A performance-based multivariable weighting framework for WRF parameterization-scheme selection
- 1China Agricultural University (Beijing, China)
Abstract
The selection of physical parameterization schemes is a major source of uncertainty in WRF simulations. A comprehensive evaluation of model performance requires the simultaneous consideration of multiple output variables, but simple multivariable averaging can dilute the influence of variables that are difficult to simulate or that respond strongly to alternative physical-process options. To address this issue, this study proposes a performance-based multivariable weighting framework for WRF parameterization-scheme evaluation. The framework combines two components: mean relative error, which represents variable-level simulation difficulty quantified by relative simulation error, and the coefficient of variation of absolute errors, which represents the scheme-dependent response of each variable within the evaluated physical-process group. The approach was applied and evaluated in assessing WRF parameterization schemes across two climatically distinct environments: the arid Northwest and the humid coastal Southeast of China. The results show that the method assigns greater weights to variables that are difficult to simulate, such as precipitation and 10 m wind, thereby helping identify physically interpretable and well-performing parameterization-scheme combinations under the prescribed evaluation framework. Compared with equal weighting and established statistical weighting methods, the proposed MRE × CV weighting helped identify parameterization-scheme combinations that achieved the lowest integrated cMIEI in both representative screening years and in the independent 2010 extreme-year validation. These results indicate that jointly considering variable-level simulation difficulty and scheme-dependent response improves the diagnostic relevance of multivariable WRF parameterization-scheme evaluation.
Versions
Related
- Basin-scale evaluation of the Noah-MP land surface model for runoff and snow generation in the Missouri River Basin: insights and recommendations for parameterization scheme selection
- Multi-data source based quantifying urban flood severity in major Chinese cities (2000–2024) using a hybrid machine-learning weighting framework
- Parameterization of bottom mixed layer and logarithmic layer heights in central Lake Erie
- A multi-scale framework for selecting CO2 storage sites in oil and gas reservoirs Multi-scale framework for CO2 storage site selection in oil and gas reservoirs
- Cumulus parameterization controls simulated rainfall more than microphysics, but not at heavy thresholds: a 3 * 2 factorial WRF study over the Western Himalaya
BibTeX
@article{gou2026performancebased,
title = {{A performance-based multivariable weighting framework for WRF parameterization-scheme selection}},
author = {Tianyu Gou and Yaoyang Deng and Jun Niu and Shaozhong Kang},
journal = {Geoscientific Model Development},
year = {2026},
month = oct,
doi = {10.5194/gmd-19-9879-2026},
url = {https://gmd.copernicus.org/articles/19/9879/2026/},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - A performance-based multivariable weighting framework for WRF parameterization-scheme selection
AU - Tianyu Gou
AU - Yaoyang Deng
AU - Jun Niu
AU - Shaozhong Kang
PY - 2026
DA - 2026/10/15
JO - Geoscientific Model Development
DO - 10.5194/gmd-19-9879-2026
UR - https://gmd.copernicus.org/articles/19/9879/2026/
AB - The selection of physical parameterization schemes is a major source of uncertainty in WRF simulations. A comprehensive evaluation of model performance requires the simultaneous consideration of multiple output variables, but simple multivariable averaging can dilute the influence of variables that are difficult to simulate or that respond strongly to alternative physical-process options. To address this issue, this study proposes a performance-based multivariable weighting framework for WRF parameterization-scheme evaluation. The framework combines two components: mean relative error, which represents variable-level simulation difficulty quantified by relative simulation error, and the coefficient of variation of absolute errors, which represents the scheme-dependent response of each variable within the evaluated physical-process group. The approach was applied and evaluated in assessing WRF parameterization schemes across two climatically distinct environments: the arid Northwest and the humid coastal Southeast of China. The results show that the method assigns greater weights to variables that are difficult to simulate, such as precipitation and 10 m wind, thereby helping identify physically interpretable and well-performing parameterization-scheme combinations under the prescribed evaluation framework. Compared with equal weighting and established statistical weighting methods, the proposed MRE × CV weighting helped identify parameterization-scheme combinations that achieved the lowest integrated cMIEI in both representative screening years and in the independent 2010 extreme-year validation. These results indicate that jointly considering variable-level simulation difficulty and scheme-dependent response improves the diagnostic relevance of multivariable WRF parameterization-scheme evaluation.
SN - 1991-9603
ER -