IMPROVING PREDICTION OF
HEAVY RAINFALL WITH
ELEVATED CONVECTION
Patrick Market, University of Missouri
Laurel McCoy, University of Missouri and NOAA/NWS, Portland, OR
Chad Gravelle, CIMSS/SSEC University of Wisconsin, Madison, WI
Charles Graves, Saint Louis University, St. Louis, MO
Presented to the National Weather Association Annual Meeting
21 October 2014, Salt Lake City, UT
Acknowledgements

Funding for PRECIP

Collaborators
 NOAA/NWS/
Weather Prediction Center
 Mike
Bodner
 SUNY-College
 Dr.
at Brockport
Scott Rochette
PRECIP Project
 Use McCoy’s forecast method to predict where heavy-rain-producing
elevated thunderstorms will occur
 Deploy teams to collect observational data from storm environment
 http://weather.missouri.edu/PRECIP
or
 https://www.facebook.com/PRECIPresearchprogram
Introduction
Colman (1990a,b)


Initiated the modern era of elevated convection
studies
Showed the preferred region of elevated
convection in US
 northeast
of a
surface cyclone
 north of its attendant
warm front
Later work…

Corfidi et al. (2006)
 examined
the nature of altocumulus castellanus
 determined
that
the “…division
between elevated
and surface-based
[convective] activity
is rarely distinct.”
A Hierarchy of Elevated Convection
Pure: τ > f -1
ex: wraparound
Surface influences on mid-level parcels reduced or
eliminated because of their 1) vertical location and/or
2) temporal history
Hybrid: τ ~ f -1
ex: north of warm front
Surface influences on mid-level parcels (if any)
mitigated by their arrival over frontal inversion
Mixed: τ < f -1
ex: warm sector castellanus
Surface influences on mid-level parcels unrestricted
A Hierarchy of Elevated Convection
Pure
Mixed
Objectives
Objectives

Analyze average environment


Compare to previous research
Create method for forecasting heavy-rainfallproducing elevated thunderstorms in this region
Methodology
Methodology

Composite events within following National
Weather Service County Warning Areas (CWAs):
 Kansas
City/Pleasant Hill (EAX)
 Springfield, MO (SGF)
 Tulsa (TSA)
 Wichita (ICT)
 Topeka (TOP)
Methodology

Event criteria:

Produced over 2” rain in 24 hrs.

Local rainfall maximum
within CWA boundary
Methodology

Used North American Regional Reanalysis (NARR)
data to find event times
 Event
time defined as NARR time-step with heaviest
rainfall occurring over next 3 hours

Used NARR data to evaluate if event was elevated
 2-meter
θe and precipitation maximum
 NARR sounding from rainfall max
Creating Composites

Lists created including:



Event time
Coordinates for local
rainfall max
Composited using
software from SLU


NARR grid layers overlaid
with coordinates centered
on centroid of CWA
Parameters averaged over
207 x 207 grid

Grid squares = 32 km2
Generate Plots


Composites show average environmental conditions
for elevated thunderstorm events
Created composites for:
 The
event time (t=0)
 6-hours prior (t-6)
 12-hours prior (t-12)
Some of the Parameters Evaluated









250-mb Wind and Divergence (Upper-Level Jet)
500-mb Absolute Vorticity
850-mb Wind (Low-Level Jet)
850-mb θe Advection and 2-meter θe (Surface
Boundary Location and Transport Maximum)
1000-500-mb Thickness
Mean-Sea-Level Pressure
Precipitable Water
Most-Unstable CAPE
K-Index
Results
Results- Kansas City

250-mb Wind (color-filled), heights (black), and
divergence (dashed)
Results- Kansas City


250-mb divergence - Interquartile Range (IQR) plot
T=00
Results- Kansas City

850-mb θe advection (color- filled) and 2-meter θe
(brown)
Results- Kansas City

850-mb θe advection & 2-m θe - IQR plots, T=00
Results- Kansas City

1000-500-mb Thickness (brown), Mean-Sea-Level
Pressure (black), and Precipitable Water (colorfilled)
Results- Kansas City

Precipitable water - IQR plots, T=00
Results- Kansas City

MUCAPE (color-filled) and K-Index (purple)
Results- Kansas City

K Index - IQR plots, T=00
Cross-Section Kansas City
“The X”
Conclusions
Conclusions

Unique patterns to look for when forecasting heavyrainfall-producing elevated thunderstorms:

Strong signal; strong variability

Upper-level jet streak to the northeast of the region


Event located within or just south of 850-mb θe advection
maximum (convergence max)


Divergence > 3 x 10-5 s-1 (lift)
Signals LLJ from the SSW (moisture; lift; instability)
Strong signal; small variability
>30 K-index values (instability)
 Precipitable water values > 1.6” (moisture)
 2-m θe pattern (confirms elevated convection)

Conclusions


Prior work largely corroborated.
However, some novel findings as well…
 MUCAPE
X
decreases while K Index increases
“marks the spot” in cross sections
 Interquartile
ranges - enhance confidence in forecasting
heavy rain events with elevated convection
Conclusions
250-mb Jet
Core > 70 kt
Moisture – PWATs > 1.6” (~40 mm)
Lifting – 250-mb DIV > 3 x 10-5 s-1
Instability – K Index > 32
Future work

Find analogs to composite grids


Find null events
Discover parameters differing for heavy-rainfall vs
non-heavy-rainfall events
PRECIP Project
 Use McCoy’s forecast method to predict where heavy-rain-producing
elevated thunderstorms will occur
 Deploy teams to collect observational data from storm environment
 http://weather.missouri.edu/PRECIP
or
 https://www.facebook.com/PRECIPresearchprogram