JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. D21, 4597, doi:10.1029/2002JD002211, 2002 Maritime cloud-to-ground lightning: The western Mediterranean Sea Luis Rivas Soriano and Fernando de Pablo Departamento de Fı́sica de la Atmósfera, Universidad de Salamanca, Salamanca, Spain Received 18 February 2002; revised 20 May 2002; accepted 22 May 2002; published 14 November 2002. [1] For the first time, the temporal and spatial distribution of cloud-to-ground lightning over the western Mediterranean Sea have been studied. The monthly variation shows a single peak in September and October while minimum values are observed in January to March. The diurnal cycle shows a maximum on 0600–0700 LT and a minimum around 1600–1700 LT. The percentage of positive flashes is 8%, although it is higher in the winter than in the summer. The average multiplicity is 2.1 for the negative flashes and 1.1 for the positive flashes, and the percentage of single-stroke flashes is 48% and 91%, respectively. The annual distribution of multiplicity reveals a maximum in the summer and a minimum in the winter for the negative flashes. The median intensity was found to be 27.5 kA for the negative flashes. Positive flashes show maximum intensities in April and minima in August. The average maximum flash density is 1.5 flashes km 2 yr 1. The geographical distribution of density shows that the highest values are associated with the coastline and decrease eastward. The lightning activity over the island of Mallorca shows characteristics associated with both the land and the sea. The results are discussed in the context of the measurements taken for the land area of the Iberian Peninsula and for INDEX TERMS: 3304 Meteorology and Atmospheric Dynamics: Atmospheric other oceanic areas. electricity; 3314 Meteorology and Atmospheric Dynamics: Convective processes; 3324 Meteorology and Atmospheric Dynamics: Lightning; KEYWORDS: maritime lightning, Mediterranean Sea Citation: Rivas Soriano, L., and F. de Pablo, Maritime cloud-to-ground lightning: The western Mediterranean Sea, J. Geophys. Res., 107(D21), 4597, doi:10.1029/2002JD002211, 2002. 1. Introduction [2] Lightning is related to a number of relevant meteorological issues. For example, it may be used as an indicator of changes in global temperature [Williams, 1992, 1994; Reeve and Toumi, 1999], it is related to convective precipitation [e.g., Petersen and Rutledge, 1998; Seity et al., 2001; Rivas Soriano et al., 2001b] and sea surface temperature [de Pablo and Rivas Soriano, 2002], and it is a source of nitrogen oxide and hence tropospheric ozone [e.g., Toumi et al., 1996; Stephen et al., 2000]. It is therefore important to gain detailed knowledge of lightning activity both over land and sea. Lightning characteristics over land are well documented [e.g., Orville, 1991, 1994; Petersen and Rutledge, 1992; Finke and Hauf, 1996; Orville and Silver, 1997; Yair et al., 1998; I. Pinto et al., 1999; O. Pinto et al., 1999; Orville and Huffines, 1999; Hotl et al., 2001; Rivas Soriano et al., 2001a; Zajac and Rutledge, 2001]. However, much less attention has been focused on lightning over the sea, perhaps because it is over land where lightning activity mainly occurs and is sensitive to deep convection [Mackerras et al., 1998]. Studies of lightning over the sea are basically restricted to tropical oceans (e.g., Lucas and Orville [1996] and Orville et al. [1997] for the TOGACOARE area; Hidayat and Ishii [1998] around Java) and Copyright 2002 by the American Geophysical Union. 0148-0227/02/2002JD002211 ACL there are few works analyzing lightning over midlatitude oceanic areas. Seity et al. [2001], e.g., studied lightning activity over the French Atlantic coast of the Bay Biscay. [3] In the present study we report an analysis of cloud-toground (CG) lightning activity over the western Mediterranean Sea, including the island of Mallorca. The information offered here constitutes one of the few contributions about lightning activity over a midlatitude water mass. It is worth mentioning here that Martin and Carretero [2001] suggested that the Mediterranean basin resembles a tropical and warm sea. [4] The paper is organized as follows. In section 2, we give a description of the lightning data analysis. In section 3, CG lightning activity over the western Mediterranean Sea is analyzed, and in section 4 CG lightning over the island of Mallorca is discussed. Conclusions are given in section 5. 2. Lightning Data [5] CG lightning data were obtained from the lightning detection network installed on the Spanish territory and belonging to the Instituto Nacional de Meteorologia. It uses ALDF (Advanced Lightning Direction Finder) model 141-T, manufactured by Lightning Location and Protection Inc. Information about this lightning detection network can be found in Rivas Soriano et al. [2001a]. The lightning sensors located on the Mediterranean coast of the Iberian Peninsula 15 - 1 ACL 15 - 2 RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA Figure 1. Area of study and locations of the ALDF sensors (indicated by crosses). and in the Balearic islands are shown in Figure 1. The detection efficiency of the lightning network has not been yet estimated and hence no attempt was made here to correct for detection efficiency and the lightning data are just the measured values. This may artificially reduce flash density on the eastern limit of the area considered in this analysis, as will be discussed in the section 3. [6] In the present study we considered the water mass of the region located between 37°N and 43°N and that between 0° and 6°E, which corresponds to the western side of the Mediterranean Sea. Only the CG lightning data for three years (1992 to 1994) were available. Thus, all data and results are related to this 3-years period. [7] The western Mediterranean Sea includes the Balearic Islands (see Figure 1). The island area is so small compared to that of water mass than the results with and without the Balearic islands were identical. Accordingly, the results for the western Mediterranean Sea discussed in the section 3 include these islands. However, we analyzed CG lightning over the island of Mallorca separately. [8] To study the temporal distribution of flash characteristics, we considered time integrals over the year, month and hour and integrated over the complete area. To analyze the geographical distribution of flash density, the lightning data were referenced to grid blocks of 0.2° longitude  0.2° latitude (368 km2) for the complete area and 0.1° longitude 0.1° latitude (92 km2) for the island of Mallorca. 3. Cloud-to-Ground Lightning in the Western Mediterranean Sea 3.1. Annual and Diurnal Cycles [9] The monthly variation in the number of CG flashes (combined positive and negative) is shown in Figure 2a. Maximum CG lightning activity was observed at the beginning of the autumn (60% of all lightning flashes were counted in September and October) and the minimum lightning occurred in the winter season (6% of all lightning flashes from January to March). This annual distribution is equivalent to those reported in the western Pacific Ocean- RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA ACL 15 - 3 Figure 2. Monthly variation for the 1992 –1994 average in (a) the number of cloud-to-ground flashes in the western Mediterranean Sea and (b) sea surface temperature corresponding to 37.5°N–0.5°E. A ‘‘month’’ refers to four consecutive periods of seven days (i.e., 13 ‘‘months’’ per year). TOGA-COARE area- [Orville et al., 1997] and the Indian Ocean [Hidayat and Ishii, 1998], where maximum CG lightning activity was found in the months January and February and minimum activity occurred in August and September. However, it is different from the values reported over land (the Iberian Peninsula), where Rivas Soriano et al. [2001a] reported maximum CG lightning activity in the summer and minimum activity in the winter. Figure 2b shows the annual distribution of the sea surface temperature (SST) (averaged over the period 1992 – 1994) corresponding to the point 37.5°N–0.5°E. The SST data were obtained from the NCEP/NCAR 40-year Reanalysis Project [Kalnay et al., 1996] (this information is available on the Internet at http://www.nic.fb4.noaa.gov). The highest SST value is reached in September. Accordingly, maximum CG lightning activity appears when the thermodynamic atmospheric background over the Mediterranean Sea is more appropriate for convection. [10] The diurnal cycle of CG lightning activity (combined number of positive and negative flashes) is depicted in Figure 3. It shows a broad peak around 0600 –0700 local time (LT) and a minimum between 1600 – 1700 LT, reflecting the typical nocturnal maximum in oceanic convection. This diurnal cycle is opposite that reported for the Iberian Peninsula [Rivas Soriano et al., 2001a], where the maximum is observed at 1700 LT and the minimum between 0900 and 1100 LT, although it is close to the annual cycle reported for the TOGA-COARE area [Orville et al., 1997] ACL 15 - 4 RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA Figure 3. Diurnal variation in the number of cloud-to-ground flashes in the western Mediterranean Sea for the 1992 – 1994 average. and the Indian Ocean [Hidayat and Ishii, 1998], where the maximum occurs at about 0200 LT and the minimum between 1000 –1200 LT. The ratio of the highest to the lowest CG flash counts over the western Mediterranean Sea is 2.3, much lower than the value 4.1 found for the Iberian Peninsula. This shows that the diurnal variation of CG lightning over land is stronger than over sea, in agreement with the findings of Lucas and Orville [1996] who compared the diurnal cycles of CG lightning in the continental United States and in the TOGA-COARE area. For the diurnal cycle over land it has been reported that the rise to the maximum is steeper than the fall from it [Finke and Hauf, 1996; Rivas Soriano et al., 2001a], owing to the diurnal cycle of the atmospheric boundary layer. However, this is not seen for the diurnal cycle over the sea, showing that initiation, evolution and dissipation of maritime storms present similar scatter of time scales. 3.2. Polarity, Multiplicity, and Intensity [11] Figure 4 shows the percentage of positive and negative flashes for each month. 8% of all detected CG flashes were positive. This number coincides with the value Figure 4. Monthly variation in the proportions of negative and positive flashes in the western Mediterranean Sea for 1992– 1994. RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA ACL 15 - 5 Figure 5. Distributions of (a) number of strokes per flash and (b) monthly mean number of strokes in a flash in the western Mediterranean Sea for 1992 – 1994. found by Rivas Soriano et al. [2001a] for the Iberian Peninsula and is lower than 5.6% reported by Orville et al. [1997] for the TOGA-COARE region. The contribution of positive flashes is higher during the winter than in the summer, in agreement with the result reported for the Iberian Peninsula [Rivas Soriano et al., 2001a], and shows that thunderstorms in midlatitude seas occurring in the colder months also mostly take place in sheared environments. It is remarkable that on comparing the CG lightning polarity over land and sea in midlatitude and tropical areas equivalent results are found. For example, Orville et al. [1997] found that the percentage of positive flashes is not a function of the month in the TOGA-COARE area, in agreement with the seasonal variation of polarity reported for Brazil [I. Pinto et al., 1999], and Seity et al. [2001] reported that positive CG flash percentages do not show any differences between land and sea for the French Atlantic coast. [12] The distribution of the number of strokes per flash is depicted in Figure 5a. The average values are 2.1 for the negative flashes and 1.1 for the positive flashes. These numbers are almost the same as those reported for the Iberian Peninsula: 2.0 and 1.1 respectively [Rivas Soriano et al., 2001a]. The percentage of single-stroke flashes is 48% for the negative flashes and 91% for the positive flashes, slightly lower than the values reported for the Iberian Peninsula (51% and 93% respectively). The multiplicity found in the western Mediterranean Sea differs from that found by Orville et al. [1997] in the TOGA-COARE area: average multiplicity of 2.3 for the negative flashes and percentage of single-stroke flashes of 40%, although it is close to the values reported by Seity et al. [2001] for the French Atlantic coast: 2.4 for positive flashes and 1.1 for negative flashes. [13] The monthly distribution of the number of strokes in a flash is depicted in Figure 5b. It exhibits a peak in the summer season and a minimum in the winter season for negative flashes. The multiplicity of positive flashes does not seem to be a function of the month. These results are in agreement with the findings of Rivas Soriano et al. [2001a] for the Iberian Peninsula, but are different from the values reported by Orville et al. [1997] in the TOGA-COARE area, ACL 15 - 6 RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA Figure 6. Distribution of (a) first stroke peak currents and (b) monthly median first stroke peak currents in the western Mediterranean Sea for 1992 – 1994. where the largest variation in multiplicity was found in the months where there was less lightning. [14] The distribution of peak current amplitudes for negative and positive flashes is depicted in Figure 6a. It is similar to that found in the Iberian Peninsula for both polarities [Rivas Soriano et al., 2001a], showing that the decay to large amplitudes is slower for the positive flashes. Median (mean) peak current amplitude is found at 27.5 kA (39.7 kA) for the negative flashes and 47.4 kA (74.1 kA) for the positive flashes. The median and mean peak current amplitudes for negative flashes are therefore close to the results of Orville et al. [1997] for the TOGACOARE area (25 kA) and those of Seity et al. [2001] for the French Atlantic coast (32.1 kA) respectively. However, the median and mean peak current amplitudes for positive flashes are much higher than the values reported by the above authors (33 kA and 59.1 kA respectively). This latter may be due to the low efficiency of the ALDF in detecting weak amplitude flashes, as discussed by Rivas Soriano et al. [2001a]. The network geometry and large station separation could also be a contributing factor. Seity et al. [2001] found that mean peak currents of CG flashes are higher over sea than over land for both polarities. The same result is seen when the mean peak currents found in the western Mediterranean Sea are compared with those reported for the Iberian Peninsula (32.4 kA for negative flashes and 69.3 kA for positive flashes) [Rivas Soriano et al., 2001a]. [15] Multiplicity is higher for negative flashes than for positive ones. This difference is opposite that observed for the peak current. This contrast is consistent, because higher peak currents transfer more charge and smaller stroke numbers are associated with the opposite effect. It could be speculated that the same effect should be found for CG RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA ACL 15 - 7 Figure 7. Geographical distribution of flash density in the western Mediterranean Sea for the 1992 – 1994 average. flashes over the sea and land. However, the multiplicity over the western Mediterranean Sea and the Iberian Peninsula is nearly the same, although peak currents are higher over the sea. Peak current amplitude is 100% higher for positive flashes than for negative ones, both over the western Mediterranean Sea and the Iberian Peninsula, but it is only 15% higher over the sea than over the land, insufficient to produce a clear statistical effect on multiplicity. [16] Figure 6b shows the monthly variation of median peak current amplitudes for both polarities. The negative flashes show little variation along the year. The positive flashes present minimum values in August and September and the maximum is found in April. These results differ from those observed over land, since for the Iberian Peninsula, for both polarities, Rivas Soriano et al. [2001a] found that peak current is higher in the summer, when CG lightning is at a maximum. However, the annual distribution of peak current amplitudes found in the western Mediterranean Sea is consistent with the findings of Orville et al. [1997], who for the region of TOGA-COARE reported little monthly variation of median peak current for negative flashes, and maxima (minima) in May (October) for positive flashes. It should be noted that the peak current amplitudes for August and September are lower for positive flashes than for negative ones. This cannot be due to a misrepresentation of the data due to a low lightning count, because September belongs to the time of the year with the highest CG lightning over the western Mediterranean Sea. 3.3. Density [17] The pattern of the average CG flash density (combined negative and positive flashes) is shown in Figure 7. It should be noted that measured data have not been corrected for detection efficiency. Thus, the flash density is the measured value. The maximum flash density is 1.5 flashes km 2 yr 1. This value compares well with the measured value of 2.0 flashes km 2 yr 1 reported by Orville et al. [1997] for the TOGA-COARE region and with the results of Seity et al. [2001] for the French Atlantic coast, but it is lower than the maximum flash density reported by Hidayat and Ishii [1998] around Java: 3.2 flashes km 2 yr 1. The maximum flash density over the land area of the Iberian Peninsula reported by Rivas Soriano et al. [2001a] was 3.3 flashes km 2 yr 1. This value was corrected assuming 70% detection efficiency. This means that the value measured in the Iberian Peninsula was on the order of 2.5 flashes km 2 yr 1, also higher than the value found over the western Mediterranean Sea. The pattern of the spatial distribution of flash density shows that maximum lightning activity tends to be fixed along the coastline. The diurnal ACL 15 - 8 RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA Figure 8. (a) Monthly and (b) diurnal variations in the number of cloud-to-ground flashes in the island of Mallorca for the 1992 – 1994 average. Figure 9. Monthly variation in the median first stroke peak current in the island of Mallorca for 1992 – 1994. RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA ACL 15 - 9 Figure 10. (a) Geographical distribution of flash density for the 1992 – 1994 average and (b) areas of highest altitude (gray) in the island of Mallorca. distribution of flashes (Figure 3) indicates that maximum CG lightning occurs in the early morning and the highest flash densities may therefore be due to land breeze. Flash density decreases eastward. Since the eastern limit of the area considered here is outside the perimeter of the lightning detection network, we expect that the number of measured flashes decreases eastward. However, this decrease may also reflect the absence of coastal effects. The maximum flash density in the area with longitude higher than 4°E is 0.7 flashes km 2 yr 1, approximately 3.5-fold lower than the maximum flash density measured over the Iberian Peninsula, which is in agreement with the ratio between maximum flash density in the continental United States and in the TOGA-COARE area [Orville et al., 1997]. 4. Cloud-to-Ground Lightning Over the Island of Mallorca [18] Figure 8 shows the annual and diurnal distributions of the number of CG flashes (combined negative and positive) over the island of Mallorca. Since Mallorca is a relatively small island (area 3600 km2), its lightning characteristics are expected to be similar to those of the sea. In fact, from Figure 8 it may be seen that maximum CG lightning activity appears in September – October and during the nocturnal hours. However, Figure 8 also shows characteristics typical of lightning over land, as the secondary maximum in the spring season and in the period 1500– 1600 LT. Land areas near the coastline are influenced by both the land and the sea. Because time variations of CG lightning over the land and sea are almost out of phase, the duration of lightning activity is longer. This was also reported by Hidayat and Ishii [1998] for the island of Java. [19] In the section 4, it was stated that the polarity, multiplicity and intensity of CG lightning flashes over the western Mediterranean Sea were similar to those measured over land (Iberian Peninsula). Accordingly, these flash characteristics over the island of Mallorca (not shown here) are similar to the values measured over the sea. The sole ACL 15 - 10 RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA difference is seen in the monthly variation of median peak current amplitudes for positive flashes (Figure 9). As expected, their annual distribution shows characteristics associated with both land (maximum in July) and sea (minimum in August). Interestingly, on comparing Figures 6b and 9, it is seen that the annual distribution of median peak current for the negative flashes only shows characteristics associated with lightning over sea. [20] The spatial distribution of the average flash density in the island of Mallorca is depicted in Figure 10a. Maximum flash density is 1.2 flashes km 2 yr 1. The pattern of flash density shows that maximum values appear near to coastline, as expected. However, the influence of orography reported over land areas [e.g., Finke and Hauf, 1996; Rivas Soriano et al., 2001a] is also clear. This is seen on comparing Figures 10a and 10b. 5. Conclusions [21] CG lightning activity over the western Mediterranean Sea and over the island of Mallorca has been analyzed. Four characteristics were studied: number of flashes, polarity, multiplicity and peak current amplitude. Temporal and geographical analyses reveal the following: 1. The monthly variation shows a single peak, with maximum activity in the September and October (60% of all lightning events) and minimum in the winter season (6% from January to March). This annual distribution is consistent with the annual variation in SST. 2. The diurnal variation also shows a single peak, with maximum activity being observed around 0600 –0700 LT and minimum between 1600 – 1700 LT. This diurnal cycle is opposed to that found for the land area of the Iberian Peninsula. Also, the diurnal variation over the Iberian Peninsula is stronger than over the western Mediterranean Sea. 3. The average percentage of positive flashes is 8%. The contribution of positive flashes is higher during the winter. These results coincide with those reported for the Iberian Peninsula. 4. The average multiplicity is 2.1 strokes per flash for negative flashes and 1.1 for positive flashes. The percentage of single-stroke flashes was found to be 48% for negative flashes and 91% for negative flashes. The monthly distribution of multiplicity reveals the existence of a peak in the summer and minimum values in the winter for negative flashes. Multiplicity of positive flashes does not seem to be a function of the month. These results are also in agreement with those found in the Iberian Peninsula. 5. The distribution of peak current amplitudes in the western Mediterranean Sea also agrees with the result reported for the Iberian Peninsula. The median (mean) peak current is 27.5 kA (39.7 kA) for the negative flashes and 47.4 kA (74.1 kA) for the positive flashes. Mean peak current is higher over sea than over land. The annual distribution shows that median peak amplitudes for negative flashes show little variation along the year. This result is different from that found in the Iberian Peninsula. 6. Maximum flash density is 1.5 flashes km 2 yr 1, which is lower than the value found for the Iberian Peninsula. The spatial pattern shows maximum values along the coastal zones and the flash density decreases eastward. 7. Polarity, multiplicity and intensity over the island of Mallorca are similar to those for the whole of the western Mediterranean Sea. The diurnal and annual cycles and the geographical distribution of flash density over Mallorca show characteristics associated with both land and sea. 8. The properties of the CG lightning over the western Mediterranean Sea are consistent with those reported for other oceanic areas. [22] Acknowledgments. This work was funded by the Junta de Castilla y Leon and the European Union (grant SA011/01). References de Pablo, F., and L. Rivas Soriano, Relationship between cloud-to-ground lightning flashes over the Iberian Peninsula and sea surface temperature, Q. J. R. Meteorol. Soc., 128, 173 – 183, 2002. Finke, U., and T. Hauf, The characteristics of lightning occurrence in southern Germany, Beitr. Phys. Atmos., 69, 361 – 374, 1996. Hidayat, S., and M. Ishii, Spatial and temporal distribution of lightning activity around Java, J. Geophys. Res., 103, 14,001 – 14,009, 1998. Holt, M. A., P. J. Hardaker, and G. P. McLelland, A lightning climatology for Europe and UK, 1990 – 99, Weather, 56, 290 – 296, 2001. Kalnay, E., et al., The NCEP/NCAR 40-year reanalysis project, Bull. Am. Meteorol. Soc., 77, 437 – 471, 1996. Lucas, C., and R. Orville, TOGA COARE: Oceanic lightning, Mon. Weather Rev., 124, 2077 – 2082, 1996. Mackerras, D., M. Darveniza, R. Orville, E. Williams, and S. Goodman, Global lightning: Total, cloud and ground flash estimates, J. Geophys. Res., 103, 19,791 – 19,810, 1998. Martin, F. and O. Carretero, Tropical-like heavy convective rains over the Spanish Mediterranean regions: A radar-based perspective, paper presented at 30th International Conference on Radar Meteorology, Am. Meteorol. Soc., Munich, Germany, 18 – 24 July 2001. Orville, R. E., Lightning ground flash density in the contiguous United States—1989, Mon. Weather Rev., 119, 573 – 577, 1991. Orville, R. E., Cloud-to-ground lightning flash characteristics in the contiguous United States: 1989 – 1991, J. Geophys. Res., 99, 10,833 – 10,841, 1994. Orville, R. E., and G. R. Huffines, Lightning ground flash measurements over the contiguous United States: 1995 – 1997, Mon. Weather Rev., 127, 2693 – 2703, 1999. Orville, R. E., and A. C. Silver, Lightning ground flash density in the contiguous United States: 1992 – 95, Mon. Weather Rev., 125, 631 – 638, 1997. Orville, R. E., E. J. Zisper, M. Brook, C. Weidman, G. Aulich, E. P. Krider, H. Christian, S. Goodman, R. Blakeslee, and K. Cummins, Lightning in the region of the TOGA COARE, Bull. Am. Meteorol. Soc., 73, 3 – 16, 1997. Petersen, W. A., and S. A. Rutledge, Some characteristics of cloud-toground lightning in tropical northern Australia, J. Geophys. Res., 97, 11,553 – 11,560, 1992. Petersen, W. A., and S. A. Rutledge, On the relationships between cloud-toground lightning and convective rainfall, J. Geophys. Res., 103, 19,777 – 19,790, 1998. Pinto, I. R. C. A., O. Pinto Jr., R. M. L. Rocha, J. H. Diniz, A. M. Carvalho, and A. Cazetta Filho, Cloud-to-ground lightning in southeastern Brazil in 1993, 2, Time variations and flash characteristics, J. Geophys. Res., 104, 31,381 – 31,387, 1999. Pinto, O., Jr., I. R. C. A. Pinto, M. A. S. S. Gomes, I. Vitorello, A. L. Padilha, J. H. Diniz, A. M. Carvalho, and A. Cazetta Filho, Cloud-toground lightning in southeastern Brazil in 1993, 1, Geographical distribution, J. Geophys. Res., 104, 31,369 – 31,379, 1999. Reeve, N., and R. Toumi, Lightning activity as an indicator of climate change, Q. J. R. Meteorol. Soc., 125, 893 – 903, 1999. Rivas Soriano, L., F. de Pablo, and E. Garcı́a Dı́ez, Cloud-to-ground lightning activity in the Iberian Peninsula: 1992 – 1994, J. Geophys. Res., 106, 11,891 – 11,901, 2001a. Rivas Soriano, L., F. de Pablo, and E. Garcı́a Dı́ez, Relationship between convective precipitation and cloud-to-ground lightning in the Iberian Peninsula, Mon. Weather Rev., 129, 2998 – 3003, 2001b. Seity, Y., S. Soula, and H. Sauvageot, Lightning and precipitation relationship in coastal thunderstorms, J. Geophys. Res., 106, 22,801 – 22,816, 2001. Stephen, W., W. Nesbitt, R. Zhang, and R. Orville, Seasonal and global NOx production by lightning estimated from the Optical Transient Detector (OTD), Tellus, Ser. B, 52, 1206 – 1215, 2000. RIVAS SORIANO AND DE PABLO: LIGHTNING OVER THE MEDITERRANEAN SEA Toumi, R., J. Haigh, and K. S. Law, A tropospheric ozone lightning climate feedback, Geophys. Res. Lett., 23, 1037 – 1040, 1996. Williams, E. R., The Schuman resonance: A global tropical thermometer, Science, 256, 1184 – 1187, 1992. Williams, E. R., Global circuit response to seasonal variations in global surface air temperature, Mon. Weather Rev., 112, 1917 – 1929, 1994. Yair, Y., Z. Levin, and O. Altaratz, Lightning phenomenology in the TelAviv area from 1989 to 1996, J. Geophys. Res., 103, 9015 – 9025, 1998. ACL 15 - 11 Zajac, B. A., and S. Rutledge, Cloud-to-ground in the contiguous United States from 1995 to 1999, Mon. Weather Rev., 129, 999 – 1019, 2001. F. de Pablo and L. Rivas Soriano, Departamento de Fı́sica de la Atmósfera, Facultad de Ciencias, Universidad de Salamanca, Pl. de la Merced s/n, E-37008 Salamanca, Spain. (ljrs@gugu.usal.es)