Scholarlink

<p>This study systematically evaluates the performance of Coupled General Circulation Models (CGCMs) in simulating the variability of Indian Summer Monsoon Rainfall (ISMR), with an emphasis on their ability to capture the Tropical Biennial Oscillation (TBO) and its interactions with large-scale climate drivers, particularly the El Niño–Southern Oscillation (ENSO) and Indian Ocean SST anomalies. This study introduces a novel classification framework based on the models’ ability to capture biennial ISMR variability, serving as a proxy for TBO skill, which reveals important performance differences beyond ENSO representation. Using a combination of spectral analysis, Taylor diagrams, and variance attribution methods, 33 models from CMIP6 are classified into two categories—CAT1 and CAT2—based on their skill in replicating the dominant biennial periodicity (2–3.5&#xa0;years) observed in ISMR. CAT1 models, representing approximately 70% of the ensemble, display stronger TBO signals in both ISMR and Niño3.4 SST time series, higher TBO-to-ENSO power ratios, and better agreement with observed rainfall variability, with higher mean spatial correlation (0.90), normalized standard deviation close to unity (0.96), and lower RMSE (2.04&#xa0;mm/day). Conversely, CAT2 models exhibit diminished biennial spectral power, weaker coupling between ISMR and SST anomalies, and poorer performance metrics (mean correlation = 0.89, normalized SD = 0.88, RMSE = 3.17&#xa0;mm/day). Spectral variance analysis reveals that CAT1 models allocate a greater proportion of variance to the TBO band (34.05%), aligning more closely with observations (42.39%), while CAT2 models exhibit stronger emphasis on ENSO (42.60%) and decadal variability (28.95%), indicating a different balance of interannual variability modes. Empirical Orthogonal Function (EOF) analysis further shows that both model categories significantly overestimate the variance of the leading modes: EOF1, which is ENSO-dominated, and EOF2, which captures meridional rainfall contrasts modulated by lagged ENSO and Indian Ocean SSTs. The overestimation of EOF1 (mean = 38.26% in CAT1; 36.96% in CAT2) relative to observations (14.26%) suggests an exaggerated ENSO–ISMR link, primarily due to model deficiencies such as weak Bjerknes feedback and a biased shortwave damping feedback that together amplify remote ENSO signals. EOF2 variance is also overestimated (observed = 8.86%; CAT1 = 11.86%; CAT2 = 12.54%), with models showing limited skill in capturing regional ISMR variability driven by Indian Ocean processes. Further analysis of model components indicates that performance differences cannot be solely attributed to shared AGCM or OGCM modules, underscoring the importance of coupling strategies and parameterization schemes. These findings reveal critical gaps in how current CGCMs simulate internal monsoon dynamics and ocean–atmosphere interactions, and highlight the necessity for targeted improvements in feedback processes and representation of teleconnections to enhance model fidelity and increase the robustness of future ISMR projections under climate change scenarios.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Understanding Indian summer monsoon rainfall variability through the lens of tropospheric biennial oscillation: a study of temporal and spatial patterns in CMIP6 models

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Anika Arora,
  • Vinu Valsala,
  • Prasanth A. Pillai,
  • R. Phani Murali Krishna

摘要

This study systematically evaluates the performance of Coupled General Circulation Models (CGCMs) in simulating the variability of Indian Summer Monsoon Rainfall (ISMR), with an emphasis on their ability to capture the Tropical Biennial Oscillation (TBO) and its interactions with large-scale climate drivers, particularly the El Niño–Southern Oscillation (ENSO) and Indian Ocean SST anomalies. This study introduces a novel classification framework based on the models’ ability to capture biennial ISMR variability, serving as a proxy for TBO skill, which reveals important performance differences beyond ENSO representation. Using a combination of spectral analysis, Taylor diagrams, and variance attribution methods, 33 models from CMIP6 are classified into two categories—CAT1 and CAT2—based on their skill in replicating the dominant biennial periodicity (2–3.5 years) observed in ISMR. CAT1 models, representing approximately 70% of the ensemble, display stronger TBO signals in both ISMR and Niño3.4 SST time series, higher TBO-to-ENSO power ratios, and better agreement with observed rainfall variability, with higher mean spatial correlation (0.90), normalized standard deviation close to unity (0.96), and lower RMSE (2.04 mm/day). Conversely, CAT2 models exhibit diminished biennial spectral power, weaker coupling between ISMR and SST anomalies, and poorer performance metrics (mean correlation = 0.89, normalized SD = 0.88, RMSE = 3.17 mm/day). Spectral variance analysis reveals that CAT1 models allocate a greater proportion of variance to the TBO band (34.05%), aligning more closely with observations (42.39%), while CAT2 models exhibit stronger emphasis on ENSO (42.60%) and decadal variability (28.95%), indicating a different balance of interannual variability modes. Empirical Orthogonal Function (EOF) analysis further shows that both model categories significantly overestimate the variance of the leading modes: EOF1, which is ENSO-dominated, and EOF2, which captures meridional rainfall contrasts modulated by lagged ENSO and Indian Ocean SSTs. The overestimation of EOF1 (mean = 38.26% in CAT1; 36.96% in CAT2) relative to observations (14.26%) suggests an exaggerated ENSO–ISMR link, primarily due to model deficiencies such as weak Bjerknes feedback and a biased shortwave damping feedback that together amplify remote ENSO signals. EOF2 variance is also overestimated (observed = 8.86%; CAT1 = 11.86%; CAT2 = 12.54%), with models showing limited skill in capturing regional ISMR variability driven by Indian Ocean processes. Further analysis of model components indicates that performance differences cannot be solely attributed to shared AGCM or OGCM modules, underscoring the importance of coupling strategies and parameterization schemes. These findings reveal critical gaps in how current CGCMs simulate internal monsoon dynamics and ocean–atmosphere interactions, and highlight the necessity for targeted improvements in feedback processes and representation of teleconnections to enhance model fidelity and increase the robustness of future ISMR projections under climate change scenarios.