3 and 78, 2 to
-------
 Table A2.   Pilot balloon measurements available for
                      July 1975

= n,: data, J = data to 200 m ur hi,jhr>rt - = data for less  than 200 m.
.ire

i*U
7?

7«
7 -

Vi
• i'i
1.'
J'l

7 «

7 -
7 ,
74

70
7 •

33
I'-
ll
76

Hit
7 i
17
7R

71
T>
1-
i -


7-i
7 '

7,\
'••'.



7 ,
~,,


~.ii

~~





L-\i;; :J::I;A,;I; Tint;;; /,-;'/•; SITC DATL
' July i. 0 11 ?) ll'*0 12'»'i 12"! 11'iu 13?0 13MO 18 July 11? / /
-j 0 0 0 / / / / 78 ° ^
7 'ulv IV ' iVO 11<"0 11J>" I1'«0 K/'O 1'.20 15UO ii.O'J 1-.20 * ?1 July 1320 ITi-)
/ / / / 0 / / / / 70 / /
/ / / / / 0 / / / / 7'J / /
l/ ^ %/ ' '* ' , , , y i .no 1 V'<)
//»•'// n ////„. < / J
, , , , , , , • 70 t »
. / » / / ')//•' •' ./M ^ /
-n ,'uiv J, i! i.""-" J3lih rn-j i i1.') J'no I'i'.'j JMO !''*'> ?•' ^ /
*'/*/!////// ?? Jujy i:';'') iri" '
»'»//// » 77 / /
l/n • nr'0 i jjo 1310 I'ton jnuo J'jO^ ii)?o 1510 7'i / /
/ • / / r' 0 / / / 71) / /
,/ ,' /////// 7H / /
,-10 July .140 2r,0 222C, 2>0 2320 2340 WOO 0020 ^ ?"-25 July 213^ 21!^
^ '',,,, 1 72 ' //
0 0 y .'//// - / /
/ / ,',//.'/ '| / /
/ °, / ^ /'/'// 7'3 / /

10-11 Julv 2140 2200 2220 224, 2320 2340 2m,o 0020 w ""^ Jul>' 21^ ?1"°
o .' a o o o o ^ v^ ^
' '/ /''//// 7t ^ •'
-','// / / ,' / 7S
12 lulv 1'"') 1400 1420 144,' i:,21' r,4') ll-OO lf;20 20-:7 ..'uly 2140 2200
.' / / / / 0 0 1 44 / /
,,','/.'/// 70 / /
,' / ,' /,''/// 72 / /
,' ,' ,','//.' 74 / /
/,'//,/ 0 71 / /
•|" J'jlv 1'J'l 1 1 iO 13r:0 I'll') !4r.T 1510 1530 1550 1G30 2'j July Ot.OO 01,2-)
///////// 70 / /
v //,'// /c 79 /
V. Julv 1..M 1340 1»00 1«2'.. 1440 1520 154C 160U 1620 ^ 2°JUly 002J °°^
'./'!///'// 7'J / /
//////// / ?'J July 12"° 130°
'i !ulv li!" 1.110 1350 1411 I'lbO 11.10 1S)0 1550 1610 73 0 /
/ '/ y • / , / 30 July OC'OO 0020
' 70 J /
/ / / / / ^ / / / 30 July 0920 0040
;i .' / ,'///// 70 o o
/ ,'///// 79 ,' /
!• .;.!-, i.-,-; .;:-> 12-0 noo 1340 IMOO 1420 1440 1500 70 30 July 12"^ "™
'- *' *^ ' J ' "j j J ~1'- / /
.' / / •)//// / 31 -July 0'J2? 03'»3
/ / -'' / .' / / / 33 TO .IP 2 / r'
. t i' / / / / / / 'J / /
1
1 / /
ii / /

MELLASL TIHCSfr.V'r;
12')0 l?2i) 1210 13.X) 1310
/ / / 0 0
/ / / / /
/ / / / 0
/ / / / /
/ / / 0 0
1100 m?n I'joo irj^o IMO
/ / / / /
////"/
11 |0 Hl.i'i 11', 0 JSno l!i'-0


/ / / / /
/ / / / /
)inn |')>o llif) ]'»?o 3'tlO
/ / / / /
/ / / / /
/,'///
/////
2210 2230 2310 2330 2350
/,'///
/ / / / 0
///'//

2200 2220 2300 2320 ?340
////!-
i i ,' / :'
222" 2240 2320 2340 f'OOO
/ / / / /
0 / •' / /
///.'/
t1 J J J J
/ / / / /
01,1)0 ,700 0740 0800 OS2'.)
/ / / / /
0 0 0 / /
1000 1020 1040 1120 11UO
/ / / / r'
////,'
1320 1340 1400 1440 1500
/ / ,' / /
01>4U 0700 0740 O8'~,0 0820
/ >' / / /
1000 1020 104'.' 1120 1140
00000
/ / / / /
1320 1340 1UOO 1420 14'H)
/ 0 / / r'
/ / / 0 0
10UO I'i20 1045 1100 1120
/ / / ,.' ,,'
,' /
/ / / / /
/ / / / t
/ ,'/,'/
,.,/ n/.'.-o 70 .fll ?.', /:.-.. ,', ew

1400 1420
0 0
/ /
0 0
/ /
0 0
1003 1020
0 /
/ /
11, .10 11.30


/ /
/ /
K'lO 11,20
/ ,'
/ /
/ /
/ /
0010
J
/
0
I/

0000
•'
^
0'J20
,'
»'
/
/
/
0-41! .MOO
l' /
/ /
UOO 1,-JO
/ ,'
1520 1540
/ ,'
lbi.0 0100
<'t >'f
12' "J ]22'J
i /
,' /
1 'i'10
'.'
115-
,''
/
/
,,'
/
• ;: .l.'k ;?'':.
                      -44-

-------
       Table A3.  Availability of data from tethered balloon,
                  boundary  layer profilers.  Times, sites when
                  useable data are available are shown by
                  /  symbol.
 STN.      DATE     	SOUNDING  TIMES

           2 July  1210   1*05
 BPI

           3 July  1115  1300   1*00   1500   1600
 BPI                        /     /     /
 BP2                  /     /

           5 July  1100  11*0   1220   1300   13*0   1*20
 BPI                  /                       /
 BP2                  //////

           7 July  1300  1330   1*00   1*30   1500   1530   1600   1630  1700  1730
 BPI                  ////////
 BP2                  J           /           //////

           8 July  1200  12*0   1320   1*00   l**0   1520   1600   16*0
 BPI                                                      /     /
 BP2                  //////
        9-10 July  2120  2200   22*0   2320   0000   00*0
 BPI                  /////'/

          1* July  1300  13*0   1*20   1500   15*0   1620   1700
 BP1                  /     /     /     /     /
 BP2                                          /     /     /

          15 July  1220  1300   13*0   1*20   1500   15*0   1620   1700
 BPI                                    /     /     /     /     /
 BP2                  ////////
          16 July  !2*0  1320   1*00   1**0
 BPI                  /     /     /     /
"BP2                  /     /     /     /

          18 July  10*0  1120   1200   12*0   1320
 BPI                  /     /     /     /     /
*BP2                              /     /     /
          21 July  1200  12*0   1320   1*00   l**0   1520   1600   16*0  1720
 BPI                  /////////
 BP2                  ////////

          22 July  1200  12*0   ^320   1*00   1**0   1520   1600   16*0  1720
 BPI                  /////////

       2*-25 July  19*0  2020   2100   21*0   2220  2300  23*0  0020  0100
 BPI                  /////////
 BP2                  /////////

       25-26 July  20*0  2120   2200   22*0   2320  0000  00*0  0120  0220  0320  0*20
 BP,                  ///////////
 BP2                  ///////     /           /

       26-27 July  2020  2100   21*0   2220   2300  23*0  0020
 BPt                        //////
 BP2                  /     /     /     /     /
          27 July  1220  1300   13*0   1*20   1500   15*0
 BPI                        /     /     /     /
          28 July  1700  17*0   1820   1900   19*0   2020  2100   21*0
 BPI                  /     /     /     >'     •'     >'     *'     '
          29 July  0020  0100   01*0   0220   0300  03*0  0*20   0500  05*0  0620  0700
 BPI                  //////     /     /     /     .'

          29 July  07*0  0820   0900   09*0   1020   1100   11*0   1220  1300  1*00  1500
 BPI                  ////////
          30 ju|y        0820   08*0   0900   0920   1000   MOO   1200  1300  1*00  1500
 BPI                        /      /           /////«
       <>; tlh-rm.wicl-cr trulfwu-li
-------
Figure A-l.  Locations of the double-theodolite pilot-balloon
observing sites (numbered)  and the two boundary-layer profiler
stations, BP1 and BP2.
                          -46-

-------
                                APPENDIX  B

        AN INVESTIGATION OF THE ERRORS  IN ESTIMATING  WIND  VELOCITY
              BY DOUBLE-THEODOLITE  PILOT-BALLOON  OBSERVATIONS


INTRODUCTION

     In studying the planetary boundary layer  (PEL) wind field over a small
meso-scale  region  such  as  the   METROMEX  area  (4000 or 5000 km ) it is
important to establish that  the  observed  variations  are larger than the
"noise"  arising  from  errors  in  the   basic   measurements.    Thus  an
investigation  of the probable error in   the  estimates  of  wind  velocity
obtained in the ISWS boundary  layer  program  has been an integral part of
the overall effort.

     Measurement of wind in the PBL over  the complex meso-region around St.
Louis has been by visual  tracking  of  pilot  balloons  in  the ISWS field
programs.  Field efforts designed to provide detailed study of the velocity
structure has utilized two theodolites, spaced a  few hundred meters apart,
to  track the balloons (double theodolite system).  The  double  theodolite-
system is an improvement over single theodolite tracking since  it does not
require the assumption of a constant rate of balloon ascent,   an assumptict,
which  is  poorly  justified  for the PBL, particularly during summer days.
Nevertheless, there  are  still  several  potential sources of error in the
wind estimate, arising from both human and equipment factors.

     Three  potential sources of error were  recognized  before  the  field
efforts began in 1971 and procedures  were  incorporated  to minimize their
effects  on the measurements.  These expected sources of  uncertainty  were
(1) observer  stress  and  fatigue,  (2)  faulty  equipment,  and (3) faulty
orientation of the theodolite in the horizontal plane (level)  and/or in the
horizontal coordinate system (azimuthal orientation).

     Human stress and mental and physical  fatigue  are factors in any type
of prolonged use of a precision instrument.   Bellucci (1960)  has discussed
the  importance  of  experience  when accurate measurements for  frequently
launched  balloons are needed.   High observational frequency (frequent dial
readings)  is  needed  for  good  vertical  resolution in the profiles  but
increases fatigue and  decreases  reading  accuracy.   Barnett and Clarkson
(1955)  have  reported  that   measurement  error decreased  rapidly  with
   The analysis of the test data was carried out by W. J. Mansell.
   This analysis was supported by funds from the National Science Foundation
     under grant number DES 74-13931.
                                    -47-

-------
decreasing observing frequency,  reaching  an  acceptable  level  at 20-sec
observation interval with no  significant improvement for longer intervals.
Rachele and Duncan  (1967) also report that a better estimate of the average
wind  in  a layer is obtained  by  several  readings  within  a  layer  but
concluded  that for a layer depth comparable  to  20-sec  reading  interval
single readings at this rate does just as well as more frequent readings.

     In order to minimize the human stress and fatigue while maximizing the
vertical and  temporal  resolution  of  the  wind  profiles  the  following
procedures were followed:  (1)  special  additional  training  and practice
were given to already  experienced  observers;  (2)  a  20-sec  measurement
interval  was  used; (3) the balloon was tracked for only 13 minutes out of
each 20-min; and  (4)  a  30-min  break was provided midway during the 3-hr
experimental period.
                                          i
     Theodolites are  precision mechanical instruments and, although hardy,
are not immune  to  problems.   The theodolites were thoroughly checked and
adjusted by skilled technicians  prior  to the field programs.  In addition
the observers checked their equipment  daily and any problems detected were
rectified before the next operation.  However  transport in military trucks
is hard on precision equipment and it  was  not  uncommon  for misalignment
between short- and  long-range optics or loosening of  the  gears  to occur.
The observers recorded when they shifted from short to  long  scope  and if
there  was misalignment  (evident as a single spike in the  processed  data)
the  resultant  error w^,~. corrected.  Errors resulting from slippage of the
gear train are  frequently  detectable  and  correctable  using  the  Thyer
computational method.

     A theodolite can lose orientation  in  the horizontal plane (level) in
several ways.  For instance, the observer  may  accidentally jar the tripod
or  the  tripod  may  settle  unevenly, particularly  in  soft  soil.    The
observers  carefully  leveled  the  theodolite  at  the  beginning  of  the
experiment and checked  the  level  periodically  during  the observational
period. Azimuthal orientation in  the fixed cartesian coordinate system was
accomplished by lining up on the marker at the opposite end of the baseline
and  on  two  other  marked landmarks  off  the  baseline  which  had  been
established   when   the   site   was   surveyed.     The   theodolite   was
azinuthally-oriented at the beginning of the period and checked for azimuth
to at least one of the markers once or twice during the experiment.

     Although  it is believed that the procedures described above helped to
reduce  errors  in  the measurements, it is unlikely that they erased them.
Consequently special  field tests were conducted in the summers of 1972 and
1975 in  order to establish the probable errors in the measurements of wind
speed, wind direction, height of  the  balloon,  and  balloon  ascent  rate
obtained with the double theodolite systems used in the boundary layer pro-
gram.    These tests were made  during  the forenoon of August 1, 1972  (Test
I), late  afternoon  of  August  9, 1972 (Test II) and forenoon of July 31,
1975 (Test III)  using the same observers and equipment as were used during
the field experiments of  those  years.   The tests were all carried out at
Scott  Air Force Base, Illinois,  about  30  km  southeast  of  St.  Louis,
Missouri.    The  local  terrain  is  mostly   level  farmland  with  sparse
population.   The upwind fetch for the test  days  was fairly flat with some
slight roll.
                                 -48-

-------
TEST PROCEDURES

     The design of the  test  was  very simple.    Several  teams  set  up along
parallel baselines tracked  a  common  balloon,   making synchronized  readings
to the balloon at 20-sec  intervals.   In  the   1972  tests  the  mini-network
consisted  of 8 observers utilized as 4  double  theodolite  teams   (Figure
B-la). All baselines were the same length (about  610 m)  and were separated
from their nearest neighbors  by 7.6 m.   The mini-network  used   during   the
1975 _ test  was  similar   (Figure   B-lb).    Ten  observers   were involved,
providing  5  double theodolite teams.   Baseline separations were 3.1 m  and
baseline  lengths  were  656.8   m.   In all tests the  balloons were  released
from  the  center of the perpendicular to the baselines at  the  downwind  end
of network.  The baselines  were aligned approximately NW-SE.

     On each test day, 30-gm  balloons were released every 20-min during  the
testing period.  On Tests   I  and   II  (August   1   and August   9,  1972), 6
releases were  made  during the testing period,  while on Test III (July  31,
1975), 8 releases  were  made.   All   teams tracked  each balloon.  Angular
measurements to the balloon  were   made by each  team  every  20 seconds using
synchronized timers.  Angles  were  read   from the   theodolite  dials to  the
nearest tenth of a degree with   an  estimate of the nearest hundreth of a
degree and were orally recorded on cassette tape recorders.  The timer and
recorder systems, the measurement  interval and  the  angular  resolutions were
all identical with procedures used during  the field experiments.

     In order  to  determine  the   magnitudes   of   errors   introduced  when
theodolites were improperly leveled or  azimuthally-oriented, one theodolite
was  purposely  misoriented by  1°  or tilted by  1 bubble during a couple  of
test runs  on  each test day.   Azimuth  orientation  errors of one degree are
small and may easily occur  due  to  slackness in  the  theodolite gears.  Error
in  vertical  orientation   (level)  of this  magnitude is   extreme  and  is
unlikely to occur except in  cases   of   settling  on   very   soft  ground or
accidental jostling of the  tripod.

     A total of 88  independent double-theodolite  estimates  of mean wind
profiles were collected, 24 on  each of  the two  test days in 1972, and 40 on
the test day in 1975.  The  basic   data were   computer processed using the
Thyer (1962) technique and  carefully  edited for  errors  due  to  obviously
misread angles.

ERROR ANALYSIS

     Since there  are  no   standards   available  against which to check the
observations, the  "true"   wind was   assumed   to   be the   average  of  the
independent estimates made  by the  double theodolite teams.   In  carrying  out
the analysis, the  square   root of the  pooled variance taken  over all teams
and all balloon releases during a  given test period was considered  the best
estimate  of  the  probable error  in  the   estimate  of  the parameter  of
interest.   Only   measurements taken  by  teams   with  properly   oriented
theodolites  were  used  in  the  calculation   of   these  estimates,  i.e.,
measurements IHvolving~~deTIberaTely misoriented or  tilted  theodolites as
mentioned in the preceding  section were excluded from the analysis.

                                    -49-

-------
        BALLOON RELEASE
            TEST I
BALLOON RELEASE.
    TEST II
                           -609.6 m-
                              (a)
 TEAM
       BALLOON RELEASE
          TEST III
Figure B-l.  The "mini-networks" used in tests of the accuracy
of  the double-theodolite pilot-balloon wind-measuring system
in (a) 1972 and (b) 1975.  The two men on each team were placed
at opposite ends of. a baseline as indicated by the heavy dots
and matched team numbers.
                             -50-

-------
     All three tests were  carried out on  days of undisturbed weather.   The
profiles of mean wind  direction (relative to baseline  coordinates,  with 0°
looking along baseline from the release end) during the  test  periods   are
shown in Figure B-2 for all three tests.   Also shown  are  bars  of   +   one
standard deviation  (la)   calculated  from  the  averages  of  the  several
estimates on each run  to indicate the temporal variability.

     The profiles of the mean   wind  speed  for  the  test  periods  and  the
la values of temporal  variability are shown in Figures B-3. The winds  dur-
ing test period I  were southerly  4  to  4.5 mps  and fairly constant with
height.   During  the  third test,   the  winds  were  southeasterly  or   ESE
throughout the lower 1.5 km, but   the  wind  speeds,   which were 5 to 6  mps
below 500 m, gradually increased  with height  to  about   12   mps at  1.5  km.
The winds were lightest during  Test II,  ranging  from  3  to   4  mps in  the
lowest kilometer.  There was a  zone of rapidly shifting   winds  during Test
II, with backing from  a northeasterly direction below 500 m to NNW at about
1.2 kilometers.

     In order to  establish estimates of  probable   errors  in wind speed,
direction, rise rate and  balloon   height,   the  pooled variances  (over all
releases) for each of  these parameters   were   calculated  for every 20-sec
layer on each  of  the three   test periods.   The square root of the pooled
variance for any one variable is  taken as  the  best  estimate of the probable
error  in  that  variable,   and  is   the   quantity  discussed  below.   The
discussion is confined to   the layer below the level  where, due to missing
data, less than 50%  of the maximum number of available degrees of freedom
were used in calculating the statistic.

Error in Wind Direction

     The  estimated error in wind  direction was only  1  or  2  degrees  for
Tests I and II and 2 or 3 degrees  for Test  III  (Fig. B-4).   The error in
wind  direction tends  to be largest  in the  first  100 m because  the   short
range  to  the balloon and  its  rapidly changing position as it responds   to
the unsteady  low  level  winds  make it   extremely   difficult to keep  the
balloon on the  cross-hairs  of the theodolite.  The  error tends to  become
larger again above  a   kilometer   or  so,   a  consequence  of the fact that
position errors for small   angular errors  are  amplified as the range  to  the
balloon increases.  Comparison  of  the error to  the temporal variability in
the mean  wind  direction   (Fig.   B-2)  clearly  shows  that  the  error is
considerably less than  the  temporal  variability,  particularly below 1 km.

Error in Wind Speed

     The estimated error in the wind speed  is  shown as a function of height
for all three tests in  Fig. B-5a.   It was  smallest  during Test II on August
9, 1972, ranging from  0.2 to 0.3 mps  in   the  first 1200 m, increasing with
height  to  0.8  mps   at  1600  m.   The average  wind   speed  for  this  test
increased from 3 mps at the surface  to about 7.5  mps  at 1600 m  (Fig.  B-3b).
Thus the percentage error increased  from  3% at  200 m,  to 5%  at 1 km to 10%
at 1600 m.
                                    -51-

-------
1.6
1.4
1.2
In
.U
"E"
^^
= 0.8
CD
i— i
UJ
n:
0.6
0.4
0.2
0
\ ' 1 '
1 	 r-l H*H
\ ( Aug 9 1972 _
~~ 1 \
\ \TEST rr
\ (*•
- i v
\ x\
.1 . \ \ i
\ "V
— \ July 31 1975 K ~~
\ TEST III \ ^x
- / 2 ^ -
/ i / • ' ^ '
"~ \ ( Aug 1 1972 ) ~~
) r TEST I ^.
— i— A-i i / 1 . / , —
y < S
i f i / ' i
 -40    -20     0      20     40     60     80     100    120
    AVERAGE WIND DIRECTION, BASELINE COORDINATES (degrees)
Figure B-2.  Mean profiles of wind direction, in baseline
coordinates, taken over the test duration for the three tests
of the accuracy of the double-theodolite pibal wind-measuring
system.  The horizontal bars are a measure of the temporal
variability (± 1 root-mean square deviation calculated from
the average for each balloon launch).
                          -52-

-------
              8 0
2    4   6   8   10 0    2
      MEAN WIND SPEED (mps)
10   12   14   16
Figure B-3.  Mean profiles of wind speed (solid line) for the
duration of  each of the three double-theodolite accuracy tests.
The small +'s  give a measure of the temporal variability as
indicated by ± 1 RMS deviation calculated from the average
profiles for each balloon launch.
                              -53-

-------
                          246
                            RMS ERROR,
                       WIND DIRECTION (deg)
Figure B-4.  Profiles of the estimated error in wind direction,
for three tests of the accuracy of the double-theodolite wind
measurement.
                           -54-

-------
i.O
1.4

1.2
1 0
x * is
s
M 0.8
CD
t— I
LU
1 0.6

0.4

0.2
o

— ^ TEST II 	 :> _
1 ">
~? Aug 1 1972 *•> ~"
/ 7 TEST I "C."
if .^ _
( ( ] July 31 1975
l-\ v TEST III _
**> /"
A .^
t\ ~
j \<
^•^
\ •
L\ (a)
^N i 1 1














(III ^
^ /
- -,< ^> -
/ TEST II .T
i <^
-1 <^
\ ^
/ , TEST I '•>
V f LS *
-V ^ ..^v —
// r
| \ i TEST III
t/ >

A x'
i(<
3^'
&
| \<
L\ cb) j
^N i i i
012345012345
             RMS ERROR,
           WIND SPEED (mps)
RMS VECTOR ERROR  (mps}
Figure B-5.  Profiles of the estimated error in wind speed and
of the estimated vector error for three tests of the accuracy
of the double-theodolite wind measurement.
                         -55-

-------
     During Test  I, the estimated error below 1 km was between 0.4 and  0.6
raps, roughly 10% of the average speed during the test.  (The spike at 200 m
was due to  the  switch from short to long scope by one team.  As mentioned
above this  is a correctable error but has been left in in this instance to
demonstrate the magnitude of the error when the two scopes are badly out of
alignment.)  The estimated error  from  Test III is quite a bit larger than
that for either of the  other  two  tests  ranging from 0.6 mps (12% of the
mean wind) at 200 m to 2 mps (20% of the average wind) at 1 km.

     It can be seen from Fig. B-3 that the estimated error was much smaller
than the temporal variability in the mean wind speed during Test II and was
less  than half the temporal variability of the mean wind  during  Test  I.
During  Test  III,  the  estimated  error  was  roughly  half  the temporal
variability below 600 or 700 m but above that,level approached the temporal
variability.

Vector Error

     The root-mean-square vector error is shown in Fig.  B-5.   Considering
the small error indicated for the wind direction, it is not surprising that
the profile for  the  RMS  vector error is similar to the error profile for
the wind speed.  The conclusions to be drawn are the same as that drawn for
wind speed, i.e., significantly smaller values for the first two tests than
for the third, and increasing vector  error above a kilometer, particularly
during the third test period.

Error in Calculated Balloon Ascent Rate

     The double  theodolite  computations  yield  the  values  of the three
components  of the balloon velocity, so that no assumptions are made as  to
constancy  of  balloon  ascent  rate,.    Nevertheless   during   the  field
experiments the amount  of  helium  used  to  inflate  the balloon was kept
constant, to within  the  accuracy  of  the  meters used to measure the gas
flow.  Thus in  still  air the ascent of the balloon would have always been
the same, and errors could have been determined from the difference between
calculated and  design  ascent  rates.   However,  the atmosphere is seldom
still and the  rate  at which the balloon rises is strongly affected by the
vertical motions of the  air.   In Figure B-6 a to c are shown the profiles
of computed ascent rate and  the  temporal variability, as indicated by the
RMS deviations of the average profiles  for  each release.  The ascent rate
tends to "settle down" at about 800  m  (in  the  mean) to about 2.1 or 2.2
mps, which was the design value, but varied  considerably  from this in the
lower levels.

     The profiles of estimated error in the balloon ascent rate is shown in
Fig.  B-6d.   Again the error was smallest for Test II, nearly uniform at 0.1
to  0.2  mps  up to 1200 m (representing 5 to 10% of the  average  rate  of
rise),  and then  increasing  slightly about 1200 m.  The estimated error in
the  first  test  period was about twice that of the second, at least up to
the 1 km level.   The profile of estimated error for ascent rate during Test
III  exhibited characteristics similar to that for  the  wind  speed.   The
errors  on  Test  III were comparable to those on the other two tests in the

                                   -56-

-------
i.O
1.2
— 1.0
~m
.s
£ 0.8
CD
UJ
0.6
0.4
0.2
n
(a) '
Aug 1 1972
- TEST I -
_ _
^
•»
^ ^
«
1 #
f
*•
— _ * f „_.
4

-------
lower  layers, but  increased  with  height so that by a kilometer  they were
larger by a factor of three.

Error  in Balloon Height

     The profiles of the estimated error in the  height  of  the balloon at
each 20-second reading are shown in Fig. B-7.  It  should  be  pointed  out
that   each  height  estimate  is  independent  since  the  calculation uses
independent angular measurements and time.  The height error increases with
height, partly  a  consequence  of  increasing  position  error for a given
angular error as the range to the balloon increases.

     By and large  the  errors  in  height  are not significant below 500 m
(less  than  20  m), only a few per cent of the height.  In fact the errors
remain insignificant for the first two  tests  to  at  least  a  kilometer.
Although the  height error estimated on Test HI was nearly the same as the
estimates on the  other  days up to 600 m, it increased rapidly above that,
reaching 10% of the height at 1200 m.

DISCUSSION

     Field tests of the double theodolite systems used in the ISWS boundary
layer  program have provided estimates of errors that must  be  taken  into
account in interpreting the results of the field experiments.  These  tests
indicate  that  for  the lowest 500 or 600 m the probable  errors  are  not
excessive and,  at least for wind direction, are not significant.  Moreover
it has been shown that up to 500 or 600 m the error  is  considerably  less
than the temporal (non-systematic) variability  expected  over two or three
hours.  Although the  errors  may remain small to a kilometer or more under
certain wind conditions, the non-uniformity of  the  results from the three
tests  indicate that they can become significant under other conditions.

     The differences in  the  results for the three test periods point up a
niaiber of factors that  need  to be considered in estimating possible error
in  wind  measurements  by the  double  theodolite  tracking  of  balloons.
AJ-\iough the basic systems were the same on the three tests, the components
ivere not identical,  i.e., both men and  equipment  were  different  on  the
three  days.   During  Test I there were more  than  the  usual  number  of
problems  with  one of the theodolites, which, despite realignment  at  the
beginning of each run, would not maintain orientation through the  10 or 12
minute tracking period.   This may help to explain why the error  was larger
on  this  test than on Test II.   However, no unusually-faulty equipment was
noted  for Test III,  which provided larger estimates of error than the other
two tests.

     A more reasonable explanation of the differences of the results of the
three tests appear to  lie  in  the  differences  in  wind speed and in the
crossing  angle   between   the   balloon   track  and  the  baseline.   In
computational techniques which use only three of the four angles available,
the advantage of the  double  theodolite  tracking  over  single theodolite
sighting decreases as the wind  direction  (and therefore the balloon track
angle)  approaches the angle of the  baseline,  and in fact the computations
fail when they are parallel.


                                   -58-

-------
             40
  80     120     160     200
RMS ERROR, HEIGHT (m)
240
Figure B-7.  Profile of the estimated error in the
height of the measurement for the three tests of
the accuracy of the double-theodolite wind
measurement.
                        -59-

-------
     Although  the  computation using Thyer's technique does not  fail  when
the crossing angle of the balloon path and the baseline is zero,  it  appears
that errors tend to be amplified when this angle is small.   It can  be seen
from Fig. B-2 that during Test III when the errors were largest,  the  wind
was almost  parallel  to  the  baseline and therefore the balloon track was
almost  along the  baseline.   Thus  the  triangle  to  be  solved was very
oblique.  On the  other hand, during Test II when the errors were smallest,
the wind  in the lowest 500 m was about 90° to the baseline, and although it
backed  to a 20° relative angle at 1200 m, by that time the balloon had been
carried out  about  2 km perpendicular to the baseline so that the triangle
to be solved  was still very acute.  The winds during Test I were about 45°
to the  baselines, and the indicated error remained fairly small.
                                             i
     Wind speed can affect accuracy in  two  ways.   The stronger the wind,
the  more difficult  it becomes to keep  the  balloon  on  the   theodolite
cross-hairs, both because of its more rapid movement  and  because   in  the
PBL,  strong  winds are frequently more turbulent, causing the  balloon  to
bounce  around.  In addition strong winds cause the range to  the  balloon to
increase  rapidly and although it is usually easier to track balloons  when
they are  far away, small errors in angle readings result in large  position
errors.

     In this set  of tests, the "crossing" angle and wind speed are  coupled
in such  a  way  that  it  is  not  possible to separate the effects of the
factors discussed above, i.e., the lowest  wind  speed  occurred  on  the day
when the  crossing  angle  was  largest  and strongest wind  on the  day the
crossing  angle was  smallest.   However,  a  close  examination of the data
indicates that the error  in  measurements of wind speed is probably more a
function  of the low-level wind  direction  (relative  to the baseline) than
the wind  speed.   This may be  true , also  for the error in wind  direction,
although  the increase in the error with  height  above a kilometer suggests
that the  range to the balloon may be a factor of some importance here.
                                REFERENCES

1.   Barnett, K. M., and 0. Clarkson, Jr.  Relation of Time Interval to the
     Accuracy of Double Theodolite Observations.   Monthly  Weather Review.
     93(6):377-379.

2.   Bellucci, R. L.  Preliminary Estimates of Variability of  Winds in the
     Lowest  500 Feet.   U.  S.  Army, USAARDL Technical Report  2122,  U.  S.
     Army  Signal  Research  and Development Laboratory.  15 June 1960.  16
     pp.

3.   Rachele, H.,  and L. D. Duncan.  Desirability of Using a Fast Sampling
     Rate for Computing Wind Velocity  from  Pilot  Balloon  Data.  Monthly
     Weather Review.  95(4):198-202.  1967.

4.   Thyer, N.  Double Theodolite Pibal Evaluation by Computer.  J. Applied
     Meteorology.  l(l):66-68.  1962.

                                    -60-

-------
                                 APPENDIX C

                          SURVEY POLICY REGARDING
                    FURNISHING DATA TO OUTSIDE AGENCIES

     In the course of conducting research or investigations the Survey  may
accumulate valuable, even unique, collections of data.  These data may have
been collected and processed into useful form either at State expense or in
connection with grants or contracts.

     It is expected that such data will be used initially and primarily for
interpretive studies and publications  by  the  Survey.   However, the data
may have values for studies of a related or quite unrelated nature by other
organizations.   Such  outside  use  may produce further  understanding  of
Illinois  resources,  represent  valuable  contributions  to  science,  and
enhance the scientific reputation of the Survey,   However, reasonable care
must be  exercised that responsible organizations or individuals propose to
make proper use  of  the  data.   It  is  also  recognized that substantial
personnel and machine time may be involved in responding to a data request.

     This  statement sets forth a policy  under  which  such  data  may  be
furnished to other agencies.

     1.    Decisions regarding  furnishing  data  to outside individuals or
           agencies shall be made by the Chief of the respective Survey.

     2.    Data are to  be  furnished  only  to  responsible scientists and
           research or user agencies  which establish a legitimate need for
           such data.

     3.    The requesting agency shall furnish a written description of the
           desired data and its proposed use.

     4.    Copies of published  or  unpublished  reports containing results
           derived from the outside  analyses  are  to  be  supplied to the
           Survey without charge.

     5.    Such  requests  from  agencies  of  Illinois  or   agencies  who
           participated  in  the  funding of the data acquisition  will  be
           without charge.   Requests  from  other  individuals or agencies
           will be assessed a reasonable charge to cover the cost of making
           the  data  available.   Such payment is to be deposited with the
           State of  Illinois or with the Survey's Indirect Cost Account as
           the Chief considers appropriate.
                                                  December 8,  1970

                                   -61-

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO                    2.

   EPA.-60Q/4 - 77-045	
 4. TITLE A\D SUBTITLE
  VERTICAL FLUXES AND EXCHANGE COEFFICIENTS IN THE
  AIR OVER ST. LOUIS
  Field Program  1975
                                                           3. RECIPIENT'S ACCESSION-NO.
                         5, REPORT DATE
                          November 1977
                         6. PERFORMING ORGANIZATION CODE
 7  AUTHOR(S)
                                                           8. PERFORMING ORGANIZATION REPORT NO
  Bernice Ackerman
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Illinois State Water Survey
  Urbana, Illinois 61801
                         10. PROGRAM ELEMENT NO.

                           1AA603
                         11. CONTRACT/GRANT NO.
                                                                R803682
 12. SPONSORING AGENCY NAME AND ADDRESS
  Environmental Sciences  Research Laboratory - RTF, N.C.
  Office of Research  and  Development
  Environmental Protection  Agency
  Research Triangle Park, N.C.  27711 	
                         13. TYPE OF REPORT AND PERIOD COVERED
                              Final	
                         14. SPONSORING AGENCY CODE

                             EPA/600/09
 15. SUPPLEMENTARY NOTES
 IS. ABSTRACT
      A field program'was carried out  in  the  greater metropolitan area of  St.  Lous,  MO
 during February and July of 1975 as part of  the Regional Air Pollution Study  (RAPS).
 Die purpose of the program was to collect atmospheric measurements needed for future
 studies of the planetary boundary layer  (PEL)  over urban and industrial areas and sur-
 rounding rural areas. The overall goals  of the PEL study are to  (1) describe  the ther-
 nodynamic, wind and turbulence fields over the region; (.2) determine the  magnitude  and
 vertical variation of the vertical fluxes of heat, moisture and momentum  as a function
 of land use; (3) obtain estimates of  the exchange coefficients of these variables;  and
 (4)  determine the dependence of turbulence intensity on land use. Pilot-balloon station;
 provided simultaneous measurements of the wind profile with vertical resolution.  Teth-
 ered-balloon sounding systems yielded thermodynamic and wind profiles. An airplane
 equipped with meteorological instruments provided measurements of the three components
 of wind velocity and -of high frequency fluctuations in velocity, temperature  and
 humidity.
      Observational periods, or missions,  were  scheduled for 3-or 4-hour durations
 during field experiments.  The objectives included (a) mapping missions to delineate
 the thermodynamic, wind and turbulent fields over the region, (b) flux missions to
 provide estimates of the true vertical fluxes  of momentum, heat and moisture  simulta-
 neously with vertical profiles of these  variables,and (c) nocturnal missions  to
    fide infer
;th  of th
:turanl heat island circulationi
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
            b.IDENTIFIERS/OPEN ENDED TERMS
                                                                        c.  cos AT I Field/Group
Air  pollution
*Meteorological Data
Measurement
Boundary  Layer
             St.  Louis, MO
                               13B
                               04B
                               20D
  . ^l?T^ISUTION STATEMENT
 RELEASE  TO PUBLIC
                                              19. SECU
                                       21. NO. OF PAGES
                                              72
                                              20. SECURITY CLASS (This page I
                                                    UNCLASSIFIED
                                       22. PRICE
E°A form 2220-1 (9-73)
                                           -62-

-------