Predicting lightning density in Mediterranean storms based on

1
Predicting lightning density in Mediterranean
storms based on the WRF model dynamic and
microphysical fields
AE31A-0258
Yoav Yair1, Barry Lynn1, Colin Price2, Vassiliki
Kotroni3, Konstantinos Lagouvardos3 Efrat Morin4,
Alberto Mugnai5 and Maria-Carmen Llasat6
1The Open University, Ra'anana, Israel, 2
Tel-Aviv University, Tel-Aviv, Israel, 3 National
Observatory of Athens, Athens, Greece, 4 Hebrew
University of Jerusalem, Jerusalem, Israel, 5
CNR-ISAS, Rome, Italy, 6 University of Barcelona,
Barcelona, Spain
A new parameter is introduced the Lightning
Power Index (LPI), which is a measure of the
potential for charge generation and separation
that leads to lightning flashes in convective
thunderstorms. The LPI is calculated within the
charge separation region of clouds, where the
non-inductive mechanism by collisions of ice and
graupel particles in the presence of super-cooled
water is most effective. As shown in several case
studies using the Weather Research and
Forecasting Model (i.e., WRF) with explicit
microphysics, the LPI is highly correlated with
observed lightning. It is suggested that the LPI
may be a useful parameter for predicting
lightning as well as a tool for improving weather
forecasting of convective storms and flash
floods. We used the FLASH project data-base
(http//flash-eu.tau.ac.il/index.php) of
Mediterranean storms to investigate the
performance of the LPI against observed
lightning, and to evaluate the evolution of
precipitation fields as predicted by the WRF
against radar and rain-gauge derived amounts
(where available). The lightning data for the
Israeli cases was based on the local LPATS
(Lightning Positioning and Tracking System)
operated by the Israel Electrical Company. For
the Italy, Spain and Greece cases we utilized the
lightning data of the ZEUS network operated by
the National Observatory of Athens, which is
comprised of six VLF receivers located around
Europe and offers a robust coverage of the
Mediterranean region.
Figure 1 3-hour averages of observed lighting,
and predicted Lightning Power Index, K-Index, and
CPTP values for 0600 to 0900 GMT on 28 October
2006 (Afula, north central Israel).
Figure 2 3-hour averages of observed lighting,
and predicted Lightning Power Index, K-Index, and
CPTP for 1800 to 2100 GMT on 8 September 2006
(Emilia Romagna, north- western Italy).
The LPI is the volume integral of the total mass
flux of ice and liquid water within the charging
zone (0 to -20C) of the cloud. It is calculated
using the simulated grid-scale vertical velocity,
and simulated hydrometeor masses of liquid water,
cloud ice, snow, and graupel. The LPI is non-zero
only within the charging zone, and further the
LPI for a particular model grid is only non-zero
when a majority of cells within a 5 grid-radius
of that grid have a vertical velocity gt 0.5 m
s-1. The LPI has the units of J kg-1 and is
defined as (1) LPI ??? e w2 dx dy dz
where the mass mixing ratios for snow (qs),
cloud ice (qi) and graupel (qg) are in units
kg/kg, and w is the vertical wind component in
m s-1. e is defined by   (2) e 2(Qi Ql)
0.5 /(QiQl) where Ql is the total
liquid water mass and Qi is the ice fractional
content defined by   (3) Qi qg ((qs qg
)0.5 /(qsqg) ) ((qi qg )0.5/
(qiqg))   Note, most bulk schemes produce mass
information and not number concentration of
various hydrometeor types. Hence, we used
relative masses of water and ice hydrometeors in
our LPI formulation.
SUMMARY When averaged over a 100 x 100 km2 area,
there is a high correlation between the LPI and
observed lightning, as well as of maximum LPI
values and maximum simulated rainfall. Since the
lightning data points presented in the case-study
analysis only represent cloud-to-ground flashes
(CG) which are detected by the various systems,
it can be expected that the correlations may be
even higher for total lightning activity, which
includes the intra-cloud (IC) component. The
superiority of the LPI against thermodynamic
indices for lightning prediction such as the KI
and CPTP was clearly demonstrated for these case
studies. Furthermore, we suggest that since the
LPI is derived from the same microphysical fields
that eventually produce model precipitation, in
principal the use of the LPI, in conjunction with
real-time lightning observations, can aid in the
prediction of the onset of heavy convective rain.
Figure 3 WRF Lightning Power Index vs. time (a)
plotted against the number of WRF rain events gt
10 mm hr-1 for area around Emilia Romagna, Italy.
(b) WRF LPI vs. number of ZEUS observed
lightning flashes, obtained after aligning the
data such that the peak time of lightning
occurrence corresponded to the peak time at which
the maximum in the curve of average LPI.
Figure 4 The October 28 2006 Afula flood (a)
The number of LPATS-detected cloud to ground
lightning flashes and the radar-derived
precipitation as a function of time around t he
time of peak rainfall. (b) The LPI values and
WRF- derived average precipitation for the same
time period.
Figure 5 Scale separation between convective
and stratiform rain (a) The number of observed
flashes vs. the WRF computed rainfall events when
the rainfall is gt10 mm hr-1. (b) The same
variables but for rainfall values between 1and 5
mm hr-1.
Acknowledgement This research was conducted in
the framework of the FLASH project, contract
number 036852 of the EU 6th framework program.