See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/264559832

Remote sensing of lightning by a ground-based microwave radiometer

Article  in  Atmospheric Research · December 2014

DOI: 10.1016/j.atmosres.2014.07.009

CITATIONS READS

6 100

5 authors, including:

Zhenhui Wang Qing Li

Nanjing University of Information Science & Technology Nanjing University of Information Science & Technology
87 PUBLICATIONS   460 CITATIONS    7 PUBLICATIONS   9 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

observation of lightning heating effect with microwave radiometer View project

radar quality control View project

All content following this page was uploaded by Zhenhui Wang on 27 December 2017.

The user has requested enhancement of the downloaded file.

 Atmospheric Research 150 (2014) 143–150

Contents lists available at ScienceDirect

Atmospheric Research
journal homepage: www.elsevier.com/locate/atmos

Remote sensing of lightning by a ground-based

microwave radiometer☆
Zhenhui Wang a,b,⁎, Qing Li b, Fangchao Hu b, Xuefen Cao b, Yanli Chu c
a
Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters, CMA Key Laboratory for Aerosol-Cloud-Precipitation,
Nanjing University of Information Science & Technology, Nanjing 210044, PR China
b
School of Atmospheric Physics, Nanjing University of Information Science & Technology, Nanjing 210044, PR China
c
CMA Institute of Urban Meteorological Research, Beijing 100089, PR China

a r t i c l e i n f o a b s t r a c t

Article history: Based on the theory of thermal radiation and its transfer in the atmosphere, the response of a
Received 26 May 2014 ground-based microwave radiometer to a lightning-superheated cylinder in the atmosphere is
Received in revised form 7 July 2014 studied and the theoretical expressions are given for the relationship between brightness
Accepted 8 July 2014 temperatures and parameters such as distance, size, duration and temperature of the lightning-
Available online 26 July 2014
superheated cylinder. The results from simulated calculations show that it is quite possible
for a lightning-superheated cylinder to be observed by a microwave radiometer working in
Keywords: the 50–60 GHz band with a sensitivity of 0.3 K, typically used for atmospheric temperature
Microwave radiometer profiling. The brightness temperature observed with any one of the channels in the band increases
Brightness temperatures
as the distance between the lightning and the radiometer decreases. Lightning at a short distance
Lightning-superheated air cylinder
would make the brightness temperature observed by the channels near to 60 GHz increase more
Lightning remote sensing
while distant lightning would make the brightness temperature observed by the channels near to
50 GHz increase more. This feature could be used to retrieve lightning distance and features of the
lightning-heated air cylinder from brightness temperature observations. One lightning observa-
tion by a ground-based radiometer, the challenge of such observations, and a theoretical analysis
are presented. Additional observations are needed for more thorough exploration of this unique
remote sensing capability.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction prediction. Current lightning detection methods include radio

frequency and visible light remote sensing (Tibor, 2006). We
The electric current discharged by lightning may approach present lightning detection by a ground-based microwave
several hundred thousand amperes and the air temperature radiometer as a case study.
along the discharging path may be as high as thousands of A microwave radiometer can detect thermal (quantum
Kelvins or larger (Prueitt, 1963; Ouyang et al., 2006). Abrupt mechanical) radiation emitted by gaseous, liquid and solid
heating creates a powerful explosion and sound waves. matter and convert it to brightness temperatures using Planck's
Lightning information is valuable for disaster monitoring and equations (Ulaby et al., 1981; Zhou, 1982). Ground-based
microwave radiometer profilers in the 20–60 GHz range are
commonly used for remote sensing of tropospheric tempera-
☆ The work is jointly supported by National Natural Science Foundation of ture, humidity, and liquid profiles (Westwater et al., 1985;
China (41275043, 41005005), and the Urban Meteorological Research Founda- Duan and Wu, 1999; Zhu et al., 1994; Solheim et al., 1998;
tion IUMKY & UMRF201101. Güldner and Spänkuch, 2001; Westwater et al., 2004; Ware
⁎ Corresponding author at: School of Atmospheric Physics, Nanjing Univer-
et al., 2003) and are followed with increasing interest (Ware
sity of Information Science & Technology, Nanjing 210044, PR China. Tel.: +86
25 58731195. et al., 2013; Ratnam et al., 2013; Sánchez et al., 2013; Xu et al.,
E-mail address: [email protected] (Z. Wang). 2014).

http://dx.doi.org/10.1016/j.atmosres.2014.07.009
0169-8095/© 2014 Elsevier B.V. All rights reserved.
144 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150

Spikes in microwave radiometer brightness temperatures and we take α = 2.5° = 0.044 rad on average, therefore the
were observed consistent with thermal radiation generated by cross-section diameter of the antenna beam is DA = αr ≈ 44 m
lightning and propagating in accordance with microwave at r = 1 km and DA ≈ 440 m at r = 10 km. In order to make it
radiative transfer equations. We estimate microwave emission easy to observe lightning from a distance, the antenna should
from lightning-heated air modeled in a cylindrical shape, and be adjusted to low elevation. And it is better to hold DA b DL so
present simulated and actual remote sensing of such emission that the beam is full of the heated air at the distance r otherwise
by a ground-based microwave radiometer. This report will corrections as discussed in Section 5.3 of the paper must be
hopefully stimulate further research and better understanding performed. The temperature inside the superheated air
of lightning-induced microwave radiation. cylinder may not be uniform (Prueitt, 1963) but for total
radiance simulation the beam intersection area is assumed to
2. Detection of microwave emission generated by lightning be uniform.
Suppose that the volumetric absorption coefficient at
According to the presentation by Dr. R. Ware of USA distance r is ka(r) (Liebe et al., 1992) and the influence of
Radiometrics Corp., there was a microwave radiometer making scattering on microwave radiance transfer is negligible, the
brightness temperature observations at the India Space Launch transmittance in the range [0, r) according to Beer's law would be
Center (Trivandrum, Kerala, India) during pre-monsoon rain at  Z r 
around 12:20 UT (5:50 LT) on March 29, 2010. The radiometer τðr Þ ¼ exp − ka ðr Þdr ð1Þ
has 14 channels in the 51–59 GHz band (Ware et al., 2013) as 0

shown by the first 2 columns in Table 1; brightness

temperature (Tb) time series in Fig. 1 show that 51.248, The radiance measured by a radiometer in terms of
51.760 and 52.280 GHz Tb's are obviously greater by 29 K, 18 K brightness temperature is defined by radiative transfer equa-
and 12 K near 12:20:20. The radiometer frequency channels tion (Ulaby et al., 1981; Yao and Chen, 2005; Zhang and Wang,
are observed sequentially at 0.7 second intervals. This spike is 1995)
generally correlated in time and direction with multiple Z ∞
lightning strokes seen by the eye. T b ð0Þ ¼ T b∞ τ ð∞Þ þ T ðr ÞW ðrÞdr ð2Þ
0

3. Analytical model
where Tb∞ is the background brightness temperature from the
out-space, τ(∞) is the transmittance from the out-space to the
Electron avalanches are the main constituent of electrical
radiometer, T(r) the temperature of the air cylinder at distance
discharges in the atmosphere, with a duration of 1–2 ns and
r, and
spectrum peak around 1 GHz (Cooray and Cooray, 2012;
Petersen and Beasley, 2014). However, superheated air may dτ ðr Þ
also contribute to total lightning-induced microwave emission. W ðr Þ ¼ − ¼ τðr Þka ðrÞ ð3Þ
dr
Though the discharging path of a cloud-to-ground lightning
may propagate in any pattern, it can be simplified as a vertical is called as weighting function. ka(r), τ(r) and W(r) for the 4
cylinder for distant observations and therefore the lightning typical channels are calculated for the case of standard
superheated area is considered as a vertical, elliptical cylinder. atmosphere at 15° elevation. The 3rd column of Table 1 lists
Suppose the radiometer is pointing at the direction (r) of the the values of ka(r) at r = 0, i.e., ka(r = 0) and Fig. 3 gives the
lightning, as shown in Fig. 2, to receive the thermal radiance profiles of τ(r) and W(r). It can be seen that:
from the superheated air cylinder with a radial diameter DR and
a cross-range diameter DL. The beam width of the radiometer (1) ka(r) increases as frequency increases for a given
antenna at 51–59 GHz varies with frequency from 2.3° to 2.7° distance, but ka(r) changes little as r increases for a

Table 1
Radiometer channel frequencies, absorption coefficient (ka) of a standard atmosphere at sea level, apparent emissivity (ε), and the simulated brightness temperature in
the case of α = 1°, R1 = 3 km, ΔT = 10,000 K, DR = 60 m, and DL = 60 m.

Channel Freq. (GHz) ka (Np/km) ε Tb (K) without lightning ΔTb (K) because of ΔTb (K) after correction
lightning for α = 2.5°

1 51.248 0.156 0.009 90.7 46.6 21.2

2 51.760 0.184 0.011 105.1 51.8 23.5
3 52.280 0.255 0.015 124.8 58.0 26.4
4 52.804 0.287 0.017 151.0 63.5 28.9
5 53.336 0.380 0.023 183.5 66.3 30.1
6 53.848 0.508 0.030 216.8 62.8 28.5
7 54.400 0.704 0.041 248.3 50.4 22.9
8 54.940 0.962 0.056 268.8 33.0 15.0
9 55.500 1.298 0.075 279.2 16.6 7.5
10 56.020 1.657 0.095 282.9 7.3 3.3
11 56.660 2.105 0.119 284.4 2.5 1.1
12 57.288 2.500 0.139 285.0 0.9 0.4
13 57.964 2.859 0.158 285.4 0.3 0.1
14 58.800 3.183 0.174 285.6 0.1 0.0
 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150 145

Fig. 1. Time series of brightness temperatures observed by a microwave radiometer installed at India Space Launch Center, Trivandrum, Kerala, India, during 12:15—
12:25 UT(5:45–5:55 LT) on Mar. 29, 2010 at 15° elevation.
(From slides presented in November, 2010, at Nanjing University of Information Science and Technology by R. Ware, Radiometrics Corporation).

certain frequency because the elevation is only 15°. Let the distance from lightning-heated cylinder to the
Therefore ka(r) is approximately independent of dis- radiometer be R1,as shown in Fig. 2, and radial diameter of
tance and is abbreviated as ka hereinafter; the cylinder is DR = R2 − R1. Let T+ be the temperature inside
(2) transmittance at r = 0 is always 1 but the weighting the range (R1, R2) after discharging. It would become
function at r = 0 is different for different frequencies.
þ
This is because W(r = 0) = ka, i.e., for higher frequencies T ¼ T ðr Þ þ ΔT ðr Þ ð4Þ
such as 57.964 GHz, atmospheric absorption is strong
and W(r = 0) would be greater than for the lower i.e., the temperature is increased by ΔT(r) due to lightning.
frequencies such as 51.248 GHz at which the atmo- Considering that the increased radiant energy because of the
spheric absorption is weak; lightning is contributed only by the heated cylinder, the
(3) both τ(r) and W(r) decrease as r increases but increment in brightness temperature measured by the radi-
(4) the decreasing rates with respect to r are different at ometer, according to Eq. (1), should be
different distances. For example, W(r) at r = 0 km are Z Z
R2 R2
1.298 and 0.380 for 55.500 and 53.336 GHz, respectively, ΔT b ¼ ΔT ðrÞW ðr Þdr ¼ ΔT ðR1 Þ  W ðr Þdr: ð5Þ
while at r = 3 km, W(r) decreases down to 0.112 for R1 R1
53.336 GHz and to less than 0.029 for 55.500 GHz. Both
the transmittance τ(r) and the weight W(r) for high The equation is further alternated as
frequency such as 57.964 GHz decrease faster than those
for low frequency such as 51.248 GHz because of the ΔT b ¼ ΔT ðR1 Þ  W ðR1 Þ  DR ð6Þ
difference in absorption coefficient. This feature makes it
possible to determine lightning distance with a multi- where W ðR1 Þ is the mean value of W(r) in the range (R1, R2).
channel radiometer as discussed in Section 5. This formula implies that the response of the radiometer to

(a) (b)

Fig. 2. Model for ground-based radiometer observation of a lightning-superheated cylinder in the atmosphere. Two dashed-line squares show the concept of beam-
filling in the cases of full filling and partial filling.
146 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150

than 0.2 at the highest frequency. After discharging, the

cylinder is at the extreme temperature, pressure, and ionization
state, ka and ε for the cylinder might change. The difference
would add some uncertainty to the estimation of ΔT from ΔTb
according to Eq. (7). Fortunately, both ε and ΔT are all
lightning-dependent and are included as a product in Eq. (7).
Therefore, take ε as a parameter independent of lightning and
ΔT retrieved from Eq. (7) would be indicative of lightning
0 5 10 15 radiance rather than physical temperature. That is, if ε after
Distance r (km) lightning increases by a percentage, retrieved ΔT must be
adjusted according to the percentage in order to obtain the
increase of physical temperature. Therefore it is assumed that
ka and ε inside the cylinder are reference constants a in the
following discussion.

4. Parameter selection for simulation and analysis

The response of radiometer to a lightning-heated air

cylinder is different at different frequencies because they are
characterized by different transmittance and weighting func-
tions as shown by Fig. 3. This can be quantitatively verified as
Fig. 3. Transmittance (upper) and weighting functions (lower) for the 4 typical the following simulative calculation.
channels of the microwave radiometer when elevation is 15° (inset vertical
In the simulation, let the antenna elevation be 15°, beam
axis = 0.5 km−1).
width α = 1°, i.e., DA ≈ 51 m at r = 3 km. A lightning-heated air
cylinder being observed by the radiometer is at the distance
lightning heating is proportional to the temperature increment R1 = 3 km and the cylinder diameter is 60 m (i.e., DR = DL =
of the air cylinder and its radial diameter with the mean value 60 m and DL ≥ DA) as shown by the sub-panel embedded in
of W(r) in the cylinder as the proportional coefficient. the left half of Fig. 2. Suppose the air from the cylinder to the
Estimation of ΔTb can be directly completed with Eq. (5) since radiometer is the same as the Standard Atmosphere and only
W(r) has been given by Fig. 3. the temperature inside the cylinder increases by ΔT = 10,000 K
On the other hand, insert W ðr Þ ¼ − dτdrðrÞ as defined by as shown in Fig. 4.
Eq. (3) into Eq. (5) and makes use of the relationship among The result from the simulation for R1 = 3 km, ΔT = 10,000 K
transmittance, absorption and emissivity (Zhang and Wang, and DR = 60 m, is given in Columns 5 and 6 of Table 1. Column 5
1995), and we have lists Tb, the brightness temperatures observed without the
existence of the lightning-heated air cylinder and Column 6 lists
ΔT b ¼ ΔT ðR1 Þ  ðτðR1 Þ−τ ðR2 ÞÞ ¼ ε  ΔT ðR1 Þ  τðR1 Þ ð7Þ ΔTb, the response of the radiometer to the lightning-heated air
cylinder. One can see that:
where
(1) brightness temperatures are all increased and 12
ε ¼ 1−τ ðR2 Þ=τ ðR1 Þ: ð8Þ channels of the 14 increase by more than 0.5 K. Since
the nominal sensitivity of operational microwave radi-
ε is the apparent emissivity of the air cylinder. Eq. (7) implies ometer is 0.3 K rms, most channels can detect signals
that the response of the radiometer to lightning heating is greater than 1 K from the lightning-heated air cylinder
proportional to both the temperature increment of the air featured by ΔT = 10,000 K and DR = 60 m at R1 = 3 km;
cylinder and the transmittance from the cylinder to the (2) brightness temperature at 53.336 GHz increases by
radiometer. The proportional coefficient is the apparent 66.3 K, as shown by the bold type in column 6 for
emissivity of the cylinder. Therefore, for the purpose of Channel 5, which is the largest amplitude relative to
response estimation, the cylinder can be taken as a radiant other channels. The increased amplitude decreases as
flat at temperature ΔT(R1) with an apparent emissivity the frequency difference of a specified channel from
determined by Eq. (8). 53.336 GHz increases. The response of 58.800 GHz
Note that 0 ≤ ε ≤ 1 and according to Eqs. (1) and (8), one channel is negligible because both the transmittance
has and weighting are nearly 0 at R1 = 3 km as shown in
Fig. 3.
ε ¼ 1− expð−ka DR Þ ð9Þ
5. Further analysis on the relationship between ΔTb and
i.e., the value of ε depends on the radial diameter of the cylinder lightning parameters
and its internal nature. A thick cylinder has a large emissivity at
high frequency channel. The 4th column of Table 1 lists the 5.1. Relationship between ΔTb and distance R1
values of emissivity for DR = 60 m computed in the condition
that ka has been given in the 3rd column for Standard Fig. 5 gives ΔTb's for 14 frequencies as a function of distance
Atmosphere. It can be seen that ε is small in general and less R1.
 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150 147

10282
360 10282

340
(a) (b)
320

T(K)

T(K)
290
300

285
280

260 280
0 5000 10000 15000 2800 2900 3000 3100 3200 3300
r(m) r(m)

Fig. 4. Air temperature distribution as a function of distance from the radiometer in the case of 15°, R1 = 3 km, ΔT = 10,000 K, and DR = 60 m. The dashed line is the
temperature following the standard atmosphere and the solid line is pulsed because of the lightning-superheating. The right panel is the same as the left but enlarged to
make the pulse outstanding.

It can be seen that 51.248 GHz channel. This feature makes lightning
distance estimates possible using simultaneous multi-
(1) lightning at a shorter distance promotes higher frequen- channel radiometer observations.
cy channels to make an outstanding response. This is
because ε in Eq. (7) is greater at higher frequency; 5.2. Relationship between ΔTb and radial diameter DR
(2) brightness temperature increment at any given frequen-
cy decreases as distance increases. This is because τ(R1) Under the condition of DL N DA, the decrease of cylinder
in Eq. (7) decreases as R1 increases as shown in Eq. (1); radial diameter leads to the decrease of ΔTb for any channel as
(3) brightness temperature increments at higher fre- shown by Eq. (5). Fig. 6 gives the calculated results in the
quency decrease faster as R1 increases, implying condition which is the same as in Fig. 5 but for DR = 20 m (Chen,
that the maximum of ΔTb at a certain distance 2006). The decrease of cylinder radial diameter is equivalent to
corresponds to a certain channel. For instance, the the decrease of ε according to Eq. (9) and further the decrease of
maximum of ΔTb at R1 ≈ 1.0 km is roughly 250 K ΔTb according to Eq. (7). So the brightness temperature
corresponding to 57.288 GHz channel, the maximum increment in Fig. 6 is less than that in Fig. 5 but the 3 points
of ΔTb at R1 ≈ 2.0 km is a little bit more than 100 K derived from Fig. 5 are all quantitatively applicable to Fig. 6.
corresponding to 54.400 GHz channel, the maximum
of ΔTb at R1 ≈ 3.0 km is approximately 60 K 5.3. Relationship between ΔTb and antenna beam width α
corresponding to 53.336 GHz channel, the maxi-
mum of ΔTb at R1 ≈ 6.0 km is about 30 K Antenna beam width at 51–59 GHz for Tb observations is
corresponding to 52.804 GHz channel, the maximum roughly 2.5° rather than 1°. This would lead to “DL b DA”,
of ΔTb at R1 ≈ 10.0 km is close to 15 K corresponding especially if the lightning-heated cylinder is quite small and
to 52.280 GHz channel, and the maximum of ΔTb distant. That is, the cylinder cannot fully fill the cross-section of
at R1 ≈ 15.0 km is about 8 K corresponding to the beam as shown by the sub-panel embedded in the right half

(a)
(c)

(b)

Fig. 5. Brightness temperature increment at 14 frequencies as a function of distance in the case of ΔT = 10,000 K and DR = 60 m. (a) General feature, (b) zoom in for
short distance, (c) zoom in for long distance (inset horizontal axis = 10,000 m).
148 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150

(a)
(c)

(b)

Fig. 6. The same as Fig. 5 but for DR = 20 m.

of Fig. 2. In this case, the response estimated with Eq. (5) should directly compared with ΔTb observation. Otherwise it must be
be corrected for partial beam filling, named as “partial beam corrected for DTL b DTA, as the following
filling correction”. Knowing the correction factor denoted by c1,
c
one has ΔT b ¼ c2  ΔT b ð12Þ
c
ΔT b ¼ c1  ΔT b ð10Þ which is called “partial time filling correction”. The correction
factor is determined by
where c1 ∈ [0,1]. The factor should be the area ratio as shown
by the sub-panel embedded in the right half of Fig. 2 but an 
DTL =DTA for DTL bDTA
approximation is the following c2 ¼ : ð13Þ
1 for DTL ≥DTA

DL =DA for DL =DA b1
c1 ¼ : The value of DTL depends on duration of flash, thickness of
1 for the others
super-heated air cylinder, wind speed and so on. Duration time
of a long stroke is typically more than 2 ms and less than 1 s
Since DA = αR1, then
(Wang et al., 2013). Therefore c2 may be any value in the range
 [0,1]. To benefit the monitoring of lightning, integration time
DL =αR1 for DL =αR1 b1
c1 ¼ : ð11Þ should be as short as possible.
1 for the others
Combining together the two corrections above and putting
them into Eq. (7), one obtains the simulated response after
For the case α = 2.5° = 0. 044 rad, R1 = 3 km, DA ≈ 132 m,
correction
and c1 = 0.455 for DL = 60 m. Substitute this into Eq. (10), the
ΔTb after correction is given in the last column of Table 1. One
can see that ΔTb (K) after correction for beam filling has been ΔT b ¼ τðR1 Þ  c1  c2  ε  ΔT ðR1 Þ: ð14Þ
decreased but still most of the channels show a large response.
Nevertheless, it is still hoped that antenna beam width for the Putting Eqs. (1), (9), (11) and (13) into Eq. (14), simplifying
purpose of monitoring lightning be as narrow as possible. ΔT(R1) as ΔT, and foot-noting frequency-dependent parame-
The scale of DL is of course very important in this issue. ters with subscript i, one has
Based on optical observations with high-speed camera system,
the spatial size of an individual return stroke is on the order of DTL D
ΔT bi ¼ ð1− expf−kai DR gÞ expf−kai R1 g L ΔT: ð15Þ
10 m, the width of a band-type flash can be greater than 10 m αDTA R1
(Chen, 2006), and the scale of a complete lightning stroke
including a series of return strokes and lasting for tens to If kaDR is small enough, one has the following approximation
hundreds of milliseconds may be tens and even nearly 100 m. If
we take these as the scale of DL, it is really possible for c1 in DTL D D
Eq. (10) to be close to 1 under certain conditions. ΔT bi ¼ k expf−kai R1 g R L ΔT: ð16Þ
αDTA ai R1

5.4. Relationship between ΔTb and integration time

It can be seen that
Let DTA be defined as the effective radiometer sky- (1) The response of brightness temperature to lightning is
observation time (including noise diode calibration and proportional to the increment of cylinder temperature,
switching times). In this case DTA = ~700 ms. Let DTL denote the duration time and the area of the cylinder cross-
the duration time of a super-heated air cylinder. In case DTL ≥ DTA section, and inversely proportional to distance in case
and DTA falls within the DTL time frame, ΔTb simulated can be DTL b DTA and DL b DA.
 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150 149

(2) The distance can be solved from the ratio of two and the proportional factor is the emissivity of the
responses obtained with 2 frequencies according to lightning-heated air cylinder.
!, (2) The Tb response at any frequency associated with the
ΔT bi kaj   radiometer channels is inversely proportional to the
R1 ¼ ln kaj −kai ð17Þ
ΔT bj kai distance between the radiometer and the cylinder.
(3) The emissivity of the lightning-heated air cylinder is
and the product of the increment of cylinder tempera- greater at higher frequency than at lower frequency.
ture, the duration time and the area of the cylinder This implies that lightning at a shorter distance results in
cross-section such as a higher frequency channel to make an outstanding
response while distant lightning leads a lower frequency
B ¼ DTL DR DL ΔT ¼ ΔT bi αDTA R1 =ðkai expf−kai R1 gÞ ð18Þ channel to make an outstanding response.
(4) Derived from points (2) and (3) above is that there is a
can be estimated and used as a parameter to indicate the maximum Tb response for a certain distance at a certain
intensity of lightning. frequency. Therefore, the distance and features of the
lightning-heated air cylinder can be estimated from Tb
6. On the possibility to detect lightning as seen in Fig. 1 observations with the radiometer.
(5) To benefit lightning monitoring with a microwave
The preliminary analysis above shows that it is really radiometer, the beam width should be as narrow as
possible for a ground-based microwave radiometer to detect possible and the integration time as short as possible. In
lightning and its observations of brightness temperature in 51– the case of distance R1 = 3 km, lightning heating ΔT =
59 GHz can be used to estimate the distance and features of 10,000 K and cylinder diameter DR = DL = 60 m, even
the lightning-heated air cylinder, especially if the brightness “partial filling” correction is 1% (i.e., c1 ∗ c2 = 0.01), there
temperature dynamic range of the radiometer is wide enough, are still a few lower-frequency channels to respond by
the antenna of the radiometer is pointing toward the lightning more than 0.5 K. But the difficulty for a ground-based
and the beam width and integration time are all adequate. microwave radiometer to observe lightning is that the
In the case given in Fig. 1, 29, 18 and 12 K spikes are seen at radiometer does not know in advance the direction of
51.248, 51.760 and 52.280 GHz respectively at 12:20:20. Since possible discharge.
the frequencies are the lowest, the position of the lightning
mentioned in Section 1 must be quite far and intense. Consider It has been assumed in the paper that the apparent
DTA = 1 s, α = 2.5° = 0.044 rad, and suppose that the value of emissivity of the lightning-heated air cylinder for thermal
ka in Table 1 is tripled for the observation site (considering that radiation is the same as before lightning. This assumption does
the water vapor content in tropical area may be 3 times that for not affect the ability to discern range according to Eq. (17) since
standard atmosphere), and the increments of 29 and 18 K for the frequency dependence of absorption outside the cylinder is
Channels 51.248 and 51.760 GHz are substituted into not altered but may add some uncertainty to the estimation of
Eqs. (17) and (18), we can obtain that R1 = 7.64 km and B = ΔT according to Eq. (18).
DTLDRDLΔT = 744 K km2 s. One of the possible condition for this As it is said that “Tossing out a brick to get a jade gem”, the
result is DTL = 1 s, DRDL = 60 m*300 m = 0.018 km2 (DA = αR1 = paper serves as a modest spur to stimulate additional research
0.044*7.64 = 336 m, DL ≈ DA) and ΔT ≈ 4.1 ∗ 104 K, implying that on lightning monitoring with microwave radiometer in order
the cylinder must be the result of a long stroke or a series of return to understand the magnitude of heating by lightning at a
strokes. specific distance, with a specific strength and microwave
We really want to make verification for this judgment but radiance. And even more, the radiometer is sensitive not only
unfortunately data needed are not available. It is therefore to thermal emission from lightning heated air, but also to
suggested that, in case the experiment for verification of this electromagnetic emission generated directly from the lightning
kind is performed in the future, lightning distance should be electric pulse, as described in the articles by Krider (2003) and
obtained with lightning locating system such as IMPACT or Gurevich et al. (1992). This should be taken into account for
SAFIR. Of course, instruments for detecting parameters like ΔT, accurate retrievals of lightning parameters from microwave
DR, DL and DTL are also needed if available. This is just the radiometer observations.
innovative sense of the study in order to use a microwave
radiometer to monitor lightning. Acknowledgment

7. Conclusion We thank Dr. R. Ware (Radiometrics Corporation, USA) for

stimulating us to start this study during his November 2010
Emissions from a super-heated air cylinder generated by presentation at Nanjing University of Information Science and
lightning can be observed with a ground-based microwave Technology and his later continuous suggestions and modifica-
radiometer currently used for atmospheric temperature, hu- tions to this paper and Dr. Tan Yongbo of NUIST/Department of
midity and liquid profiling. Simulated and preliminary obser- Lightning Sci. & Technology for his comments on the manuscript.
vation analysis shows the following:
References
(1) The Tb response of the radiometer is proportional to both
the increment of temperature in the discharging area Chen, W., 2006. The Principle of Lightning, 2nd ed. Meteorological Press,
and the atmospheric transmittance to the radiometer, Beijing, pp. 1–419 (in Chinese).
150 Z. Wang et al. / Atmospheric Research 150 (2014) 143–150

Cooray, V., Cooray, G., 2012. Electromagnetic radiation field of an electron Tibor, Horv, 2006. Understanding Lightning and Lightning Protection: A
avalanche. Atmos. Res. 117, 18–27. Multimedia Teaching Guide. John Wiley, pp. 1–220.
Duan, Y., Wu, Z., 1999. Monitoring the distribution characteristics of liquid and Ulaby, F.T.,Moore, R.K.,Fung, A.K., 1981. Microwave remote sensing. Active and
vapour water content in the atmosphere using ground-based remote Passive, vol. 1. Addison-Wesley Publishing Company [M]. Advanced Book
sensing. Q. J. Appl. Meteorol. 10 (1), 34–40 (in Chinese). Program/World Science Division, Reading, MA, USA, pp. 1–456.
Güldner, J., Spänkuch, D., 2001. Remote sensing of the thermodynamic state of Wang, Z.,Li, X.,Guo, F.,Shi, G., 2013. 101 Topics on Lightning and its Prediction,
the atmospheric boundary layer by ground-based microwave radiometry. Revised ed. Meteorological Press, Beijing, pp. 1–284.
J. Atmos. Ocean. Technol. 18, 925–933. Ware, R., Solheim, F., Carpenter, R., Güldner, J., Liljegren, J., Nehrkorn, T.,
Gurevich, A.V., Milikh, G.M., Roussel-Dupre, R., 1992. Runaway electron Vandenberghe, F., 2003. A multi-channel radiometric profiler of temper-
mechanism of air breakdown and preconditioning during a thunderstorm. ature, humidity and cloud liquid. Radio Sci. 38, 8079–8091.
Phys. Lett. A 165, 463–468. Ware, R.,Cimini, D.,Campos, E.,Giuliani, G.,Albers, S.,Nelson, M.,Koch, S.E.,Joe, P.,
Krider, Philip E., 2003. Deciphering the energetics of lightning. Science 299, Cober, S., 2013. Thermodynamic and liquid profiling during the 2010
669–670. Winter Olympics. Atmos. Res. 132, 278–290.
Liebe, H.J., Rosenkranz, P.W., Hufford, G.A., 1992. Atmospheric 60-GHz oxygen Westwater, E.R., Wang, Zhenhui, Grody, N.C., McMillin, L.M., 1985. Remote
spectrum: new laboratory measurements and line parameters. J. Quant. sensing of temperature profiles from a combination of observations from
Spectrosc. Radiat. Transf. 48, 629–643. the satellite-based microwave sounding unit and the ground-based
Ouyang, Y., Yuan, P., Jia, X., 2006. Multiple-line method used to calculate profiler. J. Atmos. Ocean. Technol. 2, 97–109.
lightning channel temperature. J. Northwest Norm. Univ. (Nat. Sci.) 42 (3), Westwater, E., Crewell, S., Matzler, C., 2004. A review of surface-based
49–53 (in Chinese). microwave and millimeter-wave radiometric remote sensing of the
Petersen, D., Beasley, W., 2014. Microwave radio emissions of negative cloud- troposphere. Radio Science Bulletin, No. 310, pp. 59–80 (Sept.).
to-ground lightning flashes. Atmos. Res. 135–136, 314–321. Xu, Guirong, Ware, Randolph (Stick), Zhang, Wengang, Feng, Guangliu, Liao,
Prueitt, M.L., 1963. The excitation temperature of lightning. Geophys. Res. 68, Kewen, Liu, Yibing, 2014. Effect of off-zenith observations on reducing the
803–811. impact of precipitation on ground-based microwave radiometer measure-
Ratnam, M.,Venkat, Y. Durga,Santhi, M.,Rajeevan, S. Vijaya,Rao, Bhaskara, 2013. ment accuracy. Atmos. Res. 140-141, 85–94.
Diurnal variability of stability indices observed using radiosonde observa- Yao, Z., Chen, H., 2005. Analysis of the remote sensing of atmospheric
tions over a tropical station: comparison with microwave radiometer temperature profiles using the seven channel microwave radiometer. Sci.
measurements. Atmos. Res. 124, 21–33. Meteorol. Sin. 25 (2), 133–141 (in Chinese).
Sánchez, J.L.,Posada, R.,García-Ortega, E.,López, L.,Marcos, J.L., 2013. A method Zhang, P., Wang, Z., 1995. Fundamentals of Atmospheric Microwave Remote
to improve the accuracy of continuous measuring of vertical profiles of Sensing. Meteorological Press, Beijing, pp. 1–412 (in Chinese).
temperature and water vapor density by means of a ground-based Zhou, Xiuji, 1982. Microwave Radiation in the Atmosphere and the Principle
microwave radiometer. Atmos. Res. 122, 43–54. for Its Remote Sensing. Science Press, Beijing, pp. 1–178 (in Chinese).
Solheim, F.,Godwin, J.,Westwater, E.,Han, Y.,Keihm, S.,Marsh, K.,Ware, R., 1998. Zhu, Y., Hu, C., Zhen, J., Zhao, B., 1994. The role of microwave radiometer in
Radiometric profiling of temperature, water vapor, and cloud liquid water weather modification research. Acta Sci. Nat. Univ. Pekin. 30 (5), 597–606
using various inversion methods. Radio Sci. 33 (2), 393–404. (in Chinese).

View publication stats