, ene oacility at Gadanki

Centre for Oceans, Rivers, Atmosphere and Land Sciencb Clouds and Convective Systems Group, National Atmo

a r t i c l e i n f o

Article history:Received 22 October 2013

turbulent kinetic energy as well as horizontal wind profiles are noticed during the different

omentum, and verticalthe development and

thes, 1982). Among all,cognized as processes

Atmospheric Research 143 (2014) 198215

Contents lists available at ScienceDirect

Atmospheric

j ourna l homepage: www.e lsproperties of the underlying surface in conjunction with thedynamics and thermodynamics of the lower atmosphere (Stull,

of central importance in the genesis and intensification ofmesoscale convective systems (MCSs). Braun and Tao (2000)stated that ABL is a critical factor because of the generation of1. Introduction

The depth and the structure of the atmospheric boundarylayer (ABL) are determined by the physical and thermal

1988). Surface fluxes of heat, moisture,mmixing in the ABL play important roles inintensification of the thunderstorms (AnABL and convection have long been reepochs of the convective activity. This work is useful in evaluating the performance of PBLschemes of mesoscale models in simulating MCS.

2014 Elsevier B.V. All rights reserved. Corresponding author. Tel.: +91 3222 281820/28255303.

E-mail addresses: anvsatya@coral.iitkgp.ernet.in,achanta.satya@gmail.com (A.N.V. Satyanarayana).

http://dx.doi.org/10.1016/j.atmosres.2014.02.0160169-8095/ 2014 Elsevier B.V. All rights reserved.radar derived horizontal wind profiles are in good comparison with observed radiosondewinds. Significant variations in the surface meteorological parameters, sensible heat flux andMST radarThunderstormDoppler Weather Radares, Indian Institute of Technology Kharagpur, Kharagpur 721 302, Indiaspheric Research Laboratory, Gadanki 517 112, India

a b s t r a c t

Mesoscale convective systems (MCSs) wreak lots of havoc and severe damage to life andproperty due to associated strong gusty winds, rainfall and hailstorms even though they lastfor an hour or so. Planetary boundary layer (PBL) plays an important role in the transportationof energy such as momentum, heat and moisture through turbulence into the upper layers ofthe atmosphere and acts as a feedback mechanism in the generation and sustenance of MCS. Inthe present study, three severe thunderstorms that occurred over mesospherestratospheretroposphere (MST) radar facility at National Atmospheric Research Laboratory (NARL),Gadanki, India, have been considered to understand turbulence, energy exchanges and windstructure during the different epochs such as pre-, during and after the occurrence of theseconvective episodes. Significant changes in the turbulence structure are noticed in the upperlayers of the atmosphere during the thunderstorm activity. Identified strong convective coreswith varying magnitudes of intensity in terms of vertical velocity at different heights in theatmosphere discern the presence of shallow as well as deep convection during initial, matureand dissipative stages of the thunderstorm. Qualitative assessments of these convective coresare verified using available Doppler Weather Radar imageries in terms of reflectivity. The MSTReceived in revised form 16 February 2014Accepted 19 February 2014Available online 26 February 2014

Keywords:Planetary boundary layerMesoscale convective systemTurbulence transportaA.N.V. Satyanarayana a,, Sabiha Sultana a, T. Narayana Rao b, S. Satheesh Kumar bEvaluation of atmospheric turbulencestructure of convective cores during thconvective systems using MST radar f1821; fax: +91 3222ergy exchanges andccurrence ofmesoscale

Research

ev ie r .com/ locate /atmoslarge fluxes of heat, moisture, andmomentum in this thin layer.Through turbulent mechanism, eddies in the ABL transport therequired moisture into the upper layers of the atmosphere forthe sustenance of these events. MCSs are an important link

199A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215between atmospheric convection and large scale atmosphericcirculation (Houze, 2004). Mesoscale shallow convections are 1to 2 km deep, and they have horizontal length scales of a fewkilometers. Pucillo andManzato (2013) have studied the skill ofone or more predictors such as wind magnitude, equivalentpotential temperature, andmoisture transport in forecasting thehighest radar for Vertical Maximum Intensity reflectivitywhich in turn assesses the signature of the storm occurrenceand intensity. Csirmaz et al. (2013) investigated two cases of thethunderstorms with rotating characteristics that are formed inan environment of relatively low ormoderate wind shear in thelowest 6 km layer of the troposphere. Mkel et al. (2014) haveinvestigated pre-monsoon thunderstorm characteristics overNepal.

Mesoscale shallow convection is worth studying given itsrole in the transportation of heat, moisture and momentuminto the free atmosphere and also in improving our under-standing of the dynamics of the ABL (Agee, 1987; Moyer andYoung, 1994; Blaskovic et al., 1991). Some attempts weremade in understanding the energy exchanges and ABLdynamics in the development of the thunderstorm over theIndian region (Tyagi et al., 2011, 2012, 2013a, 2013b; Tyagiand Satyanarayana, 2013a; Latha and Murthy, 2011) usingfield observations. Dalal et al. (2012) have shown theexistence of organizational modes of squall-type mesoscaleconvective systems during pre-monsoon season over east-ern India.

The mesospherestratospheretroposphere (MST) radaris found to be capable of detecting return signals arising fromweak fluctuations in the atmospheric refractive index (Rao etal., 1995, 1999). With certain limitations the MST techniqueis capable of continually observing winds, waves, turbulenceand atmospheric stability over the height range of 1100 kmwith excellent time and space resolution (Balsley and Gage,1980; Hocking, 1997a, 1997b). The MST radar is a valuabletool for routine monitoring of atmospheric wind field(Gregory et al., 1979; Walker, 1979; Harper and Gordon,1980; Gage and Van Zandt, 1981; Gage and Balsley, 1984)and is feasible for monitoring the atmospheric turbulence aswell. The wind profiler is capable of monitoring lower andmid-atmospheric processes like microscale turbulence andmesoscale convection (Uma and Rao, 2009; Rao et al., 2009,2010; Balsley et al., 1988; Dhaka et al., 2002, 2003; Cifelli andRutledge, 1994, 1998). Previous investigations reported thatlarge wind fluctuations in MST radar wind are mostlyassociated with enhanced wind and wind shear, generallyseen in baroclinic systems (Rao and Kirkwood, 2005)whereas the small scale variability in vertical velocity isassociated with the passage of synoptic weather disturbances(Ecklund et al., 1981; Larsen and Rttger, 1982; Sato et al.,1995; Rao et al., 1999; Dhaka et al., 2002; Kumar, 2006; Umaand Rao, 2008). Turbulence studies are made using turbulentkinetic energy (TKE), spectral widths, Signal-to-Noise Ratio(SNR) and refractivity structure constant, Cn2 (Rao et al.,2001; Rao and Rao, 2007). Spectral widths and SNRobservations indicate the intensity of turbulence and mixingdepth (height of the boundary layer) in different seasons. Asdetailed above, only a few studies exist in the literaturedescribing the vertical and spatial structures of turbulenceand convection during the different epochs of MCS oversouthern peninsular India, in particular over Gadanki. Thepresent study, therefore, focuses mainly on understandingthe wind structure, turbulence transport, and convectionmechanism during the passage of severe thunderstorms overGadanki, India.

2. Data

In the present study, three thunderstorm events thatoccurred over Gadanki region on 25 April, 4 May and 6 June2011 are chosen. Doppler Weather Radar (DWR)-derivedreflectivity imageries are used to locate the convective systemsand the duration of occurrence of the events over the studyregion. The MST radar data, the three spectral moments;received signal power, radial velocity, and spectral variance,are used to identify the convective cores and turbulent airparcels over Gadanki region during the thunderstorm period.Horizontal wind profiles derived from the MST radar for theduration of the thunderstorm activity are analyzed. Radiosondeobservations obtained at study site are utilized for validating thederived wind profiles from the MST radar.

Micro-meteorological tower data comprises of slowresponse data (1 Hz) of wind speed, wind direction, temper-ature, relative humidity and pressure at 6 heights, 2, 4, 8, 16,32 and 50 m, and fast response data (20 Hz) comprising ofsonic anemometer (zonal, meridional and vertical windcomponents and sonic temperature) at 8 m.

3. MST radar system description

The Indian MST radar is located at National AtmosphericResearch Laboratory at Gadanki, near Tirupati. Gadanki (13.5N,79.2E; 375 m above sea level) is situated in a rural area ofChittoor district (Andhra Pradesh) in the southern part of India.The MST radar is a high-power, mono-static, coherent-pulsedDoppler radar. IndianMST radar is a highly sensitive VHF phasedarray radar operating at 53 MHz, with peak power apertureproduct of 3 1010 Wm2 (Rao et al., 1995; Rao et al., 1999). Thephased array consists of 1024 (32 32) crossed three-elementantennas occupying an area of 130 m 130 m. A total trans-mitting power of 2.5 MW (peak) is provided by 32 transmitters,whose output power varies from 20 kW to 120 kW.

To measure the three components of wind (zonal, meridi-onal and vertical), a total of 4 beamswere employed (two zenithand two off-vertical tilted 10 towards east and south). Theradar switches alternatively between vertical and off-vertical inorder to obtain high resolution samples of vertical wind alongwith horizontal wind and spectral width measurements. Thereceived echo signals were sampled at interval of 300 m in theheight range of 1.5 to 20.4 km and were coherently integrated.These samples are subjected to Fast Fourier Transform (FFT) forthe on-line computation of the Doppler power spectra for eachrange bin. The off-line data processing of the Doppler spectrumincludes estimation of average noise level and computation ofthe three radar spectral moments.

4. Quality control

4.1. MST radar data

The raw data signals received from the radar required qualitycontrolling as it contains both noise and signal components.

be erroneous during the active phase of convection. The error

made (Tyagi et al., 2012; Tyagi and Satyanarayana, 2013a). First

spectral width and Signal-to-Noise Ratio (SNR) (Nastrom et

200 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215al., 2004). From the radial velocities in different directions(3 in our case), the zonal, meridional and vertical velocitiesare extracted with a temporal resolution of ~1 min. Thehorizontal winds are then estimated for the same interval.

Surface layer turbulence kinetic energy, sensible heat fluxand kinematic heat flux are calculated using eddy correlationtechnique (Businger et al., 1971).

6. Results and discussion

In order to understand the variation of the upperatmospheric turbulence and surface energy exchangeduring the different epochs of the thunderstorm activity,three thunderstorm events with varying intensity thatpassed over Gadanki on 25 April 2011 (Case I), 04 May2011 (Case II) and 06 June 2011 (Case III) are analyzed inthe present study.of all, physical check of data has been done by employing themethod suggested by Foken and Wichura (1996). Datasetsduring the rainfall hours have been discarded for possibleeffects of interference of water droplets on transducer head ofsonic anemometer which will influence the quality of the data(Thomas and Foken, 2005). The datasets have been furtherprocessed with windowing technique for 3 standard devia-tion from the mean (Viswanadham et al., 1997a, 1997b),double rotation (DR) scheme and tilt angle correction incor-poration after the physical check of data (Kaimal and Finnigan,1994; Wilczak et al., 2001; Tyagi et al., 2013a; Tyagi andSatyanarayana, 2013a).

5. Methodology

The MST radar derived three spectral moments are usedto estimate the desired parameters; Doppler velocity,may stem from the existence of bimodal spectra: one due toclear-air scattering and the other due to Rayleigh scattering(Rao et al., 1999), validity of the assumptions like homoge-neous atmosphere between the beams, large vertical winds,etc. Therefore, the variations in winds during the activeconvection (or in other words when the vertical winds arelarge) need to be treated cautiously.

4.2. Micro-meteorological fast response data

Prior to analyzing the fast response data obtained fromsonic anemometer, extensive quality checks of data have beenBefore estimating spectral parameters of the signal, it isnecessary to determine noise level as accurate as possible.The noise level at each range gate is estimated by followingHildebrand and Sekon (1974). Further, several quality checks(continuity and consensus techniques) have been performedon the data (spectral parameters and radial velocities) for theremoval of systematic and random errors (Tsuda et al., 1986;Yamamoto et al., 1988; Palmer et al., 1991; Kudeki et al., 1993).In spite of the above tests, the MST radar-derived winds could6.1. DWR imagery analysis

6.1.1. Case IOn 25 April 2011 at 17:50 LT, a convective cell was

detected near Gadanki region as observed from the DWRimageries (Fig. 1a). The convective cell was transported overto the Gadanki from the north-east direction and movedaway towards the south-westerly to southern direction. At17:50 LT, the intensity of convective cell, in terms ofreflectivity was ~28 dBZ and the height of cloud was ~5 km(Fig. 2a). The cell intensified in both magnitude and verticalextension by 18:40 LT to a maximum reflectivity of 53.4 dBZand a cloud height of ~7 km. The reflectivity as well as thecloud height decreased as the convective system movedaway by 19:30 LT from Gadanki region. The intenseconvective activity persisted for about an hour over Gadankiwith moderate to high reflectivity.

6.1.2. Case IIOn 04 May 2011 at 20:00 LT, a convective cell was noticed

over Gadanki region as seen in theDWR imageries (Fig. 1b). Thiscell was advected to Gadanki from the north and moved awaytowards the south direction. The convective cell reachedGadanki region at 20:00 LT with a reflectivity of ~47 dBZ and acloudheight of ~5 km(Fig. 2b). The reflectivity and cloudheightgradually increased to 54.7 dBZ and 10 km, respectively, by20:30 LT. Both reflectivity and cloud height decreased graduallyas the convective cell moved away from the study region. Thisevent is considered as a case of severe thunderstorm event dueto its higher reflectivity and cloud height.

6.1.3. Case IIIOn 06 June 2011, two cellular thunderstorms (convective

systems) were noticed in the evening over Gadanki region asseen in DWR imageries (Fig. 1c and d). The first convectivecell was transported to Gadanki from the northwest directionand moved away towards the southeast direction and thesecond one appears from the west direction and was movedaway in the southeast direction. As the first convectivecell appears near Gadanki region at 16:30 LT, the observedreflectivity was as high as 41.3 dBZ and the cloud heightwas about 5 km. The reflectivity increased gradually to amaximum value of 57.3 dBZ and the cloud height remains810 km during 17:30 LT to 18:40 LT over Gadanki, as seen inFig. 2c. Afterwards the magnitude of reflectivity decreasedalong with the cloud height as the cell moved away fromthe study area by 19:00 LT. Second cell arrived over theobservation facility at 19:40 LT with a reflectivity value of44 dBZ and a cloud height of 6 km. The magnitudeof reflectivity and cloud height increased gradually to amaximum value of 52 dBZ and 8 km, respectively, by 20:10LT (Fig. 2c). The reflectivity as well as the cloud heightdecreased later as the cell moved away completely by 21:10LT. During the passage of both convective cells, the intensityof reflectivity was more than 50 dBZ. It clearly indicates thatthis event can be considered as a severe thunderstorm.

6.2. Signal-to-Noise Ratio of MST radar backscattered echoes

The MST radar provides estimates of atmospheric param-eters with good spatial and temporal resolution. The quality

Fig. 1. Doppler Weather Radar imageries for (a) and (b) 06 June 2011, (c) 25 April 2011 and (d) 04 May.Data obtained from DWR Division, India Meteorological Department, Chennai.

201A.N.V.Satyanarayana

etal./

Atm

osphericResearch

143(2014)

198215

Fig. 2. Variation of reflectivity and cloud height during the thunderstorm event on (a) 25 April 2011, (b) 04 May 2011 and (c) 06 June 2011 derived from DWR imageries.

202 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215of the radar backscattered signal can be assessed from theSignal-to-Noise Ratio (SNR), which is the ratio of signalpower to noise power. The SNR plots for the three cases, i.e.,Fig. 3. Signal-to-Noise Ratio measured from the MST radar for (a) 25 April 2011, (b) 0425 April 2011, 04 May 2011 and 06 June 2011, are presented,respectively, in Fig. 3a, b, and c, portraying the quality of theradar data. Fig. 3a (Case I) shows that the SNR is quite highMay 2011, and (c) 06 June 2011 during the thunderstorm episodes over Gadanki.

(N10 dB) up to 8 km from the surface throughout theobservation period and up to 15 km after 18:15 LT, i.e.,during the passage of the thunderstorm and later. Fig. 3b(Case II) shows the SNR values N10 dBZ throughout theobservation period between 1.5 and 12 km. Fig. 3c (Case III)also depicts the SNR values N10 dBZ up to an altitude of10 km. Nevertheless, the vertical extent of high SNR valuesincreased to 15 km during the passage of the thunderstormperiod, i.e., during 17:5018:45 LT on 06 June 2011. It is clearfrom Fig. 3 that the SNR is quite high (N10 dBZ) and this highSNR is extended vertically to greater heights during thepassage of the thunderstorms, indicating good quality of datain all 3 cases.

6.3. Turbulence from spectral width

The observed spectral width includes contributions fromturbulence within the pulse volume and non-turbulent pro-cesses, like beam, shear, and wave broadening (Nastrom, 1997;Nastrom et al., 2004; Rao et al., 2001 and references therein).Though formulations for correcting non-turbulent processeshave been available, such corrections often yield negativespectral widths (particularly in the presence of strong winds).Therefore, the measured spectral widths obtained with theMST radar are used here for understanding the variation ofturbulence during the passage of the thunderstorms (Rao et al.,2010).

6.3.1. Case IFig. 4a shows the spectral width plotted for the convective

activity occurred on 25 April 2011. The spectral widthinitially in the range of 1.52 ms1 during 17:5018:00 LTin the lower altitudes 1.59 km increased in magnitude andvertical extent to 23 ms1 and up to 15 km during theintense phase of the convection, i.e., during 18:1518:30.During this period, larger patches of higher spectral width(3 ms1) extending from 3 to 17 km are noticed, indicatingthe existence of intense turbulent rolls. These turbulent rollsbegan to weaken during 18:3018:45 LT at lower altitudes.From 18:45 LT, the turbulence started decaying at higheraltitudes and the pockets of higher turbulent rolls completelydisappeared by 19:00 LT, leaving the remnants of few patchesof weaker turbulence. During the entire duration of thethunderstorm event, turbulent plumes of varying verticalextent and intensity are noticed. In the most intensive phaseof the event, we have noticed that the vertical extent of theturbulence plumes is extended from the lower layers of theatmosphere to the higher altitudes. Remnants of turbulentpatches that remained at higher altitudes are noticed evenduring the decaying phase of the thunderstorm.

6.3.2. Case IIFig. 4b depicts the spectral width during the convective

activity on 04 April 2011. In this event, unlike in other twoevents where the convection and associated turbulence

r Gada

203A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 4. Variation of turbulence with height during the thunderstorm event ove2011, and (c) 06 June 2011.nki using MST radar derived spectral width for (a) 25 April 2011, (b) 04 May

appear in cellular structure, strong turbulence appears inlayered structure. The turbulence is found to be weak tomoderate in the lower troposphere (b6 km) with values inthe range of 12 ms1. Nevertheless, strong turbulencepatches with 23 ms1 intensity prevailed in the middletroposphere and upper troposphere (915 km) during thefirst half of the observation period (till 21:20 LT) whenthe convection was active over Gadanki. After 21:20 LT, theupper tropospheric turbulence started decaying and reducedto 12 ms1 in the altitude range of 615 km. After 21:45 LT,strong turbulent patches are seen within 6 km, but the upperatmosphere is void of such strong turbulent patches in thesecond half of the observation period.

6.3.3. Case IIIFig. 4c presents the turbulence structure observed on 6 June

2011. At 17:30 LT, the turbulence intensity in the altitude rangeof 1.510 km is found to be between 1.5 and 2 ms1, whereas itis small at higher altitudes (N10 km). During 17:5518:35 LT,patches of turbulence appear to be distributed throughout thetroposphere (seen up to an altitude of 17 km) with spectralwidths as large as 23 ms1. During this period, large patches ofhighly turbulent rolls were observed between the altitudes 2and 15 kmwhen the first cell of the two thunderstorms movedover to the study site. After 18:35 LT, it is noticed that the levelsbelow 6 km appear to be less turbulent whereas the upperatmosphere remained highly turbulent. The turbulence athigher heights appears to be dissipated by 19:10 LT. Weakturbulence of the order of 0.5 ms1 was present in the entire

troposphere during 19:1019:35 LT, indicating that the first cellhas moved completely from the study region. At 19:35 LT, withthe arrival of the second cell, the atmosphere became highlyturbulent, particularly in lower altitudes up to 9 km, with amagnitude as large as 2.5 ms1.

6.4. Updrafts and downdrafts in terms of vertical wind prolesfrom the MST radar

6.4.1. Case IFig. 5a depicts the timeheight variation of vertical velocity

as measured by the MST radar during the passage of a deepconvective system over Gadanki on 25 April 2011. Weakupdrafts are noticed during 17:55 to 18:05 LT between altitudes4 and 10 km (6 km shallow convection). Weak downdraftsare seen in altitudes 1.5 to 9 km during 18:00 to 18:10 LT. Theexistence of strong updrafts is also noticed in the altitudes 1.5 to6 km with a magnitude range of 7 to 10 ms1 and very strongupdrafts of magnitude 10 ms1 between 8 and 17 km (with alayer depth of 9 km) during 18:10 to 18:25 LT, indicating thepresence of deep convective cores. Themagnitude of updrafts isdecreased to 24 ms1 during the period 18:25 to 18:45 LT andthen to normal values (within 1 ms1) by 18:50 LT.Downdrafts of magnitude ~3 ms1 are identified in the layerof 5 to 9 km during 18:35 to 18:45 LT. After 18:45 LT, weakupdrafts and downdrafts of magnitude 1 ms1 are onlypresent in the height region of 318 km, indicating the depletedstrength of convective storm.

d from

204 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 5. Variation of convective cores in terms of vertical velocity profiles deriveApril 2011, (b) 04 May 2011, and (c) 06 June 2011.the MST radar signals during the thunderstorm event over Gadanki on (a) 25

6.4.2. Case IIA convective system passed over Gadanki region in the late

evening on 4 May 2011 between 20:50 and 21:35 LT (Fig. 5b).The presence of relatively weak updraft and downdraft cores ofmagnitude 2 ms1 during 20:4020:50 LT in a layer extend-ing from 2 to 18 km is noticed (Fig. 5b). Strong updrafts ofmagnitude 4 ms1 are observed during 20:5021:00 LT in thelower layer (1.5 to 4 km). Concurrently in the layer between 9and 17 km (depth of the layer is 8 km) strong updrafts ofmagnitude 34 ms1 are observed during 20:5021:05 LT.Downdrafts of magnitude of about 4 ms1 are identified atlower altitudes, between 1.5 and 3 km, at 21:15 LT. Weakdowndrafts of magnitude 2 ms1 between altitudes 610 kmare identified during 21:0021:35 LT. After 21:35 LT, theconvective updraft cores appear to beweakened and dissipating.

6.4.3. Case IIIThe timeheight cross section of vertical wind during the

passage of convective cells on 06 June 2011, obtained fromvertical beammeasurements of MST radar is shown in Fig. 5c.During the study, two convective systems were noticed inthe evening hours on 6 June 2011 with MST radar and alsoin DWR imageries. A convective cell is noticed during18:0018:50 LT (henceforth referred to as convective cell I)

and again at 19:35 LT, which decayed after 20:40 LT (referredto as convective cell II).

(i) Convective cell I Fig. 5c depicts theMCS structure withintense updrafts and downdrafts (Houze, 2004; Uma andRao, 2008, 2009; Rao et al., 2010). Strong updrafts ofmagnitude 7 ms1 were observed during 18:00 to 18:15LT (at altitudes 27 km) and during 18:2018:30 LT (ataltitudes 915 km). Strong downdrafts of magnitude5 ms1 were present at higher altitudes between 10 and15 km during 17:45 LT18:25 LT. During 18:35 to 18:45LT updrafts of magnitudes N5 ms1 were observed athigher heights from 9 to 18 km. After 18:50 LT, onlyweak downdrafts were present at heights from 12 to15 km.

(ii) Convective cell II weak updrafts and downdrafts ofmagnitude 1 ms1 were observed during 18:50 to19:35 LT. During 19:4019:50 LT, a strong updraft ofmagnitude of about 10 ms1 was seen in lower altitudesbetween3 and 9 km(layer depth of 6 km) and is upliftedby the time 19:50 LT up to a height of 12 km, indicatingthe presence of deep convection. A strong downdraft ofintensity 4 ms1 is noticed at heights from 12 to 17 kmduring 19:45 to 19:55 LT and also at lower altitudesbetween 1.5 and 5 km during 19:5019:59 LT.

) win11) an

205A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 6. Comparison of diurnal variation of surface meteorological variables (a(e) surface pressure and (f) rainfall rate, on a thunderstorm day (25 April 20d speed, (b) wind direction, (c) surface temperature, (d) relative humidity,d a non-thunderstorm day (26 April 2011).

6.5. Variation of surface meteorological variables during thethunderstorm events

In this section, the changes in the diurnal variation of thesurface meteorological variables such as wind speed, winddirection, air temperature and relative humidity between athunderstorm (TD) and non-thunderstorm day (NTD) duringthe study period are presented.

6.5.1. Case ITypical variations were observed in surface meteorological

variables during the thunderstorm events similar to thosereported in earlier studies (Tyagi et al., 2012, 2013a). Fig. 6shows the comparison of the observed surface parameters overGadanki for 25 April 2011 (TD) and 26 April 2011 (NTD). FromFig. 6a, occurrence of gusty wind can be identified during17:0018:00 LT on TD. Fig. 6b depicts the variation of winddirection on the TD andNTD. During the TD, the reversal ofwinddirection has beennoticed and after the passage of the storm thewind direction has come back to normal direction. Fig. 6cdepicts a temperature decrease from 31.4 C to 23.2 C (drop of8.2 C) during the convective period (18:00 LT). The relativehumidity variation during the TD and NTD is presented inFig. 6d. It is seen that relative humidity rises suddenly from52.3% to 94.9% (42.6% rise) during the convective episode

(17:00 LT). The pressure variation for both days is plotted inFig. 6e, where a pre-squall low is observed in the TD pressurecurve at 16:00 LT followed by weak meso-high at 20:00 LT andthen a wake low. Fig. 6f shows a rainfall rate of 12 mm h1 at18:00 LT at the time of the thunderstorm passage over Gadanki.No such significant variations of these parameters are noticedon NTD.

6.5.2. Case IIFig. 7 depicts the comparison of the observed surface

parameters over Gadanki for 04 May 2011 (TD) and 05 May2011 (NTD). This event has occurred during night-time. FromFig. 7a, gusty winds were observed during 21:0023:00 LT onTD. Fig. 7b depicts the variation of wind direction on the TDand NTD. On the TD the wind appears to change its directionsuddenly during the thunderstorm period and again comesback to normal after the convective activity is over. Thetemperature drops from 36.7 C to 22.1 C (14.6 C drop)during the convective period (21:00 LT) and rises back to33.4 C by the time 22:00 LT again falls to 24.9 C on TD(Fig. 7c). The relative humidity variation on the TD and NTDis presented in Fig. 7d. On TD the relative humidity risessuddenly from 46.7% to 80% (33.3% rise) during theconvective episode. The RH is found to be falling back to34.9% and then again rises to 67.6% at 22:00 LT. The pressure

a) win) and

206 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 7. Comparison of diurnal variation of surface meteorological variables ((e) surface pressure and (f) rainfall rate, on a thunderstorm day (4 May 2011d speed, (b) wind direction, (c) surface temperature, (d) relative humidity,a non-thunderstorm day (5 May 2011).

207A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215variation for both days is plotted in Fig. 7e where ameso-high is observed at 21:00 LT followed by a wake lowon TD. Fig. 7f shows the rainfall rate on 04 May (TD) in whichthe rainfall activity is observed during 22:0024:00 LT(3 mm h1).

Fig. 8. Comparison of diurnal variation of surface meteorological variables (a) win(e) surface pressure and (f) rainfall rate, on a thunderstorm day (6 June 2011) and

Fig. 9. Variation of turbulent kinetic energy (a) one minute averaged instantaneoduring the thunderstorm period and (b) hourly averages on 25 April 2011.6.5.3. Case IIIFig. 8 shows the comparison of the observed surface

parameters over Gadanki during 05 June 2011 (NTD) and 06June 2011 (TD). Fig. 8a shows the occurrence of gusty windsof magnitude of about 1014 ms1 during 19:0020:00.

d speed, (b) wind direction, (c) surface temperature, (d) relative humidity,a non-thunderstorm day (5 June 2011).

us observations derived from sonic anemometer turbulence measurements

Fig. 8b depicts the variation of wind direction on TD and NTD, Interestingly occurrence of rainfall is observed during

Fig. 10. Variation of turbulent kinetic energy (a) one minute averaged instantaneous observations derived from sonic anemometer turbulence measurementsduring the thunderstorm period and (b) hourly averages on 4 May 2011.

208 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215where a reversal of wind direction is observed during the TD.Fig. 8c depicts the sudden drop of air temperature from33.5 C to 29.89 C during the first event, started at 18:00 LT,and again from 28 C to 23.8 C during the second event at20:00 LT. Similarly RH deviates from the normal diurnalvariation pattern as it rises from 45.5% to 62.9% (17.4%increase) at 18:00 LT then falls slightly and again increasesfrom 60.3% to 95.3% (35% rise) at 20:00 LT on 06 June,indicating the passage of two convective cells over the studyarea (Fig. 8d), as also seen in MST radar measurements.Fig. 8e depicts the surface pressure variation during both TDand NTD. During the TD pre-squall low during 16:30 LT andthen a meso-high are observed during 18:00 LT followed by awake low. This corresponds to the first convective cell of theTD. During the second cell meso-high was also observedduring 20:00 LT, which is followed by the wake low. Thediurnal variation of the surface meteorological variables ofNTD does not reveal any typical variation as seen in TD case.Fig. 11. Variation of turbulent kinetic energy (a) one minute averaged instantaneoduring the thunderstorm period and (b) hourly averages on 6 June 2011.20:0024:00 LT (Fig. 8f) after the passage of second cell.But no rainfall activity is observed when the first convectivecell was moving over the study region.

6.6. Turbulence kinetic energy during the thunderstorm events

In this section, significant variations in atmosphericsurface layer turbulence kinetic energy (TKE) noticed duringthe passage of the thunderstorms are presented.

6.6.1. Case IThe variation of one minute surface layer TKE (at 4 m

height) during the passage of the thunderstorm and onehour-TKE during 25 April 2011 is presented in Fig. 9a and b,respectively. A very high magnitude of TKE observed at 17:50LT indicates the initiation of convective activity at the studylocation as shown in Fig. 9a (Tyagi and Satyanarayana,2013b). In general, large magnitudes of TKE are observedus observations derived from sonic anemometer turbulence measurements

Fig. 12. Variation of (a) one minute averaged instantaneous kinematic heat flux observations derived from sonic anemometer turbulence measurements duringthe thunderstorm period and (b) hourly averages of sensible heat flux on 25 April 2011.

209A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215during 18:0018:30 LT followed by small magnitudes,indicating the decaying of the thunderstorm as seen inFig. 9a. Hourly values of TKE have shown distinct high peakindicating the intense phase of the thunderstorm activity(1800 LT) (Fig. 9b). The time of occurrence of highermagnitudes of TKE noticed in the atmospheric surface layercoincides well with the turbulent plumes extended from1.5 km to higher levels (during 17:45 to 18:15 LT) as noticedfrom the MST radar derived turbulence measurements(Fig. 4a). Even though vertically extended turbulent plumesare seen after 18:30 LT from the middle layers (above34 km) of the atmosphere, no such signature is seen after18:30 in surface layer TKE. Based on this we can concludethat significant intensity of turbulence transport is extendedfrom the surface layer to deeper layers of the atmosphereduring the intensive phase of the thunderstorm event.

6.6.2. Case IIIn general, the TKE (both oneminute as well as the hourly

averages) during the passage of the thunderstorm is higherFig. 13. Variation of (a) one minute averaged instantaneous kinematic heat flux obsthe thunderstorm period and (b) hourly averages of sensible heat flux on 4 May 20in this case (Fig. 10a and b) than in Case I. Interestingly thisevent occurred during the night. The higher TKE is attributedto the strong surface layer gusty winds (shown in Fig. 7a).During the period of the thunderstorm activity, twodistinctive peaks of TKE are noticed in the one minuteaverage plot during 20:4021:55 LT (Fig. 10a). A distinctivepeak of higher TKE is seen in hourly averages as well as at21:00 LT (Fig. 10b). During 21:0021:10 LT, the TKE is foundto be weak. As also seen in Case I, whenever turbulence rolloccurs in the upper layers of the atmosphere (depictedin Fig. 4b), the surface layer TKE is found to be high. Thisclearly indicates that during the intensive phase of thethunderstorm, the turbulence transport from surface layerto the top of the atmosphere occurs sporadically but notcontinuously.

6.6.3. Case IIIFig. 11a shows two distinctly different pockets of higher

TKE during the thunderstorm event, one during 17:30 to17:40 LT and the other one is at 19:55 LT in the 1 minuteervations derived from sonic anemometer turbulence measurements during11.

averages, coinciding with the time of occurrence of twoconvective cells (shown in Fig. 1c and d, Fig. 2c). During thetime between these two cells, no significant variation ofTKE is noticed; moreover, the magnitudes are very small.Similarly two distinguishable peaks of TKE are noticed at18:00 LT and 21:00 LT corresponding to the two cells

6.7. Sensible heat ux

6.7.1. Case IOne minute averages of instantaneous kinematic heat flux

(wT) during the duration of the thunderstorm activity aswell as hourly averages of sensible heat flux (SHF) on 25

Fig. 14. Variation of (a) one minute averaged instantaneous kinematic heat flux observations derived from sonic anemometer turbulence measurements duringthe thunderstorm period and (b) hourly averages of sensible heat flux on 6 June 2011.

210 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215identified in the hourly averages as well (Fig. 11b). Asnoticed in the above two case studies, significant variation ofTKE always coincides with the higher TKE at 1.5 km obtainedfrom the MST radar indicating the extendable transport fromsurface to upper atmosphere during the intensive phase ofthe thunderstorm activity.Fig. 15. Validation of MST radar wind profiler derived horizontal wind profiles with(a) 25 April 2011, (b) 04 May 2011, and (c) 06 June 2011.April 2011 were presented in Fig. 12a and b. Just before theintensification stage of the convective cell around 17:50 LT,rise in kinematic heat flux (0.4 K ms1) is noticed (Tyagi etal., 2013a). From 18:00 LT onwards, no significant variationsare noticed other than fluctuations due to the turbulenceactivity and the magnitude is close to zero by 18:30 LT. Thethe observations obtained from radiosonde ascents at Gadanki at ~12 UTC on

sensible heat flux has high values of about 250 Wm2 during11:0015:00 LT and at 19:00 LT, but reduced to about50 Wm2 by 19:00 LT; indicating downward transport ofheat after the thunderstorm.

6.7.2. Case IIHeat fluxes for 04 May 2011 are presented in Fig. 13a and b.

It is observed that heat flux does not show any sudden peakprior to thunderstorm as noticed in Case I, but heat flux remainspositive; just before the thunderstorm initiates. This positivevalue indicates the upward transport of heat at the starting timeof convective activity. As the 04 May thunderstorm eventoccurred during the night, the heat flux was negative before theconvection started and became positive due to the positivebuoyancy of the air. During 21:0021:30 LT, the flux has becomevery small with very marginal variation during the thunder-storm activity. Higher magnitudes of sensible heat flux of about250 Wm2 during 12:0014:00 LT were noticed, but reducedto about 50 Wm2 by 22:00 LT; indicating downwardtransport of heat after the thunderstorm activity.

6.7.3. Case IIIFig. 14a presents the kinematic heat flux during 06 June

2011 for the thunderstorm periods and the time between thetwo convective activities and Fig. 14b presents the hourly

averaged sensible heat flux for the thunderstorm day. As seenin Case I, rise in heat flux is noticed prior to the event at 17:50LT (during the first cell). It indicates positive buoyancy ofwarm air during the start of the first convective event of theday. During 17:5518:15 LT, the flux is reduced and showssmall fluctuations during the thunderstorm activity. By 18:20LT, the flux reduces to 0.3 K ms1 corresponding to thenegative buoyancy of air parcels at the dissipating stage ofthe first cell and after 18:30 LT only very small positive heatfluxes were present. At 19:25 LT, small peaks are observed inheat flux prior to the second convective event of the day(second cell). The heat flux remains small and fluctuatesaround zero until 19:45 LT before it drops suddenly to0.5 K ms1. The heat flux has further become negativeduring the second convective cell. In Fig. 14b the sensibleheat flux has high values of about 250 Wm2 at 13:00 and220 Wm2 at 15:00 LT and at 19:00 LT, but reduced to about50 Wm2; indicating downward transport of heat afterthe convective activity.

6.8. Validation of MST radar wind proler data with radiosondeobservations

The estimated mean wind profiles from the wind profilerduring the three thunderstorm cases are compared with

(a) du

211A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 16. Zonal and meridional wind profiles derived from MST wind profiler ring and (b) after the thunderstorm event on 25 April 2011 over Gadanki.

available radiosonde wind observations in Fig. 15a, b and c.There are few deviations but the variations of radar windprofiler derived wind magnitude with height are noticed to beclose to the observations (given the limitations in derivinghorizontal winds with MST radar). This gives confidence thattheMST radar basedwind profiler derivedwinds are reasonablein understanding the dynamical structure of the convectiveevents.

6.8.1. Variation of proles of horizontal wind components obtainedfrom the MST radar wind proler during the thunderstorm events

6.8.1.1. Case I. The profiles of zonal andmeridional componentsofwindduring and after the thunderstormevent occurred on25April 2011 are depicted in Fig. 16a and b, respectively. Thewindsjust before the thunderstorm occurrence (17:52 LT) werenortherly/north-easterly in the lower atmosphere and south-westerly in the upper atmosphere. The lower atmosphericwinds change drastically in both magnitude and directionduring the peak hours of the thunderstorm (18:07 to 18:30LT). Nevertheless, large change in wind speed is observedin the middleupper troposphere, where wind speeds in excess

of 40 ms1 were noticed during the thunderstorm period,producing large wind shears. Moderate winds were seen evenafter the passage of the thunderstorm (till 18:52 LT). Later, thewinds reduced following the passage of the thunderstorm overthe study region and the wind pattern became the same as thatof pre-thunderstorm.

6.8.1.2. Case II. The profiles of zonal andmeridional componentsof wind during and after the thunderstorm event occurred on 4May 2011 are depicted in Fig. 17a and b, respectively.Winds arewithin 10 ms1 before the passage of the thunderstorm(before 20:51 LT), but intensified during the mature phase ofthe thunderstorm. Also, during the mature phase significantchanges are noticed in wind direction in the upper troposphere.Whereas the wind direction remained more or less the same inthe lower troposphere, nevertheless, the wind magnitudechanged drastically. Strong jets of wind are noticed in the layersof 8 to 11 km during the mature phase of the thunderstormleading to strong wind shears in this layer.

6.8.1.3. Case III. The zonal and the meridional winds for thefirst convective cell, after the event and during the second

r (a) d

212 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 17. Zonal and meridional wind profiles derived from MST wind profile uring and (b) after the thunderstorm event on 4 May 2011 over Gadanki.

convective cell on 06 June 2011 are plotted in Fig. 18a, b andc, respectively. In general, on 06 June 2011, weak westerlywinds were present before the intensification of the firstconvective event at 17:42 LT. During the passage of the firstconvective cell, during 180018:30 LT, the winds changedsignificantly in both magnitude and direction in the lowertroposphere. As soon as the first convective core passes awayfrom the study region, the winds became weak and westerly.

The horizontal winds between 18:53 and 19:31 LT, remainedsmall and westerly but became strong again when the secondcell passed over the study location during 19:3919:55 LT.

Just like in other two cases, the winds enhanced duringthe passage of the thunderstorms and relatively weak beforethe passage of the first thunderstorm and in between thetwo convective cells. Vertical profiles of wind during andimmediately after the passage of the thunderstorm show

a) dur

213A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215Fig. 18. Zonal and meridional wind profiles derived from MST wind profiler (convective cell during the thunderstorm event on 6 June 2011 over Gadanki.ing the first convective cell, (b) after the first cell and (c) during the second

214 A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215wavy patterns (18:0018:23 LT). They are associated withgravity waves emanated from the convective cell. These wavemotions, in fact, are also seen in timeheight intensity mapsof vertical air motion as up and down draft motions duringthe above period. It is now well known that thunderstormsare one of the important sources of gravity waves andunderstanding the characteristics of such waves is an activearea of research (Uma et al., 2011).

7. Summary and conclusion

The present study has provided interesting variations invarious boundary layer parameters during the different epochsof severe thunderstorm cells passed over MST radar facility atGadanki. DWR imageries effectively suggest the arrival of thethunderstorm cell along with its severity and intensity (reflec-tivity). Interestingly it is seen that the height of convectiveclouds increases with the reflectivity. The convective coresduring the thunderstorm period can be identified from thevertical velocity profiles derived from the MST radar. It clearlyshows the signatures of shallow and deep convections in all thecases considered in the study. The existence of considerableintensity of turbulence is found in the upper layers of theatmosphere even after the dissipation/moved away of thethunderstormcells over the study region. Strong vertical plumesof turbulence are seen during the intensification period of thethunderstorms. The change of vertical velocity and turbulentkinetic energy of the atmospheric surface layer is noticed duringthe existence of the shallow convective layer above a time lagwhereas no significant changes are noticed in the presence ofdeep convective layer. Significant fall in surface layer temper-ature, rise in relative humidity, rise in atmospheric pressure(presence of meso-high), change in wind direction, andpresence of gusty winds are noticed during the thunderstormperiod. Rise in sensible heat flux is noticed just before theoccurrence of the thunderstorm activity. Typical variations ofsensible heat flux are noticed during day time and night timethunderstorms. Systematic variations in the profiles of horizon-tal wind components are seen during the different epochs of thethunderstorm such as before, during and after phases. Thepresence of strong jets is seen during the intensification stage ofthe thunderstorms in all the cases considered in the study. Theresults that emanated from the study would be useful inevaluating the performance of mesoscale models in predictingthe occurrence of the thunderstorms over the region.

Acknowledgment

Ms. Sabiha gratefully acknowledges the Indian Institute ofTechnology Kharagpur for providing assistantship to conductthe present study. Ms. Sabiha would like to thank themembers of Clouds and Convective Systems Group (CCSG),National Atmospheric Research Laboratory (NARL), Gadanki,India for their help and support during the period of her stayat NARL. The authors would like to thank Doppler WeatherRadar (DWR) Division, India Meteorological Department(IMD), Chennai, in providing the DWR imageries during thestudy period. We are thankful to the anonymous reviewersfor their critical and constructive suggestions in improvingthe quality of the manuscript.References

Agee, E.M., 1987. Meso-scale cellular convection over the oceans. Dyn.Atmos. Oceans 10, 317341.

Anthes, R.A., 1982. Tropical cyclones, their evolution, structure and effects.Meteorol. Monogr. 19.

Balsley, B.B., Ecklund, W.L., Carter, D.A., Riddle, A.C., Gage, K.S., 1988. Averagevertical motion in the tropical atmosphere observed by a radar windprofiler on Pohnpei (70N, 157E). J. Atmos. Sci. 45, 396405.

Balsley, B.B., Gage, K.S., 1980. The MST radar technique: potential for middleatmospheric studies. Pure Appl. Geophys. 118, 452493.

Blaskovic, M., Davis, R., Snider, J.B., 1991. Diurnal variation of marinestratocumulus over San Nicolas Island during July 1987. Mon. WeatherRev. 119, 14691478.

Braun, S.A., Tao, W.K., 2000. Sensitivity of high-resolution simulations ofHurricane Bob (1991) to planetary boundary layer parameterizations.Mon. Weather Rev. 128, 39413961.

Businger, J.A., Wyngaard, J.C., Izumi, Y., Bradley, E.F., 1971. Fluxprofilerelationships in the atmospheric surface layer. J. Atmos. Sci. 28, 181189.

Cifelli, R., Rutledge, S.A., 1998. Vertical motion, diabatic heating, and rainfallcharacteristics in north Australia convective systems. Quart. J. Roy.Meteor. Soc. 124, 11331162.

Cifelli, R., Rutledge, S.A., 1994. Vertical motion structure in maritimecontinent mesoscale, convective systems: results from 50-MHz profiler.J. Atmos. Sci. 51, 26312652.

Csirmaz, K., Simon, A., Pistotnik, G., Polynszky, Z., Netiak, M., Nagykovcsi,Z., Sokol, A., 2013. A study of rotation in thunderstorms in a weakly- ormoderately-sheared environment. Atmos. Res. 123, 93116.

Dalal, S., Lohar, D., Sarkar, S., Sadhukhan, I., Debnath, G.C., 2012. Organiza-tional modes of squall-type mesoscale convective systems duringpre-monsoon season over eastern India. Atmos. Res. 106, 120138.

Dhaka, S.K., Choudary, S.M., Shibagaki, Y.S., Yamanaka, M.D., Fukao, S., 2002.Observable signatures of a convectively generated wave field over thetropics using Indian MST radar, at Gadanki (13.58N, 79.28E). Geophys.Res. Lett. 29, 18721875.

Dhaka, S.K., Takahasi, M., Malik, S., Shibagaki, Y.S., Fukao, S., 2003. Observationsof deep convective updrafts in tropical convection and their role in thegeneration of gravity waves. J. Meteorol. Soc. Jpn 81, 11851199.

Ecklund, W.L., Gage, K.S., Riddle, A.C., 1981. Gravity wave activity in the verticalwinds observed by the Poker Flat MST radar. Geophys. Res. Lett. 8, 285288.

Foken, T., Wichura, B., 1996. Tools for quality assessment of surface-basedflux measurements. Agric. For. Meteorol. 78, 83105.

Gage, K.S., Van Zandt, T.E., 1981. Wind measurement techniques availablefor the middle atmosphere program. J. Geophys. Res. 86, 9591.

Gage, K.S., Balsley, B.B., 1984. MST radar studies of wind and turbulence inthe middle atmosphere. Pure Appl. Geophys. 118, 452493.

Gregory, J.B., Meek, C.E., Manson, A.H., Stephenson, D.G., 1979. Fluctuationsin tidal (24-, 12-h) characteristics and oscillations (8-h5-d) in themesosphere and lower thermosphere (70110 km): Saskatoon (52N,107W), 19791981. IEEE Trans. Geosci. Electron. GE 17, 262.

Harper, R.M., Gordon, W.E., 1980. A review of radar studies of the middleatmosphere. Radio Sci. 15, 195. http://dx.doi.org/10.1029/RS015i002p00195.

Hildebrand, P.H., Sekhon, R.S., 1974. Objective determination of the noise levelin Doppler spectra Hildebrand. J. Applied. Met. 13, 808811.

Hocking, W.K., 1997a. Recent advances in radar instrumentation andtechniques for studies of the mesosphere, stratosphere and troposphere.Radio Sci. 32, 22412270.

Hocking, W.K., 1997b. Strength and limitations of MST radar measurementsof middle atmosphere wind. Ann. Geophys. 15, 11111122.

Houze, R.A., 2004. Mesoscale convective systems. Rev. Geophys. 42 (4)RG4003.

Kaimal, J.C., Finnigan, J.J., 1994. Atmospheric Boundary Layer Flows TheirStructure and Measurement. Oxford University Press, Oxford pp. 289.

Kudeki, E., Rastogi, P.K., Src, F., 1993. Systematic errors in radar windestimation: implication for comparative measurements. Radio Sci. 28,169179.

Kumar, K.K., 2006. VHF radar observations of convectively generated gravitywaves: some new insights. Geophys. Res. Lett. 33. http://dx.doi.org/10.1029/2006GL027404 L01803.

Larsen, M.F., Rttger, J., 1982. VHF and UHF Doppler radars as tools forsynoptic meteorology. Bull. Am. Meteorol. Soc. 63, 9961008.

Latha, R., Murthy, B.S., 2011. Boundary layer signature of consecutivethunderstorms as observed by Doppler sodar over western India. Atmos.Res. 99, 230240.

Mkel, A., Shrestha, R., Karki, R., 2014. Thunderstorm characteristics inNepal during the pre-monsoon season 2012. Atmos. Res. 137, 9199.

Moyer, K.A., Young, G.S., 1994. Observations of mesoscale cellular convectionfrom the marine stratocumulus phase of FIRE. Boundary Layer Meteorol.71, 109134.

Nastrom, G.D., 1997. Doppler radar spectral width broadening due to beam-width and wind shear. Ann. Geophys. 15, 786796.

Nastrom, G.D., Rao, P.B., Sivakumar, V., 2004. Measurements of atmosphericturbulence with the dual-beamwidth method using the MST radar atGadanki, India. Ann. Geophys. 22, 32913297. http://dx.doi.org/10.5194/angeo-22-3291-2004.

Palmer, R.D., Larsen, M.F., Woodman, R.F., Fukao, S., Yamamoto, M., Suda, T.,Kato, S., 1991. VHF radar interferometry measurements of verticalvelocity and the effect of tilted refractivity surfaces on standard Dopplermeasurements. Radio Sci. 26, 417427.

Pucillo, A., Manzato, A., 2013. Usefulness and skill of station-derivedpredictors in forecasting storm occurrence and intensity. Atmos. Res.123, 3147.

Rao, P.B., Jain, A.R., Kishore, P., Balamuralidhar, P., Damle, S.H., Vishwanathan, G.,1995. Indian MST radar, part I: system description and sample vector windmeasurements in ST mode. Radio Sci. 30, 11251138.

Rao, T.N., Kirkwood, S., 2005. Characteristics of tropopause folds over arcticlatitudes. J. Geophys. Res. 110. http://dx.doi.org/10.1029/2004JD005374.

Rao, T.N., Rao, D.N., Raghavan, S., 1999. Tropical precipitating systemsobserved with Indian MST radar. Radio Sci. 34, 11251139.

Rao, D.N., Rao, T.N., Venkatratnam, M., Thulasiraman, S., Rao, S.V.B., Rao, P.B.,Srinivasulu, P., 2001. Diurnal and seasonal variability of turbulenceparameters observed with Indian MST radar. Radio Sci. 36, 14391458.

Rao, T.N., Rao, D.N., 2007. A short review on wind profiler observations oflower and middle atmospheric processes over Gadanki. Indian J. RadioSpace Phys. 36, 526542.

Rao, T.N., Uma, K.N., Satyanarayana, T.M., Rao, D.N., 2009. Differences in draftcore statistics from wet spell to dry spell over Gadanki (13.5N, 79.2 E),India. Mon. Weather Rev. 137, 42934306.

Rao, T.N., Radhakrishna, B., Satyanarayana, T.M., Satheesh Kumar, S., 2010.The exchange across the tropical tropopause in overshooting convectivecores. Ann. Geophys. 28, 113122.

Sato, K., Hashiguchi, H., Fukao, S., 1995. Gravity waves and turbulence

Tyagi, B., Satyanarayana, A.N.V., 2013b. Assessment of turbulent kineticenergy and boundary layer characteristics at a tropical Indian stationRanchi. Asia Pac. J. Atmos. Sci. http://dx.doi.org/10.1007/s13143-013-0052-8.

Tyagi, B., Naresh Krishna, V., Satyanarayana, A.N.V., 2011. Skill of thermo-dynamic indices for forecasting pre-monsoon thunderstorms overKolkata during STORM pilot phase 20062008. Nat. Hazards 56,681698. http://dx.doi.org/10.1007/s11069-010-9582-x.

Tyagi, B., Satyanarayana, A.N.V., Kumar, M., Mahanti, N.C., 2012. Surface energyand radiation budget over a tropical station: an observational study. AsiaPac. J. Atmos. Sci. 48 (4), 411421. http://dx.doi.org/10.1007/s13143-012-0037-z.

Tyagi, B., Satyanarayana, A.N.V., Naresh Krishna, V., 2013a. Thermodynam-ical structure of atmosphere during pre-monsoon thunderstorm seasonover Kharagpur as revealed by STORM data. Pure Appl. Geophys. 170(4), 675687. http://dx.doi.org/10.1007/s00024-012-0566-5.

Tyagi, B., Satyanarayana, A.N.V., Rajvanshi, R.K., Mandal, M., 2013b. Surfaceenergy exchanges during pre-monsoon thunderstorm activity over atropical station Kharagpur. Pure Appl. Geophys. http://dx.doi.org/10.1007/s00024-013-0682-x.

Uma, K.N., Rao, T.N., 2008. Characteristics of vertical velocity cores in differentconvective systems observed over Gadanki, India. Mon. Weather Rev. 137,954975.

Uma, K.N., Rao, T.N., 2009. Diurnal variation in vertical air motion over atropical station, Gadanki (13.5N, 79.2E), and its effect on the estimationof mean vertical air motion. J. Geophys. Res. 114. http://dx.doi.org/10.1029/2009JD01256 (D20106).

Uma, K.N., Kumar, K.K., Rao, T.N., 2011. VHF radar observed characteristics ofconvectively generated gravity waves during wet and dry spells ofIndian summer monsoon. J. Atmos. Sol. Terr. Phys. 73, 815824.

Viswanadham, D.V., Satyanarayana, A.N.V., Mishra, Stuti, Parthasarathi, P.,1997a. Turbulent kinetic energy budget parameters over Varanasi fromMONTBLEX-90. Proc. Indian Natl. Sci. Acad. 63, 403412.

Viswanadham, D.V., Satyanarayana, A.N.V., Srivastava, M.K., Mishra, S.,1997b. Surface layer turbulent kinetic energy budget over Kharagpur.

215A.N.V. Satyanarayana et al. / Atmospheric Research 143 (2014) 198215air-Doppler radars. J. Geophys. Res. 100, 71117119.Stull, R.B., 1988. Introduction to Boundary Layer Meteorology. Kluwer

Academic Publishers, Dordrecht, Boston, London.Thomas, C., Foken, T., 2005. Detection of long-term coherent exchange over

spruce forest using wavelet analysis. Theor. Appl. Climatol. 80, 91104.Tsuda, T., Sato, T., hirose, K., Fukao, S., Kato, S., 1986. MU radar observations

of the aspect sensitivity of backscattered VHF echo power in thetroposphere and lower stratosphere. Radio Sci. 21, 971980.

Tyagi, B., Satyanarayana, A.N.V., 2013a. Coherent structures contributions influxes of momentum and heat at two tropical sites during pre-monsoonthunderstorm season. Int. J. Climatol. http://dx.doi.org/10.1002/joc.3785.Journal of Scientific Research 47, 1120 (Banaras Hindu University).Walker, J.C.G., 1979. Radar measurement of the upper atmosphere. Science

206, 180189. http://dx.doi.org/10.1126/science.206.4415.180.Wilczak, J.M., Oncley, S.P., Stage, S.A., 2001. Sonic anemometer tilt correction

algorithms. Boundary Layer Meteorol. 99, 127150.Yamamoto, M., Sato, T., May, P.T., Tsuda, T., Fukao, S., Kato, S., 1988.

Estimation error of spectral parameters of MST radars obtained by leastsquare fitting method and its lower bound. Radio Sci. 23, 10131102.associated with cumulus convection observed with the UHF/VHF clear-

Evaluation of atmospheric turbulence, energy exchanges andstructure of convective cores during the occurrence of mesoscaleconvective systems using MST radar facility at Gadanki1. Introduction2. Data3. MST radar system description4. Quality control4.1. MST radar data4.2. Micro-meteorological fast response data

5. Methodology6. Results and discussion6.1. DWR imagery analysis6.1.1. Case I6.1.2. Case II6.1.3. Case III

6.2. Signal-to-Noise Ratio of MST radar backscattered echoes6.3. Turbulence from spectral width6.3.1. Case I6.3.2. Case II6.3.3. Case III

6.4. Updrafts and downdrafts in terms of vertical wind profiles from the MST radar6.4.1. Case I6.4.2. Case II6.4.3. Case III

6.5. Variation of surface meteorological variables during the thunderstorm events6.5.1. Case I6.5.2. Case II6.5.3. Case III

6.6. Turbulence kinetic energy during the thunderstorm events6.6.1. Case I6.6.2. Case II6.6.3. Case III

6.7. Sensible heat flux6.7.1. Case I6.7.2. Case II6.7.3. Case III

6.8. Validation of MST radar wind profiler data with radiosonde observations6.8.1. Variation of profiles of horizontal wind components obtained from the MST radar wind profiler during the thunderstorm events6.8.1.1. Case I6.8.1.2. Case II6.8.1.3. Case III

7. Summary and conclusionAcknowledgmentReferences