3.1 Optimal radius for sampling WWLLN data
The estimates of optimal radius (rounded values) and TS index obtained by two methods for all 34 weather stations in the Far East, including those previously published in (Kleshcheva et al. 2021), are summarized in Table 1, and Table 2 presents statistical characteristics. Table 2 also separately provides statistics for mainland, coastal, and island (Sakhalin) stations.
Table 1
Results of a comparison the data of thunderstorm observations at weather station and WWLLN data for the period of 2009–2018
Station and it’s localization
|
Coordinates
|
Station
Altitude
|
1st Method
|
2nd Method
|
φ,oN.
|
λ,o E.
|
h, m
|
R1,
km
|
TSI
|
Day
|
Night
|
R2, km
|
Day
|
Night
|
R1,km
|
TSI
|
R1, km
|
TSI
|
R2, km
|
R2, km
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
8
|
9
|
10
|
11
|
12
|
13
|
Mainland
|
1. Nizhnetambovskoe (KK)
|
50.92
|
138.18
|
24
|
39
|
0.60
|
27
|
0.62
|
39
|
0.51
|
29
|
26
|
39
|
2. Cekunda (KK)
|
50.87
|
132.25
|
271
|
32
|
0.69
|
28
|
0.66
|
30
|
0.59
|
34
|
33
|
32
|
3. Blagovescensk (AO)
|
50.28
|
127.48
|
168
|
30
|
0.64
|
26
|
0.61
|
29
|
0.50
|
29
|
27
|
39
|
4. Sutur (KK)
|
50.07
|
132.12
|
343
|
21
|
0.64
|
21
|
0.63
|
28
|
0.46
|
20
|
19
|
31
|
5. Konstantinovka (AO)*
|
49.62
|
127.97
|
117
|
29
|
0.63
|
27
|
0.66
|
50
|
0.54
|
36
|
27
|
51
|
6. Arhara (AO)*
|
49.42
|
130.08
|
134
|
22
|
0.62
|
22
|
0.58
|
18
|
0.52
|
25
|
23
|
23
|
7. Solekul’(KK)*
|
49.17
|
138.05
|
915
|
25
|
0.58
|
25
|
0.59
|
22
|
0.26
|
19
|
19
|
14
|
8. Elabuga (KK)*
|
48.82
|
135.88
|
62
|
28
|
0.61
|
26
|
0.58
|
36
|
0.51
|
26
|
23
|
28
|
9. Smidovich (JAO)*
|
48.6
|
133.83
|
50
|
23
|
0.58
|
23
|
0.59
|
34
|
0.47
|
26
|
21
|
31
|
10. Habarovsk (KK)*
|
48.52
|
135.12
|
88
|
23
|
0.63
|
17
|
0.64
|
25
|
0.51
|
22
|
15
|
28
|
11. Ekaterino-Nikol’skoe (JAO)*
|
47.73
|
130.97
|
71
|
13
|
0.56
|
13
|
0.52
|
16
|
0.44
|
14
|
13
|
12
|
12. Lermontovka (KK)*
|
47.15
|
134.33
|
75
|
29
|
0.67
|
26
|
0.66
|
38
|
0.52
|
34
|
28
|
44
|
13. Krasnyj Jar (PK)*
|
46.53
|
135.3
|
128
|
14
|
0.49
|
14
|
0.46
|
14
|
0.39
|
12
|
12
|
10
|
14. Dal’nerechensk (PK)*
|
45.87
|
133.73
|
100
|
27
|
0.59
|
27
|
0.56
|
37
|
0.49
|
27
|
24
|
33
|
15. Mel’nichnoe (PK)*
|
45.45
|
135.5
|
331
|
25
|
0.62
|
25
|
0.61
|
21
|
0.44
|
24
|
24
|
22
|
16. Svijagino (PK)*
|
44.8
|
133.1
|
99
|
19
|
0.57
|
17
|
0.55
|
21
|
0.50
|
19
|
16
|
25
|
17. Pogranichnyj (PK)*
|
44.4
|
131.38
|
217
|
24
|
0.67
|
22
|
0.69
|
28
|
0.48
|
26
|
24
|
27
|
18. Timirjazevskij (PK)*
|
43.88
|
131.97
|
34
|
24
|
0.60
|
24
|
0.62
|
40
|
0.44
|
23
|
21
|
31
|
Coastal
|
19. Sovetskaja Gavan’(KK)*
|
49
|
140.3
|
21
|
31
|
0.63
|
33
|
0.66
|
24
|
0.49
|
31
|
33
|
28
|
20. Ternej (PK)*
|
45
|
136.6
|
51
|
30
|
0.56
|
30
|
0.59
|
23
|
0.43
|
26
|
22
|
24
|
21. Rudnaja Pristan’(PK)*
|
44.4
|
135.9
|
27
|
22
|
0.46
|
24
|
0.41
|
22
|
0.44
|
20
|
20
|
21
|
22. Vladivostok (PK)*
|
43.1
|
131.87
|
187
|
20
|
0.51
|
15
|
0.52
|
27
|
0.41
|
18
|
17
|
17
|
23. Preobrazhenie (PK)*
|
42.9
|
133.9
|
44
|
24
|
0.51
|
16
|
0.54
|
30
|
0.52
|
23
|
18
|
25
|
24. Pos’et (PK)*
|
42.65
|
130.8
|
41
|
18
|
0.45
|
12
|
0.45
|
27
|
0.40
|
15
|
13
|
14
|
Sakhalin Island
|
25. Aleksandrovsk-Sahalinskij
|
50.9
|
142.17
|
30
|
25
|
0.43
|
21
|
0.38
|
37
|
0.42
|
26
|
26
|
25
|
26. Tymovskoe
|
50.73
|
142.72
|
94
|
30
|
0.54
|
22
|
0.50
|
31
|
0.54
|
27
|
25
|
27
|
27. Pogranichnoe
|
50.4
|
143.77
|
6
|
8
|
0.37
|
9
|
0.35
|
12
|
0.39
|
11
|
14
|
8
|
28. Poronajsk
|
49.22
|
143.1
|
7
|
28
|
0.41
|
28
|
0.52
|
19
|
0.32
|
26
|
23
|
25
|
29. Uglegorsk
|
49.07
|
142.03
|
39
|
21
|
0.50
|
21
|
0.40
|
20
|
0.45
|
25
|
23
|
28
|
30. Mys Terpenija
|
48.65
|
144.73
|
33
|
12
|
0.49
|
9
|
0.54
|
12
|
0.38
|
10
|
12
|
8
|
31. Ilyinskiy
|
47.98
|
142.2
|
17
|
21
|
0.46
|
26
|
0.41
|
17
|
0.46
|
19
|
21
|
19
|
32. Juzhno-Sahalinsk
|
46.95
|
142.72
|
22
|
18
|
0.50
|
28
|
0.48
|
13
|
0.48
|
21
|
28
|
15
|
33. Nevel’sk
|
46.67
|
141.87
|
22
|
24
|
0.43
|
24
|
0.37
|
13
|
0.40
|
20
|
25
|
14
|
34. Mys Kril’on
|
45.9
|
142.08
|
34
|
11
|
0.37
|
6
|
0.39
|
11
|
0.33
|
5
|
5
|
5
|
Note: results marked with "*" particularly were discussed in (Kleshcheva et al. 2021). |
In general for the south of the Russian Far East the optimal radius R1 vary from 8 km (Sakhalin, Pogranichnoe) to 39 km (KK, Nizhnetambovskoe), and the optimal radius R2 range from 5 km (Sakhalin, Mys Kril’on) to 36 km (AO, Konstantinovka ) (Table 1). The average (median) values of both R1 and R2 for the entire territory were 23 (24) km (Table 2). On average, the largest optimal radius were obtained at the mainland stations, where the values of \(\overline {{R1}}\) and \(\overline {{R2}}\) were 25 km each, and the smallest ones – at the Sakhalin stations, where the value of \(\overline {{R1}}\)was 20 km and value of \(\overline {{R2}}\) was 19 km (Table 2). At the coastal stations the values of \(\overline {{R1}}\) and \(\overline {{R2}}\) were 24 and 22 (km), respectively (Table 2).
The maximum values of the TS index at the weather station, from which the optimal radius R1 was determined, show the degree of consistency between the observational data and the WWLLN data, and may reflect their quality and representativeness. For the entire south of the Russian Far East, TSI values varied from 0.37 (Sakhalin, stations Mys Kril’on and Pogranichnoe) to 0.69 (KK, station Cekunda) (Table 1) with a mean (median) value of 0.55 (0.57) (Table 2). On average, the best consistency between the observational data and WWLLN was obtained at the mainland stations (mean TSI = 0.61), lower – at coastal stations (mean TSI = 0.52), and the lowest – at Sakhalin (mean TSI = 0.45) (Table 2). The relatively low consistency between weather station data and WWLLN data on Sakhalin can be explained, firstly, by the small number of days with thunderstorms per year, which distinguishes Sakhalin from the rest of the south of the Russian Far East. As can be seen from Table 2, the average 10-years annual number of thunderstorm days registered at weather stations (\(\overline {{{\text{NM}}}}\)), for Sakhalin stations on average is 6 days/year, which is approximately two times less than at coastal stations (\(\overline {{{\text{NM}}}}\)~ 14 days/year), and almost 5 times less than at mainland stations (\(\overline {{{\text{NM}}}}\)~ 29 days/year). Thus, we can say that a thunderstorm on Sakhalin is a relatively rare occurrence. Secondly, the insular position here determines foggy, wet weather with a high frequency of low cloudiness (50–60% or more (Reference book, 1968), which, possibly, significantly worsens visibility and affects the quality of visual observations for weather.
Table 2
Statistical characteristics of optimal radius and TSI for mainland, coastal, Sakhalin and all weather stations in the south of the Russian Far East for the period of 2009–2018
Region
(stations amount)
|
Base
statistics
|
1st method
|
2nd method
|
\(\overline {{{\text{NM}}}}\), days/year
|
\(\overline {{{\text{NW1}}}}\)
days/year
|
\(\overline {{{\text{NW2}}}}\)
days/year
|
\(\overline {{{\text{NM}}}}\)-\(\overline {{{\text{NW1}}}}\)
days/year
|
\(\overline {{{\text{NM}}}}\)-\(\overline {{{\text{NW2}}}}\)
days/year
|
\(\overline {{{\text{NW1}}}}\)-\(\overline {{{\text{NW2}}}}\)
days/year
|
R1,
Km
|
TSI
|
Day
|
Night
|
R2, km
|
Day
|
Night
|
R1, km
|
TSI
|
R1, km
|
TSI
|
R2, km
|
R2,km
|
Mainland (18)
|
min
|
13
|
0.49
|
13
|
0.46
|
14
|
0.26
|
12
|
12
|
10
|
17.1
|
19.4
|
16.7
|
-6.3
|
-0.1
|
-5.1
|
max
|
39
|
0.69
|
28
|
0.69
|
50
|
0.59
|
36
|
33
|
51
|
41.5
|
35.8
|
40.2
|
6.4
|
1.3
|
6.5
|
mean
|
25
|
0.61
|
23
|
0.60
|
29
|
0.48
|
25
|
22
|
29
|
28.6
|
28.0
|
28.1
|
0.6
|
0.5
|
-0.1
|
median
|
24
|
0.62
|
24
|
0.61
|
28
|
0.50
|
25
|
23
|
30
|
27.0
|
27.9
|
26.6
|
0.4
|
0.4
|
0.3
|
σ
|
6
|
0.05
|
5
|
0.06
|
10
|
0.07
|
7
|
6
|
11
|
6.8
|
5.4
|
6.8
|
3.2
|
0.4
|
3.0
|
Coastal
Stations
(6)
|
min
|
18
|
0.45
|
12
|
0.41
|
22
|
0.40
|
15
|
13
|
14
|
11.7
|
11.7
|
11.2
|
-1.3
|
0.2
|
-0.2
|
max
|
31
|
0.63
|
33
|
0.66
|
30
|
0.52
|
31
|
33
|
28
|
15.0
|
15.2
|
14.7
|
0.7
|
1.0
|
1.7
|
mean
|
24
|
0.52
|
21
|
0.53
|
25
|
0.45
|
22
|
20
|
21
|
13.8
|
14.0
|
13.2
|
-0.2
|
0.6
|
0.8
|
median
|
23
|
0.51
|
20
|
0.53
|
25
|
0.44
|
22
|
19
|
23
|
14.6
|
14.4
|
13.8
|
0.0
|
0.5
|
0.7
|
σ
|
5
|
0.07
|
9
|
0.09
|
3
|
0.05
|
6
|
7
|
5
|
1.6
|
1.3
|
1.5
|
0.8
|
0.3
|
0.7
|
Sakhalin
(10)
|
min
|
8
|
0.37
|
6
|
0.35
|
11
|
0.32
|
5
|
5
|
5
|
2.2
|
1.3
|
2.0
|
-2.7
|
-0.1
|
-1.1
|
max
|
30
|
0.54
|
28
|
0.54
|
37
|
0.54
|
27
|
28
|
28
|
9.2
|
11.1
|
9.2
|
1.5
|
0.5
|
2.6
|
mean
|
20
|
0.45
|
19
|
0.43
|
18
|
0.42
|
19
|
20
|
17
|
6.0
|
6.3
|
5.8
|
-0.3
|
0.1
|
0.5
|
median
|
21
|
0.44
|
21
|
0.41
|
15
|
0.41
|
20
|
23
|
17
|
6.6
|
6.0
|
6.3
|
-0.5
|
0.2
|
0.5
|
σ
|
8
|
0.06
|
8
|
0.07
|
9
|
0.07
|
8
|
7
|
9
|
2.7
|
2.9
|
2.6
|
1.4
|
0.2
|
1.2
|
All region
(34)
|
min
|
8
|
0.37
|
6
|
0.35
|
11
|
0.26
|
5
|
5
|
5
|
2.2
|
1.3
|
2
|
-6.3
|
-0.1
|
-5.1
|
max
|
39
|
0.69
|
33
|
0.69
|
50
|
0.59
|
36
|
33
|
51
|
41.5
|
35.8
|
40.2
|
6.4
|
1.3
|
6.5
|
mean
|
23
|
0.55
|
21
|
0.54
|
25
|
0.45
|
23
|
21
|
24
|
19.5
|
19.1
|
19.1
|
0.2
|
0.4
|
0.2
|
median
|
24
|
0.57
|
23
|
0.56
|
24
|
0.46
|
24
|
22
|
25
|
19.5
|
19.4
|
19.1
|
0
|
0.4
|
0.5
|
σ
|
7
|
0.09
|
7
|
0.10
|
10
|
0.07
|
7
|
6
|
10
|
11.8
|
10.9
|
11.7
|
2.4
|
0.4
|
2.5
|
Table 1 also shows estimates of the optimal radius and TSI for daytime and nighttime, separated by the moments of sunrise and sunset, calculated using astronomical formulas from (Khrgian, 1978), and Table 2 shows their main statistical characteristics. On average, the entire south of the Russian Far East is characterized by a decrease in the consistency between the data of weather stations and WWLLN at night (Table 2), which, as noted in (Kleshcheva et al., 2021), is most likely due to a deterioration in the quality of visual observations, since the efficiency detection of WWLLN can be two to three times higher at night than during the day (Pessi et al., 2009). For 34 weather stations, the average TSI is 0.54 at day and 0.45 at night. The most significant changes in TSI values were obtained at mainland stations, where the difference between the average daytime (0.6) and nighttime (0.48) TSI values is 2 times greater than the standard deviation (σ) of TSI (0.06 and 0.07). For coastal weather stations, the average values of daytime and nighttime TSI were 0.53 and 0.45 (Table 2), respectively, with only at three stations (Sovetskaja Gavan’, Ternej and Vladivostok stations) decreasing of TSI more than σ. For Sakhalin stations, the average daytime and nighttime TSI values were 0.43 and 0.42, respectively (Table 2), however, a decrease in TSI at night was obtained only for Poronajsk, Mys Terpenija, and Mys Kril’on stations (Table 1). At the other Sakhalin stations, the TSI either does not change (station Juzhno-Sahalinsk) or increases insignificantly within σ (Table 1).
On average, the entire south of the Russian Far East is characterized by an increase in the optimal radius at night (Table 2). The average daytime and nighttime R1 radius were 21 and 25 (km) for all weather stations, 23 and 29 (km) for mainland stations, and 21 and 25 (km) for coastal stations, respectively. The average daytime and nighttime R2 radius were 21 and 24 (km) for all weather stations, 22 and 29 (km) for mainland stations, and 20 and 21 (km) for coastal stations, respectively. Thus, the largest difference between the nighttime and daytime values of the optimal radius was obtained at mainland stations, where, on average, the values of R1 and R2 increase at night by ~ 6 and 7 (km), while at coastal stations by ~ 4 and 1 (km). In contrast to the mainland and coastal stations, on Sakhalin one can note a decrease at night of the average optimal radius R1 and R2 by ~ 1 km and 3 km (Table 2). For Sakhalin weather stations, the mean daytime radius R1 and R2 are 19 and 20 (km), respectively, the mean nighttime radius R1 and R2 are 18 and 17 (km). The largest difference between the night and day values of the radius calculated by both methods was obtained at station Juzhno-Sahalinsk, where it was − 13 km and − 15 km, and at station Nevel’sk, where it was − 11 km each.
3.2 Comparison of thunderstorm activity according to WWLLN data and weather station data
3.2.1 Intra-annual variability of thunderstorm activity
A significant indicator of the quality of WWLLN data is how the data reflects the intra-annual variability of thunderstorm activity in comparison with weather station data. Figure 1 shows graphs of the intra-annual variations of the summary number of days with thunderstorms over ten years according to the weather stations data and similar graphs obtained from lightning WWLLN samples in the stations vicinity with optimal radius R1 and R2. Note that the maximum ordinates on the graphs are 120 days for continental stations and 35 days for Sakhalin stations, since continental regions are characterized by a larger number of days with thunderstorms per year compared to the island (Table 2).
In general, it can be noted that according to visual observations, thunderstorm activity is characterized by 1 maximum at mainland stations (with the exception of Svijagino and Pogranichnyj stations), 2 maxima at coastal stations (with the exception of Ternej station), at Sakhalin stations both by 1 maximum (Uglegorsk, Poronajsk, Mys Terpenija, Ilyinskiy, Nevel’sk) and 2 maxima (Aleksandrovsk-Sahalinskij, Tymovskoe, Pogranichnoe, Mys Kril’on and Juzhno-Sahalinsk ). In most cases, graphs based on WWLLN data reflect the main qualitative features of the intra-annual course of TA according to observations at weather stations. At 4 stations, the number of peaks on the graphs does not match, when the graphs based on the data of weather stations have two maxima each, and the graphs based on the WWLLN data have only one (Svijagino), and vice versa (Mys Terpenija, Ilyinskiy, Nevel’sk). At some weather stations (Lermontovka, Krasnyj Jar, Rudnaja Pristan’, Vladivostok, Timirjazevskij, Pos’et, Aleksandrovsk-Sahalinskij, Uglegorsk, Mys Terpenija, Juzhno-Sahalinsk), one can note the asynchrony of the main or secondary maxima of TA. Quantitative difference between the graphs of the intra-annual course according to observations and according to WWLLN data in the sample with R1 varied from − 22 to + 24 days (for R2 varied from − 10 to + 17 days) and the root mean square difference (rmsd) was 4.6 (3.5) days.
For each weather station, we calculated the 10-year average values of the number of days with thunderstorms (as a measure of thunderstorm activity) according to observational data (\(\overline {{{\text{NM}}}}\) ) and with lightning according to WWLLN data, selected in the vicinity of weather stations with radius R1 (\(\overline {{{\text{NW1}}}}\)) and R2 (\(\overline {{{\text{NW2}}}}\)), and then the difference between them. The statistical characteristics of this average annual thunderstorm activity, both for the entire south of the Russian Far East, and separately for mainland, coastal and Sakhalin stations, are given in Table 2. On average, about 20 days with thunderstorms per year are observed at weather stations of the south of Russian the Far East, 29 days/year at mainland stations, 14 days/year at coastal stations and 6 days/year at Sakhalin stations (Table 2). Thus, the highest thunderstorm activity was observed at mainland stations, where the maximum value of \(\overline {{{\text{NM}}}}\) (42 days/year) was at weather station Konstantinovka, and the lowest TA was at Sakhalin, where the minimum value of \(\overline {{{\text{NM}}}}\)(2 days/year) was obtained at stations Mys Terpenija and Pogranichnoe. The difference in the average annual thunderstorm activity according to weather station data and according to WWLLN data in samples with radius R1 and R2 for the entire south of the Russian Far East varied within ± 6 days/year and up to 1.3 days/year, rmsd (σ in Table 2) were 2.4 days/year and 0.4 days/year, respectively (Table 2). For mainland stations rmsd were 3.2 days/year and 0.4 days/year, for coastal stations rmsd were 0.8 days/year and 0.3 days/year, for Sakhalin stations rmsd were 1.4 days/year and 0.2 days/year, respectively (Table 2). Thus, the second method for the determining of the optimal radius R2 gives the better results for the estimating of thunderstorm activity according to WWLLN data, if we consider the weather station data as a “reference”.
3.2.2 Diurnal variability of thunderstorm activity
The thunderstorm activity has a clearly defined diurnal variation, especially on land (Zhang et al., 2010). The diurnal variation of TA according to weather station data and according to WWLLN data will differ and their comparison is not entirely correct, since visual observation data are reduced to registering the time of the beginning (that is, the first observed lightning, thunder or flash of light during heat lightning) and the end of a thunderstorm (the last thunder or lightning), while the WWLLN data is the time of registration of each lightning in the vicinity of the station. At the same time, observations of atmospheric phenomena are influenced not only by geographical features (exposure of stations, closeness to the surrounding landscape etc.) and current weather situations (sky conditions, fog, poor or good visibility etc.), but also subjective factors, for example, the state of health of the observer’s eyesight and hearing (Instructions 1985; Guide 2008). Therefore, the marks of thunderstorms during the day are several orders of magnitude less than the number of lightnings. To illustrate this, we present three typical examples.
Figures 2a,b show the TA during the period of 7–9 August 2013 in the area of the station Konstantinovka, which, according to the surface weather maps of the Japan Meteorological Agency (JMA, http://www.jma.go.jp/), was affected by the front of a stationary cyclone moving from the southwest across the Amur Oblast to the northeast (Fig. 2c). Figure 2a shows the three-day course of the total number of lightning in 0.5 hour windows obtained from the WWLLN data in the square (65 × 65 km) limited by coordinates 49.04° N – 50.21° N and 127.07° N − 128.87° E and in the station vicinity with optimal radius R1 = 29 km and R2 = 36 km. Figure 2a also shows the moments of the beginning and end of the “thunderstorm” phenomenon registered at the station by the observer. The WWLLN graph clearly shows separate periods of lightning activity, which can be distinguished as individual thunderstorms passing over the weather station. In some cases, the marks of the thunderstorm beginning are close to the start of a sharp increase in the number of WWLLN lightning. In general, the time intervals between the marks of the beginning and end cover periods of high WWLLN lightning activity. However, it is possible to distinguish periods when there are no marks of thunderstorms at the weather station, but lightnings is detected by the WWLLN network, for example, 03:00–06:00 UTC on August 7 or from 20:30 UTC to 21:30 UTC on August 9. Figure 2b shows the spatial distribution of lightning in the vicinity of the station for the time intervals 11:00–17:00 UTC on August 7, 08:00–17:00 UTC on August 8, and 02:00–14:00 UTC on August 9. It can be seen from the figure that the areas of thunderstorms shift as the cyclone passes, while the total number of lightning (N) increases from 378 to 1711 discharges. On the map for the third time interval, areas of significant lightning density are clearly distinguished, which can be associated with mesoscale structures, such as convective complexes, a multi-cell thunderstorm, or a thunderstorm front.
The second example is thunderstorm activity in the Pos'et station area bounded by coordinates 42.07°N – 43.24°N and 130.1°E − 131.6°E on 19–20 October 2017 (Figs. 3a,b), when the weather over the southern coast of Primorsky Krai (Peter the Great Bay) was determined by the passage of atmospheric fronts with short-term precipitation and squally wind (Fig. 3c). During this period, only rain was indicated at Pos'et station, while WWLLN data show significant thunderstorm activity from 1600 UTC on October 19 to 09:00 UTC on October 20 (Fig. 3a) southeast of the station (Fig. 3b). Figure 3b shows the lightning distributions in the vicinity of the station in three different periods: 16:00–21:00 UTC and 21:00–24:00 UTC on October 19 and 00:00–03:00 UTC on October 20. It can be seen that the points of lightning discharges formed bands that can be associated with the positions of the thunderstorm front, which was moving to the southeast. Several lightnings were located within the Pos'et vicinity with optimal radius R1 = 18 km and R2 = 15 km, i.e. the observer at the weather station could see them. It is possible that the lack of records about the atmospheric phenomenon “thunderstorm” at the weather station was due to poor visibility caused by rain and sea fog, which were indicated in weather reports and marked on weather maps (Fig. 3c). It can also be assumed that the discrepancies in the data are due to the geographical features of the area and the degree of openness of the station. The Pos'et station is located on a peninsula between the Expedition bay in the west and Novgorodskaya bay in the east and is closed from the open sea (i.e. where lightning has been detected) by the Krabbe peninsula. It is possible that these reasons determine the relatively low consistency between the WWLLN data and the visual observations data at Pos'et station, for which the TSI value is only 0.45 (Table 1).
The third example, shown in Fig. 4, is the thunderstorm activity near Pos'et station on 8 September 2014 (Fig. 4a,b), caused by the passage of an occlusive sedentary cyclone (Fig. 4c). Unlike the previous case, this time the thunderstorm was detected by both the WWLLN network and the observer at the weather station, although, for example, in the interval of 11:00–12:30 UTC, only two WWLLN lightnings hit the vicinity with optimal radii, and all other lightnings were detected beyond outside of it. In the field of the spatial distribution of lightning (Fig. 4b) in the interval of 07:42–09:30 UTC to the south, southwest of the station, two areas of high density of discharges can be distinguished. Figure 4d shows part of the IR imagery of the study area received from the geostationary satellite MTSAT on 8 September 2014 at 07:32 UTC and posted on the Tropical Cyclones Archive of the University of Wisconsin—Madison/Cooperative Institute for Meteorological Satellite Studies page (http://tropic.ssec.wisc.edu/archive/). The cloudiness fields in the IR range, combined with the fields of lightning discharges, selected for an image time ± 1 hour, showed that these centers of thunderstorm activity coincide with areas of cold front intense upward movements and cumulonimbus clouds.
The above examples show that WWLLN data can well represent temporal variability of TA at diurnal scales and distinguish individual thunderstorms both in the temporal course and in the spatial distribution of lightning.