Estimation of precipitable water vapor (PWV) using generalized regression neural network (GRNN) and comparison against tomography, ECMWF, Saastamoinen, GPT3 and ANN models

Document Type : Research Article


1 Corresponding Author, Department of Surveying Engineering, Faculty of Geoscience Engineering, Arak University of Technology, Arak, Iran. E-mail:

2 Department of Surveying Engineering, Faculty of Civil Engineering, Islamic Azad University of Khoy, Khoy, Iran. E-mail:

3 Department of Surveying Engineering, Faculty of Geoscience Engineering, Arak University of Technology, Arak, Iran. E-mail:


Precipitable water vapor (PWV) is a key parameter in meteorological studies and forecasting of atmospheric events such as rain and flood. Due to the spatial limitations of GPS and meteorological stations, as well as observational discontinuities in the time domain, PWV modeling is of great importance. Obtaining PWV using direct measurements and water vapor measuring devices is a difficult task. The best way to get information on PWV variations indirectly is to use GNSS measurements. The GNSS meteorological technique can provide continuous and almost instantaneous observations of the amount of PWV around a GNSS station. Research has shown that the accuracy of weather forecasts can be improved using GNSS-dependent techniques. Models based on GNSS observations for estimating PWV are known as tropospheric analytical models. The tomographic model is one of the most famous and widely used tropospheric models. There are limitations such as a large number of unknown parameters, rank deficiency of design matrix and the inevitability of using regularization methods, assuming the amount of water vapor inside each voxel is constant and also, the need for initial amounts of water vapor inside the voxels in the voxel-based tomography (VBT) method. Such limitations have led researchers to use machine learning methods to estimate the spatio-temporal variation of PWV.
In this paper, the spatio-temporal modeling of PWV is suggested using the generalized regression neural network (GRNN) model. The GRNN model is a type of artificial neural network (ANN) that uses radial basis functions (RBF) as an activity function in the hidden layer. As a result, its accuracy is higher than the ANN model. Eight parameters of longitude, latitude and height of GPS station, day of year (DOY), time (min.), relative humidity (RH), temperature (T) and pressure (P) are considered as inputs of GRNN and ANN models and the PWVs corresponding to these eight parameters are the outputs. After the training step, to evaluate the GRNN and ANN models, the observations of two GPS networks are used. In the GPS network of north-west of Iran, observations of 23 GPS stations in the period of 300 to 314 (winter season) from 2011 have been used. For the central Alborz GPS network, observations of 11 stations at the period of 162 to 176 (summer season) in 2016 are also used. Results obtained from GRNN and ANN models in two interior control stations, one exterior control station (outside the GPS network territory) and also in Tabriz and Tehran radiosonde stations are compared and evaluated with the results of VBT, ECMWF, Saastamoinen and GPT3 models. The statistical parameters of root mean square error (RMSE), relative error and correlation coefficient (R) are used to evaluate the accuracy of the models. At the north-west GPS network, the averaged RMSE values of GRNN, ANN, VBT, ECMWF, Saastamoinen and GPT3 models in the two interior control stations are calculated as 2.14, 2.57, 3.32, 3.36, 6.31 and 4.35 mm, respectively. For the central Alborz GPS network, the averaged RMSE of two interior control stations are computed as 2.01, 2.42, 3.24, 3.26, 6.00 and 4.06 mm, respectively. For the exterior control station, the GRNN model has less error than the ANN, VBT and Saastamoinen models, but more than the ECMWF and GPT3 model. The results of this paper show that the GRNN model has a very high accuracy compared to other analytical and empirical models of the troposphere. This model has the ability to show the spatio-temporal variations of precipitable water vapor with high accuracy at the GPS network territory and; it can considered as an alternative for the other analytical and empirical models.


Main Subjects

حاجی آقاجانی، س. و وثوقی، ب. (1395). مقایسه داده‌های بازتحلیل ERA-Interim و حسگر MERIS در کاهش اثر لایة وردسپهر موجود در میدان‌های سرعت جابه‌جایی تداخل‌سنجی راداری. مجله فیزیک زمین و فضا، 42(3)،607-618.
موسوی، ز.، خرمی، ف.، نانکلی، ح.ر. و جمور، ی. (1386). تعیین مقدار بخارآب موجود در جو با استفاده از تخمین تأخیر وردسپهری سیگنال‌های جی‌پی‌اس در شبکه ژئودینامیک سراسری ایران، همایش ژئوماتیک 1386.
خسروی، م.، حسینی، ا. و قاسمی، ج. (1399). ارزیابی عملکرد منابع انسانی با رویکرد شبکه عصبی رگرسیون عمومی موردمطالعه: اعضای هیات علمی. دو فصلنامه مطالعات برنامه‌ریزی آموزشی، 9(17)، 202-222.
غفاری رزین، س.ر. (1399). ارزیابی کارایی سامانه استنتاج عصبی-فازی سازگار در مدل‌سازی بخار آب مایل وردسپهر. مجله فیزیک زمین و فضا، 47(2)، 257-272.
Aster, R., Borchers, B., & Thurber, C. (2003). Parameter estimation and inverse problems, Elsevier Academic Press, USA.
Adavi, Z., & Mashhadi-Hossainali, M. (2015). 4D-tomographic reconstruction of water vapor using the hybrid regularization technique with application to the North West of Iran. Advances in Space Research, 55(7), 1845-1854.
Askne, J., & Nordius, H. (1987). Estimation of tropospheric delay for microwaves from surface weather data. Radio Science, 22(3), 379-386.
Benevides, P., Catalao, J., Nico, G., & Miranda, P. (2018), 4D wet refractivity estimation in the atmosphere using GNSS tomography initialized by radiosonde and AIRS measurements: results from a 1-week intensive campaign. GPS Solutions, 91(2018): 22:91.
Chen, B., & Liu, Z. (2014). Voxel-optimized regional water vapor tomography and comparison with radiosonde and numerical weather model. Journal of Geodesy, 88(7), 691–703.
Davis, J.L., Herring, T.A., Shapiro, II., Rogers, E.E., & Elgered, G. (1985). Geodesy by radio interferometry: effects of atmospheric modeling errors on estimates of baseline length. Radio Sci, 20(6), 1593–1607.
Dach, R., Hugentobler, U., Fridez, P., & Meindl, M. (2007). Bernese GPS Software Version 5.0. Astronomical Institute, University of Bern, Bern.
European centre for Medium-Range Weather Forecasts (ECMWF) 2005, 2013, User Guide European center for Medium-Range Weather Forecasts (ECMWF), User Guide.
Forootan, E., Dehvari, M., Farzaneh, S., & Sam Khaniani, A. (2021). A functional modelling approach for reconstructing 3 and 4 dimensional wet refractivity fields in the lower atmosphere using GNSS measurements. Advances in Space Research, 68(10), 4024-4038.
Ghaffari Razin, M.R., & Voosoghi, B. (2022). Modeling of precipitable water vapor from GPS observations using machine learning and tomography methods. Advances in Space Research, 69(7), 2671-2681.
Ghaffari Razin, M.R., & Inyurt, S. (2022). Spatiotemporal analysis of precipitable water vapor using ANFIS and comparison against voxel‑based tomography and radiosonde. GPS Solutions. 26, 1,
Ghaffari Razin, M.R., & Voosoghi, B. (2020). Estimation of tropospheric wet refractivity using tomography method and artificial neural networks in Iranian case study. GPS Solutions 24(3), 1-14.
Ghritlahre, H., & Prasad, R. (2018). Investigation of thermal performance of unidirectional flow porous bed solar air heater using MLP, GRNN, and RBF models of ANN technique. Thermal Science and Engineering Progress. doi:
Haji‑Aghajany, S., Amerian, Y., Verhagen, S., Rohm, W., & Schuh, H. (2021). The effect of function‑based and voxel‑based tropospheric tomography techniques on the GNSS positioning accuracy. Journal of Geodesy, 95(78), 1-15.
Haji Aghajany, S., & Amerian, Y. (2017). Three dimensional ray tracing technique for tropospheric water vapor tomography using GPS measurements. Journal of Atmospheric and Solar-Terrestrial Physics, 164 (2017), 81-88.
Haji Aghajany, S., Amerian, Y., & Verhagen, S. (2020). B-spline function-based approach for GPS tropospheric tomography. GPS Solutions, 24(3), 1-12.
Haykin, S. (1994). Neural Networks, a comprehensive Foundation, Macmillan College Publishing Company, New York, 1994.
Kim, B., Lee, D.W., Park, K.Y., Choi, S.R., & Choi, S. (2004). Prediction of plasma etching using a randomized generalized regression neural network. Vacuum, 76(1), 37–43.
Landskron, D., & Böhm, J. (2017). VMF3/GPT3: refined discrete and empirical troposphere mapping functions. Journal of Geodesy, 92(4), 349–360.
Masikos, M., Demestichas, K., Adamopoulou, E., & Theologou, M. (2015). Mesoscopic forecasting of vehicular consumption using neural networks. Soft Computing, 19(1), 145-156.
Rohm, W., & Bosy, J. (2011). The verification of GNSS tropospheric tomography model in a mountainous area. Adv. Space Res., 47(10), 1721-1730
Sam Khaniani, A., Motieyan, H., & Mohammadi, A. (2021). Rainfall forecast based on GPS PWV together with meteorological parameters using neural network models. Journal of Atmospheric and Solar-Terrestrial Physics, 214(2021), 1-15.
Sam Khaniani, A., Azadi, M., & Zakeri, Z. (2017). Impact of Iranian permanent GPS network precipitable water estimates on numerical weather prediction. Earth Observation and Geomatics Engineering, 1(2), 100–111.
Sadeghi, E., Hossainali, M., & Safari, A. (2022). Development of a hybrid tomography model based on principal component analysis of the atmospheric dynamics and GPS tracking data. GPS Solutions, 26(77), 1-14.
Seeber, G. (2003). Satellite Geodesy, Foundations, Methods and Application, Walter de Gruyter, Berlin and New York, 531.
Selbesoglu, M. O. (2019). Prediction of tropospheric wet delay by an artificial neural network model based on meteorological and GNSS data. Engineering Science and Technology, an International Journal, 23(5), 967-972. DOI: 10.1016/j.jestch.2019.11.006.
Sadeghi, E., Hossainali, M., & Etemadfard, H. (2014). Determining precipitable water in the atmosphere of Iran based on GPS zenith tropospheric delays. Annals of geophysics, 57(4), 12-27.
Saastamoinen, J. (1973). Contributions to the theory of atmospheric refraction. Part II: refraction corrections in satellite geodesy. Bull. Geod, (107), 13-34.
Specht, D. F. (1991). A general regression neural network. IEEE Transactions on Neural Networks, 2(6), 568-576.
Xin, W., & Daren, L. (2005). Retrieval of Water Vapor Profiles with Radio Occultation Measurements Using an Artificial Neural Network. Advances in atmospheric science, 22(5), 759-764.
Yao, Y., Liu, C., Xu, C., Tan, Y., & Fang, M. (2020). A Refined Tomographic Window for GNSS-Derived Water Vapor Tomography. Remote Sens. 12(18), 1-15. doi:10.3390/rs12182999.
Ye, S., Xia, P., & Cai, C. (2016). Optimization of GPS water vapor tomography technique with radiosonde and COSMIC historical data. Ann. Geophys., 34, 789–799.
Yao, Y., Xin, L., & Zhao, O. (2019). An improved pixel-based water vapor tomography model. Ann. Geophys, (37), 89–100.
Yuan, Q., Xu, H., Li, T., Shen, H., & Zhang, L. (2020). Estimating surface soil moisture from satellite observations using a generalized regression neural network trained on sparse ground-based measurements in the continental US. J Hydrol, 580(2020), 1-14.
Zhang, W., Zhang, S., Ding, N., Holden, L., Wang, X., & Zheng, N. (2021). GNSS-RS Tomography: Retrieval of Tropospheric Water Vapor Fields Using GNSS and RS Observations. IEEE Transactions on Geoscience and Remote Sensing, (60), 1-13.
Zhao, Q., Du, Z., Yao, W., & Yao, Y. (2020). Hybrid precipitable water vapor fusion model in China. Journal of Atmospheric and Solar-Terrestrial Physics, 208(2020), 1-13.
Zhao, Q., Yao, Y., Cao, X., & Yao, W. (2019). Accuracy and reliability of tropospheric wet refractivity tomography with GPS, BDS, and GLONASS observations. Advances in Space Research, 63(9), 2836-2847.
Zheng, D., Hu, Y., Wang, W.S.J., & Zhu, M. C. (2015). Research on regional zenith tropospheric delay based on neural network technology. Survey Review, 47(343), 286-295.