EUVIMEDEUVIMED entdecken
Crossref

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

Priti Goyal, Nandini Sharma, Ujjwal Kumar Tiwari, Shaivi Goyal

International Journal of Innovative Science and Research Technology · 01.06.2026 · 2526

DOI: 10.38124/ijisrt/26may1453 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Innovative Science and Research Technology, online erschienen am 01.06.2026. Verfasst von Priti Goyal und 3 weiteren Autor:innen. Die Arbeit nennt 23 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

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.

Zitieren

APA

Goyal, P., Sharma, N., Tiwari, U. K., & Goyal, S. (2026). Physics-Informed Machine Learning for Air Temperature Prediction Using Surface Energy Balance–Based Feature Engineering. International Journal of Innovative Science and Research Technology, 2526. https://doi.org/10.38124/ijisrt/26may1453

Vancouver

Goyal P, Sharma N, Tiwari UK, Goyal S. Physics-Informed Machine Learning for Air Temperature Prediction Using Surface Energy Balance–Based Feature Engineering. International Journal of Innovative Science and Research Technology. 2026:2526. doi:10.38124/ijisrt/26may1453

BibTeX
@article{Goyal2026,
  author = {Goyal, Priti and Sharma, Nandini and Tiwari, Ujjwal Kumar and Goyal, Shaivi},
  title = {Physics-Informed Machine Learning for Air Temperature Prediction Using Surface Energy Balance–Based Feature Engineering},
  journal = {International Journal of Innovative Science and Research Technology},
  year = {2026},
  pages = {2526},
  doi = {10.38124/ijisrt/26may1453}
}
Kostenlos starten →