Atmospheric Research 137 (2014) 91–99

Contents lists available at ScienceDirect

Atmospheric Research

j ourna l homepage: www.e lsev ie r .com/ locate /atmos

Thunderstorm characteristics in Nepal during thepre-monsoon season 2012

Antti Mäkelä a,⁎, Rajendra Shrestha b, Ramchandra Karki b

a Finnish Meteorological Institute (FMI), Finlandb Department of Hydrology and Meteorology of Nepal (DHM), Nepal

a r t i c l e i n f o

⁎ Corresponding author at: Finnish Meteorological IFIN-00101 Helsinki, Finland. Tel.: +358 503011988.

E-mail address: antti.makela@fmi.fi (A. Mäkelä).

0169-8095/$ – see front matter © 2013 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.atmosres.2013.09.012

a b s t r a c t

Article history:Received 14 June 2013Received in revised form 4 September 2013Accepted 16 September 2013

A training period of lightning location data usage has been carried out in Nepal during thepre-monsoon season April–June 2012. The training was one part of a Finnish–Nepalese Project(FNEP) between the Department of Hydrology andMeteorology of Nepal (DHM) and the FinnishMeteorological Institute (FMI). FNEP aimed for the development of operational meteorologicalreadiness in a developing country such as Nepal. The lightning location training included theintroduction to lightning location techniques and principles and the actual hands-on training forthe operational DHM forecasters. The lightning location system used was the Vaisala long rangeGlobal Lightning Dataset 360 (GLD360), which has practically a global coverage. During the threemonths of training, a dataset of Nepalese lightningwas also collected, indicating the pre-monsoonthunderstorm characteristics of Nepal.

© 2013 Elsevier B.V. All rights reserved.

Keywords:Lightning locationNepalese thunderstormsSevere weatherFinnish–Nepalese ProjectDevelopment of meteorological readiness

1. Introduction

Global statistics of thunderstorms (Christian et al., 2003)indicate that the most abundant regions of lightning are thetropical continental areas inAfrica, South America and Indonesia,where thunderstorms occur almost on daily basis. Outside thetropics, thunderstorms have a seasonal cycle: the thunderstormseason is shorter, and therefore also less lightning occurs duringthe year. However, despite the shorter season, individual stormsat higher latitudesmay reach extreme intensity. A good exampleis the thunderstorms in the central parts of the United States(Christian et al., 2003; Mäkelä et al., 2011).

In the Himalayan region, the local climate is influenced bythe summer monsoon which brings large amounts of moistureand precipitation from the Indian Ocean towards the north(Robinson and Henderson-Sellers, 1999; Malla, 2008). Beforethe monsoon, the atmosphere changes into highly unstableand intense convection occurs. Thunderstorms during thispre-monsoon season have also been noted to reach extreme

nstitute, P.O. Box 503,

ll rights reserved.

intensity, and they have a great impact to the densely populatedcountries such as Sri Lanka, India, Bhutan, Bangladesh, andNepal. According to Christian et al. (2003), extremely highflash rates occur in the northern Pakistan in the Himalayas.Unfortunately, there are not many studies concerning thethunderstorm climate of these areas.

Nepal is without a doubt one of the most intriguingcountries in the world (Fig. 1); the Himalayan mountains liein the northern parts of the country, with eight of the tentallest mountain tops in the world situated in Nepal. Theelevation changes rapidly in the South–North direction; thelowest areas in the South are only some hundreds of metersabove the sea level, while the Mount Everest only about200 km to the North peaks at almost 9000 m. The climate ofthe lowest areas ismoist andwarm,while in themountains theair is dry and cold (Robinson and Henderson-Sellers, 1999).

Nepal is divided into five development regions (shown inFig. 1 with thick solid line): Far West, Midwest, West, Central,and East. The development areas are further subdivided into atotal of 14 zones, which are numbered in Fig. 1 and explained inFig. 1 caption.

Several kinds of meteorological studies have been madein the Himalayan region during the years. Many of them have

Fig. 1. The elevation, development areas and zones of Nepal. Development areas are (solid thick lines from West to East) Far West, Midwest, West, Central, andEast. The zones are (dashed lines): 1 = Mahakali, 2 = Seti, 3 = Karnali, 4 = Bheri, 5 = Rapti, 6 = Dhawalakiri, 7 = Lumbini, 8 = Gandaki, 9 = Narayani,10 = Bagmati, 11 = Janakpur, 12 = Sagarmatha, 13 = Kosi, and 14 = Mechi.

92 A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

dealt with the characteristics and effects of the summermonsoon (He et al., 1987; Barros et al., 2000; Krishnamurtiand Kishtawal, 2000; Shrestha, 2000; Lang and Barros, 2002;Barros and Lang, 2003), winds (Egger et al., 2000; Zängl et al.,2001), mountain effects (e.g. Kuo and Qian, 1981), precipi-tation (e.g. Shrestha et al., 2000), and weather conditions ingeneral (Ye, 1981; Bollasina et al., 2002). However, althoughsome studies of Nepalese thunderstorms exist (e.g. Rosoff andHindman, 2002) the thunderstorm characteristics of Nepalhave not been studied deeply.

One of the most detailed investigations of Nepaleselightning is by Baral and Mackerras (1992). They studiedthe lightning occurrence characteristics with a flash counternetwork for a total of 21 months (March 1987–November1988) in the KathmanduValley. Their results indicate thatwhenthe lightning activity starts in March, it intensifies quicklyreaching its peak in May, while in June the activity decreasesrapidly as themonsoon season starts. Similar findings have beenfound by Kayastha and Regmi (2008).

The pre-monsoon season thunderstorms usually approachfrom Northwest and hence are called Nor'westers. Nor'westersare mesoscale convective systems, which can develop underlarge scale envelope of the seasonal low-level trough with apossible embedded low pressure area.

The upper air flow over Gangetic West-Bengal and theadjoining areas has a shallow layer of moist southerlies/southwesterlies from the Bay of Bengal near the ground anddry westerlies aloft. Although this type of situation prevailsalmost daily during April–May, thunderstorms do not occurevery day. The atmospheric instability is released through

the action of large scale flow and the associated synoptic andsub-synoptic systems. In the lower troposphere, pressuretroughs, low pressure areas, and wind convergence areimportant. In the upper troposphere, divergent flow can causethe moist and unstable lower level air to be drawn upwardleading to convective cloud formation and hence thunder-storms. The divergent flow is usually associatedwith the uppertropospheric trough in the westerlies. Areas of upper tropo-spheric positive vorticity advection in association of large scalevertical upward motion help triggering thunderstorms.Towards mid May, anticyclonic flow starts in the middleand upper troposphere, prior to the onset of monsoon. Theanticyclonic circulation also provides upper level divergentfield, mainly near the Gangetic West Bengal and neighbor-hood which favors development of convection.

Nowadays, thunderstorm characteristics can be examinedaccuratelywith lightning location systems (LLS),which providelightning location data practically in real time. Many types ofLLSs exist (Cummins and Murphy, 2009): short-range systemsutilizing Very High Frequencies (VHF) are intended for detaileddetection of practically all discharges occurring in the thunder-cloud (Richard et al., 1986), while long-range systems moni-toring Very Low Frequency (VLF) and Low Frequency (LF)electromagnetic signals emitted by lightning return strokes aredesigned to cover large areas with only a handful of sensors(Cummins and Murphy, 2009). This is possible because VLFpulses travel long distances reflecting back and forth betweenthe ground and ionosphere. The latter types of LLSs areespecially useful because they can monitor areas which aredifficult to cover otherwise (e.g., oceans). Also, a long range

93A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

LLS can provide essential information on thunderstormsover countries in which a nation-wide network is difficult oreven impossible to establish because of a complicated terrainor of poor economical situation. Nepal fulfills both of thesearguments.

The Finnish–Nepalese Project (FNEP; Mäkelä et al., 2012)was a project in 2010–2012 aimed for the development of themeteorological capability and readiness of the government ofNepal. The project was funded by the Ministry of the ForeignAffairs of Finland, and the partners were the Finnish Meteoro-logical Institute (FMI) and the Department of Hydrology andMeteorology of Nepal (DHM). FNEP consisted of training ofthe DHMpersonnel and the development of themeteorologicalinfrastructure. For example, one outcome of FNEP was theinstallation, update and data communication and data feedarrangements of automatic surface weather stations. Thefoundations of this kind of project are easy to understandwith the data shown in Table 1, which contains disaster datafrom 2012, collected by the National Emergency OperationCenter of Nepal operating under the Nepalese Ministry ofHome Affairs. Table 1 shows how big losses and impacts theweather poses in Nepal. For example, in 2012 more than 100people were killed by lightning and twice as many wereinjured; this is much more than those found in previousstudies (e.g., Holle, 2008), but it may also be specific only foryear 2012. Because intense thunderstorms in Nepal causesannually the loss of lives and severe damage to the infrastruc-ture, a training on the use of lightning location data was alsopursued during themain thunderstorm season April–June2012.

This paper summarizes the results of the lightning locationtraining and shows what kinds of benefits a developing countrylike Nepal get from the real time information of occurringlightning. The paper is organized in the followingway. Section 2introduces the lightning location data used in the training(Vaisala GLD360) and how it was made accessible to DHM, andhow the training itself was conducted. Section 3 shows ourmain results, i.e., the lightning statistics during April–June 2012and the feedback from the DHM trainees (i.e., the operationalweather forecasters). Sections 4 and 5 contain the discussionand conclusions, respectively.

2. Data

Vaisala Global Lightning Dataset 360 (GLD360) is a relativelynew concept of locating lightning on long ranges (Said et al.,2010; Demetriades et al., 2010). The network was launched inSeptember 2009 and according to recent studies (Pohjola et al.,2011; Poelman et al., 2013; Pohjola and Mäkelä, 2013) thecoverage and performance are good considering its long range

Table 1Disaster data 2012 based on the National Emergency Operation Center of Nepal.

People

Type Death Missing Injured Affected family

Flooding 52 39 8 104Landslide 60 8 33 66Rainfall 11 0 9 0Storm 18 0 20 102Lightning 119 0 267 35Total 337 47 471 2743

nature. For example, in Belgium (Poelman et al., 2013) thedetection efficiency (DE) and the median location accuracy (LA)for cloud-to-ground (CG) flasheswere observed to be above 96%and 0.5 km, respectively, based on a ground-truth study of 57negative CG flashes; for CG strokes the DE was 70% and themedian LA 3.3 km. A statistical comparison for a larger datasetover Austria and Europe (Pohjola andMäkelä, 2013) shows thatthe GLD360 relative stroke DE was about 50%, which suggestsa CG flash DE of at least 70%. Unfortunately, no performancestudies of GLD360 are found close to Nepal. However, if weassume the DE to vary only little globally (Demetriades et al.,2010), the performance parameters discussed above can beestimated also for Nepal.

The GLD360 data consists of lightning strokes detected inthe VLF domain. At the time this study, GLD360 strokes arenot grouped into flashes or classified as CG or intracloud (IC).Therefore, the GLD360 data consists of any kind of lightningdischarge which emits a detectable signal. It is likely thatsome of the located events are IC events, although their ratiois probably small. For more discussion on this issue, we referto Pohjola and Mäkelä (2013).

Besides GLD360, there are several other long-range LLSdata providers and networks, for example, the ATDNet of theUK Met Office (Bennett et al., 2010), and the World WideLightning Location Network (WWLLN, Rodger et al., 2006).

In this study, the GLD360 data was established to the DHMoperational forecasting office at the Tribhuvan InternationalAirport in Kathmandu via a direct ssh-connection. The ssh-terminal at a laptop provided by Vaisala was open 24/7 duringthe training and the data was made visible by a visualizingsoftware (Vaisala LTS2005), also provided by Vaisala for thetraining period. In general, this configuration worked well,except when power outages or network problems occurred: inthese cases the ssh-connection had to be rebooted and themeanwhile occurring strokes were missed. Luckily, there werenot many such cases, and when these occurred, they influencedonly the real time training. In the lightning location statistics inSection 3, we have used the archived data provided by Vaisala toensure the homogeneity of the data; the data covers the area26.0°–31°N, 79.0°–88.5°E. Thunderstorm days have been calcu-lated from the LLS data with the method described in Mäkeläet al. (2011), i.e., assuming an average effective humanobservable area of thunder of 400 km2, which equals to aradius of a circle of about 11 km.

After the successful installation of the GLD360 data feed andvisualization, the practice continued with lectures on lightninglocation data usage in operational weather forecasting, withhands on training, andother issues regarding the use of lightninglocation data. These will be discussed in Subsection 3.2.

Houses etc.

Animals Completely Partly Destroyed shed

75 170 309 327 65 74 116 70 60 73 103 132 0

256 49 17 81181 4247 816 267

94 A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

The occurrence of thunderstorm has been measuredin synoptic weather stations in Nepal during the years.To have comparative data for our LLS measurements, andbecause the LLS test campaign covered only three months,we have analyzed here the synoptic weather station datafrom a total of 15 stations from Nepal from the period of2000–2009. These results are summarized shortly at the end ofSection 3.

3. Results

3.1. Lightning location statistics in Nepal in April–June 2012

During the three-month study period, a total of 460,363strokes were located over the study area. A total of 18.4% ofthe strokes occurred in April, 38.6% in May, and 43.0% in June.The daily distribution of strokes in Fig. 2 shows that lightningoccurred already during the first half of April. The activetwo-week period was followed by a quiet end-of-the month.In May, lightning activity was concentrated into three shorterperiods, each about one week long. The most stroke-abundantday was June 1, during which almost 50,000 strokes werelocated. In June, thunderstorms occurred daily; intense thun-derstorm occurred also on 5th and 14th of June.

The average hour-to-hour variation of located strokes(Fig. 3) during the study period is interesting; the largestpercentages during the local afternoon and early eveninghours are a common feature globally, but the nearly as largepercentages occurring between themidnight andmorning hours(up to 9 a.m.) indicate the presence of vigorous night-timeconvection. According to Barros and Lang (2003) this is causedby the steady increase of moisture during the day from themonsoon flow, the instability peaking during the night becauseof the interaction between increased surface moisture, anddiurnal heating and cooling of the atmosphere.

Fig. 2. The daily number of strokes in the study

The stroke density in Nepal and surroundings (Fig. 4)indicates that the largest density values occur in the southernand south-eastern parts of Nepal, near the Indian border.The southern parts of Narayani (zone 9), Sagarmatha (12),Janakpur (11), Kosi (13), and Mechi (14) experienced strokedensity values 5 km−2 or larger per three months. Especiallythe southeasternmost areas (approximately x = 850 km,y = 50 km in Fig. 4) show extremely large values; the largestobserved value is 27 strokes km−2 per three months. Whencomparing Figs. 1 and 4, it is clear how the stroke densityvalues drop rapidly starting at the elevation of about 3000 m.No lightning was observed over certain high-elevation areassuch as the northern parts of Karnali (zone 3) and Sagarmatha(12), where also Mount Everest is located.

The regional distribution of the thunderstorm days (Fig. 5)indicates that the largest values (50 days or more) occurred inGandaki (zone 8) and Dhawalakiri (6), and also in Kosi (13).The locations of these areas differ from the stroke density map(Fig. 4), i.e., the largest stroke density values and the largestnumber of thunderstorm days do not overlap. This suggests aregional difference in the intensity of individual storms. Thiscan be seen clearly when calculating the stroke density perthunderstorm day (Fig. 6), i.e., the ratio between the values inFigs. 4 and 5. This parameter has been noted to provideimportant information regarding the characteristics of individ-ual thunderstorm days (Mäkelä et al., 2011). According toFig. 6, certain parts of Nepal near the southern border wereunder the influence of extremely intense storms because of therelatively low number of thunderstorm days yet large amountof lightning. However, in the areaswith plenty of thunderstormdays as shown in Fig. 5 (Kandaki, Dhawalakiri and Kosi) thetypical thunderstorm day was “modest”.

We have also drawn the stroke density and thunderstormdaymaps individually for eachmonth (Fig. 7). Interestingly, theregional occurrence pattern of thunderstorm days is practically

area between April 1 and June 30, 2012.

Fig. 3. The average diurnal distribution of strokes over the study area between April 1 and June 30, 2012.

95A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

the same for eachmonth, only the values show larger difference.The stroke density has similar features; lightning occurredpractically over the same regions during the study period, butthe amount of strokes varies frommonth to month. A commonfeature seems to be that the southeastern corner of Nepalwas influenced by intense thunderstorms each month. Also,according to Holle (2008), tens of people get killed by lightningannually in this region.

The monthly thunderstorm statistics (Fig. 8) from thesynoptic stations show that the activity during the pre-monsoonseason is higher than that during other seasons: pre-monsoon(Mar–May) 43%, monsoon (Jun–Aug) 36%, post-monsoon(Sep–Nov) 18% and winter (Dec–Feb) 3%, respectively. Theseresults for the pre-monsoon season are highly similar to theones obtained with the LLS in this study.

3.2. Benefits of lightning location data usage in Nepal

At the moment, the main user of lightning location data inNepal is the operational weather forecasting, which useslightning data to pinpoint the thunderstormareas, estimate theirmovement and threat level. Because the FNEP lightning locationtraining was established at the Tribhuvan International Airport

Fig. 4. Stroke density (strokes km−

in Kathmandu, especially aviation showed high interest into thedata; thiswas actually the first time thatmost aviation operatorsat the airport saw this kind of data in operation. Also, thefollowing groups were also recognized as potential lightninglocation data users: aviation, military, electricity and powerplants, authorities and insurance companies.

Because at the moment there are no weather radars inNepal, lightning location data could be used as a first guess ora proxy for convective precipitation (Pessi and Businger, 2009).Naturally, this would serve only for convective precipitation.However, because intense convective precipitation is usuallythe cause for landslides and flash floods in Nepal, this methodwould provide warnings regarding these threats before theinstallation of Nepalese weather radars. Furthermore, becauselightning location data usually extends beyond the nationalborders (depending on the system), approaching intenseconvective areas will be observed before their arrival.

In a developing country such as Nepal, lightning locationdata could be economically useful: if the data, end-products, orservice would be sold for a recognized group of commercialcustomers, the national weather service could use the incomefor improving their meteorological infrastructure whichwoulddirectly improve the overall meteorological readiness.

2 in April 1–June 30, 2012).

Fig. 5. Number of thunderstorm days in April 1–June 30, 2012.

96 A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

4. Discussion

The training resulted in the following list of suggestionsand ideas for similar projects:

• lightning location data is a relatively simple and nowadayseasily available data for the monitoring and analysis ofthunderstorms,

• a long-range lightning location system makes possible theobservation of thunderstorms globally,

• long-range lightning location data can be transmitted practi-cally to any place on earth via internet; i.e., sensor installationsin the area in question are not necessary.

Our synthesis is that national hydrological and meteoro-logical services (NHMSs) of developing countries can get largebenefits from receiving long range lightning location data. Anational lightning location system owned by an NHMS is anexcellent and efficient tool for observing thunderstorms, butit needs continuous maintenance, analysis and suitable andsafe installation sites for the sensors; these may be difficult tofulfill which may lead to a network of decreasing performance(Schulz et al., 2005; Mäkelä et al., 2010). We suggest thatdeveloping countries should first use long range LLS data to gethands-on experience on the data and its usefulness in differentfields of meteorology, and afterwards make plans regarding anational medium-range network.

Fig. 6. Strokes per thunderstorm day in April 1–June 30, 2

5. Conclusions

A project aimed for the development of meteorologicalreadiness in Nepal was carried out in 2010–2012 together withthe Department of Hydrology andMeteorology of Nepal (DHM)and the Finnish Meteorological Institute (FMI); the project wasfunded by the Ministry of Foreign Affairs of Finland. One partof the project was a lightning location training; a data feedfrom the Vaisala GLD360 global lightning location system wasestablished for the DHM operational center at the TribhuvanInternational Airport in Kathmandu. The training period coveredthe pre-monsoon seasonApril–June,which is themost lightningabundant period in Nepal during a year. After the three-monthoperational usage the archived lightning location data wasanalyzed to find out the thunderstorm characteristics of thepre-monsoon season; these results are presented in this paper.

Our results show that the southern parts of regional zonesof Narayani, Sagarmatha, Janakpur, Kosi, andMechi experiencedstroke density values 5 km−2 or larger per three months; thelargest value was 27 strokes km−2 in April–June. The strokedensity decreases fast relative to the altitude, starting at about3000 m.

The local number of thunderstormdayswas largest (50 daysor more) in Gandaki, Dhawalakiri, and Kosi. Interestingly,most of these areas did not show the largest stroke densityvalues, which suggests a regional difference in the intensityof individual storms. Indeed, the ratio between the stroke

012. The unit is strokes km−2 thunderstorm day−1.

Fig. 7. Thunderstorm days (a–c) and stroke density (d–f) for April (upper row), May (middle row) and June (lowest row), respectively.

97A.M

äkeläet

al./Atm

osphericResearch

137(2014)

91–99

Fig. 8. The average monthly distribution of thunderstorm days in Nepal in 2000–2009 according to a total 15 of synoptic weather stations.

98 A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

density and thunderstorm days shows that the most intensestorms occurred near the southern and southeastern borderof Nepal.

In this study, the largest fraction of the total 460,363 locatedstrokes occurred in June (43.0%); the percentages of strokes forMay and April were 38.6% and 18.4%, respectively. The largestnumber of strokes per day, about 49,000, occurred on June 1.

The diurnal distribution of lightning showed peak at 17–18local time, which is common feature for thunderstorms globally.The interesting feature was a secondary peak during the nightuntil morning (01–09 local time). This is caused by the highnight-time instability due to the increased surface moisture anddiurnal heating and cooling of the atmosphere. This effect is notobserved in all climates.

The feedback and experience from the DHM meteorolo-gists indicate large potential of lightning location data inNepal. Because of more accurate nowcasting capability ofthunderstorms, lightning location data makes possible bettertools for warning services; this is a highly important matter,because about 30–50 persons are killed annually by lightningin Nepal. Besides the public warnings, lightning location datacould be used for commercial purposes, and the income couldbe used for improving the overall meteorological infrastructureand readiness in Nepal. Maybe themost important result of thetraining was the orientation of the Nepalese meteorologists tounderstand, use and analyze the lightning location data.

Acknowledgments

We highly appreciate the constructive criticism and com-ments of the anonymous reviewers. We thank Vaisala forproviding the lightning location data for this study. The studyhas been financed by the Ministry of the Foreign Affairs ofFinland through the Finnish–Nepalese Project (FNEP) and bythe Finnish Funding Agency for Technology and Innovation(Tekes) through the program Measurement, Monitoring andEnvironmental Assessment (MMEA/CLEEN).

References

Baral, K.N., Mackerras, D., 1992. The cloud flash-to-ground flash ratio andother lightning occurrence characteristics in Kathmandu thunderstorms.J. Geophys. Res. 97, 931–938.

Barros, A.P., Lang, T.J., 2003. Exploring spatial modes of variability of terrain–atmosphere interactions in the Himalayas during monsoon onset.Hydrosciences Rep. Series, 03-001. Division of Engineering and AppliedSciences. Harvard University (51 pp.).

Barros, A.P., Joshi, M., Putkonen, J., Burbank, D.W., 2000. A study of the 1999monsoon rainfall in a mountainous region in central Nepal using TRMMproducts and rain gauge observations. Geophys. Res. Lett. 27, 3683–3686.

Bennett, A.J., Odams, P., Edwards, D., Arason, Þ., 2010. Monitoring of lightningfrom the April–May 2010 Eyjafjallajökull volcanic eruption using a verylow frequency lightning location network. Environ. Res. Lett. 5. http://dx.doi.org/10.1088/1748-9326/5/4/044013.

Bollasina, M., Bertolani, L., Tartari, G., 2002. Meteorological observations athigh altitude in the Khumbu Valley, Nepal Himalayas, 1994–1999. Bull.Glaciol. Res. 19, 1–11.

Christian, H.J., Blakeslee, R.J., Boccippio, D.J., Boeck, W.L., Buechler, D.E.,Driscoll, K.T., Goodman, S.J., Hall, J.M., Koshak, W.J., Mach, D.M., Stewart,M.F., 2003. Global frequency and distribution of lightning as observedfrom space by the Optical Transient Detector. J. Geophys. Res. 108.http://dx.doi.org/10.1029/2002JD002347.

Cummins, K.L., Murphy, M.J., 2009. An overview of lightning locatingsystems: history, techniques, and data uses, with an in-depth look atthe U.S. NLDN. IEEE Trans. Atmos. Elec. 51, 499–518.

Demetriades, N.W.S., Murphy, M.J., Cramer, J.A., 2010. Validation of Vaisala'sglobal lightning dataset (GLD360) over the continental United States.21st International Lightning Detection Conference (ILDC) & 3rdInternational Lightning Meteorology Conference (ILMC), 19–20 April2010, Orlando, Florida (6 pp.).

Egger, J., Bajrachaya, S., Egger, U., Heinrich, R., Reuder, J., Shayka, P., Wendt,H., Wirth, V., 2000. Diurnal winds in the Himalayan Kali Gandaki valley.Part I: observations. Mon. Weather Rev. 128, 1106–1122.

He, H., McGinnis, J.W., Song, Z., Yanai, M., 1987. Onset of the Asian monsoonin 1979 and the effect of the Tibetan Plateau. Mon. Weather Rev. 115,1966–1995.

Holle, R.L., 2008. Annual rates of lightning fatalities by country. 20th Int.Lightning Detection Conf., 21–23 April, Tucson, AZ, USA (14 p., Availablefrom http://www.vaisala.com/Vaisala%20Documents/Scientific%20papers/Annual_rates_of_lightning_fatalities_by_country.pdf).

Kayastha, S.P., Regmi, S.K., 2008. Nor'westers and tornadoes over the SAARCregion and their forecasting and preparedness. Proc. South Asian Associationfor Regional Cooperation (SAARC) Seminar, 20–21 August, 2008, DhakaBangladesh, pp. 28–34.

Krishnamurti, T.N., Kishtawal, C.M., 2000. A pronounced continental-scalediurnal mode of the Asian summer monsoon. Mon. Weather Rev. 128,462–473.

99A. Mäkelä et al. / Atmospheric Research 137 (2014) 91–99

Kuo, H.L., Qian, Y.F., 1981. Influence of the Tibetan Plateau on cumulative anddiurnal changes of weather and climate in summer. Mon. Weather Rev.119, 2337–2356.

Lang, T.J., Barros, A.P., 2002. An investigation of the onsets of the 1999 and2000 monsoons in central Nepal. Mon. Weather Rev. 130, 1299–1316.

Mäkelä, A., Tuomi, T.J., Haapalainen, J., 2010. A decade of high-latitude lightninglocation: effects of the evolving location network in Finland. J. Geophys.Res. 115 (D21124).

Mäkelä, A., Rossi, P., Schultz, D.M., 2011. The daily could-to-ground lightningflash density in the contiguous United States and Finland. Mon. WeatherRev. 139 (5), 1323–1337.

Mäkelä, A., Ylikangas, I., Karki, R., Budhathoki, K.P., Perrels, A., Santti, K., Partio,M., Keranen, M., 2012. FNEP: Finnish–Nepalese Project for improvingthe meteorological readiness in Nepal. J. Hydrol. Meteorol. Soc. Hydrol.Meteorol. Nepal. 8, 71–75.

Malla, G., 2008. Climate change and its impact on Nepalese agriculture. J. Agric.Environ. 9, 10.

Pessi, A.T., Businger, S., 2009. Relationships among lightning, precipitation,and hydrometeor characteristics over the North Pacific Ocean. J. Appl.Meteorol. Climatol. 48, 833–848.

Poelman, D.R., Schulz, W., Vergeiner, C., 2013. Performance characteristics ofdistinct lightning detection networks covering Belgium. J. Atmos. Ocean.Technol. 30, 942–951.

Pohjola, H., Mäkelä, A., 2013. The comparison of GLD360 and EUCLIDlightning location systems in Europe. Atmos. Res. 123, 117–128.

Pohjola, H., Mäkelä, A., Demetriades, N.W.S., Hembury, N., Holle, R.L., 2011.The benefits of GLD360 lightning location data in operational weatherforecasting. Preprints., 6th Conference on Severe Storms. Palma de Mallorca,Spain (ESSL).

Richard, P., Delannoy, A., Labaune, G., Laroche, P., 1986. Results of spatial andtemporal characterization of the VHF–UHF radiation of lightning. J. Geophys.Res. 91, 1248–1260.

Robinson, P.J., Henderson-Sellers, A., 1999. Contemporary climatology.Addison Wesley Longman, Singapore (317 pp.).

Rodger, C.J., Werner, S., Brundell, J.B., Lay, E.H., Thomson, N.R., Holzworth,R.H., Dowden, R.L., 2006. Detection efficiency of the VLF World-WideLightning Location Network (WWLLN): initial case study. Ann. Geophys.24, 3197–3214.

Rosoff, Y.N., Hindman, E.E., 2002. Mt. Everest, 10 May 1996: study of a highelevation thunderstorm. 10th Conf. on Mountain Meteorology, Ameri-can Meteorological Society, 17–21 June, Park City, USA (4 pp.).

Said, R.K., Inan, U., Cummins, K., 2010. Long-range lightning geolocationusing a VLF radio atmospheric waveform bank. J. Geophys. Res. 115,1–19.

Schulz, W., Cummins, K., Diendorfer, G., Dorninger, M., 2005. Cloud-to-groundlightning in Austria: a 10-year study using data from a lightning locationsystem. J. Geophys. Res. 110 (D09101).

Shrestha, M.L., 2000. Interannual variation of summer monsoon rainfall overNepal and its relation to the Southern Oscillation Index. Meteorol. Atmos.Phys. 75, 21–28.

Shrestha, A.B., Wake, C.P., Dibb, J.E., Mayewski, P.A., 2000. Precipitationfluctuations in the Nepal Himalaya and its vicinity and relationship withsome large scale climatological parameters. Int. J. Climatol. 20, 317–327.

Ye, D., 1981. Some characteristics of the summer circulation over the Qinghai–Xizang (Tibet) Plateau and its neighborhood. Bull. Am. Meteorol. Soc. 62,14–19.

Zängl, G., Egger, J., Wirth, V., 2001. Diurnal winds in the Himalayan KaliGandaki valley. Part II: modeling. Mon. Weather Rev. 129, 1062–1080.