In the frame of this paper, 1-year of lightning data from the experimental network ZEUS operated by National Observatory of Athens is analyzed. The area of interest is the Mediterranean and surrounding countries. At a first stage, the ZEUS data are compared with the data provided by the UK Met Office long-range lightning detection network for the warm period of the year. It was found that although ZEUS system underestimates the nighttime and early morning activity, during the rest of the day ZEUS system detects increased number of lightning compared to ATD that suffers from saturation when the number of flashes exceeds a certain threshold. Furthermore, the possible relationship between lightning and elevation, terrain slope and vegetation is investigated. The analysis showed that during spring and summer there is a positive relationship of lightning activity with elevation, while this feature is not evident during the rest of the year. The lightning activity was found to be positively correlated with the elevation slope throughout the year except winter. As it concerns the vegetation cover, it was found that over bareground the lightning “yield” is low during the whole year, while the inverse is true for woodland areas. During the warm period of the year, due to drying of the Mediterranean surfaces, the forested and wooded areas that keep soil moisture present increased lightning “yield” in contrast with the shrubland areas that, for the same period, present a decreased lightning yield.
 The number of studies that discuss the relationship between lightning and surface characteristics such as elevation, slope and vegetation is quite limited. Dissing and Verbyla  found over Alaska a positive correlation between lightning and elevation up to 1100–1200 m and also that the differential heating between adjacent surfaces with different vegetation cover may trigger mesoscale circulations that are likely to produce convection. Recently, Kilinc and Beringer  studied among other issues the relationship of lightning strikes with elevation and vegetation type in the Northern Territory of Australia, and found a distinct increase in lightning strike density from woodland to shrub to grassland which was attributed to greater surface heating of grasslands than the other vegetation types.
 With the aim to contribute to the improvement of understanding of lightning activity in the Mediterranean, this work explores the possible relationship between lightning and elevation, terrain slope and vegetation. One-year data from the experimental long-range lightning detection network ZEUS operated by the National Observatory of Athens, are used. Especially for the summer period (June–August 2006), ZEUS data are compared with the data provided by the long-range lightning detection network operated by UK Met Office.
2.1. ZEUS Lightning Detection Network
 ZEUS is a long-range lightning detection network with receivers located at five sites over Europe (Birmingham in the United Kingdom, Roskilde in Denmark, Iasi in Romania, Larnaka in Cyprus and Lisbon in Portugal). The system was manufactured by Resolution Displays, Inc. ZEUS receivers record the radio noise (sferics) emitted by cloud-to-ground lightning discharges in a very low frequency (between 7 and 15 kHz). The VLF signal is preamplified at each receiver site and the signal is synchronized to geographic positioning system time. At each receiver site an identification algorithm is executed that detects a probable sferics candidate, excludes weak signal and noise and is capable of capturing up to 70 sferics per second. Then the lightning location is retrieved (at the central station of the network) using the arrival time difference triangulation technique. The arrival time difference values represent positions between two outstations with the same time difference, and their intersection defines a sferic fix. ZEUS locating algorithm requires a minimum of four receivers to record the same event. Further details on ZEUS locating algorithm can be found in the study of Chronis and Anagnostou . As it concerns the location accuracy of ZEUS system a work is underway that compares its measurements with those from LINET lightning detection network [Betz et al., 2004] over Europe. Preliminary results have shown that the location accuracy of ZEUS is of the order of 4–5 km over the study area (Betz, personal communication).
 In the frame of this study ZEUS data from June 2006 up to May 2007 are analyzed. Although ZEUS is an experimental network the data availability during this period was quite high (90%). The 10% gaps within the data set may have some impact on the results presented in the following, and it is in the aims of the authors to extend this study when a larger and more complete database will be available in the future.
2.2. UK Met Office ATD Lightning Detection Network
 As ZEUS is an experimental network a comparison with another quite established long-range cloud-to-ground detection system is of interest. Data for a 3-month period (June–August 2006) were provided by the Arrival Time Difference (ATD) lightning location system operated by UK Met Office. The ATD system observes vertically polarized lightning discharges centered near 10 kHz. The waveforms are received by 8 ground stations (at Exeter, Camborne in Cornwall and Lerwick in the Shetland Islands in the United Kingdom and in Iceland, Finland, Germany, Gibraltar and Cyprus). Time differences are obtained by correlating the waveforms from the different outstations. Three time differences (satisfactory quality waveforms received from a minimum of four outstations) are required to compute a flash location. Errors associated with each flash location are also estimated and reported along with the flash location. The operation and characteristics of the ATD network, including its detection efficiency, have been analytically described by Holt et al.  and Keogh et al. .
2.3. Elevation and Vegetation Data
 The GLOBE orography database with a resolution of 1/120° derived by www.ngdc.noaa.gov/mgg/topo/globe.html [Hastings et al., 1999] and the global vegetation cover data set with the same resolution derived from the Global Land Cover Facility (www.landcover.org [Hansen et al., 2000]) have been used in this study. The domain of the analysis has been restricted in the area 0–32°E, 31–46°N (Figure 1). Within this area that encounters the Mediterranean Sea and the surrounding countries the surface of land areas equals the surface of sea areas so as to facilitate the discussion about the distribution of lightning over land compared to that over the sea. The vegetation database includes 14 categories (evergreen needleleaf forest, evergreen broadleaf forest, deciduous needleleaf forest, deciduous broadleaf forest, mixed cover, woodland, wooded grassland, closed shrubland, open shrubland, grassland, cropland, bare ground, urban, and water bodies). From the orography database slopes in percent have been calculated. The method uses a 3 × 3 grid centered over the cell where the slope is calculated as described by Burrough .
3.1. ZEUS and ATD Data Comparison
 The comparison of ZEUS with ATD data aims at a certain point to reveal any advantages or shortcomings of this experimental network as compared to a well-established network. During the analyzed period, that is June through August 2006, most of the electrical activity is concentrated over the land surfaces. Figure 2 presents the hourly variation of the percentage of the total number of lightning recorded over land over the total number of lightning (over both land and water) as sensed by both ZEUS and ATD systems. The two curves are quite close to each other and they show that from 1200 to 1900 UTC (local time is UTC + 1 and UTC + 2 in the studied domain) more than 80% of the lightning was detected over land, while before 0900 UTC the percentage of lightning over land drops to less than 45%. This feature was expected as during the warm period of the year strongly electrified convection over land is favored as excessive daytime heating over land produces larger and more buoyant boundary layer parcels which more efficiently transform convective available potential energy (CAPE) to kinetic energy of the updrafts in the moist stage of conditional instability.
Figure 3 presents the total number of lightning per hour for the whole three month period as sensed by the two networks over the land and over the sea. It is evident that the daily variation of lightning during summer presents a clear maximum at 1400 UTC, and this is common for both networks. Over land the largest number of lightning occurs from 1000 to 1900 UTC (84% and 72% of daily records over land of ZEUS, and ATD respectively). A similar behavior for the diurnal flash distribution during the warm season is reported for other areas, e.g., Spain [Soriano et al., 2001], Greece [Mazarakis et al., 2009]. Over the sea, the two time series present a common maximum at 0900 UTC. Analysis of the temporal distribution of cloud-to-ground lightning over the western Mediterranean Sea performed by Soriano and Pablo  for the whole year showed that the diurnal cycle has also a maximum during the morning hours but around two hours earlier. This difference can be explained by the fact that the authors have analyzed data from only the western part of the Mediterranean and for a whole year.
 From Figure 3 some characteristics about the detection efficiency of the two networks can be pointed out. Namely, ZEUS system clearly underestimates the nighttime and early morning activity (from 2000 UTC up to 0300 UTC) both over land and sea. Thunderstorm activity over remote areas as South America during their local day hours may produce stronger sferics than the flashes over Europe leading thus to underestimation of the number of flashes detected over Europe during the night hours. During the rest of the day ZEUS system detects increased number of lightning compared to ATD and this difference is maximized at the hour of maximum activity. This feature can be attributed to the difference in detection efficiency of the two systems. Indeed, the detection efficiency of the ATD system varies from 20 to 90% in Europe during summer with the low detection efficiency percentages attributed to the saturation of recording which has an upper limit of 12,000 flashes per hour [Keogh et al., 2006]. ZEUS system is capable of capturing up to 70 flashes per second that is more than 200,000 flashes per hour [Chronis and Anagnostou, 2006].
3.2. Relationship of the Lightning Distribution With Land Surface Properties
 The data from ZEUS system recorded during the period June 2006–May 2007 (1 year) are analyzed in this section. It is reminded that the area of interest (Figure 1) encounters almost equal surfaces of land and sea areas. Figure 4 shows the monthly distribution of the total number of flashes expressed as percentage (solid bars in Figure 4). It is obvious that the percentage of yearly lightning is increased during summer, with the peak on July when 25% of the yearly number of lightning have been observed. On the contrary, the lightning activity is minimized during January and February when only 1% of the yearly number of lightning have been detected. On the same figure the percentage of land flashes compared to sea flashes for each month are given (open and shaded bars in Figure 4). From October through March most of the flashes are detected over the sea with the percentage of sea flashes over the total monthly values varying from 66% to 82%. The percentage of land flashes over the total monthly values increases from April up to July with percentages greater than 70% and is suppressed during the rest of the year.
 The results shown in Figure 4 support the findings published in the literature that the difference in the thermal properties of land compared to sea surfaces explain the much greater number of lightning observed over land than over the sea [Williams and Stanfill, 2002]. Moreover, the excessive heating from sunlight over land during the warm period of the year also explains the early afternoon maximum of lightning activity (shown in Figure 3). Concerning the early morning maximum of lightning activity over the sea, shown also in Figure 3, the reason is not clear. Some authors [e.g., Altaratz et al., 2003] attribute this feature to land breeze development near the coast lines, which is more developed during the morning when the temperature difference between land and sea is maximized, but this issue needs further investigation.
 As aforementioned (section 1) the number of studies that discuss the relationship between lightning and surface characteristics such as elevation, slope and vegetation is quite limited. For that reason, in this section the 1-year ZEUS lightning data that occurred over land surfaces have been distributed in bins of elevation, terrain slope and vegetation type. As the analyzed period includes only one year, the lightning data have been grouped in seasons of three months, namely December-January-February (DJF), March-April-May (MAM), June-July-August (JJA) and September-October-November (SON).
 The reason for the investigation of a possible relationship between lightning activity and elevation is the known role of orographic lifting to the enhancement of convection and consequently to lightning. Figure 5 shows the cumulative frequency percentage of the number of land points and the number of lightning for various bins of elevation within the studied domain for each season of the year. During winter (DJF) which is the period with the lowest percentage of activity over land, most of the lightning occurs over the lowest elevation bins (Figure 5). Indeed, while the 53% of land points have elevation heights less than 400 m the percentage of lightning observed in these areas is ∼70%. This behavior could be expected as during this period of the year most of the convective activity develops within organized low pressure systems and accompanied frontal activity and thus the orographic forcing is less important. Enhancement of convection within frontal thunderstorms is not related with the elevation itself as enhancement of precipitation has been reported in various studies, midway between the mountain peaks and the shore due to forced preorographic lifting (e.g., Ogura et al.  over Japan, Kotroni et al.  over Greece, and Kotroni et al.  over Turkey). A similar behavior is evident during autumn (SON) but to a lesser degree as this is a transient period between summer and winter when convection maybe also associated with localized thunderstorms (during the beginning of the period) as well as with frontal thunderstorms. During spring (MAM) and summer (JJA), when 81% of the yearly lightning over land has been observed, it seems that enhanced lightning activity is associated with higher elevation.
Figure 6 shows the cumulative frequency percentage of the number of land points and the number of lightning for various bins of terrain slope within the studied domain. From this graph it is evident that the cumulative percentages of the number of lightning are well below the cumulative percentage of the number of land points for the all slope bins during spring, summer and autumn, indicating that sloping terrain is a contributing factor to increased lightning activity. For example, during summer over the 0 slope points that represent 36% of the land points only 20% of the lightning is observed while over the 25% of the land points that have slopes greater than 5%, the corresponding lightning is 35%. During winter (DJF) the cumulative percentages of the number of lightning almost coincides with the cumulative distribution of the land points, showing that winter time lightning activity is not influenced by sloping terrain. Comparison of Figures 5 and 6 shows that spring but mainly summer the lightning activity is much more positively related with the terrain slope, than it is with the terrain height, through enhancement of vertical air motions which are vital to deep convection.
 In order to investigate the relationship of lightning activity with the vegetation type, the flash rate (number of flashes/km2/season) per vegetation category has been calculated for each season of the studied period within the studied domain shown in Figure 1. As the number of flashes over land presents an important seasonal variation (with the warm period lightning number being one order of magnitude larger than that during the cold period of the year) and also because the total area of each vegetation category varies also a lot, the above mentioned flash rates have been scaled with the total number of lightning recorded each season and by the total land surface of the studied domain. The resulting numbers could be viewed as the “potential” of each vegetation category in the “production” of lightning per month and will be referred to as “lightning yield” in the following (Table 1). In that case, a value of ∼1 would mean that the percentage of seasonal lightning over the areas of this category is equal to the percentage of the area of this category to the total area, and thus there is “no preference” of lightning occurrence for this vegetation category. Inspection of this table shows the following:
Table 1. Seasonal “Lightning Yield” per Vegetation Categorya
Evergreen Needleaf Forest
Deciduous Broadleaf Forest
For details see in the text.
 1. Over the forest areas the lightning “yield” is quite low (less than 1), during autumn and winter, while it is quite elevated during spring and especially summer.
 2. Over the woodland and wooded grassland areas the lightning “yield” is important (>1) during almost the whole year.
 3. Over the shrubland areas the lightning “yield” is important, during autumn and winter, while it is suppressed during spring and especially summer.
 4. Over the cropland areas the lightning “yield” is important (>1.3) during spring and summer.
 5. Over the bare ground surfaces the lightning “yield” is very low during the whole year, with the lowest values during summer.
 The relation of vegetation type with lightning activity could be viewed in the light of the Mediterranean climate but also of the aforementioned results about the relation of lightning and terrain. During the cold period of the year the forested areas have decreased lightning “yield” as during this period convection over land mainly takes place near the shore and at lower elevation areas that usually are less forested. During the warm period of the year the Mediterranean surfaces are quite dry due to decrease of precipitation. The areas that keep soil moisture during the dry period are the forested and wooded areas and this could explain the increased lightning “yield” during the dry period of the year over these areas. In addition, frictional lift may play a role in increasing storm initiation and lightning activity over forested areas that may act as a barrier to the synoptic or mesoscale wind flow, enhancing upward motion on the windward side of the forest. At this point one could ask if it is not the coverage of an area with forests that is related with more lightning activity during summer, but the fact that during this period the higher elevations (that are also more forested) are related with the increased lightning activity. For that reason the lightning “yield” of the forested areas for various bins of elevations has been calculated (not shown). It was found that for all elevation bins the forested areas during summer present very similar lightning “yield” values supporting thus the idea of positive “preference” of lightning occurrence over the forested areas during the warm period of the year.
 Over bareground the lightning yield is low over the whole year and especially during the warm and dry period due to excessive dryness. At this point it should be noted that most of the bareground areas in the studied domain are found over Northern Africa coasts where convective activity is limited during spring and almost suppressed during summer.
4. Concluding Remarks
 In the frame of this paper the possible relationship between lightning and elevation, terrain slope and vegetation was explored. For that reason, 1-year data (June 2006 up to May 2007) from the experimental long-range lightning detection network ZEUS operated by the National Observatory of Athens, were used.
 As ZEUS is an experimental network, at a first stage, a comparison of its data with those of the established long-range lightning detection network operated by the UK Met Office for the summer period (June-July-August 2006) has been undertaken in order to reveal its advantages or shortcomings. It was found that ZEUS clearly underestimates the nighttime and early morning activity both over land and sea. During the rest of the day ZEUS detects increased number of lightning compared to ATD that was attributed to the increased detection efficiency of ZEUS compared to ATD.
 The investigation of the possible relationship between lightning and elevation, terrain slope and vegetation has shown the following: (1) during spring and summer there is a positive relationship of lightning activity with elevation while this feature is not evident during the rest of the year and especially during winter; (2) the lightning activity was found to be positively correlated with the terrain slope throughout the year except winter; (3) over bareground the lightning “yield” is low during the whole year while the inverse is true for woodland areas. During the warm period of the year, due to drying of the Mediterranean surfaces, the forested and wooded areas that keep soil moisture present increased lightning “yield” in contrast with the shrubland areas that for the same period present a decreased lightning yield.
 The authors are conscious that the analyzed period is quite short in order to derive definitive conclusions and the results of this study should be considered as a first indication on the lightning activity preferences in relation with the topographic features in the Mediterranean. It is in the authors' plans to extend this study when a larger database of ZEUS data will be available.
 This work has been partially supported by the Greek–non-EU countries cooperation program, financed by the Greek General Secretariat for Research and Technology, and by the EU financed project FLASH (contract 036852). I. Koletsis is thanked for his support on the ZEUS database. The UK Met Office is thanked for provision of ATD data.