Sahu et al. (2026) Evaluation of microphysics and boundary layer schemes for simulating extreme rainfall events over Saudi Arabia using WRF-ARW

Identification

• Journal: Natural hazards and earth system sciences
• Year: 2026
• Date: 2026-01-06
• Authors: Rajesh Kumar Sahu, Hamza Kunhu Bangalath, Suleiman Mostamandi, Jason Evans, Paul A. Kucera, Hylke E. Beck
• DOI: 10.5194/nhess-26-21-2026

Research Groups

• Physical Science and Engineering Division, King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia
• Climate Change Research Centre, University of New South Wales, Sydney, Australia
• COMET Program, University Corporation for Atmospheric Research, Boulder, Colorado, USA

Short Summary

This study evaluates 36 combinations of planetary boundary layer (PBL) and cloud microphysics (MP) schemes within the WRF-ARW model to simulate 17 extreme rainfall events (EREs) over the Arabian Peninsula, identifying the Thompson-Yonsei University (MP8_BL1) combination as generally the best performer for rainfall and other meteorological variables.

Objective

• To evaluate planetary boundary layer (PBL) and cloud microphysics (MP) schemes for simulating Extreme Rainfall Events (EREs) in the Arabian Peninsula (AP) using the Advanced Research version of the Weather Research and Forecasting (WRF-ARW) model.
• To determine the best combination of PBL and MP schemes for ERE simulation at convection-permitting resolution.

Study Configuration

• Spatial Scale: Arabian Peninsula (21–65° E, 2–40° N), with two-way nested domains: D01 (9000 m resolution, 493 × 418 grid cells) and D02 (3000 m resolution, 1012 × 889 grid cells). Model top pressure of 3000 Pa.
• Temporal Scale: 17 Extreme Rainfall Events (EREs) from 2010 to 2022. Each simulation ran for 84 hours, including a 48-hour spin-up, with analysis focused on a 24-hour peak rainfall window.

Methodology and Data

• Models used:
  • Weather Research and Forecasting (WRF-ARW) model V4.4 (non-hydrostatic, fully compressible, terrain-following coordinate system).
  • Planetary Boundary Layer (PBL) schemes tested: Mellor-Yamada Nakanishi Niino (MYNN) Level 2.5 (BL5), MYNN Level 3 (BL6), Yonsei University (YSU; BL1), Bougeault-Lacarrère (BouLac; BL8).
  • Cloud Microphysics (MP) schemes tested: Kessler (MP1), Purdue Lin (MP2), WRF Single-Moment 3-class (MP3), WRF Single-Moment 5-class (MP4), Eta Ferrier (MP5), WRF Single-Moment 6-class (MP6), Goddard (MP7), Thompson (MP8), Morrison 2-Moment (MP10).
  • Cumulus parameterization: Kain–Fritsch (D01), no cumulus scheme (D02).
  • Surface layer parameterization: Noah Land Surface scheme.
  • Surface layer physics: Revised MM5.
  • Radiation schemes: RRTMG (shortwave and longwave).
• Data sources:
  • Initial and Boundary Conditions: ERA5 pressure-level data (37 levels, 0.25° spatial resolution, 3-hour time step) from Copernicus Climate Data Store (CDS).
  • Rainfall Observations: Integrated Multi-satellite Retrievals for GPM (IMERG) Final V07 product (30-minute, 0.1° resolution, aggregated to hourly).
  • Meteorological Observations: 2 m air temperature, 2 m relative humidity, and 10 m wind speed (metre per second) from IOWA Environmental Mesonet (METAR) data.

Main Results

• The WRF-ARW model generally captured observed rainfall patterns well, though with some over- and underestimations.
• The Yonsei University (YSU; BL1) scheme performed best among PBL schemes (mean temporal Kling–Gupta Efficiency (KGE) of 0.43, mean spatial KGE of 0.29), attributed to its non-local mixing approach.
• The Thompson (MP8) and Goddard (MP7) schemes performed best among MP schemes (mean temporal KGE of 0.42 for both, mean spatial KGE of 0.31 and 0.33 respectively), due to their sophisticated handling of mixed-phase and ice-phase microphysics.
• The Thompson-YSU (MP8_BL1) combination yielded the highest mean KGE for rainfall across the 17 EREs (mean temporal KGE of 0.48) and performed statistically significantly better than 21 other combinations.
• MP8_BL1 also performed best for 2 m air temperature (mean temporal KGE of 0.47), 2 m relative humidity (mean temporal KGE of 0.31), and 10 m wind speed (mean temporal KGE of 0.29).
• For rainfall, correlation and variability were the dominant components affecting the final KGE scores.
• Performance rankings showed weak consistency across different meteorological variables, suggesting that different physical processes or model components might govern their simulation.

Contributions

• Conducted the most comprehensive evaluation of WRF-ARW PBL and MP schemes for simulating EREs over the Arabian Peninsula to date, testing 36 combinations across 17 events at convection-permitting resolution.
• Identified optimal PBL (YSU; BL1) and MP (Thompson; MP8, Goddard; MP7) schemes and their best combination (Thompson-YSU; MP8_BL1) for ERE simulation in arid regions.
• Provided practical guidance for scheme selection to improve ERE forecasting and regional climate modeling over the Arabian Peninsula.
• Highlighted the importance of analyzing multiple EREs with high-quality reference data for robust scheme evaluation.
• Quantified the statistical significance of performance differences between scheme combinations.

Funding

• Resources from the Shaheen supercomputer at King Abdullah University of Science and Technology (KAUST).

Citation

@article{Sahu2026Evaluation,
  author = {Sahu, Rajesh Kumar and Bangalath, Hamza Kunhu and Mostamandi, Suleiman and Evans, Jason and Kucera, Paul A. and Beck, Hylke E.},
  title = {Evaluation of microphysics and boundary layer schemes for simulating extreme rainfall events over Saudi Arabia using WRF-ARW},
  journal = {Natural hazards and earth system sciences},
  year = {2026},
  doi = {10.5194/nhess-26-21-2026},
  url = {https://doi.org/10.5194/nhess-26-21-2026}
}