# A Survey of Deep Learning-Based Lightning Prediction

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Numerical Prediction Methods

## 3. Traditional Machine Learning Methods

## 4. Deep Learning Methods

#### 4.1. Convolutional Neural Network Methods

#### 4.2. Recurrent Neural Network Methods

## 5. Hybrid Neural Network Methods

## 6. Discussion

## 7. Conclusions

## Author Contributions

## Funding

## Data Availability Statement

## Conflicts of Interest

## References

- Hinton, G.E.; Osindero, S.; Teh, Y.W. A fast learning algorithm for deep belief nets. Neural Comput.
**2006**, 18, 1527–1554. [Google Scholar] [CrossRef] - Taud, H.; Mas, J.F. Multilayer perceptron (MLP). In Geomatic Approaches for Modeling Land Change Scenarios; Camacho Olmedo, M., Paegelow, M., Mas, J.F., Escobar, F., Eds.; Springer: Cham, Switzerland, 2018; pp. 451–455. [Google Scholar] [CrossRef]
- LeCun, Y.; Bengio, Y.; Hinton, G. Deep learning. Nature
**2015**, 521, 436–444. [Google Scholar] [CrossRef] [PubMed] - Krogh, A. What are artificial neural networks? Nat. Biotechnol.
**2008**, 26, 195–197. [Google Scholar] [CrossRef] [PubMed] - Hinton, G.; Deng, L.; Yu, D.; Dahl, G.E.; Mohamed, A.R.; Jaitly, N.; Senior, A.; Vanhoucke, V.; Nguyen, P.; Sainath, T.N.; et al. Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. IEEE Signal Process. Mag.
**2012**, 29, 82–97. [Google Scholar] [CrossRef] - Skamarock, W.C.; Klemp, J.B.; Dudhia, J.; Gill, D.O.; Liu, Z.; Berner, J.; Wang, W.; Powers, J.G.; Duda, M.G.; Barker, D.; et al. A description of the advanced research WRF model version 4. In NCAR Technical Note NCAR/TN-556+STR; Mesoscale and Microscale Meteorology Laboratory, National Center for Atmospheric Research: Boulder, Colorado, USA, 2019; Volume 145. [Google Scholar]
- Yair, Y.; Lynn, B.; Price, C.; Kotroni, V.; Lagouvardos, K.; Morin, E.; Mugnai, A.; Llasat, M.D.C. Predicting the potential for lightning activity in Mediterranean storms based on the Weather Research and Forecasting (WRF) model dynamic and microphysical fields. J. Geophys. Res. Atmos.
**2010**, 115, D04205. [Google Scholar] [CrossRef] - Price, C.; Rind, D. A simple lightning parameterization for calculating global lightning distributions. J. Geophys. Res. Atmos.
**1992**, 97, 9919–9933. [Google Scholar] [CrossRef] - Price, C.; Rind, D. What determines the cloud-to-ground lightning fraction in thunderstorms? Geophys. Res. Lett.
**1993**, 20, 463–466. [Google Scholar] [CrossRef] - Price, C.; Rind, D. Possible implications of global climate change on global lightning distributions and frequencies. J. Geophys. Res. Atmos.
**1994**, 99, 10823–10831. [Google Scholar] - Lynn, B.; Yair, Y. Prediction of lightning flash density with the WRF model. Adv. Geosci.
**2010**, 23, 11–16. [Google Scholar] [CrossRef] - Gharaylou, M.; Farahani, M.M.; Hosseini, M.; Mahmoudian, A. Numerical study of performance of two lightning prediction methods based on: Lightning Potential Index (LPI) and electric POTential difference (POT) over Tehran area. J. Atmos. Sol.-Terr. Phys.
**2019**, 193, 105067. [Google Scholar] [CrossRef] - Jordan, M.I.; Mitchell, T.M. Machine learning: Trends, perspectives, and prospects. Science
**2015**, 349, 255–260. [Google Scholar] [CrossRef] [PubMed] - Kingsford, C.; Salzberg, S.L. What are decision trees? Nat. Biotechnol.
**2008**, 26, 1011–1013. [Google Scholar] [CrossRef] [PubMed] - Keerthi, S.S.; Shevade, S.K.; Bhattacharyya, C.; Murthy, K.R.K. Improvements to Platt’s SMO algorithm for SVM classifier design. Neural Comput.
**2001**, 13, 637–649. [Google Scholar] [CrossRef] - Cutler, A.; Cutler, D.R.; Stevens, J.R. Random forests. In Ensemble Machine Learning Methods and Applications; Springer: New York, NY, USA, 2012; pp. 157–175. [Google Scholar]
- Webb, G.I.; Keogh, E.; Miikkulainen, R. Naïve Bayes. Encycl. Mach. Learn.
**2010**, 15, 713–714. [Google Scholar] - Azad, M.A.K.; Islam, A.R.M.T.; Rahman, M.S.; Ayen, K. Development of novel hybrid machine learning models for monthly thunderstorm frequency prediction over Bangladesh. Nat. Hazards
**2021**, 108, 1109–1135. [Google Scholar] [CrossRef] - Wang, W.C.; Chau, K.W.; Xu, D.M.; Chen, X.Y. Improving forecasting accuracy of annual runoff time series using ARIMA based on EEMD decomposition. Water Resour. Manag.
**2015**, 29, 2655–2675. [Google Scholar] [CrossRef] - Sharpley, R.C.; Vatchev, V. Analysis of the intrinsic mode functions. Constr. Approx.
**2006**, 24, 17–47. [Google Scholar] [CrossRef] - Schön, C.; Dittrich, J.; Müller, R. The error is the feature: How to forecast lightning using a model prediction error. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Anchorage, AK, USA, 4–8 August 2019; pp. 2979–2988. [Google Scholar]
- Pakdaman, M.; Naghab, S.S.; Khazanedari, L.; Malbousi, S.; Falamarzi, Y. Lightning prediction using an ensemble learning approach for northeast of Iran. J. Atmos. Sol.-Terr. Phys.
**2020**, 209, 105417. [Google Scholar] [CrossRef] - Johari, D.; Rahman, T.K.A.; Musirin, I. Artificial neural network based technique for lightning prediction. In Proceedings of the 2007 5th Student Conference on Research and Development, Selangor, Malaysia, 11–12 December 2007; pp. 1–5. [Google Scholar]
- Moon, S.H.; Kim, Y.H. Forecasting lightning around the Korean Peninsula by postprocessing ECMWF data using SVMs and undersampling. Atmos. Res.
**2020**, 243, 105026. [Google Scholar] [CrossRef] - Gu, J.; Wang, Z.; Kuen, J.; Ma, L.; Shahroudy, A.; Shuai, B.; Liu, T.; Wang, X.; Wang, G.; Cai, J.; et al. Recent advances in convolutional neural networks. Pattern Recognit.
**2018**, 77, 354–377. [Google Scholar] [CrossRef] - Brodehl, S.; Müller, R.; Schömer, E. End-to-End Prediction of Lightning Events from Geostationary Satellite Images. Remote Sens.
**2022**, 14, 3760. [Google Scholar] [CrossRef] - Ronneberger, O.; Fischer, P.; Brox, T. U-net: Convolutional networks for biomedical image segmentation. In Proceedings of the Medical Image Computing and Computer-Assisted Intervention—MICCAI 2015: 18th International Conference, Munich, Germany, 5–9 October 2015; Proceedings, Part III 18. Springer International Publishing: Berlin/Heidelberg, Germany, 2015; pp. 234–241. [Google Scholar]
- He, K.; Zhang, X.; Ren, S.; Sun, J. Identity mappings in deep residual networks. In Proceedings of the Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, 11–14 October 2016; Proceedings, Part IV 14. Springer International Publishing: Berlin/Heidelberg, Germany, 2016; pp. 630–645. [Google Scholar]
- Bao, R.; Zhang, Y.; Ma, B.J.; Zhang, Z.; He, Z. An Artificial Neural Network for Lightning Prediction Based on Atmospheric Electric Field Observations. Remote Sens.
**2022**, 14, 4131. [Google Scholar] [CrossRef] - Kullback, S.; Leibler, R.A. On information and sufficiency. Ann. Math. Stat.
**1951**, 22, 79–86. [Google Scholar] [CrossRef] - Sashiomarda, J.A.; Aji, A.S.; Handoko, D. Implementation of convolutional neural networks to determine lightning location. J. Phys. Conf. Ser.
**2020**, 1528, 012009. [Google Scholar] [CrossRef] - Lu, M.; Zhang, Y.; Ma, Z.; Yu, M.; Chen, M.; Zheng, J.; Wang, M. Lightning strike location identification based on 3D weather radar data. Front. Environ. Sci.
**2021**, 9, 714067. [Google Scholar] [CrossRef] - LaValley, M.P. Logistic regression. Circulation
**2008**, 117, 2395–2399. [Google Scholar] [CrossRef] - Peterson, L.E. K-nearest neighbor. Scholarpedia
**2009**, 4, 1883. [Google Scholar] [CrossRef] - Saito, T.; Rehmsmeier, M. The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLoS ONE
**2015**, 10, e0118432. [Google Scholar] [CrossRef] - Graves, A. Long short-term memory. In Supervised Sequence Labelling with Recurrent Neural Networks. Studies in Computational Intelligence; Springer: Berlin/Heidelberg, Germany, 2012; pp. 37–45. [Google Scholar] [CrossRef]
- Bao, R.; He, Z.; Zhang, Z. Application of lightning spatio-temporal localization method based on deep LSTM and interpolation. Measurement
**2022**, 189, 110549. [Google Scholar] [CrossRef] - Wackernagel, H. Ordinary kriging. In Multivariate Geostatistics: An Introduction with Applications; Springer Science and Business Media: Berlin/Heidelberg, Germany, 2003; pp. 79–88. [Google Scholar] [CrossRef]
- Fukawa, M.; Deng, X.; Imai, S.; Horiguchi, T.; Ono, R.; Rachi, I.; A, S.; Shinomura, K.; Niwa, S.; Kudo, T.; et al. A Novel Method for Lightning Prediction by Direct Electric Field Measurements at the Ground Using Recurrent Neural Network. Ieice Trans. Inf. Syst.
**2022**, 105, 1624–1628. [Google Scholar] [CrossRef] - Shi, X.; Chen, Z.; Wang, H.; Yeung, D.Y.; Wong, W.K.; Woo, W.C. Convolutional LSTM network: A machine learning approach for precipitation nowcasting. Adv. Neural Inf. Process. Syst.
**2015**, 28, 802–810. [Google Scholar] - Geng, Y.A.; Li, Q.; Lin, T.; Jiang, L.; Xu, L.; Zheng, D.; Yao, W.; Lyu, W.; Zhang, Y. Lightnet: A dual spatiotemporal encoder network model for lightning prediction. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Anchorage, AK, USA, 4–8 August 2019; pp. 2439–2447. [Google Scholar]
- Geng, Y.A.; Li, Q.; Lin, T.; Zhang, J.; Xu, L.; Yao, W.; Zheng, D.; Lyu, W.; Huang, H. A heterogeneous spatiotemporal network for lightning prediction. In Proceedings of the 2020 IEEE International Conference on Data Mining (ICDM), Sorrento, Italy, 17–20 November 2020; pp. 1034–1039. [Google Scholar]
- Essa, Y.; Ajoodha, R.; Hunt, H.G.P. A LSTM recurrent neural network for lightning flash prediction within southern africa using historical time-series data. In Proceedings of the 2020 IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE), Gold Coast, Australia, 16–18 November 2020; pp. 1–6. [Google Scholar]
- Guastavino, S.; Piana, M.; Tizzi, M.; Cassola, F.; Iengo, A.; Sacchetti, D.; Solazzo, E.; Benvenuto, F. Prediction of severe thunderstorm events with ensemble deep learning and radar data. Sci. Rep.
**2022**, 12, 20049. [Google Scholar] [CrossRef] [PubMed] - Zhou, X.; Geng, Y.A.; Yu, H.; Li, Q.; Xu, L.; Yao, W.; Zheng, D.; Zhang, Y. LightNet+: A dual-source lightning forecasting network with bi-direction spatiotemporal transformation. Appl. Intell.
**2022**, 52, 11147–11159. [Google Scholar] [CrossRef]

**Figure 1.**End-to-end CNN-based lightning prediction network architecture [26].

**Figure 2.**Time-series observation data visualization: feature values converted into numbers from 0 to 255 corresponding to a binary image [29].

**Figure 3.**CNN architecture [32].

**Figure 4.**Typical LSTM block [36]. Graves, A. Long short-term memory. Supervised Seq. Label. Recurr. Neural Netw. 2012, 37–45, Springer Link, reproceduced with permission from SNCSC.

**Figure 5.**Network architecture [39]. Fukawa, M.; Deng, X.; Imai, S.; Horiguchi, T.; Ono, R.; Rachi, I.; A, S.; Shinomura, K.; Niwa, S.; Kudo, T.; et al. A Novel Method for Lightning Prediction by Direct Electric Field Measurements at the Ground Using Recurrent Neural Network. Ieice Trans. Inf. Syst. 2022, 105, 1624–1628, 8 June 2022, Copyright (c) 2022 IEICE.

**Figure 6.**HSTN architecture [42]. 2020 IEEE. Reprinted, with permission from Geng, Y.A.; Li, Q.; Lin, T.; Zhang, J.; Xu, L.; Yao, W.; Zheng, D.; Lyu, W.; Huang, H. A heterogeneous spatiotemporal network for lightning prediction. In Proceedings of the 2020 IEEE International Conference on Data Mining (ICDM), Sorrento, Italy, 17–20 November 2020; pp. 1034–1039.

**Figure 7.**DCLSTM architecture: PE represents the result after subsampling [45]. Zhou, X.; Geng, Y.A.; Yu, H.; Li, Q.; Xu, L.; Yao,W.; Zheng, D.; Zhang, Y. LightNet+: A dual-source lightning forecasting network with bi-direction spatiotemporal transformation. Appl. Intell. 2022, 52, 11147–11159, reproduced with permissions from SNCSC.

**Table 1.**LightNet architecture [41]. Used with permission of ResearchGate from Geng, Y.A.; Li, Q.; Lin, T.; Jiang, L.; Xu, L.; Zheng, D.; Yao,W.; Lyu,W.; Zhang, Y. Lightnet: A dual spatiotemporal encoder network model for lightning prediction. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Anchorage, AK, USA, 4–8 August 2019; pp. 2439–2447, 2019, permission conveyed through Copyright Clearance Center, Inc.

Module | Notation | Size | Stride |
---|---|---|---|

WRF Encoder | $Con{v}_{1}$ | [7 × 7, 64] | 2 |

ConvLSTM | [5 × 5, 128] | 1 | |

Obs. Encoder | $Con{v}_{2}$ | [7 × 7, 4] | 2 |

ConvLSTM | [5 × 5, 8] | 1 | |

Fusion Module | $Con{v}_{3}$ | [1 × 1, 64] | 1 |

$Con{v}_{4}$ | [1 × 1, 64] | 1 | |

Pred. Decoder | $Con{v}_{5}$ | [7 × 7, 4] | 2 |

ConvLSTM | [5 × 5, 64] | 1 | |

DeConv | [7 × 7, 64] | 2 | |

$Con{v}_{6}$ | [1 × 1, 1] | 1 |

**Table 2.**Results of [44].

Strategy | TSS | CSI | wFP | wFN | wTSS | wCSI |
---|---|---|---|---|---|---|

wTSS | 0.78 (±0.04) | 0.17 (±0.02) | 243.88 (±41.34) | 6.79 (±1.64) | 0.68 (±0.04) | 0.10 (±0.02) |

TSS | 0.77 (±0.05) | 0.17 (±0.03) | 240.99 (±60.57) | 7.24 (±2.60) | 0.67 (±0.06) | 0.10 (±0.02) |

Methods | Generalize | Specific Methods in Articles | Data Source |
---|---|---|---|

Numerical Prediction Methods | In accordance with the principles of lightning genesis, relevant observation parameters are used to calculate the lightning potential index (LPI) and positive ratio (PR), determining the probability of lightning occurrence. | LPI [7] | Total mass flux of liquid water and ice |

PR92 [8] | The intensity of updrafts and the intensity of updrafts | ||

PR93 [9] | Radiosonde data | ||

PR94 [10] | Global convective cloud data | ||

Grid LPI [11] | Total mass flux of liquid water and ice from 1 to 4 km | ||

POT + LPI [12] | Total mass flux of liquid water and ice | ||

Traditional Machine Learning Methods | Employing manual calculations, key lightning data features are hand-extracted, and then traditional machine learning methods such as support vector machines (SVM) and simple artificial neural networks (simple ANN) are utilized based on the extracted features for classification. | EEMD + SVM/ANN [18] | Observation meteorological parameters |

Gradient boosting machine learning [21] | Binary meteosat satellite images and lightning detected data | ||

Undersampling + shallow neural network/DT [22] | Observation meteorological parameters | ||

Back-propagation neural network [23] | Observation meteorological parameters | ||

Undersampling + SVM/RF [24] | Observation meteorological parameters | ||

Convolutional Neural Network Methods | Integrating image data such as satellite images, electromagnetic and acoustic signals, or converting lightning data into image form allows the utilization of convolution to extract features and conduct predictions. | U-Net + ResNet-v2 [26] | Geostationary satellite images |

KL + CNN + MLP [29] | Time-series observation meteorological parameters | ||

Spectrograms + CNN [31] | Random noise, lightning sounds, and background noise | ||

Sliding window + CNN [32] | 3D weather radar data | ||

Recurrent Neural Network Methods | The most commonly used deep learning method for lightning prediction primarily processes sequential data, often combining with existing methodologies in a variant form to achieve more accurate predictions. | OK + LSTM [37] | Time-series electric field data |

Convolution + LSTM [39] | Time-series electric field data | ||

ConvLSTM [40] | Spatio-temporal data | ||

LightNet (WRF + ConvLSTM) [41] | Spatio-temporal WRF simulation data | ||

HSTN (Gaussian diffusion + WRF + CNN + ConvLSTM) [42] | The weather station observation data, real four-dimensional data, the WRF simulation data and binary three-dimensional data | ||

LSTM [43] | Lightning flash density | ||

Hybrid Neural Network Methods | Combining multiple neural network models often involves the use of one network model for feature extraction and another for prediction. With more information contained in the features, this approach aids in improving the accuracy of predictions. | LRCN (CNN + LSTM) [44] | Radar data |

CNN + ConvLSTM + Dual ConvLSTM (DCLSTM) + Deformable CNN + deConv [45] | WRF simulation data |

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |

© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Wang, X.; Hu, K.; Wu, Y.; Zhou, W.
A Survey of Deep Learning-Based Lightning Prediction. *Atmosphere* **2023**, *14*, 1698.
https://doi.org/10.3390/atmos14111698

**AMA Style**

Wang X, Hu K, Wu Y, Zhou W.
A Survey of Deep Learning-Based Lightning Prediction. *Atmosphere*. 2023; 14(11):1698.
https://doi.org/10.3390/atmos14111698

**Chicago/Turabian Style**

Wang, Xupeng, Keyong Hu, Yongling Wu, and Wei Zhou.
2023. "A Survey of Deep Learning-Based Lightning Prediction" *Atmosphere* 14, no. 11: 1698.
https://doi.org/10.3390/atmos14111698