High-Resolution Evaluation of WRF Cumulus Parameterization Schemes for Rainfall Prediction in Thailand

Abstract

This study evaluates the performance of cumulus parameterization schemes under a default SBU-YLin within the weather research and forecasting (WRF) model to enhance rainfall forecasting in Thailand's tropical climate. Using a high-resolution nested domain setup (8.1 × 8.1 km, 2.7 × 2.7 km, and 0.9 × 0.9 km), simulations are validated against 128 Thai Meteorological Department stations from September 21 to 23, 2024. The New Tiedtke (RAIN-16) scheme achieved the most balanced performance, with underestimation and overestimation rates of approximately 50% and mean underestimation and overestimation of −12.75 mm and 13.30 mm, respectively. However, the SBU-YLin scheme reduced biases, root mean square error (RMSE), and mean absolute error (MAE) but could not resolve heavy rainfall events, especially in Domains 2 and 3, which are considered high-resolution domains. In Domain 1, very heavy rainfall was overestimated with an RMSE of 98.50 mm; Domain 2 showed an overestimation of rainfall, and Domain 3 had a substantial overestimation of rainfall associated with localized convective activity. However, the SBU-YLin scheme reduced the biases, RMSE, and MAE, but at higher computational demands. Results showed that Domains 2 and 3 increased accuracy with finer spatial resolution, but it was challenging to simulate heavy rainfall events. The selected 3 days are an intense rainfall event that is characteristic of monsoonal convection in Thailand. This brief time span enables WRF schemes to be examined at a resolution in extreme conditions. The temporal coverage of future research is going to be extended to a seasonal-level assessment.

Received: 14 August 2025 | Revised: 3 February 2026 | Accepted: 23 April 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data used in this study were obtained from TMD with permission and are not publicly available.

Author Contribution Statement

Usa Wannasingha Humphries: Conceptualization, Methodology, Software, Formal analysis, Investigation, Resources, Writing – review & editing, Supervision, Project administration, Funding acquisition. Muhammad Waqas: Validation, Investigation, Data curation, Writing – original draft, Visualization. Boonlert Archevarahuprok: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation. Porntip Dechpichai: Validation, Visualization. Angkool Wangwongchai: Conceptualization, Methodology, Software, Formal analysis, Writing – review & editing, Supervision, Project administration. Boobphachard Chansawang: Writing – review & editing, Visualization.