Environmental Management Hazards

Environmental Management Hazards

Dew Point Temperature Modeling Based on Machine Learning Algorithms

Document Type : Applied Article

Author
Nikrooz Bagheri
Associate Professor, Agricultural Engineering Research Institute, Agricultural Research, Education and Extension Organization (AREEO), Karaj, Iran
10.22059/jhsci.2026.423021.958
Abstract
One of the most important applications of the dew point in agriculture is forcasting radiation frost. The dew point can play a role in conditions conducive to frost occurrence and in its management. Radiation frost is one of the major hazards affecting the agricultural sector, causing substantial annual losses in crop yield and quality, particularly in fruit trees. Accurate and timely prediction of radiation frost is therefore of great importance for effective damage management. The occurrence of frost is not solely dependent on air temperature; atmospheric moisture-related variables also play an important role in its development. Accordingly, this study aimed to evaluate the short-term prediction of dew point temperature and investigate the effects of meteorological–physical features on the performance of machine learning-based models using meteorological data from Tehran, Iran, covering the period from 2005 to 2025. Air temperature, relative humidity, and dew point temperature were considered as the baseline variables. In addition, saturation vapor pressure, actual vapor pressure, vapor pressure deficit (VPD), and the difference between air temperature and dew point temperature were incorporated into the models as physical features. Seven algorithms, including Ridge Regression, Support Vector Regression (SVR), K-Nearest Neighbors (KNN), Random Forest (RF), Extremely Randomized Trees (Extra Trees), Gradient Boosting Regression, and XGBoost, were evaluated over two forecasting horizons of 6 and 12 hours, and their performance was compared with that of the baseline model. The results showed that, when meteorological–temporal features were used, Ridge Regression achieved the best overall performance at both the 6- and 12-hour forecasting horizons, with mean absolute error (MAE) values of 2.19 and 2.76 °C and root mean square error (RMSE) values of 3.01 and 3.77 °C, respectively. With the addition of meteorological–physical features, XGBoost demonstrated superior performance at both forecasting horizons in terms of reducing numerical prediction errors. Specifically, at the 6-hour horizon, the MAE and RMSE were 2.14 and 3.03 °C, respectively, whereas at the 12-hour horizon, the corresponding values were 2.61 and 3.62 °C. Support Vector Regression achieved better classification performance at the 6-hour horizon, with an F1-score of 0.911 and a Critical Success Index (CSI) of 0.837, while Ridge Regression performed better at the 12-hour horizon, with an F1-score of 0.878 and a CSI of 0.783. Furthermore, increasing the forecasting horizon from 6 to 12 hours resulted in a decrease in prediction accuracy across all models. Comparison of the models with the baseline model also indicated that machine learning models were capable of significantly improving prediction performance, although the differences among some of the top-performing models were not statistically significant. The findings demonstrate that machine learning algorithms can provide a suitable framework for short-term dew point temperature prediction and forecasting conditions conducive to radiation frost.
Keywords
Dew Point Temperature
Forecasting
Frost
Intelligent Agriculture
Machine Learning
Abdolshah, M. (2024). Identification and assessment of environmental hazards using the COCOSO technique - Excavation case: East Semnan industrial town. Environmental Hazards Management, 11 (3), 175-192. DOI: http//doi.org/10.22059/jhsci.2024.381982.840
Alizamir, M., Kim, S., Kisi, O., & Zounemat-Kermani, M. (2020). Deep echo state network: A novel machine learning approach to model dew point temperature using meteorological variables. Hydrological Sciences Journal, 65(7), 1173–1190. https://doi.org/10.1080/02626667.2020.1735639
Bazgir, S., Mohammadi, H., Sharifi, L., & Soleimani, N. (2016). Analysis of frost and freezing hazards to horticultural crops in West Azerbaijan Province. Environmental Hazards Management, 3(4), 365–378. https://doi.org/10.22059/jhsci.2016.62371
Bhakare, S., Matiu, M., Crespi, A., & Zardi, D. (2025). Spatial Downscaling of Daily Temperature Minima Using Machine Learning Methods and Application to Frost Forecasting in Two Alpine Valleys. Atmosphere, 16, 38. https://doi.org/10.3390/atmos16010038
Bhatti, S., Jeranyama, P., Kennedy, C. D., Buda, A. R., Millar, D. J., Wiegman, A. R. H., & Zalapa, J. E. (2026). Forecasting spring frost events in agriculture using machine learning: A case study from southeastern Massachusetts, United States. Smart Agricultural Technology, 13, 101720. https://doi.org/10.1016/j.atech.2025.101720
Cadenas, J. M., Garrido, M. C., Martínez-España, R., & Guillén-Navarro, M. A. (2020). Making decisions for frost prediction in agricultural crops in a soft computing framework. Computers and Electronics in Agriculture, 175, 105587. https://doi.org/10.1016/j.compag.2020.105587
Diniz, E.S., Lorenzon, A.S., Castro, N.L.M., Marcatti, G.E., Santos, O.P., Júnir, J.C., Cavalcante, R.B.L., Fernandes-Filho, E.I., Amaral, C.H. (2021). Forecasting frost risk in forest plantations by the combination of spatial data and machine learning algorithms. Agric. For. Meteorol. 306, 108450. https://doi.org/10.1016/j. agrformet.2021.108450.      
Hand, D. J., Christen, P., & Kirielle, N. (2021). F*: an interpretable transformation of the F-measure. Machine Learning, 110(3), 451-456.https://doi.org/10.1007/s10994-021-05964-1.
Irmak, S., & Kukal, M. S. (2022). Temporal trends in agriculturally relevant climate indicators across nine agroecosystems of Turkey. Journal of Applied Meteorology and Climatology, 61(6), 631–649. https://doi.org/10.1175/JAMC-D-21-0209.1
Omazić, B., Anić, M., Telišman Prtenjak, M., Kvakić, M., & Blašković, L. (2024). Analysis of different existing measurement-based methods and a new approach for frost probability detection. Agricultural and Forest Meteorology, 347, 109898. https://doi.org/10.1016/j.agrformet.2024.109898
Rowlandson, T., Gleason, M., Sentelhas, P., Gillespie, T., Thomas, C., & Hornbuckle, B. (2015). Reconsidering leaf wetness duration determination for plant disease management. Plant Disease, 99(3). https://doi.org/10.1094/PDIS-05-14-0529-FE Rozante, J. R.,
Song, H., Hu, Y., Lu, Y., Wang, J., Pan, Q., & Li, P. A. (2021). Review of Methods and Techniques for Detecting Frost on Plant Surfaces. Agriculture, 11, 1142. https://doi.org/10.3390/agriculture11111142
Talsma, C.J., Solander, K.C., Mudunuru, M.K., Crawford, B., & Powell, M.R. (2023) Frost prediction using machine learning and deep neural network models. Front. Artif. Intell. 5:963781. doi: 10.3389/frai.2022.963781
Tarraga, J. P., Castillo-Cara, M., Arias-Antúnez, E., & Dujovne, D. (2025). Frost forecasting through machine learning algorithms. Earth Science Informatics, 18, 183. https://doi.org/10.1007/s12145-025-01710-6
Torres-Cruz, F., Yana-Yucra, D. M., & Vilca-Solorzano, R. A. (2025). Comparative analysis of statistical, machine learning, and deep learning approaches for frost prediction in the Peruvian Altiplano. International Journal of Advanced Computer Science and Applications, 16(9).
Xu, J., Guga, S., Rong, G., Riao, D., Liu, X., Li, K., Zhang, J., 2021. Estimation of frost hazard for tea tree in zhejiang province based on machine learning. Agriculture 11, 607. https://doi.org/10.3390/agriculture11070607.
Yang, C., & Song, H. (2026). Short-term frost prediction during apple flowering in Luochuan using a 1D-CNN–BiLSTM network with attention mechanism. Horticulturae, 12(1), 47. https://doi.org/10.3390/horticulturae12010047
Zhang, Z., Li, P., & Wang, J. (2026). An Interpretable Machine Learning Framework for Next-Day Frost Forecasting in Tea Plantations Using Multi-Source Meteorological Data. Horticulturae, 12(3), 392. https://doi.org/10.3390/horticulturae12030392
Zhang, G., Bateni, S. M., Jun, C., Khoshkam, H., Band, S. S., & Mosavi, A. (2022). Feasibility of random forest and multivariate adaptive regression splines for predicting long-term mean monthly dew point temperature. Frontiers in Environmental Science, 10, 826165. https://doi.org/10.3389/fenvs.2022.826165
Environmental Management Hazards
Volume 13, Issue 2
Summer 2026
Pages 163-179