Assessment of Ocean-Atmosphere Interactions for the Indian Summer Monsoon Intraseasonal Oscillations in CMIP5 and CMIP6 Models

Abstract

The Asian monsoon system (AMS) is a largescale phenomenon resulting in the strong coupling of ocean and atmosphere and is highly associated with the remote forcings (e.g., El-Niño southern oscillations, Indian Ocean dipole, Pacific decadal oscillation). The present dissertation focuses on the Indian summer monsoon (ISM) system. ISM covers a large area and about 80% of the annual rainfall occurs over the Indian mainland during this season. Earlier studies reported that the ISM rainfall (ISMR) is highly variable from intraseasonal through interannual to decadal timescales. On intraseasonal timescales, convection anomalies develop in the equatorial Indian Ocean (EIO) and propagate towards north (foot of the Himalayan Mountains in northern India) and eastward to the west Pacific Ocean (WNP). The northward propagation of intraseasonal oscillations is also known as the Boreal Summer Intraseasonal Oscillations (BSISO) and are the prominent South Asian summer monsoon features, mainly governed by the atmospheric internal dynamics and air-sea interactions. However, air-sea interactions role in modulation of the BSISO in the coupled climate models was not well understood. The present study utilizes the extended empirical orthogonal function-based bimodal ISO index to assess the phase-relationship of air-sea fluxes in the North Indian Ocean (NIO) using observations and reanalysis products. These phase-relationships are also evaluated in the historical outputs of Coupled Model Intercomparison Project (CMIP) phase-5 and phase-6. On intraseasonal timescales, enhanced deep convection leads by a phase of 850 hPa westerly winds and negative sea surface temperature (SST) anomalies; and deep convection (clear sky conditions) leads (~ 5-10 days) enhanced westerly (easterly) winds and cool (warm) SST over the NIO. Most CMIP5 models represents the northward propagation of precipitation and zonal winds at 850 hPa. However, models bias of BSISO variance shows significant spatial heterogeneity over the regions of the Arabian Sea (AS), Sub-Continent of India (SCI), and Bay of Bengal (BoB). The CMIP5 models, which shows significant biases in the mean state, causes the failure of models in representing the BSISO propagation. However, the majority of the models show large uncertainty to represent this prominent feature over AS and SCI. Further, improper representation of the lead-lag relationship of SST and precipitation on intraseasonal timescales over the NIO in the CMIP5 models contributes to significant bias variances. The new release of CMIP6 models provides an opportunity to examine models' ability to simulate the characteristic features of BSISO and its associated air-sea interactions. Most CMIP6 models underestimate the precipitation over central India and overestimate the precipitation over the eastern equatorial region. In the observations, the precipitation anomalies propagate northward from the equatorial latitudes to the northern latitudes over the ISM region. However, the initiation of northward propagating convection shows a significant variation with time in the CMIP6 models, and a prominent representation of BSISO propagation is evident over the BoB and the SCI. The phase speed of BSISO over the AS is underestimated by many models, which led to the failure of models in representing the northwest-southeast tilt of convection. Surface turbulent fluxes and zonal winds lag the deep convection over the NIO on intraseasonal timescales. However, misrepresentation of air-sea fluxes in the CMIP6 models leads to the significant biases of intraseasonal variances. Earlier studies suggest that ISOs can modulate the rainfall over India. The role of BSISO during excess/deficit ISM years is analysed in the observations and the CMIP6 models. Models overestimated the moisture transport from the west Indian Ocean to the mainland of India during deficit monsoons, which plays a crucial role in modulating the precipitation and ISOs. During deficit monsoon years, the faster moving 20-100 day oscillations are stronger than the excess monsoon years, which affects the duration of convection activity and causes dry conditions over the regions. Further, a stronger response from underlying SST increases the phase speed of BSISO activity. During excess monsoon years, BoB (AS) responds more strongly (weakly) and quickly (slowly) to the atmosphere than deficit monsoon years. However, models are failed to represent the ocean's response to the atmosphere over BoB. The improvement of freshwater forcing in the models may simulate the ocean's atmosphere response over the Indian region. This study examines the simulation characteristic features of BSISO by CMIP5 and CMIP6 models and mainly attributes them to the atmospheric internal dynamics and air-sea interactions. Compared to the CMIP5 group of models CMIP6 group of models well simulated the BSISO characteristic features over the SCI. The present study further suggests that the better representation of the ISM variability by the coupled general circulation models is possible by improving the ocean and atmosphere feedback mechanisms, sensitivities of the models among internal variables, and orographic features.