Intercomparison of bias correction methods for precipitation of multiple GCMs across six continents
Young Hoon Song, Eun‐Sung Chung
Seoul National University of Science and Technology
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
This study proposed a Comprehensive Index (CI) that jointly considers bias correction performance metrics and uncertainty to guide the selection of quantile mapping methods. This approach reveals not only a performance-based ranking of bias correction methods but also how optimal method choices shift as the uncertainty weight varies. This study evaluated daily precipitation performance from 11 CMIP6 GCMs corrected by Quantile Delta Mapping (QDM), Empirical Quantile Mapping (EQM), and Detrended Quantile Mapping (DQM) using ten evaluation metrics and applied TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution) to compute performance-based rankings. Furthermore, Bayesian Model Averaging (BMA) was used to quantify both individual model and ensemble prediction uncertainties. Moreover, entropy based weighting of the ten evaluation metrics reveals that error based measures such as RMSE and MAE carry the highest information content (weights 0.13–0.28 and 0.15–0.22, respectively). By aggregating TOPSIS performance scores with BMA uncertainty measures, this study developed CI. Results show that EQM achieved the best performance across most metrics 0.30 (RMSE), 0.18 (MAE), 0.98 ( R 2 ), 0.87 (KGE), 0.93 (NSE), and 0.99 (EVS) and exhibited the lowest uncertainty (variance = 0.0027) across all continents. QDM outperformed other methods in certain regions, reaching its lowest model uncertainty (variance = 0.0025) in South America. EQM was selected most frequently under all weighting scenarios, while DQM was least chosen. In South America, DQM was preferred more often than QDM when performance was emphasized, whereas the opposite occurred when uncertainty was emphasized. These findings suggest that incorporating uncertainty leads to spatially heterogeneous and parameter dependent changes in optimal bias correction method choice that would be overlooked by metric only selection.
逐年被引趋势
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
物理Climate variability and models
Precipitation Measurement and Analysis · Hydrology and Watershed Management Studies