THUNDERSTORM INITIATION AND EVOLUTION DURING IHOP: IMPLICATIONS FOR AVIATION THUNDERSTORM NOWCASTING

THUNDERSTORM INITIATION AND EVOLUTION DURING IHOP:
IMPLICATIONS FOR AVIATION THUNDERSTORM NOWCASTING

James W. Wilson and Rita D. Roberts

National Center for Atmospheric Research
jwilson@ucar.edu

1. Introduction                                      area, location of surface stations, radiosonde
In this paper we summarize some findings from        locations and radar locations. Visible and IR
the International H2O Project (IHOP) which was       data were available from GOES-8 and 11. These
conducted over a 400,000 km2 area in Kansas,         data sets were used to identify storm initiation
Oklahoma and Texas during the spring of 2002.        locations and times, as well as, to identify and
These findings on thunderstorm initiation and        characterize boundaries. Radar mosaics were
evolution are presented in a paper prepared by       prepared at 10 min intervals for the entire 44 day
the above authors (Wilson and Roberts 2005).         period of IHOP. Their primary use was to
Here we examine those findings in relationship       identify storm initiation locations, identify and
to implications for nowcasting thunderstorms for     track boundaries and to monitor storm evolution.
aviation purposes.
                                                     3. Analysis
A primary goal of IHOP was to study                  Storm initiations that clustered in time and space
thunderstorm initiation using a very impressive      were identified and called storm initiation
collection of radars, lidars, surface mesonet        episodes. An episode consisted of two or more
stations, soundings and rapid scan satellite.        cell initiations whose close appearance in time
Wilson and Roberts (2005) examined all               and space suggested a common forcing
convective storm initiations (radar cells            mechanism. The number of cells in an initiation
exceeding 40 dBZ) occurring during the 44 days       episode varied between 2 and 55 over time
of IHOP. These events were identified and            periods varying between 20 and 200 min. A total
grouped into 112 initiation episodes. As part of     of 112 initiation episodes were identified during
this study all surface convergence lines were        the 44 day study period. For each initiation
identified and tracked during IHOP so that their     episode the following were recorded: location,
impact could be examined on storm initiation         number of individual cells that initiated between
and evolution. A variety of environmental            the beginning and end of the episode, the
variables were then used to characterize the         orientation and size of the episode and the
conditions in the area of each initiation episode.   suspected initiation mechanism.
Environmental variables included surface
convergence, CAPE and CIN.                           Forcing mechanisms for initiation episodes were
                                                     divided into two groups: surface based and
The 3 and 6 h forecasts from the operational         elevated. The classification of surface based
numerical forecast model called the Rapid            required the observation of a nearby boundary.
Update Cycle (RUC, Benjamin et al. 2004a, b)         Boundaries were classified into 7 categories;
were examined for its ability to forecast            frontal, gust fronts, trough lines, dry lines,
thunderstorm initiation movement and evolution       colliding, bores and unknown. If no surface
during IHOP. A special 10 km version of the          convergence feature was identified it was
operational RUC (RUC10) was specifically run         classified as an elevated initiation episode.
by the NOAA Forecast System Laboratory for
IHOP.                                                The evolution of each initiation episode was
                                                     documented and was classified at its mature
2. Data                                              stage as a multi-cell complex, linear feature or
Particularly important for the above studies were    squall line. A squall line was differentiated from
the mesonet, radiosonde, radar and satellite data.   a linear feature by the presence of a gust front.
Within the IHOP study area there were about
275 surface stations generally reporting wind,       Computer programs were developed for
temperature and dewpoint at time intervals           obtaining a) high resolution near surface
between 1 and 60 min. Fig. 1 shows the study         divergence fields from the surface stations, and

b) high resolution fields of Convective Available        again reinforced. Thus it is essential to
Potential Energy (CAPE) and Convective                   continue to vigorously pursue human
Inhabitation (CIN) derived from the IHOP                 efforts to insert boundaries into the
sounding data set and a lifted surface parcel            nowcast systems. The observation that
based on the mesonet data.                               new storm initiation tended to occur in
                                                         close time and space proximity to each
The evolution of the initiation episodes covered a       other reinforces the desirability to
great range of situations including intense              include in the nowcast systems a means
supercells, squall lines, mesoscale convective           to increase storm initiation likelihood
systems, short lived convective lines, and short         near locations that have recently had
lived unorganized groups of storms. The                  storm initiation.
evolution and lifetime of the initiation episodes
was closely tied to the development of gust              Even though elevated nocturnal
fronts. Elevated systems were much less likely to        convection and its initiation is a
develop gust fronts thus they were less likely to        common event over much of the central
become well organized with long lifetimes.               and eastern U.S., the Auto-nowcaster,
                                                         Growth and Decay Tracker and NCWF
4. IHOP findings and implications for                    nowcast systems have no procedures in
aviation weather nowcasting                              place to nowcast these events. This is
                                                         primarily because of a lack of basic
For a number of years NCAR and Lincoln                   understanding of how these events are
Laboratory have been involved in developing,             triggered and factors influencing their
testing and implementing operational 0-2 hr              evolution. New understanding from
thunderstorm forecasting systems for aviation            IHOP discussed below may provide
enroute and terminal weather. These systems are          some insight.
called the Auto-nowcaster (Mueller et al 2003),
National Convective Weather Forecast product         2) The cause of many of the elevated
(NCWF, Megenhardt et. al 2000) and the                  initiations during IHOP appeared to be
Growth and Decay Tracker (Wolfson et al.                associated with synoptic or mesoscale
1998). Most recently there has been a desire to         wind convergence or confluence at mid-
extend this period to 6 h. While these systems          levels (between 900 and 600 hPa). This
make extensive use of radar extrapolation               convergence was often observed in the
techniques they also forecast convective storm          RUC analysis wind fields. However,
initiation, growth and dissipation. These               there is no known method for
techniques combine forecast parameters from             anticipating the specific time of
numerical models, statistical methods and               initiation with these features. It was
heuristic methods. Future significant forecast          uncertain if the RUC had the skill to
improvements will likely be closely tied to             forecast these convergence features or
improved basic understanding and improved               the initiation of elevated convection. In
observations. Summarized in this section are            addition bores were frequently observed
findings from the IHOP research reported in             during the night which infrequently
Wilson and Roberts (2005) and their possible            initiated convective storms.
implications for nowcasting
                                                         Implication: Consideration should be
    1) The initiation episodes during IHOP               given toward producing an elevated
       were almost evenly divided between                storm initiation interest field that
       surface based and elevated. The surface           identifies      intense         prolonged
       based initiations occurred mostly during          convergence fields between 900 and
       the afternoon and early evening and the           600 hPa. Improved basic understanding
       elevated initiations during the night and         of elevated storm initiation is in need of
       early morning. The surface based                  research. However means for directly
       episodes were triggered mostly by                 observing detailed wind and stability
       synoptic fronts and gust fronts.                  parameters are not yet possible.

         Implication: The importance of              3) While individual storms typically have
         boundaries in initiating convection is         lifetimes of only 10 to 60 min the

complex of storms associated with an       5) Given the observed importance of
initiation episodes had lifetimes of       gust fronts and their associated
hours.     The lifetime of the storm       convergence on the evolution and
complex was related to whether the         motion of the initiated storm complexes
convective system produced a gust          it is essential for very short period
front. Those systems that did not          forecasting techniques to anticipate
produce a gust front lived between 2       which storms will produce gust fronts
and 4 h whereas those that did produce     and the strength of the cold pools; this
a gust front typically lived >6 h.         is a major research challenge for
                                           observational and numerical model
Implication: This finding reinforces the   scientists. Precipitation microphysics
notion that while the precise details of   probably plays a key role in
movement and intensity of individual       determining       the     timing     and
storms within a complex have little        characteristics of the downdraft and
predictability beyond about 30-60 min      associated gust front. This suggests that
the storm complex and its statistical      precipitation particle type and drop size
characteristics can be extrapolated for    distributions derived from polarimetric
many hours particularly if a gust front    radar should prove a profitable avenue
develops.                                  for research.

4) Understanding processes that            Implication: It is likely microphysical
determine the precise location and         precipitation    parameters      strongly
timing of initiation was an objective of   influence the emergence and strength of
IHOP. While this particular study did      gust     fronts   thus    the    planned
not directly address this objective, our   implementation of polarization on the
study of two cases on 12-13 June and       WSR-88D’s will provide an excellent
15-16 June using high resolution           opportunity to identify precipitation
convergence and stability parameters       particle types and size distributions for
points to the importance of small scale    input into the nowcasting systems.
variations in convergence and CIN.         Additional research focusing on the
Surface based storm initiation in high     specifics for utilizing polarimetric
CIN areas is unlikely, however once a      observations in these nowcast systems
strong gust front develops storms can      should be pursued.
continue in regions of high CIN.
Initiation of convection was not as        6) The evolution, movement and
sensitive to the magnitude of the          lifetime of the initiated storms appeared
corresponding CAPE values.                 to be primarily influenced by the
                                           emerging gust fronts and their
Implications: Observations of the          characteristics rather then stability
necessary resolution of the convergence    parameters. Organized squall lines
field would seem to require station        propagated with the motion of the gust
spacing at least as dense as in            front and not the steering level flow. In
Oklahoma combined with the WSR-            addition the RUC10 showed little skill
88D reflectivity and Doppler velocity      in propagating large well developed
observations. Observation of stability     storm complexes. This was likely
variations may require high resolution     because of difficulties with the model in
water      vapor    measurements     as    producing realistic gust fronts.
demonstrated in IHOP by radar
refractivity measurements. Software        Implication: Large storm complexes
should be installed on the WSR-88D         should be extrapolated with their
and TDWR operational radars to             observed motions or that of the gust
provide refractivity measurements so       front rather with steering level winds or
that near surface water vapor              numerical model motions.
measurements can be obtained as
described by Fabry et al. (1997).          7) The RUC10 convective precipitation
                                           forecasts were able to correctly forecast

precipitation for 62% of the initiation          isentropic-terrain-following   coordinate
        episodes given a tolerance of 250 km in          model. Mon. Wea Rev, 132, 473-494.
        space and 1-5 h in time; if no tolerance
        is allowed 15% were correct. The             Benjamin, S. G., D. Devenyi, S. S. Weygandt, K.
        highest accuracy was with synoptic               J. Brundage, J. M. Brown, G. A. Grell, D.
        frontal systems. There was a tendency            Kim, B. Schwartz, T. G. Smirnova, T. L.
        for the RUC to over forecast the amount          Smith and G.S. Manikin, 2004b: An hourly
        of convection and the time length of             assimilation-forecast cycle: The RUC. Mon.
        convection.                                      Wea Rev. 132, 495-518.

         Implication: Particularly for frontal       Fabry, F. C. Frush, I. Zawadzki and A. Kilambi,
        situations the RUC can provide skill in          1997: On the extraction of near surface
        defining broad scale regions for storm           index of refraction using radar phase
        initiation. The NCWF system already              measurements from ground targets.        J.
        produces a storm initiation interest field       Atmos. Ocean. Tech., 14, 978-987.
        based on regions identified from RUC
        as likely frontal regions.                   Megenhardt, D., C.K. Mueller, N. Rehak, G.
                                                        Cunning, 2000: Evaluation of the National
        8) During IHOP NOAA Severe Storm                Convective Weather Forecast Product.
        Prediction Center forecasters provided          Preprints, Ninth Conf. On Aviation, Range
        special convective storm outlooks for           and Aerospace Meteorology,
        project scientists that identified primary
        boundary (fronts, dry line and outflow       Mueller, C., T. Saxen, R. Roberts, J. Wilson, T.
        boundaries) locations and assigned              Betancourt, S. Dettling, N. Oien and J. Lee,
        forecast probabilities for thunderstorm         2003: NCAR Auto-nowcast system, Wea
        likelihood. These daily outlooks proved         Forecasting, 18, 545 – 561.
        very beneficial for planning project
        operations.                                  Roberts, R., T. Saxen, C. Mueller and D. Albo,
                                                        2003: A forecaster-computer interactive
        Implications: Inclusion of this type of         capability of the NCAR Auto-nowcaster
        forecaster information into automated           system for improved storm initiation
        thunderstorm nowcasting systems will            nowcasts. Preprints, 31st Conf. On Radar
        likely improve the nowcasts from these          Meteorology, Seattle, WA., Amer. Meteoro.
        systems. This concept has already been          Soc.,
        demonstrated in Roberts et al. (2003)
        where the Auto-nowcaster system was          Wilson, J. W. and R. D. Roberts, 2005:
        modified to accept forecaster input.             Summary of Convective Storm Initiation
        Improved      automated        convection        and Evolution During IHOP: Observational
        initiation   nowcasts using         NWS          and Modeling Perspective, submitted Mon.
        forecaster input will ultimately benefit         Wea. Rev.
        the aviation community. A rigorous test
        of this concept will be initiated in 2005    Wolfson, M. M., G. E. Forman, R. G. Hallowell,
        in Fort Worth Texas which will                  and M. P. Moore, 1998: The growth and
        incorporate     NWS       and      CWSU         decay tracker. Preprints, Eighth Conf. on
        forecasters.                                    Aviation, Range and Aerospace Meteor.,
                                                        Dallas, TX, Amer. Meteor. Soc., 58-62.
References
Benjamin, S. G., D., G. A. Grell, J. M. Brown,
    and T. G. Smirnova, 2004a: Mesoscale
    weather prediction with the RUC hybrid

Fig 1. Location of surface stations (wind barbs), radars
(stars) and radiosondes (squares) within the IHOP study
area.