Evaluation of MICE-based machine learning models for reconstructing missing climate data in the Urmia Lake basin

Document Type : Research/Original/Regular Article

Authors

Department of Water Science and Engineering, College of Agriculture, Isfahan University of Technology, Isfahan, Iran

Abstract

Extended Abstract
Introduction
Complete and continuous climatic datasets are fundamental for reliable analyses in hydrology, climate change assessment, water resources management, and environmental modeling. However, observational climate records frequently suffer from missing values due to instrument malfunction, station relocation, data transmission errors, or long-term interruptions in measurements. If not appropriately addressed, missing data can introduce bias, reduce statistical power, and compromise the reliability of subsequent modeling and decision-making processes. This challenge is particularly critical in regions with complex climatic variability and environmental sensitivity, such as the Lake Urmia Basin in northwestern Iran. Traditional approaches for handling missing climatic data, including listwise deletion or simple statistical substitution (e.g., mean or median imputation), are computationally convenient but often distort the statistical structure of the data and fail to capture inter-variable dependencies. In response to these limitations, advanced multivariate and machine-learning-based imputation methods have gained increasing attention. Among them, Multiple Imputation by Chained Equations (MICE) has emerged as a robust framework that accounts for uncertainty and exploits relationships among multiple variables.
Recent studies suggest that integrating MICE with machine learning algorithms can further enhance imputation accuracy, particularly for non-linear and highly interdependent climatic variables. Nevertheless, comprehensive evaluations comparing different MICE-based hybrid models across multiple climatic variables and stations remain limited. Therefore, this study aims to systematically assess and compare the performance of standard MICE and four hybrid approaches MICE-Linear Regression (MICE–LR), MICE- Decision Tree (MICE–DT), MICE-K-Nearest Neighbor (MICE–KNN), and MICE- Support Vector Machine (MICE–SVM), across a wide range of climatic variables and meteorological stations within the Lake Urmia Basin.
Materials and Methods
This study was conducted using daily climatic data from six synoptic meteorological stations located in the Lake Urmia Basin. The dataset includes a diverse set of climatic variables representing thermal conditions, atmospheric moisture, cloudiness, wind characteristics, radiation and energy balance, and sea-level pressure. To ensure consistency and robustness, all variables were preprocessed through quality control procedures, including outlier detection and temporal consistency checks. Missing data were reconstructed using five imputation approaches: standard MICE and four hybrid MICE-based models (MICE-LR, MICE-DT, MICE-KNN, and MICE-SVM). The imputation procedure was implemented iteratively within the chained equations framework to ensure convergence and stability of the reconstructed values.
Model performance was evaluated using multiple complementary statistical metrics, including the coefficient of determination (R²), normalized root means square error (NRMSE), Kling–Gupta Efficiency (KGE), and percent bias (PBIAS). These metrics collectively assess accuracy, variability representation, correlation structure, and systematic bias. In addition to predictive performance, computational efficiency was assessed by measuring the average execution time of each model. The evaluation framework was designed to enable comparisons from three perspectives: climate-variable-based, model-based, and station-based analyses.
Results and Discussion
The comparative analysis revealed substantial differences in imputation performance among the evaluated models, depending on the type of climatic variable and station characteristics. Overall, hybrid MICE-based models demonstrated superior performance compared to the standard MICE approach, particularly for temperature-related variables and atmospheric moisture parameters. Among the hybrid models, MICE–DT achieved comparatively higher KGE values for several variables, highlighting its ability to model non-linear interactions. Nevertheless, both MICE–DT and MICE–LR provided a more balanced trade-off between reconstruction accuracy and computational efficiency.
In contrast, MICE–KNN and MICE–SVM exhibited variable performance, with notable sensitivity to station-specific conditions and variable type. While these models performed reasonably well for certain variables, their performance deteriorated for others, especially in cases involving higher variability or weaker spatial coherence. Standard MICE and MICE–LR showed comparable results, suggesting that linear assumptions may be insufficient for fully representing the dynamics of complex climatic systems.
The station-based analysis highlighted spatial heterogeneity in model performance, emphasizing the influence of local climatic and topographic conditions. Furthermore, the computational analysis indicated that while hybrid models generally required longer execution times than standard MICE, MICE–DT provided a favorable balance between accuracy and computational efficiency. These findings underscore the importance of selecting imputation methods based on both data characteristics and practical constraints.
Conclusion
This study provides a comprehensive evaluation of standard and hybrid MICE-based imputation methods for reconstructing missing climatic data in a multi-variable and multi-station framework. The results demonstrate that incorporating machine learning algorithms within the MICE framework substantially improves reconstruction accuracy, particularly for variables characterized by non-linear behavior. Among the evaluated models, MICE–DT emerged as the most robust and efficient approach, offering consistently high performance across different climatic variables and stations. Despite these strengths, certain limitations were identified. The performance of some hybrid models, particularly MICE–KNN and MICE–SVM, showed sensitivity to station-specific conditions and increased computational demand, which may limit their applicability in large-scale studies. These findings suggest that no single imputation method is universally optimal, and model selection should be tailored to the characteristics of the dataset and research objectives. From a practical perspective, the proposed framework provides valuable guidance for researchers and practitioners seeking reliable methods for handling missing climatic data. The results have direct implications for hydrological modeling, climate trend analysis, and environmental impact assessments in data-scarce regions. Future research should explore the integration of deep learning approaches within the MICE framework and assess model performance under varying missing-data scenarios and spatial scales.

Keywords

Main Subjects

References

  1. منابع

    ستاری، م. ت. شیرینی، ک. و جاویدان، س. (1403). ارزیابی کارائی روش‌های کاهش بعد در بهبود دقت مدل‌سازی شاخص کیفی آب با استفاده از الگوریتم‌های یادگیری ماشین. مدل‌سازی مدیریت آب و خاک، 4(2)، 89-104. doi: 10.22098/mmws.2023.12434.1241

    محتشم، جاوید. 1402. بازسازی داده‌های بارندگی روزانه دارای دوره‌های مفقودی، با استفاده از روش‌های محاسبه چندگانه با تطبیق میانگین پیش‌بینی و جنگل تصادفی. پایان نامه کارشناسی ارشد، دانشگاه صنعتی اصفهان. 278 ص.

    وفائی ممقانی، آ. اسدی، ا. دربندی، ص. و ستاری، م. ت. (1405). کاربست و مقایسه روش‌های درونیابی داده‌های گمشده‌ی تراز آب زیرزمینی با تأکید بر عملکرد DeepMVI (منطقه مورد مطالعه: دشت عجبشیر). مدلسازی مدیریت آب و خاک، 6 (1)، 17-36. doi: 10.22098/mmws.2025.17457.1601

    References

    Alejo-Sanchez, L. E., Márquez-Grajales, A., Salas-Martínez, F., Franco-Arcega, A., López-Morales, V., Acevedo-Sandoval, O. A., González-Ramírez, C. A., & Villegas-Vega, R. (2025). Missing data imputation of climate time series: A review. MethodsX, 15, 103.455. doi:10.1016/j.mex.2025.103455

    Amekudzi, L. K., Gyasi-Agyei, Y., Obuobie, E., & Addi, M. (2022). Evaluation of imputation techniques for infilling missing daily rainfall records on river basins in Ghana. Taylor & Francis.

    Azur, M. J., Stuart, E. A., Frangakis, C., & Leaf, P. J. (2011). Multiple imputation by chained equations: what is it and how does it work? International Journal of Methods in Psychiatric Research, 20(1), 40-49. doi:10.1002/mpr.329

    Chivers, B. D., Wallbank, J., Cole, S. J., Sebek, O., Stanley, S., Fry, M., & Leontidis, G. (2020). Imputation of missing sub-hourly precipitation data in a large sensor network: A machine learning approach. Journal of Hydrology, 588, 125126. :doi:10.1145/2939672.2939785

    Davari, S., Eslamian, S., Jamali, M., & Safavi, H. R. (2025). Application of machine learning algorithms for groundwater level prediction in the Najafabad plain. Scientific Reports. doi:10.47176/jwss.24.4.42931  

    Diouf, S., Deme, A., El Hadji Deme, P. F., & Diouf, I. (2023). An evaluation of the performance of imputation methods for missing meteorological data in Burkina Faso and Senegal. Afr. J. Environ. Sci. Technol, 17, 252-274. doi:10.47176/jwss.24.4.42933

    Farzandi, M., Sanaeinejad, H., Rezaei-Pazhan, H., & Sarmad, M. (2022). Improving estimation of missing data in historical monthly precipitation by evolutionary methods in the semi-arid area. Environment, Development and Sustainability, 24(6), 8313-8332. doi:10.1007/s10668-021-01784-4

    Géron, A., (2022). Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow.  O'Reilly Media, Inc.

    Ghanavati, R., Salajegheh, A., Pourghasemi, H. R., Khalighi Sigaroodi, S., & Keshtkar, H. (2025). Evaluation of machine learning techniques (SVM, GLM, FDA, RF) in preparing flood susceptibility map of a part of Khuzestan province. Water and Soil Management and Modeling, 5(1), 231-246. doi:10.47176/jwss.24.4.42905

    Golkhatmi, N. S. N., & Farzandi, M. (2024). Enhancing Rainfall Data Consistency and Completeness: A Spatiotemporal Quality Control Approach and Missing Data Reconstruction Using MICE on Large Precipitation Datasets. Water Resources Management, 38(3), 815-833. doi:10.1007/s11269-023-03567-0

    Gupta Hoshin, V., Sorooshian, S., & Yapo Patrice, O. (1999). Status of Automatic Calibration for Hydrologic Models: Comparison with Multilevel Expert Calibration. Journal of Hydrologic Engineering, 4(2), 135-143. doi:10.1061/(ASCE)1084-0699

    Hamzah, F. B., Mohamad Hamzah, F., Mohd Razali, S. F., & El-Shafie, A. (2022). Multiple imputations by chained equations for recovering missing daily streamflow observations: A case study of Langat River basin in Malaysia. Hydrological Sciences Journal, 67(1), 137-149.

    Hassanzadeh, E., Zarghami, M. and Hassanzadeh, Y., (2012). Determining the main factors in declining the Urmia Lake level by using system dynamics modeling. Water Resources Management, 26(1), pp.129-145.

    Jääskeläinen, E., Manninen, T., Hakkarainen, J., & Tamminen, J. (2022). Filling gaps of black-sky surface albedo of the Arctic sea ice using gradient boosting and brightness temperature data. International Journal of Applied Earth Observation and Geoinformation, 107, 102701. doi:10.1016/j.jag.2022.102701

    Jamali Jezeh, M., Shayannejad, M., & Hejazi, S. M. (2020). Evaluation the Performance of Filters Made of BC, PET and PP Textiles in Removing Oil Contaminants from Water [Research]. Journal of Water and Soil Science, 24(4), 29. (In Persian) doi:10.47176/jwss.24.4.42931

    Jamali, M., & Eslamian, S. (2023). Parametric and nonparametric methods for analyzing the trend of extreme events. In Handbook of Hydroinformatics (pp. 223-237). Elsevier.

    Jamali, M., & Eslamian, S. (2024). Climate Adaptation and Water–Environment–Energy Nexus. In Handbook of Climate Change Impacts on River Basin Management (pp. 225-236). CRC Press.

    Jing, X., Luo, J., Wang, J., Zuo, G., & Wei, N. (2022). A multi-imputation method to deal with hydro-meteorological missing values by integrating chain equations and random forest. Water Resources Management, 36(4), 1159-1173.

    Karbasi, M., 2016. Reconstruction of missing data of monthly total sunshine hours using artificial neural networks. Iranian Journal of Irrigation & Drainage, 10(5), pp.570-580.

    Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., & Liu, T.-Y. (2017). Lightgbm: A highly efficient gradient boosting decision tree. Advances in Neural Information Processing Systems, 30.

    Knoben, W. J. M., Freer, J. E., & Woods, R. A. (2019). Technical note: Inherent benchmark or not? Comparing Nash–Sutcliffe and Kling–Gupta efficiency scores. Hydrol. Earth System. Science., doi:10.5194/hess-23-4323-2019

    Kumar, G. P., & Dwarakish, G. (2025). Comparison of the multiple imputation approaches for imputing rainfall data: A humid tropical river basin case study. Water Conservation Science and Engineering, 10(2), 87.

    Legates, D. R., & McCabe Jr, G. J. (1999). Evaluating the use of “goodness-of-fit” Measures in hydrologic and hydroclimatic model validation. Water Resources Research, 35(1), 233-241. doi:10.1029/1998WR900018

    Little, R., & Rubin, D. (1987). Multiple imputation for nonresponse in surveys. Wiley, 10, 9780470316696.

    Little, R. J., & Rubin, D. B. (2019). Statistical analysis with missing data. John Wiley & Sons.

    Liu, J., Jiang, L., Zhang, X., Druce, D., Kittel, C. M., Tøttrup, C., & Bauer-Gottwein, P. (2021). Impacts of water resources management on land water storage in the North China Plain: Insights from multi-mission earth observations. Journal of Hydrology, 603, 126933.

    Matinzadeh, M. m., Fattahi, R., Shayanzadeh, M., & Abdollahi, K. (2013). Estimation and Reconstruction of Annual Maximum 24-H Rainfall Data Using Combination of Genetic Algorithm and Artificial Neural Networks Models (Case Study: Chaharmahal va Bakhtiyari Province) ijwmse, 7(22), 53. http://jwmsei.ir/article-1-245-fa.html

    1. Moriasi, D., G. Arnold, J., W. Van Liew, M., L. Bingner, R., D. Harmel, R., & L. Veith, T. (2007). Model Evaluation Guidelines for Systematic Quantification of Accuracy in Watershed Simulations. Transactions of the ASABE, 50(3), 885-900. doi:10.13031/2013.23153

    O'Sullivan, B., & Kelly, G. (2024). Infilling of high‐dimensional rainfall networks through multiple imputation by chained equations. International Journal of Climatology, 44(9), 3075-3091.

    Plein, M., Feigel, G., Zeeman, M., Dormann, C. F., & Christen, A. (2025). Using Gradient Boosting for gap-filling to analyze temperature and humidity patterns in an urban weather station network in Freiburg, Germany. Urban Climate, 62, 102496. doi:10.1016/j.uclim.2025.102496

    Sahoo, A. and Ghose, D. K. 2022. Imputation of missing precipitation data using KNN, SOM, RF, and FNN. Soft Computing. 26: 5919-5936.

    Qaraghuli, K., Murshed, M., Said, M. A. M., Mokhtar, A., & Rousta, I. (2024). Univariate and multivariate imputation methods evaluation for reconstructing climate time series data: A case study of Mosul station-Iraq. Journal of Agrometeorology, 26(3), 318-323.

    Tashman, L.J., (2000). Out-of-sample tests of forecasting accuracy: an analysis and review. International Journal of Forecasting, 16(4), pp.437-450.

    Van Buuren, S. (2000). Multivariate imputation by chained equations: MICE V1. 0 User's Manual. Leiden: TNO.

    Wilks, D.S., 2011. Statistical methods in the atmospheric sciences (Vol. 100). Academic press.

    Willmott, C. J. (1981). ON THE VALIDATION OF MODELS. Physical Geography, 2(2), 184-194. doi:10.1080/02723646.1981.10642213

     

     

Water and Soil Management and Modelling
Volume 6, Issue 3
August 2026
Pages 181-199

Files

History

  • Receive Date: 26 January 2026
  • Revise Date: 10 March 2026
  • Accept Date: 09 April 2026

Share

How to cite

Statistics