Abstract
EXtreme Gradient Boosting over Decision Trees (XGBoost or XGBDT) is a powerful tool to model a wide range of processes. We propose a new approach to create a global total electron content model, using machine-learning-based techniques, in particular, gradient boosting. The model is based on the Global Ionospheric Maps computed by Universitat Politècnica de Catalunya with a tomographic-kriging combined technique (UQRG). To reduce the problem complexity, we used empirical orthogonal functions (EOFs). The created model involves the first 16 spatial EOFs. For training and validation we used the 1998–2016 data sets, and the 2017 data as a test data set. To drive the model, we used the following features: (1) geomagnetic activity indexes (Kp, Ap, AE, AU, AL) and solar activity indexes (R, F10.7); (2) derivative values from these indexes such as the mean value and standard deviations within the last 12 h, last 11 days, and last 40 days; (3) day of the year (DOY); (4) averaged EOFs for given Kp and UT, and those for a given DOY and UT. The validation data set revealed the following hyperparameters for XGBoost learning: number of trees is 100, tree depth is 6, and learning rate is 0.1. Comparisons with the NeQuick2, Klobuchar, and GEMTEC models show that machine learning achieves higher accuracy for the 2017 test data set. The global averaged root-mean-square errors and mean absolute percentage errors were about 2.5 TECU and 19% for the nonlinear GIMLi-XGBDT model, about 4 TECU and 30–40% for NeQuick2, GEMTEC, and the linear model GIMLi-LM, and about 5.2 TECU and 73% for the Klobuchar model. A 4-fully-connected-layer artificial neural network provided a higher error (3.28 TECU and 27.7%) as compared to GIMLi-XGBDT. For all models mentioned, the error peaked in the equatorial anomaly region. The solar activity increase does not affect the error of the nonlinear GIMLi-XGBDT model. However, an increase in geomagnetic activity strongly affects that model.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Data availability
GIM data is available through ftp://cddis.gsfc.nasa.gov/gps/products/ionex. Indexes of solar and geomagnetic activity are available through OMNI database (https://omniweb.gsfc.nasa.gov/).
Change history
05 February 2021
A Correction to this paper has been published: https://doi.org/10.1007/s10291-020-01063-1
References
Bergstra JS, Bardenet R, Bengio Y, Kegl B (2011) Algorithms for hyper-parameter optimization. Advances in neural information processing systems: 2546–2554. https://hal.inria.fr/hal-00642998
Breiman L (2001) Random forests. Mach Learn 45(1):5–32. https://doi.org/10.1023/A:1010933404324
Cander LR, Milosavljevic MM, Stankovic SS, Tomasevic S (1998) Ionospheric forecasting technique by artificial neural network. Electron Lett 34(16):1573–1574. https://doi.org/10.1049/el:19981113
Cesaroni C, Spogli L, Aragon-Angel A, Fiocca M, Dear V, De Franceschi G, Romano V (2020) Neural network based model for global total electron content forecasting. J Space Weather Space Clim 10:11. https://doi.org/10.1051/swsc/2020013
Chen T, Guestrin C (2016) Xgboost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, pp 785–794. https://doi.org/10.1145/2939672.2939785
Cherrier N, Castaings T, Boulch A (2017) Forecasting ionospheric total electron content maps with deep neural networks. Proc Conf Big Data Space (BIDS), ESA workshop, 2017
Feng J, Han B, Zhao Z, Wang Z (2019) Total electron content empirical model. Remote Sens 11(6):706. https://doi.org/10.3390/rs11060706
Friedman JH (2001) Greedy function approximation: a gradient boosting machine. Ann Stat 29(5):1189–1232
Geurts P, Ernst D, Wehenkel L (2006) Extremely randomized trees. Mach Learn 63(1):3–42. https://doi.org/10.1007/s10994-006-6226-1
Golub GH, Van Loan CF (1996) The singular value decomposition and unitary matrices. Matrix Computations. Johns Hopkins University Press, Baltimore, MD, pp 70–71
Habarulema JB, McKinnell L-A, Opperman BDL (2011) Regional GPS TEC modeling; attempted spatial and temporal extrapolation of TEC using neural networks. J Geophys Res Space Phys 116(A4):A04314. https://doi.org/10.1029/2010JA016269
Hernandez-Pajares M, Juan JM, Sanz J, Orus R, Garcia-Rigo A, Feltens J, Komjathy A, Schaer SC, Krankowski A (2009) The IGS VTEC maps: a reliable source of ionospheric information since 1998. J Geod 83(3):263–275. https://doi.org/10.1007/s00190-008-0266-1
Hernandez-Pajares M, Juan JM, Sanz J (1997) Neural network modeling of the ionospheric electron content at global scale using GPS data. Radio Sci 32(3):1081–1089. https://doi.org/10.1029/97RS00431
Hochegger G, Nava B, Radicella S, Leitinger R (2000) A family of ionospheric models for different uses. Phys Chem Earth Part C 25(4):307–310. https://doi.org/10.1016/S1464-1917(00)00022-2
ICD-Galileo (2016) European GNSS (Galileo) open service–ionospheric correction algorithm for Galileo single frequency user. European Commission Tech Rep 1.2, November 2016. https://www.gsc-europa.eu/sites/default/files/sites/all/files/Galileo_Ionospheric_Model.pdf
ICD-GLONASS (2016) GLONASS. Interface control document. General description of code division multiple access signal system. Russian space systems, Tech. Rep., 2016. https://russianspacesystems.ru/wp-content/uploads/2016/08/ICD-GLONASS-CDMA-General.-Edition-1.0-2016.pdf
Ivanov V, Gefan G, Gorbachev O (2011) Global empirical modelling of the total electron content of the ionosphere for satellite radio navigation systems. J Atmos Sol-Terr Phys 73(13):1703–1707. https://doi.org/10.1016/j.jastp.2011.03.010
Jakowski N, Mayer C, Hoque MM, Wilken V (2011) Total electron content models and their use in ionosphere monitoring. Radio Sci. https://doi.org/10.1029/2010RS004620
Kingma D, Ba J (2015) Adam: a method for stochastic optimization. https://arxiv.org/abs/1412.6980
Klobuchar JA (1987) Ionospheric time-delay algorithm for single-frequency GPS users. IEEE T Aero Elec Sys AES 23(3):325–331. https://doi.org/10.1109/TAES.1987.310829
Li W, Zhao D, He C, Hu A, Zhang K (2020) Advanced machine learning optimized by the genetic algorithm in ionospheric models using long-term multi-instrument observations. Remote Sens 12(5):866. https://doi.org/10.3390/rs12050866
Llewellyn SK, Bent RB (1973) Documentation and description of the Bent ionospheric model. Technical report AFCRL-TR-73–0657, July 1973, AD. 772733
Miikkulainen R (2017) Topology of a neural network. In: Sammut C, Webb GI (eds) Encyclopedia of machine learning and data mining. Springer, Boston. https://doi.org/10.1007/978-1-4899-7687-1_843
Montenbruck O, Gonzalez Rodriguez B (2019) Nequick-G performance assessment for space applications. GPS Solut 24:13. https://doi.org/10.1007/s10291-019-0931-2
Nava B, Coisson P, Radicella SM (2008) A new version of the NeQuick ionosphere electron density model. J Atmos Sol-Ter Phys 70(15):1856–1862. https://doi.org/10.1016/j.jastp.2008.01.015
Orus-Perez R (2019) Using TensorFlow-based neural network to estimate GNSS single frequency ionospheric delay (IONONet). Adv Space Res 63(5):1607–1618. https://doi.org/10.1016/j.asr.2018.11.011
Orus-Perez R, Hernandez-Pajares M, Juan J, Sanz J (2005) Improvement of global ionospheric VTEC maps by using kriging interpolation technique. J Atmos Sol-Terr Phys 67(16):1598–1609. https://doi.org/10.1016/j.jastp.2005.07.017
Srivani I, Siva Vara Prasad G, Venkata Ratnam D (2019) A deep learning-based approach to forecast ionospheric delays for GPS signals. IEEE Geosci Remote S Lett 16(8):1180–1184. https://doi.org/10.1109/LGRS.2019.2895112
Wang N, Yuan Y, Li Z, Huo X (2016a) Improvement of Klobuchar model for GNSS single-frequency ionospheric delay corrections. Adv Space Res 57(7):1555–1569. https://doi.org/10.1016/j.asr.2016.01.010
Wang Y, Yao H, Zhao S (2016b) Auto-encoder based dimensionality reduction. Neurocomputing 184:232–242. https://doi.org/10.1016/j.neucom.2015.08.104
Yuan Y, Wang N, Li Z, Huo X (2019) The BeiDou global broadcast ionospheric delay correction model (BDGIM) and its preliminary performance evaluation results. Navigation 66(1):55–69. https://doi.org/10.1002/navi.292
Zhang Z, Pan S, Gao C, Zhao T, Gao W (2019) Support vector machine for regional ionospheric delay modeling. Sensors 19(13):2947. https://doi.org/10.3390/s19132947
Zhukov AV, Sidorov DN, Mylnikova AA, Yasyukevich YV (2018) Machine learning methodology for ionosphere total electron content nowcasting. Artif Intell 16(1):144–157
Acknowledgement
The authors thank D.A. Zatolokin for his help in preparing the GEMTEC and Klobuchar model data, and Dr. B. Nava for making the source code of NeQuick2 available. We acknowledge the Universitat Politècnica de Catalunya and the International GNSS Service for the GIM data and OMNI database on solar and geomagnetic activities. The study is supported by the Russian Foundation for Basic Research Grant No. 18-35-20038 and partly by the Ministry of Education and Science (Basic Research program II.16).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Zhukov, A.V., Yasyukevich, Y.V. & Bykov, A.E. GIMLi: Global Ionospheric total electron content model based on machine learning. GPS Solut 25, 19 (2021). https://doi.org/10.1007/s10291-020-01055-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10291-020-01055-1