NASA Technical Reports Server (NTRS) 19990099283: Local Data Integration in East Central Florida

12 Pages·1.2 MB·English

by  NASA Technical Reports Server (NTRS)

P2.11
.s t,¢ -,J
gi-- --_¢ I
LOCAL DATA INTEGRATION IN EAST CENTRAL FLORIDA
Jonathan L. Case* ar,dJohn T. Manobianco
NASA Kennedy Space Center / Applied Meteorology Unit / ENSCO, Inc.
1980 N. Atlantic Ave. Suite 230
Cocoa Beach, FL 32931
1. INTRODUCTION 4 km), and KSC/CCAS tower observations (4 km). The
greatest frequency of observations are provided by WSR-88D,
The Applied Meteorology Unit has configured a Local KSC/CCAS towers, profilers, ACARS (Aeronautical Radio,
Data Integration System (LDIS) for east central Florida which Inc. [ARINC] Communications, Addressing, and Reporting
assimilates in-situ and remotely-sensed observational data into System), and GOES-8 satellite imagery (< 15 minutes). The
a series of high-resolution gridded analyses. The ultimate largest variability in horizontal/vertical coverage and density
goal for running LDIS is to generate products that may occurs with aircraft data, satellite soundings, and satellite-
enhance weather nowcasts and short-range (< 6 h) forecasts derived winds. The maximum density of near-surface wind
issued in support of the 45_ Weather Squadron (45 WS), and temperature observations occurs within KSC/CCAS as a
result of the tower network. Manobianco and Nutter (1998)
Spaceflight Meteorology Group (SMG), and the Melbourne
National Weather Service (NWS MLB) operational provide further details on the data distribution, coverage, and
resolution used for this specific configuration of LDIS.
requirements.
LDIS has the potential to provide added value for
nowcasts and short-term forecasts for two reasons. First, it 3. ANALYSIS CONFIGURATION
incorporates all data operationally available in east central
Florida. Second, it is run at finer spatial and temporal The software utilized for LDIS is the Advanced Regional
resolutions than current national-scale 6perational models Prediction System (ARPS) Data Analysis System (ADAS;
such as the Rapid Update Cycle (RUC; Benjamin et al. 1998) Brewster 1996; Carr et al. 1996) available from the University
and Eta models. LDIS combines all available data to produce of Oklahoma. The analyses in ADAS are produced following
grid analyses of primary variables (wind, temperature, etc.) at Bratseth (1986) who developed an iterative Successive
specified temporal and spatial resolutions. These analyses of Corrections Method (SCM; Bergthorsson and Doos 1955) that
primary variables can be used to compute diagnostic converges to statistical or optimum interpolation (OI). In
quantities such as vorticity and divergence. general, Ol schemes are superior to SCM methods because
This paper demonstrates the utility of LDIS over east they can account for variations in data density, errors in the
central Florida for awarm season case study. The evolution data, and dynamical relationships between variables such as
of a significant thunderstorm outflow boundary is depicted the wind components and pressure.
through horizontal and vertical cross section plots of wind The configuration of ADAS follows the layout used for
speed, divergence, and circulation. In combination with a the terminal wind analysis in the Integrated Terminal Weather
suitable visualization tool, LDIS may provide users with a System (ITWS; Cole and Wilson 1995). ADAS is run every
more complete and comprehensive understanding of evolving 15 minutes at 0, 15, 30, and 45 minutes past the hour, over
mesoscale weather than could be developed by individually outer and inner grids with horizontal resolutions of 10km and
examining the disparate data sets over the same area and time. 2 km, respectively. The 10-km (2-km) analysis domain covers
an area of 500 x 500 km (200 x 200 km) and consists of 30
2. DATA COVERAGE AND RESOLUTION vertical levels that extend from the surface to about 16.5 km
above ground level. The vertical levels are stretched, with the
The utility of LDIS depends to a large extent on the finest resolution near the surface (20 m spacing) and the
reliability and availability of both in-situ and remotely-sensed coarsest resolution at upper levels (~1.8 km spacing).
observational data. Therefore, it is important to document all The background fields used on the 10-km analysis
existing meteorological data sources around central Florida domain are the RUC grids of temperature, wind, relative
that can be incorporated by LDIS. humidity, and geopotential height at 25-mb intervals from
1000 to 100 mb. The RUC analyses are available at a
Data density and coverage in east central Florida vary
considerably depending on the level in the atmosphere and horizontal resolution of 60-km every 3-h for the ease study
distance from KSC/CCAS. The data that are currently presented in this paper. The RUC grids are linearly
incorporated into LDIS include surface, buoy, ship, Kennedy interpolated in time every 15 minutes for each 10-km ADAS
Space Center/Cape Canaveral Air Station (KSC/CCAS) tower cycle. The resulting 10-km analysis grids are then used as
and wind profiler, GOES-8, WSR-88D, rawinsonde, and background fields for analyses on the 2-km domain. This
commercial aircraft observations. The data sources that nested-grid configuration and cascade-of-scales analysis
contain the finest horizontal resolution are WSR-88D follows that used for terminal winds in ITWS. With such an
(< 1km), GOES-8 visible and infrared imagery (1 km and approach, it is possible to analyze for different temporal and
spatial scales of weather phenomena.
The observational data are incorporated into ADAS using
*Corresponding author address: Jonathan Case, ENSCO, multiple passes of the Bratseth scheme to account for the
Inc., 1980 N. Atlantic Ave., Suite 230, Cocoa Beach, FL varying spatial resolution of the data sources. Five
3293 I. E-mail: [email protected].
computatiopnaaslseasreusedforthe10-kmgridandfour
Using the fine-scale results of the 2-km ADAS analyses,
passeasreutilizedonthe2-kmgrid. Datawithsimilar
the evolution of the wind speeds and wind vectors at 480 m
resolutioanrsegroupetodgetheinrthesamceomputational
(Figs. 2a-d) illustrates the formation and intensification of the
passsuchthatADASincorporaeteaschdatasourcweithout
outflow boundary during the late afternoon of 26 July. Wind
excessivsemlyoothitnhgeresolvabmleeteorologfeicaatul res.
speeds greater than 8m s"_develop over the Brevard/Osceola
Thismethodoloegnysurtehsaetachdatasourciesutilizeidn
county border at 2215 UTC 26 July (Fig. 2a) and spread
ADAStoitsmaximupmotentibaalseodnthemeteorological
radially over the next 45 minutes. The maximum winds (> 12
featurethsatthedatacanresolve.
m s"t) move northeastward into KSC/CCAS and offshore
regions of central Brevard county by 2245 UTC (Fig. 3c), the
4.
DIAGNOSIS OF AN OUTFLOW BOUNDARY approximate time that the Atlas launch was postponed.
Examination of level I1 radar reflectivity (Fig. I) and
A warm season case was selected from 26-27 July 1997 radial velocity data (not shown) indicates that features present
in order to investigate the capabilities and utility provided by in the high-resolution wind analyses (Fig. 2) are consistent
ADAS. Both 10-km and 2-km grid analyses were generated with the scale and motion of patterns associated with the
for the warm season case at 15-minutes intervals from 1800 observed thunderstorm. It should be noted that the detailed
UTC 26 July to 0200 UTC 27 July using all available data. structure of horizontal winds associated with this outflow
However, only results from the 2-km analysis grid are boundary would likely be easier to visualize in real-time using
presented in this section. ADAS rather than WSR-88D radial velocity displays alone.
A typical, undisturbed warm season environment The evolution of the thunderstorm outflow can also be
characterized the 26-27 July 1997 case. Early in the examined through the divergence of the horizontal wind on
afternoon, scattered thunderstorms developed across the the 2-km ADAS grid. In Figure 3, plots of convergence
peninsula and a sea-breeze boundary was evident along the (shaded), divergence (dashed lines), and horizontal winds are
east coast (not shown). By 2212 UTC, strong thunderstorms shown at 650 m for the same times as in Figure 2. At 2215
developed southwest of KSC/CCAS and generated an outflow UTC, an area of low-level convergence is found over much of
boundary that propagated northeastward as indicated by the central Brevard county (Fig. 3a). Weak divergence is also
Melbourne WSR-88D level II base reflectivity (Fig. la). The found over northeastern Osceola county at this time. Fifteen
leading edge of the outflow boundary intersected the eastern minutes later, an extensive area of low-level divergence
tip of KSC/CCAS by 2242 UTC as shown in Figure lb develops in southern Brevard county, behind the developing
(counties and location of KSC/CCAS are given in Fig. 2a). outflow boundary (Fig. 3b). Convergence along the outflow
This outflow boundary caused wind gusts greater than 15m s"1 boundary spreads outward into central Brevard county (just
as noted on the KSC/CCAS mesonet towers around 2245 south of KSC/CCAS), offshore of southern Brevard county,
UTC (not shown). This case was chosen because the strong and into Indian River county. Divergence continues to
winds associated with the outflow boundary forced Atlas intensify across much of interior Brevard county over the next
launch operation AI393 to be scrubbed for the day. 30 minutes beneath the dissipating thunderstorm (Figs. 3c and
d), while the band of convergence spreads out radially along
the leading edge of the outflow boundary.
Figure 1. Base reflectivity images are shown from the Melbourne WSR-88D on 26 July at a) 2212 UTC and b) 2242 UTC. The
leading edge of the outflow boundary is indicated by "OB" in panel b).
By2300UTCa,well-definbeadnodfconvergeanrcces
KSC/CCAS (Fig. la). The onset of the dissipating stage of
fromnortherOnsceoclaountya,crosesasterOnrangaend
the convection occurs at 2230 UTC when convergence
northeBrnrevarcdountiyn,totheoffshorweaterosfBrevard
transitions to divergence in the lowest 1000 m directly beneath
countya,ndthenbackonshoirneIndiaRnivercount(yFig.
the deep layer of maximum convergence (Fig. 4b). The
3d).Strondgivergenecxeceedi8nxg10s"-_occurosveerast-
leading edge of the outflow boundary is evident in Figure 4b
centraBlrevacrdounty.
given by the low-level convergence below 1000 m on either
Theverticasltructuoreftheconvectiiosnindicatebdy
side of the newly-developed low-level divergence.
north-souctrhosssectionosfdivergenacnedthevertical
By 2245 UTC, the area of low-level divergence expands
velocity(w)fieldthroughthedevelopinthgunderstorm
laterally and upward to 3000 m (Fig. 4c). The leading edge of
outflow(Fig.4and5respectivelAy)c.cordintogBrewster
the outflow boundary spreads rapidly north and southward
(1996),
w is derived within ADAS from the analyzed (left and right, primarily below 1000 m) on either side of the
horizontal winds (via continuity) and a constraint that the intensifying low-level divergence maximum. Meanwhile,
wind velocity normal to the bottom and top boundaries is zero. mid-level convergence remains strong, exceeding -6x 10"_s" in
Deviations from these boundary conditions are treated as the 3500--5500-m layer. By 2300 UTC, a well-developed
errors in the horizontal divergence that vary linearly with signature exists for a downburst-producing thunderstorm (Fig.
height. The w field is adjusted once these errors are removed 4d). Strong mid-level convergence in the 3000-6000-m layer
and the horizontal wind field is relaxed such that total mass overlies a maximum in low-level divergence which exceeds
divergence domain-wide is zero. 8x10' s".
A deep column of convergence is prevalent at2215 UTC
(Fig. 4a) associated with the strong convection southwest of
Figure 2. A display of the 2-km ADAS wind speed and wind vectors at 480 m. Wind speed is contoured every 2 m s", shaded
above 6 m s', while wind vectors are denoted by arrows. Valid times are a) 2215 UTC, b) 2230 UTC, c) 2245 UTC, and d) 2300
UTC 26 July. Counties and the location of KSC/CCAS are labeled in a).
Notice
that the low-level convergence slopes upward convective cell with values of w exceeding 10cm s"_from 750
from near the surface in the middle of Figure 4c to 500 m at m up to the top of the cross section. Also, a southerly wind
the noah/south (letVright) edges. This structure results from component prevails across the southern (fight) half of Figure
the incorporation of WSR-88D radial velocity data which 5a (most notably below 3000 m) which feeds the updraft of
slopes upward away from the radar site located approximately the storm.
in the middle of the cross section. Above 200 m, surface data During the transition phase of the thunderstorm, an area
do not influence the analyses. Therefore changes in the of sinking motion is indicated in the lowest 2-3 km beneath
horizontal winds and corresponding divergence field above the still strong updraft (Fig 5b). Notice that the prevailing
the surface develop primarily in response to WSR-88D radial southerly wind component has weakened below 1000 m in the
velocities in areas where radar reflectivity targets are southern (right) portion of Figure 5b. By 2245 UTC, a
available. significant but narrow downdraft has developed from near the
Finally, the vertical structure of the kinematically- surface up to 6000 m between two updrafts (Fig. 5c). Sinking
derived vertical velocity field associated with the dissipating motionreaches 75 cm s-_by this time at roughly 3500 m. In
convection is depicted in Figure 5. Also shown in Figure 5 is the lowest 1000 m of Figure 5c, southerlies increase in
the evolution of the derived circulation (arrows) in the plane intensity to the north (left) of the strengthening downdrafl
of the cross section. At 2215 UTC, a strong updraft greater whereas winds have turned around to a slight northerly
than 200 cm s_ occurs above 5500 m (Fig. 5a) in conjunction component below 1000 m to the south (right) of the
with the deep column of convergence in Figure 4a. Rising downdraft.
motion prevails throughout much of the depth of the
Figure 3. Divergence of the horizontal wind (x l0_ s") and wind vectors at 650 m derived from the 2-km ADAS grids. Dashed
lines indicate divergence and shading indicates convergence. Valid times are a) 2215 UTC, b) 2230 UTC, c) 2245 UTC, and d)
2300 UTC 26 July.
Thematurpehasoeftheoutflowboundaarty230U0TC withmesoscaplheenomethnaatcanbeusedto
initialize
featureasmassivdeowndrainftthecenteorfFigur5ed.The
numerical weather prediction models.
circulatiovnectorinsdicattehatasthedowndranfetartshe
surfac(ebelow1000m),thewindstrongldyivergeasnd
5. SUMMARY
spreadlasteralljyustabovtehegrounrdesultiningstrong
southerltioetshenorth(left)ofthedowndraanftdanortherly
Results from the 26-27 July 1997 case demonstrate that
windcomponteonthtesout(hrighto)fthemaidnowndraft.
subsequent 15-min analyses of horizontal winds and its
It isinterestitnognotethatsoutherlipersevabilelow
associated divergence and derived vertical motion fields on
100m0ovetrhelengtohfthecrosssectioant221U5TC(Fig.
the 2-km ADAS domain can depict the formation and
5a).Howevethr,eoutflowboundaaryctstostrengthtehne
propagation of a thunderstorm outflow boundary. LDIS has
low-levseolutherltioetshenorth(left)ofthedowndrwafhtile
thesoutherliweesakeanndturntonortherliteosthesouth the potential to provide added value because it can incorporate
data which are currently available only at KSC/CCAS and run
(righto)fthedowndrbayft230U0TC(Fig5.d).
at finer spatial and temporal resolutions over smaller domains
Forthemethodolougsyedinthiscaseth, eanalyzed
than current national-scale, operational models such as the
thermodynavmariciableasrenotconsistewnitththederived
RUC and Eta. Furthermore, itis noticeably easier to diagnose
kinemaftiieclds(i.en.ocoldpooils.analyziendthevicinityof
and visualize the vertical motion fields and outflow boundary
thestrongthunderstodrmowndraft)H.owevear,more
using ADAS analyses rather than strictly radial velocity data
sophisticatetecdhniquiseavailabwleithinADAStoretrieve
atmultiple elevation angles.
thethree-dimenstieomnaplerataunredpressudriestributions
fromDopplerradardata.Thisschemperovidemsore
representaatinvaelyseosfmasasndwindfieldsassociated
Divergence 2215 UTC Divergence 2245 UTC
i7000 700(
600(
6000 1
5000 500(
"t
4000 400(
-2
3000 ,_00(
-4
2000 • -6 200(
to0o • -6 100(
O.
28.7;-00.7 27.7;-80.5 a 20.7;-00.7 27.7;-80.5 C
I
Divergence 2230 UTC Divergence 2300 UTC
7000 ,
6000 -I
5000 q
4000
,_000
lOOO
o
29.7;40.7 27.7;40.5 _) 26.7;-e0.7 27.7i-00.5 d
Figure 4. Cross sections of divergence of the horizontal wind (x 10" s-', derived from the 2-km ADAS grids) along north-south
oriented lines indicated in Figure 3. Dashed lines indicate divergence while shading represents convergence according to the scale
provided. Valid times are a) 2215 UTC, b) 2230 UTC, c) 2245 UTC, and d) 2300 UTC 26 July. The ordinate ranges from 0 to
7000 m whereas the latitude and longitude are labeled at the endpoints of the abscissa.
6.
REFERENCES Cart, F. H., J. M. Krause, and K. Brewster, 1996: Application
of the Bratseth scheme to high-resolution analyses in
Benjamin, S. G., J. M. Brown, K. J. Brundage, D. Devenyi, B. inhomogeneous data regimes. Preprints, 15" Conf. on
E. Schwartz, T. G. Smirnova, T. L: Smith, L. L. Morone, Weather Analysis and Forecasting, Norfolk, VA, Amer.
and G. J. DiMego, 1998: The operational RUC-2. 16'_ Meteor. Soc., 231-234.
Conf. on Weather Analysis and Forecasting, Phoenix,
Cole, R. E., and W. W. Wilson, 1995: ITWS grldded winds
AZ, Amer. Meteor. Soc., 249-252.
product. Preprints, 6'_ Conf. on Aviation Weather
Bergthorsson, P., and B. Doos, 1955: Numerical weather map Systems, Dallas, TX, Amer. Meteor. Soe., 384-388.
analysis. Tellus, 7, 329-340.
Manobianco, J. and P. Nutter, 1998: A local data integration
Bratseth, A. M., 1986: Statistical interpolation by means of system configured for weather support in east central
successive corrections. Tellus, 38A, 439-447. Florida. Preprints, 2" Syrup. on Integrated Observing
Systems, Phoenix, AZ, Amer. Meteor. Sot., 64-67.
Brewster, K., 1996: Application of aBratseth analysis scheme
including Doppler radar data. Preprints, 15'h Conf. on
Weather Analysis and Forecasting, Norfolk, VA, Amer.
Meteor. Soc., 92-95.
W (cm/s), 2215 UTC w (om/s), 2245 UTC
, , , _.. _,,_df[t
- _ "TT_ t_'i " t__
_I. !i),,rr_,_LI ]2![ .... ] L:
5000 ....... _.- ,,. _,._,'t)___+,_ll'_.J........ _,J...,L ' i 5000 ' _ ' :' ,-._!_._,_.-._m......_*_._- ,.
..... _ _ "1_ '.'t!ltl zr., r fe -10
4000 ......... ---,.\ i,,'_._,'d,;_'_' .... ,_,.L-,,i- -25
"v'_ L ._.... ,,_;; _ ,_nL_._j._x'._;_l_ .i_,_ .25
.............. _ _.,,,._,_ '_...,,_--,.-Lso
3000 r'_.... ,_ , ,_ '/ -_, ' _i -75
........... _ ".""! ___ _o- "T;-_,'T'_
2000 i ....._ ....n.s .................. -100
........... .L.,_L.... I,,,.... _%,,_,,_, .... ,"......... -150
t000 :: .........•._..........._...... "_'_ /....... +-_. -200
II II Illllll Illlql IIIFIII lit 14 IIt [till ;lllit I_jl|l_ [|1 _
28.7;-00,7 27.7;40.5 a 2B.7;-BO.7 27.7;-80,5 C
W (cm/s), 2230 UTC W (omls), 2300 UTC
7ooo,_"Y'"'":,, Y"__s'_.,;'r,,_":.':....._.r,,_,_i;_ 1_,1__ 7000 _" ";.... _'_'_.... _:- ":':-'"'"_" ,",'"
_vvv _ I _ q _ i_
5000 _''` _ '_"' /_ '_'J_ _-T_'_L'_'i-_' _ooo_' "_£.
I
-lO . I lO
4 00 4 . z ' ._ -25 4000 " .25
3000 -so 50
-75 75
-lOO 3000 ' lOO
.15o 2000 ._.._,, _.... _S._,_,_:,.;...'_" -_._. ..... _:__ 150
-2oo lOOO 200
28.7;-110.7 27.7;40.5 b 211,7;-80.7 27,7;-80,_ d
Figure 5. Cross sections of vertical velocity (cm s", derived from the 2-km ADAS grids) along north-south oriented lines
indicated in Figure 3. Dashed lines indicate rising motion whereas shading represents sinking motion according to the scale
provided. Valid times are a) 2215 UTC, b) 2230 UTC, c) 2245 UTC, and d) 2300 UTC 26 July. The ordinate ranges from 0 to
7000 m whereas the latitude and longitude are labeled at the endpoints of the abscissa.
P2.11
LOCAL DATA INTEGRATION IN EAST CENTRAL FLORIDA
Jonathan L. Case* and John T.Manobianco
NASA Kennedy Space Center / Applied Meteorology Unit / ENSCO, Inc.
1980 N. Atlantic Ave. Suite 230
Cocoa Beach, FL 32931
I. INTRODUCTION 4 km), and KSC/CCAS tower observations (4 km). The
greatest frequency of observations are provided by WSR-88D,
The Applied Meteorology Unit has configured a Local KSC/CCAS towers, profilers, ACARS (Aeronautical Radio,
Data Integration System (LDIS) for east central Florida which Inc. [ARINC] Communications, Addressing, and Reporting
assimilates in-situ and remotely-sensed observational data into System), and GOES-8 satellite imagery (< 15 minutes). The
a series of high-resolution gridded analyses. The ultimate largest variability in horizontal/vertical coverage and density
goal for running LDIS is to generate products that may occurs with aircraft data, satellite soundings, and satellite-
enhance weather nowcasts and short-range (< 6 h) forecasts derived winds. The maximum density of near-surface wind
issued in support of the 45'hWeather Squadron (45 WS), "and temperature observations occurs within KSC/CCAS as a
Spaceflight Meteorology Group (SMG), and the Melbourne result of the tower network. Manobianco and Nutter (1998)
National Weather Service (NWS MLB) operational provide further details on the data distribution, coverage, and
requirements. resolution used for this specific configuration of LDIS.
LDIS has the potential to provide added value for
nowcasts and short-term forecasts for two reasons. First, it 3. ANALYSIS CONFIGURATION
incorporates all data operationally available in east central
Florida. Second, it is run at finer spatial and temporal The software utilized for LDIS is the Advanced Regional
resolutions than current national-scale operational models Prediction System (ARPS) Data Analysis System (ADAS;
such as the Rapid Update Cycle (RUC; Benjamin et al. 1998) Brewster 1996; Cart etal. 1996) available from the University
and Eta models. LDIS combines all available data to produce of Oklahoma. The analyses in ADAS are produced following
grid analyses of primary variables (wind, temperature, etc.) at Bratseth (1986) who developed an iterative Successive
specified temporal and spatial resolutions. These analyses of Corrections Method (SCM; Bergthorsson and Doos 1955) that
primary variables can be used to compute diagnostic converges to statistical or optimum interpolation (OI). In
quantities such as vorticity and divergence. general, Ol schemes are superior to SCM methods because
This paper demonstrates the utility of LDIS over east they can account for variations in data density, errors in the
central Florida for a warm season case study. The evolution data, and dynamical relationships between variables such as
of a significant thunderstorm outflow boundary is depicted the wind components and pressure.
through horizontal and vertical cross section plots of wind The configuration of ADAS follows the layout used for
speed, divergence, and circulation. In combination with a the terminal wind analysis in the Integrated Terminal Weather
suitable visualization tool, LDIS may provide users with a System (ITWS; Cole and Wilson 1995). ADAS is run every
more complete and comprehensive understanding of evolving 15 minutes at 0, 15, 30, and 45 minutes past the hour, over
mesoscale weather than could be developed by individually outer and inner grids with horizontal resolutions of 10km and
examining the disparate data sets over the same area and time. 2 km, respectively. The 10-km (2-kin) analysis domain covers
an area of 500 x 500 km (200 x 200 km) and consists of 30
2. DATA COVERAGE AND RESOLUTION vertical levels that extend from the surface to about 16.5 km
above ground level. The vertical levels are stretched, with the
The utility of LDIS depends to a large extent on the finest resolution near the surface (20 m spacing) and the
reliability and availability of both in-situ and remotely-sensed coarsest resolution atupper levels (-1.8 km spacing).
observational data. Therefore, it is important to document all The background fields used on the 10-km analysis
existing meteorological data sources around central Florida domain are the RUC grids of temperature, wind, relative
that can be incorporated by LDIS. humidity, and geopotential height at 25-mb intervals from
Data density and coverage in east central Florida vary I000 to 100 mb. The RUC analyses are available at a
considerably depending on the level in the atmosphere and horizontal resolution of 60-km every 3-h for the case study
distance from KSC/CCAS. The data that are currently presented in this paper. The RUC grids are linearly
incorporated into LDIS include surface, buoy, ship, Kennedy interpolated in time every 15 minutes for each 10-km ADAS
Space Center/Cape Canaveral Air Station (KSC/CCAS) tower cycle. The resulting 10-km analysis grids are then used as
and wind profiler, GOES-8, WSR-88D, rawinsonde, and background fields for analyses on the 2-km domain. This
commercial aircraft observations. The data sources that nested-grid configuration and cascade-of-scales analysis
contain the finest horizontal resolution are WSR-88D follows that used for terminal winds in ITWS. With such an
(< 1km), GOES-8 visible and infrared imagery (1 km and approach, it is possible to analyze for different temporal and
spatial scales of weather phenomena.
The observational data are incorporated into ADAS using
*Corresponding author address: Jonathan Case, ENSCO, multiple passes of the Bratseth scheme to account for the
Inc., 1980 N. Atlantic Ave., Suite 230, Cocoa Beach, FL varying spatial resolution of the data sources. Five
32931. E-mail: [email protected].
computatiopnaaslseasreusedforthe10-kmgridandfour
passeasreutilized Using the fine-scale results of the 2-km ADAS analyses,
on the 2-km grid. Data with similar the evolution of the wind speeds and wind vectors at 480 m
resolutions are grouped together in the same computational (Figs. 2a-d) illustrates the formation and intensification of the
pass such that ADAS incorporates each data source without outflow boundary during the late afternoon of 26 July. Wind
excessively smoothing the resolvable meteorological features. speeds greater than 8 m st develop over the Brevard/Osceola
This methodology ensures that each data source is utilized in county border at 2215 UTC 26 July (Fig. 2a) and spread
ADAS to its maximum potential based on the meteorological radially over the next 45 minutes. The maximum winds (> 12
features that the data can resolve. m s") move northeastward into KSC/CCAS and offshore
regions of central Brevard county by 2245 UTC (Fig. 3c), the
4. DIAGNOSIS OF AN OUTFLOW BOUNDARY approximate time that the Atlas launch was postponed.
Examination of level II radar reflectivity (Fig. 1) and
A warm season ease was selected from 26-27 July 1997 radial velocity data (not shown) indicates that features present
in order to investigate the capabilities and utility provided by in the high-resolution wind analyses (Fig. 2) are consistent
ADAS. Both 10-km and 2-km grid analyses were generated with the scale and motion of patterns associated with the
for the warm season case at 15-minutes intervals from 1800 observed thunderstorm. It should be noted that the detailed
UTC 26 July to 0200 UTC 27 July using all available data. structure of horizontal winds associated with this outflow
However, only results from the 2-km analysis grid are boundary would likely be easier to visualize in real-time using
presented in this section. ADAS rather than WSR-88D radial velocity displays alone.
A typical, undisturbed warm season environment The evolution of the thunderstorm outflow can also be
characterized the 26-27 July 1997 case. Early in the examined through the divergence of the horizontal wind on
afternoon, scattered thunderstorms developed across the the 2-km ADAS grid. In Figure 3, plots of convergence
peninsula and a sea-breeze boundary was evident along the (shaded), divergence (dashed lines), and horizontal winds are
east coast (not shown). By 2212 UTC, strong thunderstorms shown at 650 m for the same times as in Figure 2. At 2215
developed southwest of KSC/CCAS and generated an outflow UTC, an area of low-level convergence is found over much of
boundary that propagated northeastward as indicated by the central Brevard county (Fig. 3a). Weak divergence is also
Melbourne WSR-88D level II base reflectivity (Fig. la). The found over northeastern Osceola county at this time. Fifteen
leading edge of the outflow boundary intersected the eastern minutes later, an extensive area of low-level divergence
tip of KSC/CCAS by 2242 UTC as shown in Figure Ib develops in southern Brevard county, behind the developing
(counties and location of KSC/CCAS are given in Fig. 2a). outflow boundary (Fig. 3b). Convergence along the outflow
This outflow boundary caused wind gusts greater than 15m s1 boundary spreads outward into central Brevard county (just
as noted on the KSC/CCAS mesonet towers around 2245 south of KSC/CCAS), offshore of southern Brevard county,
UTC (not shown). This case was chosen because the strong and into Indian River county. Divergence continues to
winds associated with the outflow boundary forced Atlas intensify across much of interior Brevard county over the next
launch operation A1393 to be scrubbed for the day. 30 minutes beneath the dissipating thunderstorm (Figs. 3c and
d), while the band of convergence spreads out radially along
the leading edge of the outflow boundary.
Figure I. Base reflectivity images are shown from the Melbourne WSR-88D on 26 July at a) 2212 UTC and b) 2242 UTC. The
leading edge of the outflow boundary is indicated by "OB" inpanel b).
By2300UTC,
a well-defined band of convergence arcs KSC/CCAS (Fig. la). The onset of the dissipating stage of
from northern Osceola county, across eastern Orange and the convection occurs at 2230 UTC when convergence
northern Brevard county, into the offshore waters of Brevard transitions to divergence in the lowest 1000 m directly beneath
county, and then back onshore in Indian River county (Fig. the deep layer of maximum convergence (Fig. 4b). The
3d). Strong divergence exceeding 8x104 s"_occurs over east- leading edge of the outflow boundary is evident in Figure 4b
central Brevard county. given by the low-level convergence below 1000 m on either
The vertical structure of the convection is indicated by side of the newly-developed low-level divergence.
north-south cross sections of divergence and the vertical By 2245 UTC, the area of low-level divergence expands
velocity (w) field through the developing thunderstorm laterally and upward to 3000 m (Fig. 4c). The leading edge of
outflow (Fig. 4 and 5 respectively). According to Brewster the outflow boundary spreads rapidly north and southward
(1996), w is derived within ADAS from the analyzed (left and right, primarily below 1000 m) on either side of the
horizontal winds (via continuity) and a constraint that the intensifying low-level divergence maximum. Meanwhile,
wind velocity normal to the bottom and top boundaries is zero. mid-level convergence remains strong, exceeding -6x 104 s1 in
Deviations from these boundary conditions are treated as the 3500-5500-m layer. By 2300 UTC, a well-developed
errors in the horizontal divergence that vary linearly with signature exists for adownburst-producing thunderstorm (Fig.
height. The w field is adjusted once these errors are removed 4d). Strong mid-level convergence in the 3000-6000-m layer
and the horizontal wind field is relaxed such that total mass overlies a maximum in low-level divergence which exceeds
divergence domain-wide is zero. 8xI0"4s".
A deep column of convergence is prevalent at 2215 UTC
(Fig. 4a) associated with the strong convection southwest of
'_%,_,2215 UTC 26 July _tT1 ',,_,t,2245 UTC 26 July T_I
' r t _ K_r t .................... _ _ I I 9[ ' t'vJ_,r _ ................. t I t 1
,,_,, ..... ,.,,.-,r• , _.-_...,tll_l-,i/ll/_ _'i,_,_'_'_','_",7_I , ,b , • , __, ,____3).Z_!_t_ili_ _.3_ 't'_,i'ti'_'t
' 't' ' ' __°_"---'--'' ' '_ '_ _YI"_ '_'t" ¢'l" '__'7"1 ''t ..... " __'___)q."_E_ _] _" '_'t'q''
'J[ ......... _" "/'"_'_ i_______ ____"_1 /''i ...... _ +' __ ]__-_( ( T_x" t,',
• ' N" ""'_"'J"'_'_'/'S_.,.,_KT _ ",,",,"_,",," ' ".,_I1,_.",,i| _",_,•/"• _..'tji__, _.
..... ''"''" '1_'/ .... _ ' _' "_--'__<.,_" ' ' ' '_'
,,....._..". "__,_'_ i _ ,_.__,_.____1't_ ...... _ " '\ .... "_;-_' ' ____\_
.......... _'_'_ ................... x_ • "'-'-J'l ,f_,\,l, __,,_,_• _
t _/ I t rl_ / / tl_ _%t t r t l] 1 i t i 1 t_ ##_x.t.*tt_,,,_,,,t,A,t,,l_t"_l_!,.1, lfll I
.le uZa, II III :;',_',J-l_°° V.Te. 2tl a'uly IIiii
fir \l I .................... Nil I 1]
_lll r_tl I I I lltllkt llt Ill _1111111111 ,,>i,,_,77 ;ii_tli, i_i_i iii'_.i ill li _
?i)l 11211 I I I 11 L._1%I I I I I 1%%I t I I I I II , _#,1_, ftt__t f I t, iti_/ t t t I I t i 1%1 i i i I 1
I It1 i I 1"%l 1'i I 1%tllx% _ I t %%1%I I I I 1X III ,, ,I,,, ,'-_,,,, _, ,i.k, i ____I _,\_ ___ _I
--l,-f-ll I I I i'll I I I I \ 'i _, '_1 I _ 1% I '_ t ¶ I I ', _ II ,--_--_.i, I. t _ fl / t t t I i I _t_l< I I I i I i i I I,_I I I I 1
I t l I 1r<..l"t_.'i ".%1'_ k\tt'_%ttt'_%t_It tl _ I I I I tp'_'._ / r t I i _ _k\l 1% I I I 1% _1 I I I 1
? _,I.-L-L-L_ I t_t.,l,c_hl L-II__\'_\'_ I ¶ ¶ ¶ t t ¶ %%I t \ tI _ _ t t_L..iDh _ 7_li\l I 1% I I I 1%\11 1% 1
I *tl.l _;I t %X%'.t'_ ' %litX%It%"_"_',t ',t t I t _i ,,/,,,,,,_ ___,_L,I__/_.a.llt','_V'_A.__I _1 t %'
11111\t_i%ltl%t%% _'%t'_lf'X_'\t\\\'_\'_ti ,,I,,,,,,, _ _t _'_. '__].I1 11%9%
' _ __,/-' __/_____1 ,,-...
,,,_,_,_l_, t';-,_--_%,,k_'_)/'_t',_,_.', I .............
, , t , , ,, , , ,'t,, • _: ___,,t_ _.t __j) I ............... ,,
Figure 2. A display of the 2-km ADAS wind speed and wind vectors at 480 m. Wind speed is contoured every 2 m s_, shaded
above 6 ms", while wind vectors are denoted by arrows. Valid times are a) 2215 UTC, b) 2230 UTC, c) 2245 UTC, and d) 2300
UTC 26 July. Counties and the location of KSC/CCAS are labeled ina).
Notice that the low-level convergence slopes upward convective cell with values of wexceeding 10cm s" from 750
from near the surface in the middle of Figure 4c to 500m at m up to the top of the cross section. Also, a southerly wind
the north/south (left/right) edges. This structure results from component prevails across the southern (right) half of Figure
the incorporation of WSR-88D radial velocity data which 5a (most notably below 3000 m) which feeds the updraft of
slopes upward away from the radar site located approximately the storm.
inthe middle of the cross section. Above 200 m, surface data During the transition phase of the thunderstorm, an area
do not influence the analyses. Therefore, changes in the of sinking motion is indicated in the lowest 2-3 km beneath
horizontal winds and corresponding divergence field above the still strong updraft (Fig 5b). Notice that the prevailing
the surface develop primarily inresponse to WSR-88D radial southerly wind component has weakened below 1000m inthe
velocities in areas where radar reflectivity targets are southern (right) portion of Figure 5b. By 2245 UTC, a
available. significant but narrow downdraft has developed from near the
Finally, the vertical structure of the kinematically- surface up to6000 m between two updrafts (Fig. 5c). Sinking
derived vertical velocity field associated with the dissipating motion reaches 75cm s" by this time at roughly 3500 m. In
convection isdepicted in Figure 5. Also shown in Figure 5is the lowest I000 m of Figure 5c, southerlies increase in
the evolution of the derived circulation (arrows) in the plane intensity to the north (left) of the strengthening downdraft
of the cross section. At 2215 UTC, a strong updraft greater whereas winds have turned around to a slight northerly
than 200 cm s-'occurs above 5500 m (Fig. 5a) in conjunction component below 1000 m to the south (right) of the
with the deep column of convergence in Figure 4a. Rising downdraft.
motion prevails throughout much of the depth of the
Figure 3. Divergence of the horizontal wind (x 10"_ and wind vectors at 650m derived from the 2-km ADAS grids. Dashed
S"1)
lines indicate divergence and shading indicates convergence. Valid times are a)2215 UTC, b) 2230 UTC, c) 2245 UTC, and d)
2300 UTC 26July.