⚠ Official Notice: www.ijisrt.com is the official website of the International Journal of Innovative Science and Research Technology (IJISRT) Journal for research paper submission and publication. Please beware of fake or duplicate websites using the IJISRT name.



Physics-Informed Machine Learning for Air Temperature Prediction Using Surface Energy Balance–Based Feature Engineering


Authors : Priti Goyal; Nandini Sharma; Ujjwal Kumar Tiwari; Shaivi Goyal

Volume/Issue : Volume 11 - 2026, Issue 5 - May

Google Scholar : https://tinyurl.com/4veyvapd

Scribd : https://tinyurl.com/5es9rb7w

DOI : https://doi.org/10.38124/ijisrt/26May1453

PlumX Metrics

Semantic Scholar

ResearchGate

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Precise predictions of air temperature are critical to environmental studies and climate research. However, conventional machine learning algorithms have a disadvantage in that, while performing well on capturing non-linear relationships, their physics interpretations can be poor. In this study, a physics-informed machine learning (PIML) model is proposed using Surface Energy Based (SEB) -based feature engineering techniques applied to ensemble models. The meteorological dataset for the Delhi region from NASA POWER for 2023–2025 was employed for this analysis. For testing purposes, two types of ensemble models are considered, including Random Forest and Gradient Boosting Machine (GBM) in both pure data-driven and physics-informed configurations. The physics-based features such as net radiation, sensible and latent heat proxies, vapor pressure deficit (VPD), and energy imbalance were included through feature engineering. Performance evaluation was done based on Root Mean Square Error (RMSE) and coefficient of determination (R²). The results reveal a marked improvement in the accuracy of predictions, where the physics-based GBM model lowers RMSE from 1.54°C to 1.17°C and attains an R² of 0.968. It can be seen that integrating physical knowledge in machine learning models is beneficial for enhancing predictive accuracy and robustness making it a promising approach for environmental data analysis.

Keywords : Physics-Informed Machine Learning, Surface Energy Balance, Air Temperature Prediction, Gradient Boosting, Random Forest, Vapor Pressure Deficit, Environmental Data Analysis.

References :

  1. Breiman, L. (2001) Random forests. Machine Learning, 45(1), 5–32. https://doi.org/10.1023/A:1010933404324
  2. Adeyinka, O. A., et al. (2025). Analysis and evaluation of ensemble methods in weather classification. Signal, Image and Video Processing. https://doi.org/10.1007/s11760-025-04322-1
  3. McGovern, A., Lagerquist, R., Gagne, D. J., et al. (2019) Making the black box more transparent: Understanding the physical implications of machine learning. Bulletin of the American Meteorological Society, 100(11), 2175–2199. https://doi.org/10.1175/BAMS-D-18-0195.1
  4. Culf, A.D., Foken, T., Gash, J.H.C. (2004). The Energy Balance Closure Problem. In: Kabat, P., et al. Vegetation, Water, Humans and the Climate. Global Change — The IGBP Series. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-18948-7_13
  5. Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021) Physics-informed machine learning. Nature Reviews Physics, 3(6), 422–440. https://doi.org/10.1038/s42254-021-00314-5
  6. Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019) Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378, 686–707. https://doi.org/10.1016/j.jcp.2018.10.045
  7. Kashinath, K., Mustafa, M., Albert, A., et al. (2021) Physics-informed machine learning: Case studies for weather and climate modelling. Philosophical Transactions of the Royal Society A, 379(2194), 20200093. https://doi.org/10.1098/rsta.2020.0093
  8. McCandless, T., Gagne, D.J., Kosović, B. et al. Machine Learning for Improving Surface-Layer-Flux Estimates. Boundary-Layer Meteorol 185, 199–228 (2022). https://doi.org/10.1007/s10546-022-00727-4
  9. Ma J., Shen H., Jiang M., Lin L., Meng C., Zeng C.,  Li H., Wu P. (2023), A physics-constrained machine learning method for mapping gapless land surface temperature, Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning https://doi.org/10.48550/arXiv.2307.04817
  10. Aslam L., Zou R., Li G. Awan E.S. (2025) PIEL-NET: Physics-Informed Ensemble Learning for city-center grid cell temperature prediction during thermal extremes, Urban Climate 64:102669, DOI:10.1016/j.uclim.2025.102669
  11. Yang H., Xue X.(2026) Physics-informed data-driven model for the prediction of river water temperature, Engineering Applications of Artificial Intelligence, Volume 172,  114386, https://doi.org/10.1016/j.engappai.2026.114386
  12. Kara, A.B., Rochford, P.A., & Hurlburt, H.E. (2000). Efficient and Accurate Bulk Parameterizations of Air-Sea Fluxes for Use in General Circulation Models. Journal of Atmospheric and Oceanic Technology, 17, 1421-1438 https://doi.org/10.1175/1520-0426(2000)017%3C1421:EAABPO%3E2.0.CO;2
  13. Brunel, J.P. (1989). Estimation of sensible heat flux from measurements of surface radiative temperature and air temperature at two meters: application to determine actual evaporation rate. Agricultural and Forest Meteorology, 46, 179-191 https://doi.org/10.1016/0168-1923(89)90063-4
  14. Kim, Y., Garcia, M., Morillas, L., Weber, U., Black, T. A., & Johnson, M. S. (2021). Relative humidity gradients as a key constraint on terrestrial water and energy fluxes. Hydrology and Earth System Sciences, 25(9), 5175- 5191. https://doi.org/10.5194/hess-25-5175-2021
  15. Cuxart, J., L. Conangla, and M. A. Jiménez (2015), Evaluation of the surface energy budget equation with experimental data and the ECMWF model in the Ebro Valley, J. Geophys. Res. Atmos., 120, 1008–1022, doi:10.1002/2014JD022296
  16. Stoy, P. C., Mauder, M., Foken, T., et al. (2013) A data-driven analysis of energy balance closure across FLUXNET research sites: The role of landscape scale heterogeneity. Agricultural and Forest Meteorology, 171–172, 137–152. https://doi.org/10.1016/j.agrformet.2012.11.004
  17. Lisan Yu. 2019. Global Air–Sea Fluxes of Heat, Fresh Water, and Momentum: Energy Budget Closure and Unanswered Questions. Annual Review Marine Science. 11:227-248. https://doi.org/10.1146/annurev-marine-010816-060704
  18. Beucler, T., Rasp, S., Pritchard, M., & Gentine, P. (2023) Physics-constrained deep learning for postprocessing of temperature and humidity. Artificial Intelligence for the Earth Systems, 2(4). https://doi.org/10.1175/AIES-D-22-0089.1
  19. Breiman, L. (1996). Bagging predictors. Machine Learning, 24(2), 123–140. https://doi.org/10.1007/BF00058655
  20. Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5), 1189–1232. https://doi.org/10.1214/aos/1013203451
  21. L. Gu, T. Meyers, S. Pallardy, P. Hanson, Bai Yang, M. Heuer, K. P. Hosman, J. Riggs, D. Sluss, S. Wullschleger (2006), Direct and indirect effects of atmospheric conditions and soil moisture on surface energy partitioning revealed by a prolonged drought at a temperate forest site, Journal of Geophysical Research,  111, D16102, doi:10.1029/2006JD007161
  22. Detto, M., C.Still, and A.Porporato. 2026. “Atmospheric Boundary Layer Control on Forest Thermal Properties.” Global Change Biology, 32, no. 4: e70841. https://doi.org/10.1111/gcb.70841
  23. Brayan-Leonardo Sierra-Forero, Julio Barón-Velandia, Sebastian-Camilo Vanegas-Ayala (2024), Assessment of the relevance of features associated with corn crop yield prediction in Colombia, a country in the Neotropical zone, Int. j. inf. tecnol. 16(4):2129–2138, https://doi.org/10.1007/s41870-024-01762-9

Precise predictions of air temperature are critical to environmental studies and climate research. However, conventional machine learning algorithms have a disadvantage in that, while performing well on capturing non-linear relationships, their physics interpretations can be poor. In this study, a physics-informed machine learning (PIML) model is proposed using Surface Energy Based (SEB) -based feature engineering techniques applied to ensemble models. The meteorological dataset for the Delhi region from NASA POWER for 2023–2025 was employed for this analysis. For testing purposes, two types of ensemble models are considered, including Random Forest and Gradient Boosting Machine (GBM) in both pure data-driven and physics-informed configurations. The physics-based features such as net radiation, sensible and latent heat proxies, vapor pressure deficit (VPD), and energy imbalance were included through feature engineering. Performance evaluation was done based on Root Mean Square Error (RMSE) and coefficient of determination (R²). The results reveal a marked improvement in the accuracy of predictions, where the physics-based GBM model lowers RMSE from 1.54°C to 1.17°C and attains an R² of 0.968. It can be seen that integrating physical knowledge in machine learning models is beneficial for enhancing predictive accuracy and robustness making it a promising approach for environmental data analysis.

Keywords : Physics-Informed Machine Learning, Surface Energy Balance, Air Temperature Prediction, Gradient Boosting, Random Forest, Vapor Pressure Deficit, Environmental Data Analysis.

Paper Submission Last Date
30 - June - 2026

SUBMIT YOUR PAPER CALL FOR PAPERS
Video Explanation for Published paper

Never miss an update from Papermashup

Get notified about the latest tutorials and downloads.

Subscribe by Email

Get alerts directly into your inbox after each post and stay updated.
Subscribe
OR

Subscribe by RSS

Add our RSS to your feedreader to get regular updates from us.
Subscribe