Calibration of MODIS Water vapor data using multi-Layer perceptron neural ‎network and radiosonde observations ‎

Ghobadi, K.; Javadnia, E.; Jalili, H.; Zandkarimi, A.

doi:10.22061/jrsgr.2025.11605.1090

Articles in Press

Document Type : Original Research Paper

Authors

¹ Department of Geomatics Engineering, Faculty of Engineering, University of Zanjan, Zanjan, Iran

² Iranian Space Research Center, Tehran, Iran

https://doi.org/10.22061/jrsgr.2025.11605.1090

Abstract

Background and Objectives: Water vapor in the atmosphere is one of the most critical meteorological parameters, significantly influencing climate studies, weather forecasting, and climate change modeling. Precipitable Water Vapor (PWV) serves as a key indicator in atmospheric studies and is measured using satellite data, including Moderate Resolution Imaging Spectroradiometer (MODIS) sensor products. MODIS provides two PWV products: Near-Infrared (NIR), which is available only during the day, and Infrared (IR), which provides data for both day and night. Due to its broader temporal coverage, the IR product is widely utilized in various applications. However, the accuracy of this product, especially under varying atmospheric conditions during day and night, remains a major challenge. This study aims to enhance the accuracy of MODIS IR PWV data using machine learning and to assess the calibration's impact on day and night conditions.
Methods: This study utilized data from 10 radiosonde stations in Iran during the 2019-2020 period as reference ground-truth data. Three datasets were prepared: raw MODIS data, fitted data, and modified data. A Multi-Layer Perceptron (MLP) model was employed for calibration and to evaluate its performance for day and night data separately. Standard machine learning methods were applied to design and implement the model. The model's accuracy was evaluated using the Root Mean Square Error (RMSE) and correlation coefficient (R) metrics.
Findings: The results demonstrated that the MLP model significantly improved the accuracy of MODIS PWV data. During the day, RMSE decreased from 3.72 mm in the raw data to 2.63 mm in the calibrated model, while the correlation coefficient increased from 0.81 to 0.86. At night, RMSE reduced from 4.9 mm to 3.16 mm, and the correlation coefficient improved from 0.76 to 0.78. Overall, RMSE in raw MODIS data was 4.48 mm, which was reduced to 2.92 mm in the fitted model and 3.03 mm in the modified model. The correlation coefficient also improved from 0.77 to 0.87 and 0.85, respectively.
Conclusion: This study confirmed that the MLP model effectively enhances the accuracy of MODIS PWV data and reduces existing errors under different atmospheric conditions. The primary innovation of this research is the application of the MLP model to calibrate satellite-derived PWV data for day and night conditions. By improving the precision of satellite data, this method enhances its reliability for practical applications, particularly in weather forecasting and climate studies. Limitations include dependency on radiosonde data as the reference and the absence of analysis on specific atmospheric factors influencing modeling. This approach can also be

Keywords

Main Subjects

Remote Sensing

COPYRIGHTS
© 2025 The Author(s). This is an open-access article distributed under the terms and conditions of the Creative Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) (https://creativecommons.org/licenses/by-nc/4.0/)

References

[1] Bevis, M., et al., GPS meteorology: Remote sensing of atmospheric water vapor using the global positioning system. Journal of Geophysical Research: Atmospheres, 1992. 97(D14): p. 15787-15801. DOI: 10.1029/92JD01517

[2] Mockler, S., Water vapor in the climate system, special report, American Geophysical Union (AGU), 2000 Florida Ave., NW, Washington, DC 20009. 1995, December.

[3] Gao, B.-C., et al., Measurements of water vapor and high clouds over the Tibetan Plateau with the Terra MODIS instrument. IEEE Transactions on geoscience and remote sensing, 2003. 41(4): p. 895-900.

DOI: 10.1109/TGRS.2003.810704

[4] Held, I.M. and B.J. Soden, Water vapor feedback and global warming. Annual review of energy and the environment, 2000. 25(1): p. 441-475. DOI: 10.1146/annurev.energy.25.1.441

[5] urre, I., R.S. Vose, and D.B. Wuertz, Overview of the integrated global radiosonde archive. Journal of Climate, 2006. 19(1): p. 53-68. DOI: 10.1175/JCLI3594.1

[6] Vaquero-Martínez, J. and M. Antón, Review on the role of GNSS meteorology in monitoring water vapor for atmospheric physics. Remote Sensing, 2021. 13(12): p. 2287.

DOI: 10.3390/rs13122287

[7] Gao, B.-C. and Y.J. Kaufman, The MODIS near-IR water vapor algorithm. Algorithm Theoretical Basis Document, ATBD-MOD, 1998. 5.

[8] Van Malderen, R., et al., A multi-site intercomparison of integrated water vapour observations for climate change analysis. Atmospheric Measurement Techniques, 2014. 7(8): p. 2487-2512. DOI: 10.5194/amt-7-2487-2014

[9] Javadnia, E., N. Beygi, and S. Abbasy, Comparison and Evaluation of MODIS and Sun photometer Integrated water vapor using GPS observations at IASBS AERONET site (Zanjan, Iran). Journal of Geomatics Science and Technology, 2021. 11(1): p. 117-128.

https://dor.isc.ac/dor/20.1001.1.2322102.1400.11.1.9.7

[10] Liu, H., et al., Evaluation of MODIS water vapour products over China using radiosonde data. International Journal of Remote Sensing, 2015. 36(2): p. 680-690.

DOI: 10.1080/01431161.2014.999880

[11] Norouzi Razieh, J.E., Validation of precipitable water vapor derived from MODIS infrared product using GPS observations, in The Second National Conference on Data Mining in Earth Sciences. 2021: Arak, Iran.

[12] Lindenbergh, R., et al. Validating time series of a combined GPS and MERIS Integrated Water Vapor product. in Proceedings 2nd MERIS/(A) ATSR User Workshop. 2008.

[13] Li, Z. Production of Regional 1 km× 1 km Water Vapor Fields through the Integration of GPS and MODIS Data. in Proceedings of the 17th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2004). 2004.

[14] Zhang, B., et al., Precipitable water vapor fusion: an approach based on spherical cap harmonic analysis and Helmert variance component estimation. Journal of Geodesy, 2019. 93(12): p. 2605-2620. DOI: 10.1007/s00190-019-01322-0

[15] Zhao, Q., et al., Precipitable water vapor fusion method based on artificial neural network. Advances in Space Research, 2022. 70(1): p. 85-95.

DOI: 10.1016/j.asr.2022.04.016

[16] Zhang, B. and Y. Yao, Precipitable water vapor fusion based on a generalized regression neural network. Journal of geodesy, 2021. 95(3): p. 36. DOI: 10.1007/s00190-021-01482-3

[17] Xu, J. and Z. Liu, Enhanced all-weather precipitable water vapor retrieval from MODIS near-infrared bands using machine learning. International Journal of Applied Earth Observation and Geoinformation, 2022. 114: p. 103050.

DOI: 10.1016/j.jag.2022.103050

[18] Ghaderi, M. and M. Rahimzadegan, Improving precipitable water vapor estimations of the Moderate Resolution Imaging Spectroradiometer (MODIS) using metaheuristic algorithms. Advances in Space Research, 2022.

DOI: 10.1016/j.asr.2022.06.046

[19] Xiong, Z., et al., Modify the accuracy of MODIS PWV in China: A performance comparison using random forest, generalized regression neural network and back-propagation neural network. Remote Sensing, 2021. 13(11): p. 2215.

DOI: 10.3390/rs13112215

[20] Xu, J., et al., A New Machine Learning Based Calibration Scheme for MODIS Thermal Infrared Water Vapor Product Using BPNN, GBDT, GRNN, KNN, MLPNN, RF, and XGBoost. IEEE Transactions on Geoscience and Remote Sensing, 2024.

DOI: 10.1109/TGRS.2024.3350476

[21] Lee, Y., et al., Retrieval of total precipitable water from Himawari-8 AHI data: A comparison of random forest, extreme gradient boosting, and deep neural network. Remote Sensing, 2019. 11(15): p. 1741. DOI: 10.3390/rs11151741

[22] Amiri, M.J. and S. Eslamian, Investigation of Climate Change in Iran. Journal of Environmental Science and Technology, 2010. 3. https://doi.org/10.3923/jest.2010.208.216

[23] asakereh, H. and m. dostkamian, Survey of role of spatial factors on distribution- dispersion of maximums perceptible water in atmosphere of Iran. Journal of Applied Research in Geographical Sciences, 2015. 15(36): p. 7-24.

http://dorl.net/dor/20.1001.1.22287736.1394.15.36.2.8

[24] Seemann, S.W., et al., Operational retrieval of atmospheric temperature, moisture, and ozone from MODIS infrared radiances. Journal of applied meteorology, 2003. 42(8): p. 1072-1091.

DOI: 10.1175/1520-0450(2003)042<1072:OROATM>2.0.CO;2

[25] Borbas, E. and P. Menzel, MODIS Atmosphere L2 Atmosphere Profile Product, NASA MODIS Adaptive Processing System, Goddard Space Flight Center [data set], USA. 2015, August. http://dx.doi.org/10.5067/MODIS/MOD07_L2.006

[26] Niell, A., et al., Comparison of measurements of atmospheric wet delay by radiosonde, water vapor radiometer, GPS, and VLBI. Journal of Atmospheric and Oceanic Technology, 2001. 18(6): p. 830-850.

DOI: 10.1175/1520-0426(2001)018<0830:COMOAW>2.0.CO;2

[27] Bolton, D., The computation of equivalent potential temperature. Monthly weather review, 1980. 108(7): p. 1046-1053.

DOI: 10.1175/1520-0493(1980)108<1046:TCOEPT>2.0.CO;2

[28] Wang, J. and L. Zhang, Systematic errors in global radiosonde precipitable water data from comparisons with ground-based GPS measurements. Journal of Climate, 2008. 21(10): p. 2218-2238. DOI: 10.1175/2007JCLI1944.1.

DOI: 10.1016/j.atmosres.2017.07.021

[29] Gui, K., et al., Evaluation of radiosonde, MODIS-NIR-Clear, and AERONET precipitable water vapor using IGS ground-based GPS measurements over China. Atmospheric Research, 2017. 197: p. 461-473. DOI: 10.1016/j.atmosres.2017.07.021

[30] Wang, H., et al., Prediction models of soil heavy metal (loid) s concentration for agricultural land in Dongli: A comparison of regression and random forest. Ecological Indicators, 2020. 119: p. 106801.

DOI: 10.1016/j.ecolind.2020.106801

[31] Vapnik, V. and A. Vashist, A new learning paradigm: Learning using privileged information. Neural networks, 2009. 22(5-6): p. 544-557. DOI: 10.1016/j.neunet.2009.06.042

[32] Larya, D.J., et al., Machine learning in geosciences and remote sensing. Geoscience Frontiers, 2015. 30: p. 1e9.

DOI: 10.1016/j.gsf.2015.07.003

[33] Li, J. and A.D. Heap, A review of comparative studies of spatial interpolation methods in environmental sciences: Performance and impact factors. Ecological Informatics, 2011. 6(3-4): p. 228-241. DOI: 10.1016/j.ecoinf.2010.12.003

[34] Maxwell, A.E., T.A. Warner, and F. Fang, Implementation of machine-learning classification in remote sensing: An applied review. International journal of remote sensing, 2018. 39(9): p. 2784-2817. DOI: 10.1080/01431161.2018.1433343

[35] Scornet, E., G. Biau, and J.-P. Vert, Consistency of random forests. 2015. DOI: 10.1214/15-AOS1321

[36] Charu C, A., Neural networks and deep learning: a textbook. 2018: Springer. DOI: 10.1214/15-AOS1321

[37] Nair, V. and G.E. Hinton. Rectified linear units improve restricted boltzmann machines. in Proceedings of the 27th international conference on machine learning (ICML-10). 2010. June 21-24, Haifa, Israel

[38] Weisstein, E.W., Hyperbolic Functions., 2003.

[39] Chauvin, Y. and D.E. Rumelhart, Backpropagation: theory, architectures, and applications. 2013: Psychology press.

https://doi.org/10.4324/9780203763247

[40] Kingma, D.P., Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.

https://doi.org/10.48550/arXiv.1412.6980

[41] Hastie, T., The elements of statistical learning: data mining, inference, and prediction. 2009, Springer.

https://doi.org/10.1007/978-0-387-84858-7