Seasonal-Intraseasonal Coupling and Systematic CMIP6 Biases in the Indian Summer Monsoon 

The Indian Summer Monsoon (ISM) supplies nearly 80% of annual rainfall over the Indian mainland during June–September and exhibits variability across multiple timescales. Intraseasonal variations, especially the timing and intensity of active and break spells, are critical for water resources and agriculture. However, how well CMIP6 models capture the observed link between the seasonally persistent background state and intraseasonal variability remains unexamined. We apply Multichannel Singular Spectrum Analysis (MSSA) to IMD rainfall observations (1979–2014) and CMIP6 historical simulations over the Indian mainland to evaluate how well models represent the observed spatial structure and amplitude of the dominant intraseasonal oscillation (ISO) modes: a low-frequency mode (20–60 days) with poleward propagation from the equatorial Indian Ocean and a high-frequency mode (10–20 days) with northwestward propagation from the Bay of Bengal. Across CMIP6 models, systematic biases are evident in both the spatial structure and amplitudes of these modes. Most models also fail to reproduce the observed relationship between seasonal rainfall and ISO intensity: observations show a negative correlation between all-India summer monsoon rainfall and the low-frequency ISO and a positive correlation with the high-frequency ISO, whereas many models simulate the opposite. These errors suggest that widely reported JJAS rainfall biases, particularly dry biases over the monsoon core region, may be closely linked to deficiencies in simulated intraseasonal variability. To investigate further and diagnose processes, we introduce a moisture budget framework that decomposes the total variability into contributions from the daily climatology, daily anomalies, and a seasonally persistent component defined as the seasonal mean of daily anomalies. By combining this persistent component with the daily climatology to construct an augmented mean state, we quantify interannual variability embedded within the mean advection terms, which incorporates the seasonally persistent component of daily anomalies, and isolate residual transient anomalies upon subtracting both the daily climatology and the seasonally averaged daily anomalies. The seasonally persistent component of both wind and moisture anomalies emerges as the key term differentiating flood and drought years with respect to both horizontal and vertical moisture advection.  We extend the same framework to analysis of vorticity budgets and examine biases in moisture and vorticity budget terms to understand biases in the rainfall-weighted latitude of precipitation (ITCZ) i.e. assess the ability of a model to realistically simulate this parameter vis-a-vis observations. Some models simulate a northward-displaced ITCZ, while others show a southward bias relative to the climatological mean ITCZ position of 23.8° N derived from IMD data. These analyses help elucidate mechanisms governing intraseasonal ITCZ migration. Finally, phase composites of budget terms conditioned on low- and high-frequency ISO phases identify the dominant dynamical and thermodynamical contributions to northward and westward propagation, respectively, and highlight the processes CMIP6 models fail to represent accurately. Overall, the analysis provides a systematic assessment of intraseasonal variability dynamics and their biases in CMIP6. By linking ISO dynamics to persistent large-scale circulation and background moisture fields, this study advances diagnostics of interannual variations in active and break spell occurrence across models.